---
slug: loom-of-the-singer
passed: 2026-10-18
by: the assay line (research/assay-line/claude-1018b/), Claude fixer (claude-opus-5-5) directed by claude-assaying-boyle-1014; source record by a Claude agent (claude-sonnet-5-5), checked by a Claude agent (claude-opus-5-5)
checked: 79
confirmed: 58
wrong: 19             # 3 WRONG + 16 MINOR
unverifiable: 1
observed: 1
stale: 0
---

This was a directed repair pass over the 19 triage items in research/assay-line/claude-1018b/loom-of-the-singer/triage.json (3 WRONG, 16 MINOR, tier B); the claim lines below are the checker's (N2-claude.md, 79 claims against the N1 source record), with each WRONG and MINOR line's outcome appended. The three WRONG items were re-read at the source (Parry, The Making of Homeric Verse, archive.org text; Wiktionary; the Perseus Book 1 XML) and fixed at their source, research/loom-of-the-singer/verify.mjs, whose regenerated loom-data.json was re-pasted into the page; three dated Corrected lines were added. One MINOR in the same sentence as a WRONG was fixed with it; the other fifteen MINOR lines are listed, page unchanged. The unverifiable superlative ("the most-quoted sentence in Homeric scholarship") was not in the triage and is left as it stands.

## Claims
- CONFIRMED | [west-two-cola] M.L. West: "we may regard the hexameter as consisting of two cola" | https://archive.org/download/west-1987-introduction-to-greek-metre/West%201987%20Introduction%20to%20Greek%20Metre_djvu.txt | "In view of the regular word-break in the third foot we may regard the hexameter as consisting of two cola." Verbatim.
- CONFIRMED | [west-feminine-formulaic] it is the feminine break that "is more often presupposed by formulaic phrases." | https://archive.org/download/west-1987-introduction-to-greek-metre/West%201987%20Introduction%20to%20Greek%20Metre_djvu.txt | "In the third foot the 'feminine' caesura ... predominates over the 'masculine' ... in the proportion 4 : 3, and is more often presupposed by formulaic phrases." Verbatim.
- CONFIRMED | [parry-formula-def] "a group of words which is regularly employed under the same metrical conditions to express a given essential idea." attributed to "Studies in the Epic Technique of Oral Verse-Making I" (1930) | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | MHV p. 272, HS section 2 "The Formula": "The formula in the Homeric poems may be defined as a group of words which is regularly employed under the same metrical conditions to express a given essential idea." HS is HSCPh 41, 1930 (MHV contents).
- MINOR | [parry-1928-def] "Milman Parry's answer, in 1928, was that the verse is not remembered but woven, live, from ready-made, exactly-measured word-groups he called formulae." | https://en.wikipedia.org/wiki/Milman_Parry | The 1928 theses argued formulaic, traditional diction; the oral, in-performance claim came later. Wikipedia: "In his American publications of the 1930s, Parry introduced the hypothesis that the formulaic structure of Homeric epic is to be explained as a characteristic feature of oral composition". The TE translation in MHV never uses the word "oral" (searched; "formula" occurs 689 times in the same span). The page itself says the in-performance inference was "added later". Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [parry-he-said-to-him] Parry's own example of a formula was one of these very half-lines, τὸν δʼ αὖτε προσέειπε, whose essential idea, he wrote, is merely "he said to him." | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | MHV p. 272: "that in τὸν δ᾽ αὖτε προσέειπε is 'he said to him'." The rack's thread is τὸν δʼ αὖτε προσέειπε θεὰ (Parry's phrase plus θεά), close enough to "one of these very half-lines".
- CONFIRMED | [parry-thrift-def] "The thrift of a system lies in the degree in which it is free of phrases which, having the same metrical value and expressing the same idea, could replace one another." (Collected Papers, 276) | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | Verbatim, on the page headed "276 The Epic Technique of Oral Verse-Making".
- CONFIRMED | [parry-37-characters] "in a list of 37 characters who have formulas of this type… there are only three names which have a second formula which could replace the first." | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | MHV p. 277; the ellipsis drops "which includes all those having any importance in the poems,", a fair cut. "This type" is the nominative noun-epithet formula from the trochaic caesura to the verse-end.
- CONFIRMED | [parry-disproof] "If one wishes to think that Homer composed his poems orally, and then sat down and wrote them out, there is little that can be said in disproof, and little that needs to be said." | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | MHV p. 322, verbatim; Parry's sentence goes on ", since the question ceases to be one of the oral style".
- MINOR | [parry-inference-not-proof] "That the poems were therefore composed in performance by a non-writing singer is an inference he added later, and he knew it was not a proof" | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | The quoted sentence concedes only that an oral poet might have written his poems out himself ("the question ceases to be one of the oral style, and becomes that of the way in which the spoken poetry was recorded"); the next sentence reads "We now come to the last of proofs that the diction of the Iliad and the Odyssey is oral". Parry presented orality as proved; he conceded the writing-down. The page turns his concession about writing into a disclaimer about oral composition. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [parry-proved-formulaic] Parry proved the diction is formulaic and traditional. | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | MHV p. 277: "the length and the thrift of the system are striking enough to be sure proof that only the very smallest part of it could be the work of one poet"; Page, quoted in the introduction: "a fact capable of proof in detail".
