--- slug: the-knot-says-how-many passed: 2026-10-14 by: the assay line (research/assay-line/cited-1014b/), Codex fixer directed by claude-assaying-boyle-1014; source record by Codex, checked by Codex, repaired by Codex on 2026-10-14 checked: 74 confirmed: 63 wrong: 11 unverifiable: 0 observed: 0 stale: 0 --- This was the directed correction pass for the twelve items in research/assay-line/cited-1014b/the-knot-says-how-many/triage.json, against the layer read at f5e5f84f34. The fixer re-read each triage item's evidence before editing, checked the pinned database and arithmetic, and carried forward the 62 CONFIRMED source readings recorded by the Codex checker in research/assay-line/cited-1014b/the-knot-says-how-many/N2-codex.md, explicitly attributed below. After repair, the 11 findings comprise eight WRONG and three MINOR claims; MINOR is included in wrong only for the gate's count arithmetic. Ten triage items were fixed, X7 was rejected on evidence, and X8 remains open under shared census note d24c18. The repair records X8 as WRONG with an open note under the user's shared-class instruction; it also records X2's share-card copy for the landing rebuild. ## Claims - MINOR | The shape of a knot fixes a digit; a person's eye never has to guess it. | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | [X1; convention-02] Re-read Locke's discussion of the Huando specimen: damaged cords made long knots difficult to read and he presents first and probable readings. The page now limits the shape rule to intact, conventionally tied knots. MINOR, no dated line. - MINOR | The shape of a knot fixes a digit, which a reader never has to guess; the position of that knot along the cord fixes its power of ten, which means seeing columns, and empty ones, in the spacing. | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | [X2; convention-13] Same evidence as X1, independently re-read in Locke. Replaced the absolute in the markdown dek with the intact, conventionally tied qualification. MINOR, no dated line. - WRONG | Every long knot in the corpus, bucketed by its number of turns. | git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 108-119; https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/data/khipu.db | [X3; hist-all] Re-read analyse.mjs and queried the 39,170,048-byte database after matching SHA-256 607913034a0f478c5c792025d0c3cab8a9a8f05223ed6bc9c2990122a8c1c4db. TYPE_CODE='L' gives 26,380 rows, 21,893 with NUM_TURNS and 4,487 with NULL turns. The histogram now names its recorded-turn subset and omitted rows. Existing percentages retained. - MINOR | Take every cord that carries at least two knot-columns and exactly one units-type knot, read only the knot shapes and their order down the cord, and check where that units knot lands. | git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 155-164; https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/data/khipu.db | [X4; eligible-rule] Re-read clusterType and the eligibility loop: unknown types are excluded, then L/E clusters are counted. The wording now requires at least two clusters, all of classified types, and exactly one units-type cluster; the caption and other descriptions use the same terminology. MINOR, no dated line. - WRONG | Nearly all of the rest differ for one reason only: the recorder saw an empty column, a place with no knots, and read it as a zero, where the naive left-to-right reading has no way to know a column was skipped. | git show f5e5f84f34:research/the-knot-says-how-many/data/findings.json; git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 183-210 | [X5; rest-zero] Re-read the comparison: recStr.replace(/0/g, '') === mine. The pinned data give clean 32,035, exact 23,178, zeroOnly 6,514, other 2,343. Computed 6,514/(32,035-23,178) = 73.5463475%. The page now says about three quarters of mismatches match after zeros are removed, without attributing the recorder's reasoning. Duplicate causal descriptions are qualified too. - WRONG | The remainder do not: cords with more than one units-type knot, mixed clusters, knots the code has no digit for. | git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 216-234; https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/data/khipu.db | [X6; remainder-causes] Re-read clusterDigit and the cleanNumber loop: each cluster is tested, with no cord-wide single-units restriction. Read-only SQL returns cord 3000049 as cluster 1000071, L, two turns, value 2; cluster 1000072, E, value 1. Both clusters pass and concatenate to 21, while recorded values sum to 3. The remainder now means at least one cluster the parser cannot read as a digit. - CONFIRMED | Reproduce it: node research/the-knot-says-how-many/fetch.mjs then analyse.mjs, verify.mjs. | research/the-knot-says-how-many/README.md; research/the-knot-says-how-many/verify.mjs; scripts/check-placard.mjs; https://artwaste.land/checks/research/the-knot-says-how-many/verify.mjs | [X7; reproduce-instructions] REJECTED. Re-read the footer's repository-relative commands, the README recipe, and verify.mjs input paths; all repository inputs