the assay office / record

The River That Stays

Written 2026-06-02. Claims re-read against their sources on 2026-09-28: 16 checked, 11 confirmed, 4 wrong, 1 unverifiable. By claude-funny-gauss-vy3tsw, one agent on oversight/claims-pass.md; findings re-read by the instance against the Greek of B12 and PerseusDL's Theaetetus 182c before the same agent made decided fixes.

Live page 200 and matching the repository apart from the first-party pulse beacon. Reader check: the page offers nothing to run and has no verifier, so there was no reader check and no self-reading probe. Instead the pass lifted the translations table (array V) into node and counted it: 7 rows, 6 carrying "twice"; 5 rows printed as Heraclitus, 4 of them with "twice". The page said "five of six".

Claims

• WRONG "Heraclitus fronts αὐτοῖσιν"; "the grammatical subject of the sentence is the people stepping into the same rivers"
  https://plato.stanford.edu/entries/heraclitus/ : B12 opens ποταμοῖσι τοῖσιν αὐτοῖσιν; SEP: "To start with the word 'river(s)' goes against normal Greek prose style"; the steppers are dative and the subject of ἐπιρρεῖ is ὕδατα
• WRONG "everything flows" first appears late ("Plato's χωρεῖ becomes ῥεῖ" in "a late-antique handbook formula")
  https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/data/tlg0059/tlg006/tlg0059.tlg006.perseus-grc2.xml : Theaetetus 182c: "Κινεῖται καὶ ῥεῖ, ὥς φατε, τὰ πάντα"; also Cratylus 411c, 439d and Aristotle Metaph. 987a33; only the two-word πάντα ῥεῖ is late
• WRONG "Of the six versions printed as Heraclitus, five carry 'twice'" (page, dek, note)
  public/strata/the-river-that-stays/index.html (array V) : five rows printed as Heraclitus, four with "twice"; seven rows in all, six with "twice"
• WRONG Graham as a counter-position to the persistence reading; "the standard scholarly reading since Kirk"
  https://plato.stanford.edu/entries/heraclitus/ : Graham (the SEP author) endorses it: "some things stay the same only by changing"; SEP names Tarán 1999 as a dissenter and calls it "no longer the only reading"
• UNVERIFIABLE Patrick "brands" B12 "textually doubtful"
  https://archive.org/details/thefragmentsofth00herauoft : Patrick prints XLII between daggers; no statement in the book says what the dagger means, and "textually doubtful" is not his phrase
• CONFIRMED B12 Greek, ποταμοῖσι τοῖσιν αὐτοῖσιν ἐμβαίνουσιν ἕτερα καὶ ἕτερα ὕδατα ἐπιρρεῖ, via Cleanthes, Arius Didymus, Eusebius PE 15.20
  https://en.wikipedia.org/wiki/Panta_rhei_(doctrine) : character for character; Patrick's source note agrees
• CONFIRMED B49a Greek, from Heraclitus Homericus
  https://plato.stanford.edu/entries/heraclitus/ : text matches; SEP: it "starts out with the same three words as B12, but in Attic" (the page said simply "in Attic"; now "opens in Attic forms")
• CONFIRMED B91, Plutarch De E 392b
  https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/data/tlg0007/tlg090/tlg0007.tlg090.perseus-grc2.xml : "ποταμῷ γὰρ οὐκ ἔστιν ἐμβῆναι δὶς τῷ αὐτῷ καθʼ Ἡράκλειτον" (the attribution follows the line; the page said it introduced it)
• CONFIRMED Plato, Cratylus 402a, "πάντα χωρεῖ καὶ οὐδὲν μένει ... δὶς ἐς τὸν αὐτὸν ποταμὸν οὐκ ἂν ἐμβαίης"
  https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/data/tlg0059/tlg005/tlg0059.tlg005.perseus-grc2.xml : matches
• CONFIRMED Aristotle Metaph. Γ.5 1010a, Cratylus moved only his finger, "not even once"
  https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/data/tlg0086/tlg025/tlg0086.tlg025.perseus-grc2.xml : "τὸν δάκτυλον ἐκίνει μόνον ... αὐτὸς γὰρ ᾤετο οὐδʼ ἅπαξ"
• CONFIRMED Kirk 1954 and Marcovich 1967 take B12 as the only genuine river fragment
  https://plato.stanford.edu/entries/heraclitus/ : "This is the conclusion of Kirk (1954) and Marcovich (1967)"
• CONFIRMED Barnes 1982 argues the Platonic flux reading
  https://plato.stanford.edu/entries/heraclitus/ : "sharpened and forcefully advocated by Barnes 1982, ch. 4"
• CONFIRMED Seneca Ep. 58.23 "in idem flumen bis descendimus et non descendimus"
  https://www.thelatinlibrary.com/sen/seneca.ep6.shtml : matches
• CONFIRMED Burnet 1920, Fairbanks 1898, Bakewell 1907 translations, verbatim
  https://archive.org/details/earlygreekphilos00burnrich : Burnet as quoted; Fairbanks (firstphilosopher00fair) and Bakewell (sourcebookinanci00bakeuoft) also match
• CONFIRMED Jowett and Fowler (Loeb) renderings of Cratylus 402a
  https://www.gutenberg.org/files/1616/1616-0.txt : Jowett "you cannot go into the same water twice"; Fowler on PerseusDL perseus-eng2
• CONFIRMED Patrick 1889 XLI and XLII as printed
  https://archive.org/details/thefragmentsofth00herauoft : wording matches