- CONFIRMED | [parry-extension-thrift] Parry named them extension (...) and thrift, or economy | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | Parry's words vary: TE "great extension and by great simplicity"; HS p. 276 "two factors, that of length and that of thrift"; FM p. 234 "too wide an extension and too rigorous an economy". All three terms are his.
- MINOR | [parry-same-words] "Milman Parry's discovery is that Homeric Greek gave its singers a stock of shapes so precisely engineered that a poet could build a flawless six-beat line, on the fly, out of parts that were guaranteed to fit." | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | Parry's other 1928 thesis (FM) is about the flaws formulae cause. Adam Parry, introduction p. xxxix: Parry "shows how most of the metrical flaws in Homer can be explained as uncommon juxtapositions of traditional formulae ... a complex, but not infinitely adjustable, system." "Flawless" and "guaranteed to fit" overstate the discovery. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [parry-one-thread] for a given hero and a given metrical slot the tradition keeps essentially one | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | MHV p. 277: "there are in only a very few cases more than one such formula for a single character"; 3 of 37.
- MINOR | [parry-flagged-doublets] "Parry flagged such doublets himself;" (said of ἄναξ ἀνδρῶν Ἀγαμέμνων and εὐρὺ κρείων Ἀγαμέμνων) | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | Parry did flag equivalent formulae (p. 277 n. 2: "The formulas joined by a bracket are those which can replace one another"; TE p. 40: "only 9 are not unique in both sense and metre"), but the Agamemnon pair is not among them: his lists give ἄναξ ἀνδρῶν Ἀγαμέμνων alone in the feminine-caesura slot. "Such doublets" rests on the wrong claim at d-agamemnon-two-long. FIXED 2026-10-18 with the WRONG item beside it: "such doublets" now points at the pair Parry names (MHV p. 234), "this one by name".
- CONFIRMED | [parry-hard-reading] The hard reading says the epithet is chosen for metre, not sense: ... and the audience did not mind. | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | HG, MHV p. 249: the fixed epithet "is used inattentively by the poet, and heard by the auditor in a like manner"; "this semantic consideration is dominated by that of metrical convenience".
- CONFIRMED | [making-of-homeric-verse] Parry, The Making of Homeric Verse (1971) | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | Title page: "THE MAKING OF HOMERIC VERSE ... Edited by Adam Parry ... OXFORD AT THE CLARENDON PRESS 1971".
- CONFIRMED | [lord-singer-of-tales] Lord, The Singer of Tales (1960) | https://en.wikipedia.org/wiki/Albert_Lord | "Lord authored the book The Singer of Tales, first published in 1960."
- CONFIRMED | [finkelberg] Finkelberg, "Oral Theory and the Limits of Formulaic Diction," Oral Tradition 19 (2004) | https://api.crossref.org/works/10.1353/ort.2005.0004 | Crossref: author Margalit Finkelberg, Oral Tradition vol. 19 issue 2, pp. 236-252, issued 2004-10.
- CONFIRMED | [shive-citation] David Shive (Naming Achilles, 1987) | https://archive.org/metadata/namingachilles0000shiv | "creator":"Shive, David","date":"1987" (tool FOUND).
- CONFIRMED | [shive-argument] argued there is more choice and less economy than Parry claimed | https://www.labyrinthbooks.com/naming-achilles | "there is more extension and less economy than Milman Parry claimed to find"; "more choice" is a fair gloss.
- CONFIRMED | [hainsworth-citation] J.B. Hainsworth (The Flexibility of the Homeric Formula, 1968) | https://archive.org/metadata/flexibilityofhom0000jbha | "creator":"J.B. Hainsworth","date":"1968" (tool FOUND).
- CONFIRMED | [hainsworth-argument] showed formulae bend and move | https://bmcr.brynmawr.edu/?p=26004 | "contest the idea that formulas were immovable and unchanging blocks ... explored various ways in which formulas changed and evolved".
- CONFIRMED | [friedrich-argument] counted the breaches and argued many are deliberate art | https://www.steiner-verlag.de/Formular-Economy-in-Homer/9783515090650 | "reveals a high frequency of breaches of economy in Homer, and demonstrates that these are for the most part motivated by poetic considerations."
- CONFIRMED | [friedrich-citation] Rainer Friedrich (2007) | https://www.steiner-verlag.de/Formular-Economy-in-Homer/9783515090650 | Hermes Einzelschriften 100, Copyrightjahr 2007.
- CONFIRMED | [aa-parry-citation] Anne Amory Parry (Blameless Aegisthus, 1973) | https://api.crossref.org/works/10.1163/9789004327351 | Brill monograph, published-print 1973 (DOI MATCH).
- CONFIRMED | [epithets-not-dead] Anne Amory Parry (Blameless Aegisthus, 1973), Norman Austin, and Paolo Vivante argued the epithets are not semantically dead: they can be contextually alive. | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt ; https://etheses.bham.ac.uk/id/eprint/10259/7/grey2020MRes.pdf | N1 found no source; two exist. MHV intro p. lv n. 2: "the forthcoming monograph of A. Amory, Blameless Aegisthus, which, accepting the formulary nature of such epithets, argues cogently that they, like other words in Homer, had none the less precise and ascertainable meanings." Grey (Birmingham MRes thesis, 2019): "For Vivante, the notion that meter is the raison d'etre for Homeric language is a presumption that must be rejected outright in favour of the primary question of semantics"; "Lowenstam, like Austin and Stanford before him, feels that we are entitled to look for wider significance beyond the purely formulaic".
