Expositio de tabernaculo

Merged from two agent batches (chunks 0000–0007 and 0008–0014), 2026-07-28.

Work-wide conventions (reconciled at merge)

• The column's four members keep their Latin, in italic: stylus, epistylium (pl. epistylia). Richard argues the vocabulary etymologically at 0233A — epistylium dici non solet nisi quod stylo inhaeret — so an English pair like "epistyle"/"shaft", which share no root, destroys the very thing being argued. The two batches disagreed here (first half Latin, tail anglicized) and the tail's 22 instances were converted to the first half's scheme at merge. Same principle as the Theotocon precedent in 7561.
• capitellum → "capital" in both halves, deliberately: it is a standard English architectural term and carries no etymological load in Richard's argument.
• "shaft" is reserved for the LAMPSTAND's shaft (chunk 0000, with its branches, cups and lily) — a different object. It is never the column member.
• Other fixed terms: cortina / sagum / velum / operimentum; annulus vs circulus (load-bearing at 0222B–D); net / row / order kept distinct (Richard's argument at 0240A turns on versus vs ordines not collapsing); laver / basin / sea; propitiatorium, oraculum.

Anchors: column 0241 genuinely recurs — NOT a defect

Chunk 0010 runs 0240D → 0241A → 0242A; chunk 0011 then opens 0241 (band-less) → 0241B → 0241C → 0241D → 0242B. Column 241 is revisited. This is the plate's own order — a table set across the tops of both columns, with the prose resuming below it — and nothing was reordered. Verified against the Latin at merge.

Every entry is an individual judgment call, logged per instance. Pattern numbers refer to translation-style.md. Chunks 0008–0014 are logged separately in cruces-0008.md; the two files are to be merged.

Terminology fixed for this batch

Richard's tropological scheme rests on a technical vocabulary he uses with care, and several terms have no obvious English default. The renderings below are held constant across chunks 0000–0007 and should be held constant in the rest of the work.

• cortina → "curtain"; sagum → "hair-cloth"; velum → "veil"; operimentum → "covering". Four distinct words for four distinct textiles, all of which English wants to call "curtain." Velum is used both for the assembled sheet of ten curtains ("velum cortinarum") and for the sanctuary veil; "veil" carries both without collapsing into cortina. Saga cilicina is rendered "hair-cloths of goats' hair" — redundant in English, but cilicina is a separate printed word and Richard argues from its roughness (0217A).
• annulus → "ring"; circulus → "circle". The distinction is load-bearing at 0222B–D, where Richard's whole argument is that the annuli of the grate and the circuli of the altar must be the same objects. Rendering both "ring" would answer his question for him. The same convention is kept for the golden circuli of the curtains at 0215A.
• craticula → "grate"; arula → "hearth"; retiaculum → "network"; stipes → "post"; vectis → "bar"; basis → "base".
• propitiatorium → "propitiatory." Not "mercy-seat": Richard derives the name from propitiatio in the same sentence (0213A, "Merito ergo propitiatorii nomen accepit … propitiationis suae praerogativam"), and the etymology has to survive.
• oraculum → "oracle." Likewise: 0227A is an extended note on the word's three senses.
• tabulatum → "boarding"; tabula → "board"; mansio → "storey"; coenaculum → "upper room"; deambulacrum → "walkway"; laquear → "ceiling."
• stylus, epistylium are kept as italic Latin (0232D–0233B). Richard's four-member analysis of columna turns on the etymology epi + stylus ("Nam epistylium idem sonat quod super. Inde epistylium, quod ponatur supra stylum"), which "shaft"/"architrave" destroys. Basis and capitellum have unproblematic English ("base", "capital") and are translated. Note the consequence: hastile of the candlestick (0214A–B) is "shaft," and stylus of the column is not.
• interrasilis (0213D), coclea and vitis (0225B), cameron (0231C) are kept as italic Latin — in each case Richard is naming the word, not using it.
• holocaustum → "holocaust" throughout, including the chapter head Descriptio altaris holocausti.