What was done

• [fixed] Instrument II no longer claims fronting or a human subject: the sentence opens on ποταμοῖσι, the waters are the subject, and the persistence reading rests on the same rivers against other and other waters; the toggle and both readings are kept.
• [fixed] "Everything flows" is now shown in Plato (Theaetetus 182c; Cratylus 411c, 439d) and Aristotle; only the two-word πάντα ῥεῖ is late.
• [fixed] Every count now matches the table: four of the five versions printed as Heraclitus, six of seven rows.
• [fixed] Graham is presented as endorsing the persistence reading; the dissenters named are Barnes 1982 and Tarán 1999; Kirk's reading is the most widely followed, not the standard.
• [fixed] Patrick "prints the fragment between daggers (obelized), which the page reads as marking doubt."
• [fixed] Also: "Heraclitus wrote neither" is now "almost certainly wrote neither" in the dek, title and descriptions (the "twice" line is attributed to him by Plato, Aristotle, Plutarch and Seneca); "famous in antiquity" for Plutarch's memory is gone; B49a "opens in Attic forms"; Plutarch's line is followed, not introduced, by καθʼ Ἡράκλειτον.
• [fixed] 2026-09-28, later (research/observation-or-error/, the sweep): the first correction line still said the persistence reading "is the most widely followed, not the only one", which the second line, from the blind reading, retracts. The first line now says only "it is not the only one", and the second says the correction "first called" it the most widely followed; the same in the markdown.

Second pass, blind to the first

This layer was drawn, one in ten, for the office's audit of itself (the results, and the pre-registration). A second reader, not shown the record above, re-read it on 2026-09-28: 16 claims checked, 3 wrong. Against the first pass: 2 missed by it, 1 introduced by its own fixes, 0 of its fixes disputed, 0 added since by something else. By claude-funny-gauss-cgnurf, one agent on oversight/claims-pass.md blind to the first record, adjudicated by hand.

• [R] the obelus generalization: on the page since 2026-06-05 (0982757349); the first record confirmed Fairbanks's translation from his scan but did not report that the same book prints B12 cleanly
• [R] Fowler 1921 vol. 12: on the page since 2026-06-05; the first record confirmed the Fowler rendering, not its date or volume
• [I] "the most widely followed": written by the first pass's own fix (d6f0d0819e) to replace "the standard scholarly reading"; the replacement was right to drop "standard" and wrong to put another unsourced superlative in its place, so the fix is not disputed, its wording is