- CONFIRMED | [ornamental-dispute] The "ornamental / meaning-free epithet" is a live dispute (A. A. Parry, Austin, Vivante) | https://etheses.bham.ac.uk/id/eprint/10259/7/grey2020MRes.pdf | Grey (2019) describes the "century-long revision of Parry's structuralist theories" and the "archetypal schism" that has "divided us all into Parryists 'of one shade or another'", citing all three.
- CONFIRMED | [achilles-swift-footed-sitting] Achilles is "swift-footed" even while sitting still in his tent | research/loom-of-the-singer/iliad-monro-allen.xml | Il. 9.218 "αὐτὸς δʼ ἀντίον ἷζεν Ὀδυσσῆος θείοιο" (he sits facing Odysseus in his hut), then 9.307 "τὸν δʼ ἀπαμειβόμενος προσέφη πόδας ὠκὺς Ἀχιλλεύς·". Also 1.349 he sits apart, 1.364 πόδας ὠκὺς Ἀχιλλεύς.
- CONFIRMED | [achilles-epithet-wiki] Achilles is "swift-footed" even while sitting still in his tent, because that is the thread that fits, and the audience did not mind. | https://en.wikipedia.org/wiki/Epithets_in_Homer | Wikipedia confirms the usage ("not called 'swift-footed' only when he runs") and reads it differently; the page offers its reason as "the hard reading", which is Parry's (see parry-hard-reading), and gives the opposing view next.
- CONFIRMED | [line-count-nearly-28000] Homer's two epics run to nearly 28,000 lines | https://en.wikipedia.org/wiki/Odyssey | "Of the 27,803 lines in the original texts"; Perseus numbering gives 15,693 + 12,110 = 27,803.
- MINOR | [line-count-allen] "This line count is the two poems' combined length (Allen edition, 27,793 lines)." | https://en.wikipedia.org/wiki/Odyssey ; research/loom-of-the-singer/iliad-monro-allen.xml ; https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/data/tlg0012/tlg002/tlg0012.tlg002.perseus-grc2.xml | No source gives 27,793. The standard numbering is 27,803 (Iliad 15,693 + Odyssey 12,110, the books' highest line numbers in Perseus); the Perseus line elements sum to 27,794 (Monro-Allen Iliad 15,687 + Odyssey 12,107, also 12,107 at the June 2026 Perseus commit 0c97c913d6). The Perseus Odyssey is A. T. Murray's Loeb (1919), not Allen. The repository has no derivation (git grep; the build log says "~27,000 lines"). Listed 2026-10-18 (tier B MINOR), page unchanged.
- MINOR | [line-count-iliad] (same sentence as line-count-allen: "Allen edition, 27,793 lines") | https://en.wikipedia.org/wiki/Iliad | "It contains 15,693 lines in its standard edition"; with the Odyssey's 12,109/12,110 the total is 27,802/27,803, not 27,793. Same finding as line-count-allen. Listed 2026-10-18 (tier B MINOR), page unchanged.
- MINOR | [line-count-odyssey] (same sentence as line-count-allen: "Allen edition, 27,793 lines") | https://en.wikipedia.org/wiki/Odyssey | "The epic has 12109 lines"; does not yield 27,793. Same finding as line-count-allen. Listed 2026-10-18 (tier B MINOR), page unchanged.
- MINOR | [composed-orally] "Homer's two epics ... were composed without writing, at speaking speed, in performance, by a poet who did not have them by heart" (dek and opening, stated as fact) | public/strata/loom-of-the-singer/index.html | Wikipedia (Odyssey) supports oral composition, but the page's own §V says "Treat oral composition here as a strong, well-supported hypothesis, not a settled fact" and lists "dictation to a scribe, a literate master using traditional craft" as still open; Parry (MHV p. 322) also leaves writing open. The dek states as settled what §V says is not. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [book1-611-caesura] "In Iliad Book 1's 611 lines the break falls after the first short of the third foot (the "feminine" caesura) in 356 lines and at the strong "masculine" break in 247." | https://dcc.dickinson.edu/node/28775 | Monro §367: "In the first book of the Iliad, which contains 611 lines, the trochaic caesura of the third foot occurs in 356, and the corresponding strong caesura in 247." The Perseus Book 1 has 611 lines.
- CONFIRMED | [caesura-third-foot] Almost every hexameter has a word-break, a caesura, near the middle, in the third foot | https://dcc.dickinson.edu/node/28775 | "There is nearly always a caesura in the third foot."
- CONFIRMED | [fifth-sixth-foot] The fifth foot is almost always a dactyl; the sixth is always two syllables, the last of them free. | https://dcc.dickinson.edu/node/28775 | §366: "The fifth thesis nearly always consists of two short syllables"; "except the last thesis, which consists of one syllable, either long or short."
- CONFIRMED | [six-feet-equal] same duration, since a long counts as two shorts | https://dcc.dickinson.edu/node/28775 | "six feet, of equal length"; the thesis is "one long or two short syllables".
- CONFIRMED | [yugoslavia-1930s] Parry and Albert Lord went to Yugoslavia in the 1930s | https://en.wikipedia.org/wiki/Milman_Parry | "Between 1933 and 1935 ... two visits to Yugoslavia ... with the help on his second visit of his assistant Albert Lord".
- CONFIRMED | [guslari-illiterate] recorded illiterate guslari composing epic in performance | https://en.wikipedia.org/wiki/Milman_Parry ; https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | "illiterate farmers who sang epic songs of prodigious length"; Lord in MHV: singers "Accustomed to compose their verses rapidly to the rhythm of the instrumental accompaniment".