Note: these are editorial choices, not defects. Italic Latin here means "term deliberately left in Latin" (register section of translation-style.md), never [sic: …].

Chunk 0000

0212B — eroganero for erogavero. Printed: "et esurientibus atque petentibus libenter eroganero." No such Latin verb form exists; the future perfect of erogare is erogavero, and it is required by the parallel collegero in the same clause. Non-word type under Pattern 7: carried into the English untranslated, in italics, marked [sic: eroganero]. Compositor's n for v.

0214B — excepta ad junctione calamorum for excepta adjunctione. Printed: "Similem vero ornatum excepta ad junctione calamorum habebat in summo." excepta governs an ablative, and ad + ablative junctione does not construe; the reading required is adjunctione, "except for the attachment of the branches" — which is exactly Richard's point, that the topmost ornament is the same as the others minus the branch-junctions. Split type under Pattern 10, and the WHOLE run is carried (including ad, which is a real word), marked [sic: ad junctione].

0212B — nunquam credendus est nullos … impertiri? — double negative, and a question mark on a statement. Printed: "Sed qui assidue producit novos fructus … nunquam credendus est nullos de caetero scientiae profectus impertiri, ad innovandos sensus hominis interioris?" The argument requires nunquid ("is he to be believed to impart NO further advances?"); nunquam + nullos makes the sentence assert the opposite of the prologue's whole thesis, that God still gives new understanding. nunquam is a real word and can be rendered, so under Pattern 12 it takes no marker: rendered literally, "is he … never to be believed to impart no further advances." The ? is preserved per Pattern 8. Near-certain nunquid.

0211B–D — Bede's exposition quoted with no quotation marks on the plate. Printed: "dum inter exponendum adjungit. Haec, inquit, de re difficillima … si quis nos docere voluerit. In quibus tamen omnibus patet allegoriae sensus perlucidus." Migne prints neither guillemets nor italics; the quotation is marked only by inquit. Per Pattern 6 no « » are supplied (they would falsely claim the plate prints them) and none are added at all here, since the inquit already marks it and the quotation's end is not typographically determined. The full stop after adjungit — which breaks the sentence immediately before the quotation it introduces — is preserved per Pattern 8.

0211D → 0212B — the column band A is skipped. The anchors run …[0211D] … [0212B]…, with no [0212A]. Reproduced as printed; column 0212A is occupied on the plate by matter our TEI does not carry at this point. Not an omission of ours.

0213A — nisi quia non aliud est ejus operculum, et aliud propitiatorium! A question closed with an exclamation mark. Preserved per Pattern 8.

0212D — the lemma Coram, inquit, propitiatorio … ubi loquar tibi has no terminal stop. Printed: "… ubi loquar tibi [n: (Exod. XXX)] Testimonium quod in arca ponitur…". The italic run closes and the next sentence begins with no punctuation between. Mirrored exactly; no stop supplied (Pattern 8).

0213B — Sicut itaque illi de templo, sic nec isti … stare poterant. The sicut clause has no verb; the sense is "just as those of the temple [could not], so neither could these." Rendered with the ellipsis carried as English syntax allows, nothing invented.

0213B — measurement pair: joined wings of ten cubits (Solomon's cherubim) against five (the tabernacle's), over an ark of bicubitum et semissem. Reproduced as printed, including the plural verb tenebant with the singular subject extensio. Richard's argument depends on the two-and-a-half-cubit ark being too small to stand on; the figures agree with Exod. 25:10 and III Reg. 6:24 and are not emended.