exist. On 2026-10-14, bounded HTTP GETs returned verify.mjs 404 (4,404 bytes), fetch.mjs 200 (2,349 bytes), analyse.mjs 200 (14,320 bytes) at the assumed /checks/ prefix. That confirms an omitted public mirror, not a false repository recipe. scripts/check-placard.mjs deliberately leaves unpublished verifiers as bare paths. The layer verifier was rerun successfully in this repository. - WRONG | Not yet put to that test. | memory/log.d/2026-08-21T2130Z-the-knot-says-how-many.md; scripts/check-placard.mjs; research/verifier-independence/verdicts.mjs | [X8; browser-test-status] Open under the shared census class, per the repair instruction. Re-read the 2026-08-21 log's ordinary browser pass and scripts/check-placard.mjs measuredSentence: this fallback refers to a dependence/mutation census, not an ordinary browser run. The repair rechecked the loader's 20 accepted JSON files: 2,013 records and no record for this layer. The visible sentence does not explain which test it means. Leave the generated line for shared census board note d24c18; no Corrected or Updated line. - WRONG | digit 1: 5 of 21,893 long knots (0.023%). digits 2–9: 98.42%. Every long knot in the corpus, bucketed by its number of turns. | git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 108-119; https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/data/khipu.db | [X9; hist-note] Same independently re-read code and hash-matched SQL as X3. The 5/21,893, 0.023% and 98.42% figures are correct for recorded turn counts. The dynamic histogram note now identifies that subset; the caption discloses the 4,487 missing counts. Shares X3's dated correction. - WRONG | Every long knot in the corpus, bucketed by its number of turns. The bar at 1 is all but empty: 5 knots in 21,893. You cannot tie a long knot of one turn, and the cords agree. | git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 108-119; https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/data/khipu.db | [X10; hist-caption] Same independently re-read code and hash-matched SQL as X3. Replaced the every-long-knot caption with the L-type recorded-turn subset and 4,487 omitted rows. Missing turn counts remain missing, not zero. Shares X3's dated correction; the separate physical-knot convention is unchanged. - WRONG | 72.352% of the 32,035 clean cords match the recorders’ number exactly; another 20.334% differ only because the recorder read an empty column as a zero. Together, 92.686%: the digits are objective, and almost all the disagreement is the placement of zeros. | git show f5e5f84f34:research/the-knot-says-how-many/data/findings.json; git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs lines 183-210 | [X11; read-caption] Same independently re-read comparison and arithmetic as X5. The caption retains 72.352%, 20.334%, and 92.686% with all 32,035 clean cords as denominator, and separately gives 6,514 of 8,857 mismatches (73.546%). It reports numerical agreement after removing zeros without a causal diagnosis. Shares X5's dated correction. - WRONG | Linear B was cracked from the internal shape of a consonant-vowel grid before any sound was known | https://academic.oup.com/bics/article/67/1/86/7977031; https://oup.silverchair-cdn.com/article-minimal/7977031 | [X12; linear-b-grid] Re-read Ventris' Grids, The Syllabic Grids, at the Oxford article and its cited CDN text: the second state maps vowel/consonant relationships without phonetic values; the third has values before the announced solution. The relationship note now states that the grid revealed consonant-vowel relationships before phonetic values were assigned. Structural analysis was a stage toward decipherment. - CONFIRMED | An Inka khipu records a number in two separate acts. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-01] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | The position of that knot along the cord fixes its power of ten, and reading that means seeing columns, and empty columns, in the spacing. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-03] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | single knot = one unit of its column | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-04] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | long knot, n turns = the units digit n | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-05] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | figure-eight = the units digit 1 | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-06] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | Every column but the last is a little bundle of single knots; you just count them. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-07] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | The last column, nearest the free end, is the units, and it is tied differently: a long knot whose turns you count, or, for a lone 1, a figure-eight. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-08] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | It is the one rule that makes the reading start in the right place. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-09] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | If the position of a knot fixes its place value, then the one knot that marks the units, the long knot or figure-eight, should sit at the end of the cord farthest from the top, below all the higher-place bundles. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-10] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | The digit is objective; the zero is a judgment | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-11] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | An Inka khipu writes a number in two acts: a knot's shape gives the digit, and where it sits along the cord gives its power of ten, so the spacing carries the columns and the empty ones too. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-12] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | Inka khipus record numbers in knots. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [convention-14] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | Why a long knot can never say “one” | https://courses.cit.cornell.edu/quipu/data/intro-1.pdf | N2 source reread, carried forward: [long-knot-01] Read the actual scan of Ascher introduction printed p. 17, then OCRed that page only: "A long knot of one turn (1L) is not possible." The figure-eight construction appears directly below. The code convention supplies the units-one sign. - CONFIRMED | A long knot is made by wrapping the cord through a loop several times before pulling it tight; the digit is how many turns you can see. | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [long-knot-02] Locke describes passing the end through the loop "as many times as there are units to be denoted"; one turn therefore leaves the overhand knot. Ascher p. 17 states the impossibility explicitly. This is a construction argument, not a verbatim Locke quotation. - CONFIRMED | But you cannot wrap once and get a long knot. | https://courses.cit.cornell.edu/quipu/data/intro-1.pdf | N2 source reread, carried forward: [long-knot-03] Read the actual scan of Ascher introduction printed p. 17, then OCRed that page only: "A long knot of one turn (1L) is not possible." The figure-eight construction appears directly below. The code convention supplies the units-one sign. - CONFIRMED | One turn is just an overhand knot, a different thing. | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [long-knot-04] Locke describes passing the end through the loop "as many times as there are units to be denoted"; one turn therefore leaves the overhand knot. Ascher p. 17 states the impossibility explicitly. This is a construction argument, not a verbatim Locke quotation. - CONFIRMED | So the code has a built-in hole exactly at the digit 1, which is why the units place needs a second sign, the figure-eight, to fill it. | https://courses.cit.cornell.edu/quipu/data/intro-1.pdf | N2 source reread, carried forward: [long-knot-05] Read the actual scan of Ascher introduction printed p. 17, then OCRed that page only: "A long knot of one turn (1L) is not possible." The figure-eight construction appears directly below. The code convention supplies the units-one sign. - CONFIRMED | Locke states the physical fact directly: a long knot is an overhand knot with the free end passed through the loop once per unit, so a “long knot of one turn” is just an overhand knot, and cannot be a long knot at all. | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [long-knot-06] Locke describes passing the end through the loop "as many times as there are units to be denoted"; one turn therefore leaves the overhand knot. Ascher p. 17 states the impossibility explicitly. This is a construction argument, not a verbatim Locke quotation. - CONFIRMED | a “long knot of one turn” is just an overhand knot | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [locke-one-turn-quotation] The physical inference is supported by Locke and Ascher. Source-record defect: kind=quote incorrectly treats "long knot of one turn" as Locke verbatim. Full-text search finds zero occurrences of that phrase; the same search finds "single or overhand knot" once and "The long knot" three times. In context the page uses a hypothetical label after "so", rather than attributing a verbatim sentence. Change the record to fact/inference. - CONFIRMED | The decimal reading, single knots for the higher places and a long knot for the units, with no more than nine to a place, was published by L. Leland Locke | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [locke-decimal-publication] The two excerpt spans occur in the article: the decimal rows and the Huando specimen convention. They support the described numerical reading; the page separately identifies the figure-eight as the modern convention. - CONFIRMED | L. Leland Locke (“The Ancient Quipu, a Peruvian Knot Record,” American Anthropologist 14, 1912 | https://api.crossref.org/works/10.1525%2Faa.1912.14.2.02a00070; https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [locke-1912-citation] Publisher curl returned 403, so opened the complete Wikisource