- CONFIRMED | [avdo-12311] One, Avdo Međedović, dictated a song of 12,311 lines | https://en.wikipedia.org/wiki/Avdo_Me%C4%91edovi%C4%87 | Listed under "Dictated in 1935": "The Wedding of Meho, Son of Smail ... (12,311 lines)". MHV intro: "taken down by dictation ... has over 12,000 lines".
- MINOR | [avdo-12323] dictated a song of 12,311 lines | https://en.wikipedia.org/wiki/Avdo_Me%C4%91edovi%C4%87 | The same article quotes Lord: "Avdo's in 1935 had 12,323 lines". Two published counts for the same text; the page gives one without a note. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [avdo-illiterate] from a man who could not read | https://en.wikipedia.org/wiki/Avdo_Me%C4%91edovi%C4%87 | "Međedović was illiterate his whole life."
- CONFIRMED | [avdo-odyssey-length] Odyssey-scale | https://en.wikipedia.org/wiki/Avdo_Me%C4%91edovi%C4%87 | Lord: "Avdo could sing songs of about the length of Homer's Odyssey." (Odyssey 12,109).
- CONFIRMED | [proves-possibility] That proves a poem of Homeric length can be made orally. | https://en.wikipedia.org/wiki/Avdo_Me%C4%91edovi%C4%87 | "convincing him that someone Homer-like could produce a poem so long".
- CONFIRMED | [collected-material-dictated] much of the collected material was itself dictated | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | N1 found no source for the collection as a whole; MHV has it: Parry "collected songs, some recorded on phonograph discs, some taken down by dictation"; "nearly 13,000 Serbocroatian texts, including those on more than 3,500 phonograph discs"; Lord: "many instances of the same song from a single singer both in a recorded and in a dictated version".
- CONFIRMED | [perseus-monro-allen] re-extracted from the Monro-Allen Oxford Iliad (Perseus edition perseus-grc2) | research/loom-of-the-singer/iliad-monro-allen.xml | TEI header: "Homeri Opera / David B. Monro / Thomas W. Allen / Editio Tertia / Oxford / Clarendon Press".
- OBSERVED | [d-verifier-47] "Verifier 47/47, offline." | recreated: node verify.mjs (scratch copy of research/loom-of-the-singer at 9ddc1e9045) | Printed "ALL CHECKS PASSED: 47 passed, 0 failed", both on the pinned XML and on Perseus master; the build log memory/log.d/2026-07-19-loom-of-the-singer.md records 47/47. Re-run 2026-10-18 after the fixes: "ALL CHECKS PASSED: 47 passed, 0 failed".
- MINOR | [d-economy-611] "Then the economy principle is re-counted from all 611 lines of Book 1 before your eyes" (also the §IV heading "Economy, re-derived live from Book 1" and the markdown "re-counted, live") | public/strata/loom-of-the-singer/index.html | Nothing is counted in the browser: the §IV table is drawn from the DATA.economy constant (index.html lines 327-335), a copy of loom-data.json written offline by verify.mjs. The only live work is re-scanning the 8 attested lines. The numbers are computed, but not live and not before the reader's eyes. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [d-seven-figures] each hero carries a small kit of formulae of different metrical lengths, one per slot, plus a scatter of ordinary namings | research/loom-of-the-singer/loom-data.json | True as a tendency of the computed table. In the table recomputed 2026-10-18 Hera has two 8-syllable formulae of one shape, θεὰ λευκώλενος Ἥρη 4× and βοῶπις πότνια Ἥρη 2× (the breach §V names), so the dek and the markdown body now read "essentially one per slot" (MINOR, fixed 2026-10-18 by the assay line's repair stage, no page line). The Odysseus overcount of ordinary namings (see the added Odysseus line) is fixed 2026-10-18.
- MINOR | [d-every-time-seven] "Below is every time one of seven principal figures is named at the end of a line in Book 1 of the Iliad" | research/loom-of-the-singer/verify.mjs ; research/loom-of-the-singer/iliad-monro-allen.xml | Not every time: verify.mjs line 335 counts only lines whose last word, accents stripped, equals the nominative stem, so it skips line-final vocatives and oblique cases (1.131 "θεοείκελʼ Ἀχιλλεῦ", 1.508 "μητίετα Ζεῦ", Apollo at 1.14, 21, 370, 373, 438). Because the stripped dative matches "ηρη", it does count one dative as an "ordinary syntax" one-off (1.572 "λευκωλένῳ Ἥρῃ"). Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [d-achilles-two] Achilles is πόδας ὠκὺς Ἀχιλλεύς to fill the long slot and δῖος Ἀχιλλεύς to fill the short one, and almost nothing else. | research/loom-of-the-singer/iliad-monro-allen.xml | Recounted independently: πόδας ὠκὺς Ἀχιλλεύς at 1.58, 84, 148, 215, 364, 489; δῖος Ἀχιλλεύς at 1.7, 121, 292 (1.121 is the longer ποδάρκης δῖος Ἀχιλλεύς); one-offs 1.54, 330, 348.