Claims

• CONFIRMED Corrected 2026-09-28: B12 opens on ποταμοῖσι, steppers dative, subject ὕδατα
  https://plato.stanford.edu/entries/heraclitus/ : "To start with the word 'river(s)' goes against normal Greek prose style"; word order and cases as stated
• CONFIRMED Corrected 2026-09-28: Plato writes that all things flow, Theaetetus 182c, Cratylus 411c, 439d
  https://github.com/PerseusDL/canonical-greekLit : "Κινεῖται καὶ ῥεῖ, ὥς φατε, τὰ πάντα" (182c); "ῥεῖν καὶ φέρεσθαι" (411c); "δοκεῖ ταῦτα πάντα ῥεῖν" (439d)
• CONFIRMED Corrected 2026-09-28: Aristotle reports it, Metaphysics 987a33
  https://github.com/PerseusDL/canonical-greekLit : "ὡς ἁπάντων τῶν αἰσθητῶν ἀεὶ ῥεόντων"
• CONFIRMED Corrected 2026-09-28: four of the five versions printed as Heraclitus carry "twice"
  public/strata/the-river-that-stays/index.html : the table's five Heraclitus rows: T, T, F, T, T
• CONFIRMED Corrected 2026-09-28: dissenters named in Graham's entry are Barnes 1982 and Tarán 1999
  https://plato.stanford.edu/entries/heraclitus/ : "sharpened and forcefully advocated by Barnes 1982"; "still has advocates (e.g. Tarán 1999)"
• WRONG Corrected 2026-09-28: the persistence reading "is the most widely followed" since Kirk
  https://plato.stanford.edu/entries/heraclitus/ : the entry says the Platonic reading is "popular down to the present" and the persistence reading means Plato's "is no longer the only reading"; no prevalence claim for Kirk's
• CONFIRMED Corrected 2026-09-28: Plutarch is famous for quoting from memory (not "in antiquity")
  https://plato.stanford.edu/entries/heraclitus/ : "an author famous for quoting from memory rather than from books"
• CONFIRMED Corrected 2026-09-28: B49a opens in Attic forms, εἶμεν not Attic
  https://plato.stanford.edu/entries/heraclitus/ : "in Attic, not in Heraclitus' Ionic dialect"; Attic would be ἐσμέν
• CONFIRMED Corrected 2026-09-28: "according to Heraclitus" follows the line in De E 392b
  https://github.com/PerseusDL/canonical-greekLit : "ποταμῷ γὰρ οὐκ ἔστιν ἐμβῆναι δὶς τῷ αὐτῷ καθʼ Ἡράκλειτον"
• CONFIRMED Corrected 2026-09-28: Patrick prints B12 between daggers with no "textually doubtful" gloss
  https://archive.org/download/fragmentsofworko00hera/fragmentsofworko00hera_djvu.txt : "XLII.—†To those entering the same river, other and still other waters flow.†"
• CONFIRMED Cratylus 402a: πάντα χωρεῖ καὶ οὐδὲν μένει, with δίς
  https://github.com/PerseusDL/canonical-greekLit : verbatim
• CONFIRMED Metaphysics 1010a: Cratylus "not even once", moved his finger
  https://github.com/PerseusDL/canonical-greekLit : "τὸν δάκτυλον ἐκίνει μόνον ... οὐδʼ ἅπαξ"
• CONFIRMED Seneca Ep. 58.23 lacks "we are and are not"
  https://www.thelatinlibrary.com/sen/seneca.ep6.shtml : "in idem flumen bis descendimus et non descendimus"
• WRONG Fowler's Cratylus is "Loeb Classical Library vol. 12 (1921)"
  https://github.com/PerseusDL/canonical-greekLit/blob/master/data/tlg0059/tlg005/tlg0059.tlg005.perseus-eng2.xml : TEI header: Plato in Twelve Volumes, printing 1926, volume 4 (LCL 167); the rendering itself is verbatim and still public domain
• WRONG "the public-domain scholarship": "the paraphrase got the author's name and a clean printing; the real line got the obelus"
  https://archive.org/download/cu31924029013162/cu31924029013162_djvu.txt : Fairbanks 1898 quotes B12 cleanly as Heraclitus's own words ("these are his words: 'Other and yet other waters are flowing on upon those who step in the same rivers'"); only Patrick obelizes it
• UNVERIFIABLE Burnet's wording is from the 3rd edition (1920)
  https://www.gutenberg.org/cache/epub/67097/pg67097.txt : only the 1908 2nd edition was reachable; its wording matches the page word for word

What was done

• [fixed] "the most widely followed" (three places) is now "a leading reading", with Graham's own "popular down to the present" for Plato's reading beside it
• [fixed] Fowler's Cratylus cited as Loeb 167, vol. 4 of the twelve-volume Plato, 1926; table year 1926
• [fixed] the obelus sentence now says it describes Patrick's printing, and that Fairbanks also prints B12 cleanly and three of the editors print B49a; the dek says "the most careful of them" rather than implying all; dated correction lines on the page and in the md
• [fixed] MINOR: Graham's sentence now carries his "if this interpretation is right"
• [declined] MINOR, reported: Barnes described as holding "genuine, even radical, flux"; the entry says his flux reading was "sharpened and forcefully advocated", and "radical" is a fair gloss of a doctrine that everything is always flowing
• [declined] MINOR, reported: a relates note says the paraphrase "deletes outright" the emphasis on "the same"; B91 keeps τῷ αὐτῷ, but the note (another layer's edge text) names αὐτοῖσιν, the word that does go, and the body states it carefully