article and Crossref metadata. Crossref has title "THE ANCIENT QUIPU, A PERUVIAN KNOT RECORD", author L. LELAND LOCKE, year 1912, American Anthropologist volume 14, pages 325-332. - CONFIRMED | elaborated in his 1923 monograph of the almost-same name | https://www.khipufieldguide.com/guidebook/Introduction.html | N2 source reread, carried forward: [locke-1923-monograph] The field guide names the 1923 book "The Ancient Quipu or Peruvian Knot Record"; Medrano also describes Locke following the 1912 article with the 1923 monograph. The distinction between "a" and "or" is accurate. - CONFIRMED | Marcia and Robert Ascher, Code of the Quipu, 1981 | https://openlibrary.org/books/OL4110455M/Code_of_the_quipu | N2 source reread, carried forward: [ascher-citation] Opened the catalogue through the web tool after curl was redirected to a human check. It lists Marcia Ascher and Robert Ascher, "Code of the quipu", and publication date 1981. The Medrano/Khosla article bibliography independently gives the same authors, title and year. - CONFIRMED | Manuel Medrano, 2021 | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [medrano-citation] The transcript heading gives "Manuel Medrano" and "20 October 2021". The numerical convention is in its opening section. - CONFIRMED | The figure-eight sign reserved for the units digit 1 is the standard modern statement of the convention | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [modern-figure-eight] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | it is not named in Locke's 1912 text | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [locke-no-figure-eight] Read the complete article. Searches for figure-eight, figure eight and figure-8 return zero; positive controls "single or overhand knot" and "The long knot" return 1 and 3 matches. Locke does discuss two forms of single knot; the page narrowly says the figure-eight is not named. - CONFIRMED | The khipu remains, as a whole, undeciphered: the numbers are readable, but what they count, and whatever the many non-numerical khipus record, is not settled. | https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/README.md | N2 source reread, carried forward: [undeciphered] The README explicitly says "the khipu recording system remains undeciphered". Medrano distinguishes readable canonical numerals from unsettled nonnumerical interpretation. This does not deny individual contextual interpretations. - CONFIRMED | Knot direction (Z vs S) is recorded in the data and has been argued to carry further, binary information | https://revista.drclas.harvard.edu/signs-of-the-inka-khipu/ | N2 source reread, carried forward: [direction-binary] The review presents Urton's binary theory through successive material decisions; the recorded excerpt specifically discusses spin and ply, so it alone is narrower than the claim. Medrano explicitly identifies S/Z long-knot direction, and the database DIRECTION field recreates the reported split. Treat the binary reading as a theory, as the page does. - CONFIRMED | That khipus are decimal, and that most of them carry readable numbers, is a century old and not in question. | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record | N2 source reread, carried forward: [century-old] Locke 1912 describes decimal orders. Medrano says this can be reliably interpreted across most Inka-style examples. This supplements the record's excerpt, which proves decimal orders but does not by itself establish "most". - CONFIRMED | The Khipu Field Guide reports that non-Lockean knot typologies appear on about 12% of all cords (in roughly 80% of khipus) | https://www.khipufieldguide.com/databook/DatabaseBuild.html | N2 source reread, carried forward: [kfg-non-lockean] Read section 4.14: "80% (535 of 668)" khipus and "12% (7306 of 60858)" of all cords. The page preserves those two denominators. - CONFIRMED | Sabine Hyland (“Knot Anomalies on Inka Khipus,” 2024) | https://digitalcommons.unl.edu/pctix/8/ | N2 source reread, carried forward: [hyland-citation] The university repository names Sabine Hyland, the full title "Knot Anomalies on Inka Khipus: Revising Locke’s Knot Typology", version 2024, and DOI 10.32873/unl.dc.zea.1617. The abbreviated title on the page identifies the same work. - CONFIRMED | finds “nether knots” below the units position on at least one pendant in over 20% of khipus. | https://digitalcommons.unl.edu/pctix/8/ | N2 source reread, carried forward: [hyland-nether] The abstract defines the position below the unit register and reports "over 20%" of repository khipus. The term is in the source. This is a khipu-level fraction, not a cord-level fraction. - CONFIRMED | “nether knots” below the units position | https://digitalcommons.unl.edu/pctix/8/ | N2 source reread, carried forward: [hyland-term] The abstract defines the position below the unit register and reports "over 20%" of repository khipus. The term is in the source. This is a khipu-level fraction, not a cord-level fraction. - CONFIRMED | A 2026 preprint (ALBA Project) reports that 5.4% of knotted cords are incompatible with decimal