- WRONG | [d-agamemnon-two-long] "Agamemnon carries two long formulae of almost equal metrical value, ἄναξ ἀνδρῶν Ἀγαμέμνων and εὐρὺ κρείων Ἀγαμέμνων, where strict thrift wants one." offered under "Economy is a tendency with documented breaches. Even the count above shows it." | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt ; research/loom-of-the-singer/loom-data.json | Not a breach of thrift as Parry defines it (the page quotes him: phrases "having the same metrical value"). The shapes differ: SLLLSSLL (13 morae) against LLLLSSLL (14 morae), so the first fills the verse from the feminine caesura and the second from the masculine. The two cannot replace each other. Parry's TE Table I (MHV p. 39) gives ἄναξ ἀνδρῶν Ἀγαμέμνων as Agamemnon's one formula "Between the feminine caesura and the end of the line", and κρείων Ἀγαμέμνων for the hephthemimeral slot. TE p. 80 says εὐρὺ κρείων "owes its existence to its presence in two formulae of greater length: (thrice) ἥρως Ἀτρείδης εὐρὺ κρείων Ἀγαμέμνων (7 times) ... Ἀτρείδης εὐρὺ κρείων Ἀγαμέμνων". All three Book-1 cases are "Ἀτρεΐδης εὐρὺ κρείων Ἀγαμέμνων" (1.102, 355, 411). The page's own verifier passes "no two RECURRING noun-epithet formulae for a figure share a metrical shape" on this pair. FIXED 2026-10-18: §V now gives Parry's own named doublet, θεὰ λευκώλενος Ἥρη ~ βοῶπις πότνια Ἥρη (MHV p. 234), both in Book 1; Corrected line added.
- CONFIRMED | [d-one-formula-per-length] And within any one length there is essentially one formula, used again and again, and that is economy. | research/loom-of-the-singer/loom-data.json | Holds as "essentially" in the recomputed table: the one same-shape pair, θεὰ λευκώλενος Ἥρη 4× ~ βοῶπις πότνια Ἥρη 2×, is the breach §V names; Agamemnon's two eight-syllable formulae fill different slots.
- CONFIRMED | [d-verbatim] Every Greek word is verbatim from the Perseus Monro-Allen text | research/loom-of-the-singer/verify.mjs | Section [A] passes on the pinned XML; I checked that the page's inline DATA equals the regenerated loom-data.json, and matched all 40 thread pairings against the XML.
- CONFIRMED | [d-scanned] every displayed line is machine-scanned | research/loom-of-the-singer/verify.mjs | Section [B]: the 12 shown lines scan. Il. 1.1 is in loom-data.json but is not displayed on the page (no μῆνιν in index.html).
- CONFIRMED | [d-attested] Every combination the loom marks attested is asserted to be the exact word-sequence at that line of Book 1. | research/loom-of-the-singer/iliad-monro-allen.xml | Section [C] passes. Running the page's own scanner over all 40 pairings and matching each against all 611 Book-1 lines found no Book-1 line the loom fails to mark.
- MINOR | [d-sha-pinned] "the file's SHA-256 is pinned so a future reader sees the same bytes" | research/loom-of-the-singer/verify.mjs ; https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/data/tlg0012/tlg001/tlg0012.tlg001.perseus-grc2.xml | The hash is recorded, not enforced. verify.mjs prints the hash (line 97) and compares it with nothing. Its download hint fetches Perseus master, now ae8b617018c6..., not the pinned ad44f6d75acd..., and on that file it still passes 47/47 without comment. A reader without the repository does not see the same bytes, and nothing tells them so. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [d-same-scanner] the same scanner the loom runs live in your browser | public/strata/loom-of-the-singer/index.html | I compared syllabify/scan in index.html with verify.mjs lines 186-260: the same algorithm, LEXQ, DIPH and correption rule, hand-ported. The 2026-10-18 circumflex and lexical-table change was made in both copies; all 40 loom pairings give the same verdicts before and after.
- CONFIRMED | [d-one-sourced-quantity] (as amended 2026-10-18) "two sourced lexical quantities, θεά's long final ᾱ and the long ῡ of χρυσόθρονος" | https://en.wiktionary.org/wiki/%CE%B8%CE%B5%CE%AC | Wiktionary headwords "θεᾱ́" and (https://en.wiktionary.org/wiki/χρυσόθρονος) "χρῡσόθρονος", from "χρῡσός". The page's Cunliffe/LSJ citation for θεά could not be fetched, as N1 found.
- CONFIRMED | [d-last-verified] Last verified 19 July 2026, when it was written. | src/content/strata/loom-of-the-singer.md | date: 2026-07-19T22:30:00Z; the layer commit is 41acc997311 of 2026-07-19.
- MINOR | [d-note-wine] There the formulaic epithet οἶνοψ ('wine-faced') is followed across a dozen English hands | src/content/strata/wine-dark-sea.md ; public/strata/wine-dark-sea/index.html | That page has six: "Six public-domain translators decided" and section "II · One word, six hands". Its own record lists "a dozen ways" as a MINOR that it fixed. Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [d-note-first-word] The first-word page weighs μῆνιν, the poem's opening word, against every English translator who could not keep it first. | public/strata/the-first-word-is-rage/index.html | Its eight public-domain translators put the word at positions 2 to 9; the modern ones who kept it first are described, not weighed.
- CONFIRMED | [d-note-sixth-letter] There the meter is a detector: a sound the alphabet dropped, the digamma, is still legible in lines that stop scanning without it | src/content/strata/the-sixth-letter.md | Its dek: "It still left a fingerprint in the meter: in hundreds of lines that don't scan unless you put the letter back."