encoding | https://alba-project.org/khipu-reader/ | N2 source reread, carried forward: [alba-figure] The ALBA page gives "5.4% of knotted cords (1,992 out of 37,206)". This confirms what the preprint reports, not the proposed decipherment or its derivation. - CONFIRMED | A 2026 preprint (ALBA Project) | https://alba-project.org/khipu-reader/ | N2 source reread, carried forward: [alba-preprint-citation] The ALBA page calls the work "a preprint (Sivan, 2026)" and says it is open for peer review; its scholarly note also describes readings as hypotheses. The page accurately labels the cited evidence as a preprint. - CONFIRMED | it is not peer-reviewed | https://alba-project.org/khipu-reader/ | N2 source reread, carried forward: [alba-peer-review] The ALBA page calls the work "a preprint (Sivan, 2026)" and says it is open for peer review; its scholarly note also describes readings as hypotheses. The page accurately labels the cited evidence as a preprint. - CONFIRMED | Medrano and Khosla, ~74% of khipus; the Khipu Field Guide | https://www.cambridge.org/core/journals/latin-american-antiquity/article/how-can-data-science-contribute-to-understanding-the-khipu-code/E2D0E68A0515F3F9582A23C63693F4CE | N2 source reread, carried forward: [sums-citation] Read Cambridge authors and abstract: Manuel Medrano and Ashok Khosla, 650-khipu corpus, Ascher formulae in "at least 74% of khipus". This concerns internal sums and is distinct from units position and digit concatenation. - CONFIRMED | Corpus-scale work on khipu sums (Medrano and Khosla, ~74% of khipus; the Khipu Field Guide) is a separate, occupied question | https://www.cambridge.org/core/journals/latin-american-antiquity/article/how-can-data-science-contribute-to-understanding-the-khipu-code/E2D0E68A0515F3F9582A23C63693F4CE | N2 source reread, carried forward: [sums-research] Read Cambridge authors and abstract: Manuel Medrano and Ashok Khosla, 650-khipu corpus, Ascher formulae in "at least 74% of khipus". This concerns internal sums and is distinct from units position and digit concatenation. - CONFIRMED | github.com/khipulab/open-khipu-repository and cite DOI 10.5281/zenodo.18025748. | https://api.datacite.org/dois/10.5281/zenodo.18025748; https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/README.md | N2 source reread, carried forward: [okr-citation] README requests this DOI. Crossref returned 404 because the DOI is registered at DataCite; DataCite resolves it to "The Open Khipu Repository", OKR Team, version v2.1.0. That 404 is not evidence of a bad DOI. - CONFIRMED | now maintained by the Open Khipu Research Laboratory | https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/README.md | N2 source reread, carried forward: [okr-maintenance] The README identifies the Open Khipu Research Laboratory as administrator under Mackinley FitzPatrick. - CONFIRMED | The data reflect decades of fieldwork by many scholars and, as the repository's own README states plainly, the uneven history of the discipline that produced it. | https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/README.md | N2 source reread, carried forward: [okr-ethics] The README describes "power structures" and "funding disparities" in collection of the data and lists contributions by many scholars. The page fairly summarizes that institutional statement. - CONFIRMED | khipu.db, MIT-licensed, DOI 10.5281/zenodo.18025748 | https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/LICENSE | N2 source reread, carried forward: [okr-mit] Read the repository LICENSE headed "MIT License". The README identifies the SQLite database and gives the DOI. License alone does not establish the DOI, so both were checked. - CONFIRMED | the snapshot Contreras 2026 also analysed | https://arxiv.org/abs/2607.00185 | N2 source reread, carried forward: [contreras-snapshot] Read arXiv abstract and heading: Maria Contreras, submitted 30 June 2026; its OKR corpus is 619 khipus, 54,403 cords and 110,677 knots, matching the page's snapshot counts. This supports the corpus identification, not an independently published byte hash. - CONFIRMED | Contreras 2026 | https://arxiv.org/abs/2607.00185 | N2 source reread, carried forward: [contreras-citation] Read arXiv abstract and heading: Maria Contreras, submitted 30 June 2026; its OKR corpus is 619 khipus, 54,403 cords and 110,677 knots, matching the page's snapshot counts. This supports the corpus identification, not an independently published byte hash. - CONFIRMED | a later Zenodo release adds more. | https://zenodo.org/records/18025748 | N2 source reread, carried forward: [later-zenodo] Opened Zenodo through the web tool and the DataCite record after curl timed out. Release v2.1.0 reports "72 new KH numbers" plus corresponding files. Later refers to expansion beyond the legacy SQLite holdings, not a release after this page's 2026 fetch. - CONFIRMED | Recording conventions vary between investigators | https://www.khipufieldguide.com/databook/DatabaseBuild.html | N2 source reread, carried forward: [recording-variation] The field guide describes reconciling Harvard SQL, Ascher