- MINOR | [d-note-sixth-scanner] Both scan real Book-1 lines live with the same conservative scanner | public/strata/the-sixth-letter/index.html | They are different scanners. The sixth-letter page says "Epic correption, unmarked hiatus, and muta-cum-liquida ambiguity all give a competent ancient reciter more freedom than the page's scanner allows", and the loom's scanner adds epic correption and a lexical table (build log: "ported ... and extended"). Listed 2026-10-18 (tier B MINOR), page unchanged.
- CONFIRMED | [d-note-bench] The scansion bench teaches English iambic pentameter, where you perform a line over the metronome grid and watch the stresses the dictionary fixes. | src/content/strata/prosody-scansion-bench.md | "Perform a line and watch three layers at once: the metronome grid, the stresses the dictionary fixes".
- CONFIRMED | [d-note-grid] The Linear B grid cracks a Bronze Age syllabary with no sound-values (...) until real Greek words fall out. | src/content/strata/the-grid-that-spoke-greek.md | "sorted into a grid by pure structure ... with no sound values at all ... spelling real Greek words".
- CONFIRMED | [d-recounted-by-scanner] recounted from the Perseus text by the same scanner that runs the loom | research/loom-of-the-singer/verify.mjs | Section D counts by word-matching and shapes by the shared scanner, offline (the "live" part is the d-economy-611 finding).
- CONFIRMED | [d-table-not-typed] The table in §IV is not typed by hand: it is computed from all 611 lines and printed by the verifier. | research/loom-of-the-singer/verify.mjs (recreated: node verify.mjs in a scratch copy) | The regenerated loom-data.json equals the committed one except the hash, and the inline DATA.economy equals it. The computed values carried the errors in the two added WRONG lines; regenerated 2026-10-18 and re-pasted, inline DATA.economy equals loom-data.json.
- CONFIRMED | [d-same-slot] Swap πόδας ὠκὺς Ἀχιλλεύς for νεφεληγερέτα Ζεύς after the same opening and the line still scans | research/loom-of-the-singer/loom-data.json | Both SSLSSLL; the page's scanner marks every gold thread plus νεφεληγερέτα Ζεύς valid, and Il. 1.560 and 20.19 attest the swap.
- CONFIRMED | [d-long-agamemnon-overflows] Reach for Agamemnon's long form after that same opening and it overflows the line | public/strata/loom-of-the-singer/index.html | The page's scanner rejects all five gold threads plus ἄναξ ἀνδρῶν Ἀγαμέμνων; 1.130 and 1.285 use κρείων Ἀγαμέμνων.
- CONFIRMED | [west-citation] M.L. West, Introduction to Greek Metre (1987), 19–22. | https://archive.org/download/west-1987-introduction-to-greek-metre/West%201987%20Introduction%20to%20Greek%20Metre_djvu.txt | Title page "OXFORD • CLARENDON PRESS 1987"; contents "A. THE DACTYLIC HEXAMETER ' . ig" (OCR for 19) and "B. THE ELEGIAC COUPLET 23".
- CONFIRMED | [monro-citation] Caesura counts for Book 1: Monro, Homeric hexameter, via Dickinson College Commentaries. | https://dcc.dickinson.edu/node/28775 | "D.B. Monro, A Grammar of the Homeric Dialect. Carlisle, Pennsylvania: Dickinson College Commentaries, 2014."
- CONFIRMED | [gloss-swift-footed] πόδας ὠκὺς Ἀχιλλεύς (swift-footed Achilles) | https://en.wikipedia.org/wiki/Epithets_in_Homer | "swift-footed (πόδας ὠκύς".
- WRONG | (added) §IV table metrical shapes: δῖος Ἀχιλλεύς shown as ⏑⏑⏑–– (DATA shape "SSSLL", 5 syl; index.html lines 328 and 332) and χρυσόθρονος Ἥρη as ⏑–⏑⏑–– ("SLSSLL"), under "with their metrical shape (– long, ⏑ short)" | https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt ; https://en.wiktionary.org/wiki/%CF%87%CF%81%CF%85%CF%83%CF%8C%CE%B8%CF%81%CE%BF%CE%BD%CE%BF%CF%82 | Both are mis-scanned. δῖος carries a circumflex, so its ι is long. Parry's TE Table I (MHV p. 39) puts δῖος Ἀχιλλεύς "Between the bucolic diaeresis and the end of the line": –⏑⏑––, and three shorts in a row cannot stand in a hexameter. χρυσόθρονος has a long ῡ (Wiktionary "χρῡσόθρονος"), so the true shape is ––⏑⏑––, the same measure as λευκώλενος Ἥρη (Parry p. 277 n. 2 lists it among formulae "which have the same measure"). Cause: the scanner treats α, ι, υ as always short outside diphthongs (verify.mjs line 188; index.html line 340). The page's own scanner prints δῖος Ἀχιλλεύς "SSSLL" and χρυσόθρονος Ἥρη "S?SSLL". FIXED 2026-10-18: the scanner (verify.mjs and the page) reads α ι υ under a circumflex as long and carries χρῡσόθρονος in its sourced table; table now δῖος Ἀχιλλεύς LSSLL, χρυσόθρονος Ἥρη LLSSLL; Corrected line added.