databooks and original Excel measurements, including changes in order and repairs. The excerpt only states multiple sources; the surrounding construction account supplies the recording differences. - CONFIRMED | about a fifth of cords are broken | https://www.khipufieldguide.com/databook/DatabaseBuild.html | N2 source reread, carried forward: [broken-fifth] Read section 4.12, not just the compressed summary excerpt: "20% (12412 of 60858) of All cords are Broken". The fifth refers to cords, not khipus. - CONFIRMED | This page does not argue the convention; it measures how well the corpus obeys it. | git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs | N2 source reread, carried forward: [convention-scope] The analysis explicitly uses the published convention as a hypothesis and computes its corpus conformity; the page does not claim to discover the numerical convention. - CONFIRMED | The nearest published figures are consistent with what is measured here. | https://www.khipufieldguide.com/databook/DatabaseBuild.html; git show f5e5f84f34:research/the-knot-says-how-many/data/findings.json | N2 source reread, carried forward: [comparison-near] KFG reports 12% of all cords; this parser yields 13.613% of its knotted cords. The page names the different denominators and makes a qualitative proximity statement, not an equality. Different corpora and predicates mean this is context, not independent validation. - CONFIRMED | close to the 13.6% of knotted cords that do not read as clean decimal numbers in finding 4 | https://www.khipufieldguide.com/databook/DatabaseBuild.html; git show f5e5f84f34:research/the-knot-says-how-many/data/findings.json | N2 source reread, carried forward: [kfg-near-remainder] KFG reports 12% of all cords; this parser yields 13.613% of its knotted cords. The page names the different denominators and makes a qualitative proximity statement, not an equality. Different corpora and predicates mean this is context, not independent validation. - CONFIRMED | a real rate of ambiguous or incomplete cords is expected and is part of what finding 4 measures. | https://www.khipufieldguide.com/databook/DatabaseBuild.html; git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs | N2 source reread, carried forward: [remainder-ambiguity] KFG documents recording errors and damage. The pinned data directly contain missing turn counts and an incomplete knot-cluster record, which the parser cannot read. The additional possibility of nonnumerical use is explicitly qualified as "may". - CONFIRMED | An immersive page: `public/strata/the-knot-says-how-many/index.html`. | git show f5e5f84f34:research/the-knot-says-how-many/build-page.mjs | N2 source reread, carried forward: [page-location] Pinned git tree and scratch archive contain the named immersive page, findings, analysis and both verification programs. This confirms repository location only; public retrieval is tested separately. - CONFIRMED | Reading data, verifier, and the reproducible analysis live in `research/the-knot-says-how-many/`. | git show f5e5f84f34:research/the-knot-says-how-many/README.md | N2 source reread, carried forward: [research-location] Pinned git tree and scratch archive contain the named immersive page, findings, analysis and both verification programs. This confirms repository location only; public retrieval is tested separately. - CONFIRMED | the khipu's knots read as base-10 numbers though the wider code, and whatever the non-numerical khipus hold, is undeciphered. | https://raw.githubusercontent.com/khipulab/open-khipu-repository/master/README.md | N2 source reread, carried forward: [related-undeciphered] The README explicitly says "the khipu recording system remains undeciphered". Medrano distinguishes readable canonical numerals from unsettled nonnumerical interpretation. This does not deny individual contextual interpretations. - CONFIRMED | the khipu's number is read from the shape of a knot and its column before anyone knows what it counts. | https://www.gresham.ac.uk/watch-now/knot-just-numbers | N2 source reread, carried forward: [related-khipu-form] Medrano transcript, canonical Inka-style convention: numeral from knot type/count, magnitude from horizontal level; 2,035 illustrates an unknotted hundreds level. The claim is about the convention, not an assertion that every surviving cord is conventional. Read through the web fetch after curl returned 403; online checker found the recorded excerpt. - CONFIRMED | Open Khipu Repository, born as the Harvard Khipu Database Project | https://khipufieldguide.ghost.io/counting-the-khipus-how-many-are-there-to-study/ | N2 source reread, carried forward: [okr-origin] Read footnote 13: "Initially designed and managed by Carrie Brezine at Harvard." It says the project was rebadged as OKR in 2021. - CONFIRMED | The recorders' place assignment (from cluster alignment across the whole khipu) is a trained interpretation, not a mechanical fact | https://en.wikisource.org/wiki/The_Ancient_Quipu,_a_Peruvian_Knot_Record; https://www.gresham.ac.uk/watch-now/knot-just-numbers; https://www.khipufieldguide.com/databook/ReadingKhipu.html | N2 source reread, carried forward: [recorder-place-interpretation] The recorded KFG excerpt establishes learned skill, not the full alignment claim. Supplemented with Locke's "arranged roughly in rows across the quipu" and Medrano's horizontal levels and blank-place example. These support contextual place interpretation; the excerpt alone is insufficient. - CONFIRMED | Some are damage or recording gaps; some may be the part of the khipu that was never counting at all. | https://www.khipufieldguide.com/databook/DatabaseBuild.html; git show f5e5f84f34:research/the-knot-says-how-many/analyse.mjs | N2 source reread, carried forward: [remainder-damage] KFG documents recording errors and damage. The pinned data directly contain missing turn counts and an incomplete knot-cluster record, which the parser cannot read. The additional possibility of nonnumerical use is explicitly qualified as "may". - CONFIRMED | Linear A's decimal sums add to KU-RO though the Minoan tongue is lost | https://discovery.ucl.ac.uk/10098748/1/out.pdf | N2 source reread, carried forward: [linear-a-sums] Opened the PDF myself. Text on printed p. 38 says 81-02 precedes 66, the sum of prior numerals, and means total. Printed p. 117 identifies "81-02 KU-RO" and total contexts. Extracted only the relevant pages; the source record's excerpts match. - CONFIRMED | the Minoan tongue is lost | https://www.repository.cam.ac.uk/items/f50c0df4-f355-4bc0-be2a-8e960b2bb5da | N2 source reread, carried forward: [linear-a-language] Opened Cambridge myself after the online checker timed out. The abstract says "the still poorly understood Minoan language" and "remains undeciphered". The page's "lost" is a compressed description of that undeciphered language, not a claim no signs or texts survive. - CONFIRMED | There, why leading digits in real data follow Benford; here, how a non-Western place-value system encodes those same digits in knots, and how well it obeys its own rule across a whole corpus. | git show f5e5f84f34:src/content/strata/the-first-digit-is-a-one.md | N2 source reread, carried forward: [related-benford] Read the pinned linked layer: its dek and plain text describe leading digits, Benford's logarithmic distribution, illustrative sequences and the exceptions. This is a fair statement of the linked page's topic. - CONFIRMED | Nothing here may lie about anything real; that rule never bends. | git show f5e5f84f34:WAKE.md | N2 source reread, carried forward: [site-truth] WAKE.md rule 2 is "It must never lie about anything real." This is a stated project rule, not a claim that this audit finds zero mistakes. ## Findings - [fixed] X1 MINOR: qualified the HTML dek to intact, conventionally tied knots after rereading Locke; no page correction line. - [fixed] X2 MINOR: applied the same qualification to the markdown dek; no page correction line. The PNG share card and its WebP thumbnail retain a truncated copy of the original dek. Landing disposition: share card rebuilt at landing from the corrected dek. No image was rendered or re-encoded during repair. - [fixed] X3 WRONG: histogram heading, caption, surrounding descriptions, README and OG source now identify L-type knots with recorded turns; the analysis exposes the 4,487 missing-turn rows and the verifier checks that count. Added the histogram Corrected line to HTML and markdown. The existing OG raster was not re-rendered under the task restriction. - [fixed] X4 MINOR: eligibility and position descriptions now refer to classified clusters and exactly one units-type cluster; no page correction line. - [fixed] X5 WRONG: replaced nearly all with 6,514 of 8,857 (73.546%, about three quarters), and replaced causal claims with the actual zero-removal comparison throughout this layer. Added the readback Corrected line to HTML and markdown. - [fixed] X6 WRONG: the unreadable remainder now requires at least one unreadable cluster; multiple units-type clusters alone do not exclude a cord. Added the parser Corrected line to HTML and markdown. - [open note:d24c18] X8 WRONG: leave the ambiguous generated census sentence for the shared census correction. The ordinary browser run does not contradict the absence of a dependence/mutation census result, but the visible sentence does not explain that referent. Reclassified from MINOR [declined] under the repair instruction for shared classes. No local placard edit or dated line. - [fixed] X9 WRONG: dynamic histogram note names the recorded-turn subset, with omissions disclosed in its caption; covered by the histogram Corrected line. - [fixed] X10 WRONG: dynamic histogram caption no longer says every long knot; it names the recorded-turn subset and 4,487 omitted rows; covered by the histogram Corrected line. - [fixed] X11 WRONG: dynamic readback caption gives both denominators explicitly and makes no diagnosis of recorder intent; covered by the readback Corrected line. - [fixed] X12 WRONG: relationship note distinguishes prephonetic grid relationships from completed decipherment. Added the Linear B Corrected line to HTML and markdown.