- WRONG | (added) §IV Odysseus row: "named at line-end 4× in Book 1", πολύμητις Ὀδυσσεύς 1×, and 3× "named by ordinary syntax (verb / particle + name)" (the page: "the 'one-off' lines, where the poet names the hero not with a formula but with ordinary syntax") | research/loom-of-the-singer/iliad-monro-allen.xml | Two of the three are stock formulae: 1.145 "ἢ Αἴας ἢ Ἰδομενεὺς ἢ δῖος Ὀδυσσεὺς" and 1.440 "τὴν μὲν ἔπειτʼ ἐπὶ βωμὸν ἄγων πολύμητις Ὀδυσσεὺς". So πολύμητις Ὀδυσσεύς occurs 2× (1.311, 1.440), δῖος Ὀδυσσεύς 1×, and only 1.430 "αὐτὰρ Ὀδυσσεὺς" is ordinary syntax. Both formulae are in verify.mjs's own KNOWN list (line 321). The matcher (line 340) compares accented forms exactly, so the grave-accented Ὀδυσσεὺς of a run-on line misses the acute in KNOWN. Recreated with the verifier's own output: "· one-off: …δῖος Ὀδυσσεὺς", "· one-off: …πολύμητις Ὀδυσσεὺς". FIXED 2026-10-18: verify.mjs matches the known formulae with grave read as acute; the row now reads πολύμητις Ὀδυσσεύς 2×, δῖος Ὀδυσσεύς 1×, one-off 1×; Corrected line added.
- MINOR | (added) dek: "Snap one of each together and a live scanner tells you whether it is a real line of Homer or a well-formed hexameter the system merely licenses." | research/loom-of-the-singer/iliad-monro-allen.xml | The loom checks Book 1 only. Four pairings it labels "system-valid (not in Book 1)" are real lines of the Iliad: 10.382 "τὸν δʼ ἀπαμειβόμενος προσέφη πολύμητις Ὀδυσσεύς" (also 10.423, 10.554, 19.154, 19.215), 20.19 "τὸν δʼ ἀπαμειβόμενος προσέφη νεφεληγερέτα Ζεύς", 5.888 "τὸν δʼ ἄρʼ ὑπόδρα ἰδὼν προσέφη νεφεληγερέτα Ζεύς", 4.349 and 14.82 "τὸν δʼ ἄρʼ ὑπόδρα ἰδὼν προσέφη πολύμητις Ὀδυσσεύς". The verdict label itself ("not in Book 1") is true; the dek's either/or is not. Listed 2026-10-18 (tier B MINOR), page unchanged.
- MINOR | (added) §I: the third-foot caesura is "the seam the singer weaves across: build the first half up to the break, then drop in a subject-phrase cut to exactly fill the rest", and §II's gold threads are "speech-introductions that run from the line-start to the caesura" | https://dcc.dickinson.edu/node/28775 ; https://archive.org/download/MilmanParryTheMakingOfHomericVerse1971/Milman%20Parry%2C%20edited%20by%20Adam%20Parry%20-%20The%20Making%20of%20Homeric%20Verse-Oxford%20at%20the%20Clarendon%20Press%20%281971%29_djvu.txt | All five gold threads run past the third foot to the princeps of the fourth (the verifier's scansions: τὸν δʼ ἀπαμειβόμενος προσέφη is 10 syllables of –⏑⏑ –⏑⏑ –⏑⏑ –). So the loom's seam is the hephthemimeral caesura, which Monro names separately ("a strong caesura in the fourth foot (τομὴ ἑφθημιμερής)"). Parry's Table I files πόδας ὠκὺς Ἀχιλλεύς, νεφεληγερέτα Ζεύς, κρείων Ἀγαμέμνων, γλαυκῶπις Ἀθήνη, λευκώλενος Ἥρη and πολύμητις Ὀδυσσεύς "Between the hephthemimeral caesura and the end of the line". The 356/247 third-foot figures describe a different seam from the one the loom weaves. Listed 2026-10-18 (tier B MINOR), page unchanged.
- UNVERIFIABLE | (added) "Parry's definition of the unit is the most-quoted sentence in Homeric scholarship" | tried: no citation count or survey for a single sentence exists to check against; Wikipedia and the MHV introduction call the definition central but rank nothing | An unsourced superlative. "One of the most-quoted" or "the classic definition" would be checkable.

## Findings
- [fixed] d-agamemnon-two-long (WRONG): §V offered ἄναξ ἀνδρῶν Ἀγαμέμνων / εὐρὺ κρείων Ἀγαμέμνων as a thrift breach; by Parry's definition (same metrical value) they are not one (Table I, MHV p. 39; TE p. 80). Now Parry's named doublet θεὰ λευκώλενος Ἥρη ~ βοῶπις πότνια Ἥρη (MHV p. 234; Book 1 at 1.55, 195, 208, 595 and 1.551, 568). The checker's suggested Hera pair (λευκώλενος / χρυσόθρονος) was not used: Parry p. 277 n. 2 says the initial sound changes χρυσόθρονος's metrical value. verify.mjs KNOWN gains θεὰ λευκώλενος Ἥρη so the table shows the pair, and its economy check, which asserted no recurring same-shape pair, now asserts that this doublet is the only one (made to go red by removing the formula). Corrected line on the page and in the markdown.
- [fixed] parry-flagged-doublets (MINOR, in the sentence fixed above): "Parry flagged such doublets himself, this one by name;" now points at a pair Parry did name. No line on the page (MINOR).
- [fixed] (added) shapes (WRONG): δῖος Ἀχιλλεύς was ⏑⏑⏑–– and χρυσόθρονος Ἥρη ⏑–⏑⏑––; true –⏑⏑–– (Parry Table I, bucolic slot) and ––⏑⏑–– (Wiktionary χρῡσόθρονος). Scanner in verify.mjs and index.html now reads α ι υ under a circumflex as long and has χρυσόθρονος in LEXQ; the check list on the page now says "two sourced lexical quantities". loom-data.json regenerated, DATA.economy re-pasted. Corrected line added.
- [fixed] (added) Odysseus row (WRONG): 1.145 δῖος Ὀδυσσεὺς and 1.440 πολύμητις Ὀδυσσεὺς were counted as ordinary syntax because the matcher compared accents exactly; verify.mjs now reads grave as acute when matching, row is πολύμητις 2×, δῖος 1×, one-off 1× (made to go red by undoing the change). Corrected line added.
- [declined] [parry-1928-def] "Milman Parry's answer, in 1928, was that the verse is not remembered but woven, live, from r (MINOR, parry-1928-oral-dating): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; the checker's note holds on its face (Wikipedia: Parry's oral-composition hypothesis came in "his American publications of the 1930s"); the 1928 sentence is loose dating of the 'live' part, not a false figure.
- [declined] [parry-inference-not-proof] "That the poems were therefore composed in performance by a non-writing singer is (MINOR, parry-concession-reframed): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; the quoted words are verbatim (MHV p. 322); the checker's point is that the page frames his concession about writing-down as a concession about orality, a slant in the gloss, not a false quotation.
- [declined] [parry-same-words] "Milman Parry's discovery is that Homeric Greek gave its singers a stock of shapes so preci (MINOR, flawless-guaranteed-overstated): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; 'flawless' and 'guaranteed to fit' overstate Parry's finding (Adam Parry, MHV p. xxxix, on metrical flaws from formula juxtaposition); loose wording.
- [declined] [line-count-allen] "This line count is the two poems' combined length (Allen edition, 27,793 lines)." (MINOR, line-count-27793): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; 27,793 has no source (standard numbering 27,803; Perseus elements 27,794) and the Perseus Odyssey is Murray, not Allen; a precision slip in a footnote whose 'nearly 28,000' headline is confirmed.
- [declined] [line-count-iliad] (same sentence as line-count-allen: "Allen edition, 27,793 lines") (MINOR, line-count-27793): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; same sentence and same error as line-count-allen (K5).
- [declined] [line-count-odyssey] (same sentence as line-count-allen: "Allen edition, 27,793 lines") (MINOR, line-count-27793): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; same sentence and same error as line-count-allen (K5).
- [declined] [composed-orally] "Homer's two epics ... were composed without writing, at speaking speed, in performance, by (MINOR, oral-composition-stated-as-fact): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; the dek states oral composition flatly while the page's own section V hedges it as "a strong, well-supported hypothesis"; the page itself names the edge, so this is loose wording in the dek.
- [declined] [avdo-12323] dictated a song of 12,311 lines (MINOR, avdo-two-counts): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; 12,311 is a published count (Wikipedia list); Lord's 12,323 is a second count; giving one without the other is a precision point.
- [declined] [d-economy-611] "Then the economy principle is re-counted from all 611 lines of Book 1 before your eyes" (also (MINOR, economy-live-overclaim): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; 're-counted ... before your eyes' / 're-derived live' overclaims the page's own checking: the table is computed offline by verify.mjs and pasted; the brief classes such overclaims as MINOR.
- [declined] [d-every-time-seven] "Below is every time one of seven principal figures is named at the end of a line in Book (MINOR, every-time-overclaim): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; 'every time' is really 'every line ending in the nominative name' (verify.mjs line 335); the dative 1.572 counted as a one-off is a small slip. Loose wording about the page's own method.
- [declined] [d-sha-pinned] "the file's SHA-256 is pinned so a future reader sees the same bytes" (MINOR, sha-recorded-not-enforced): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; verify.mjs prints the hash and compares it with nothing (line 97); 'pinned so a future reader sees the same bytes' overstates the page's own check.
- [declined] [d-note-wine] There the formulaic epithet οἶνοψ ('wine-faced') is followed across a dozen English hands (MINOR, wine-dozen-vs-six): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; relates note says 'a dozen English hands'; the wine-dark-sea page has six. A slip of precision in a cross-reference note.
- [declined] [d-note-sixth-scanner] Both scan real Book-1 lines live with the same conservative scanner (MINOR, same-scanner-overstated): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; the loom's scanner is the sixth-letter scanner extended (epic correption, a lexical table), not 'the same'; loose wording in a relates note.
- [declined] (added) dek: "Snap one of each together and a live scanner tells you whether it is a real line of Homer or a w (MINOR, real-line-of-homer-vs-book1): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; the loom checks Book 1 only, so some 'system-valid' pairings are real lines elsewhere in the Iliad (10.382, 20.19, 5.888, 4.349); the verdict label 'not in Book 1' is true, the dek's either/or is loose.
- [declined] (added) §I: the third-foot caesura is "the seam the singer weaves across: build the first half up to the break (MINOR, caesura-seam-imprecise): Tier B MINOR, not in a sentence being fixed for a WRONG item: listed, page left as it is; the gold threads run to the fourth-foot (hephthemimeral) break, not the third-foot caesura the 356/247 counts describe (Parry Table I files these namings 'Between the hephthemimeral caesura and the end of the line'); imprecise wording about which seam.