Chunk 0001

0216B — in satis for in sagis. Printed: "Littera sic legatur: Ex medietate autem ejus quod superfuerit in satis quae parantur tecto, id est sagum unum." This is Richard re-reading the lemma he has just quoted correctly two lines above ("Quod autem superfuerit in sagis quae parantur tecto"), so the intended word is beyond doubt sagis. satis is a real Latin word but is indeclinable and cannot govern the relative quae; the clause does not construe. Under Pattern 12's real-word-merely-wrong rule it takes no [sic: …] marker and is rendered as ordinary English — "in enough which are prepared for the roof." This is the entry most in need of a human's eye in this batch: the boundary case Pattern 12 flags (11081 col. 0030B) is exactly this shape, a literal rendering that is actively misleading rather than merely odd.

0214C–D — the whole opening paragraph is an attack on Josephus, with no [n: …] anywhere. Migne supplies no citation for the claim that Jeremiah hid the tabernacle (II Macc. 2:4–8). None invented.

0215A — nec tamen dicitur quod sint in summitatibus contradicts the lemma just quoted. Printed lemma: "Ansulas hyacinthinas in lateribus, ac summitatibus facies cortinarum"; Richard then: "Ecce dicitur quod ansulas habeant in omni parte; nec tamen dicitur quod sint in summitatibus, sed tantum quod habeant in utroque latere." As printed he denies what he has just quoted. The argument works if both quods are quot ("it is not said HOW MANY are at the tops, but only how many they have on either side") — which is what the following lemma supplies, fifty per side and no number for the tops. quod is a real word: rendered literally, no marker, per Pattern 12.

0215D — arithmetic of the two coverings. Curtains 28 × 40 cubits; hair-cloths 30 × 42. Reproduced exactly, including the genitive quadraginta duorum. The one-cubit overhang on each dimension is Richard's own conclusion and the figures are internally consistent.

Chunk 0002

0217C / 0217D — the same Exodus lemma printed two different ways within one chapter. First: "Facies et tabulas stantes tabernaculi de ligno setim, quae singulae denos cubitos…"; second: "Facies, inquit, et tabulas stantes tabernaculis de lignis setim, quae singulae denos cubitos…" — tabernaculi → tabernaculis, ligno → lignis. Each rendered as printed per the scripture policy ("of the tabernacle" / "for the tabernacles"; "of setim wood" / "of setim woods"); no harmonization. The second is the reading Richard is arguing FROM, which is why it matters that it not be silently aligned with the first.

0217D — Sunt qui tradunt … habere. sed unde hoc probetur. Full stop followed by lowercase sed. The stop is preserved per Pattern 8 and the English sentence begins "But"; capitalization is English orthography, not a printed mark, and is not a silent correction. Same class as 11542 col. 1041D.

0218D — note [n: (Ezech. IV)] on a passage from Ezekiel 41. The quoted matter ("et mensus est frontes sex cubitos latitudinis hinc, et sex cubitos latitudinis inde") is Ezech. 41:1, not chapter IV. Migne's citation reproduced verbatim per the marker rule; flagged only so the indexer's refDisplay is not mistaken for a resolution.

0218B–C — the corner-board arithmetic is Richard's own and is internally consistent. Six western boards at 1½ cubits = 9; the corner boards each contribute half a cubit inward to make the tenth, and project one cubit outward on each side; hence 10 cubits within and 12 without. Rendered exactly, with no reconciliation against Exod. 26:22–23, which does not give these figures. His rejection of the half-cubit corner board (0217D, "dimidium tantummodo cubitum") and of the five-cubit bar (0218D, "de quinque cubitorum vectibus") are both positions he attributes to others; the numbers are theirs, not errors.

0219A — per medias tabulas secundum librum de angulo in angulum. secundum librum, "according to the book," is Richard's standing formula for the biblical text; it recurs at 0221A and 0227C. Rendered uniformly.

Chunk 0003

0219D — a posteriori par te for a posteriori parte. Printed with the word split. Both fragments are real words (par, te), which is precisely why Pattern 10 requires the whole broken run to be carried: [sic: par te]. Rendering it "from the hinder part" would look like a considered reading and would conceal the plate's defect.

0219D — novam autem hinc, et novem inde for novem … novem. Printed: "decem medii cubiti expandantur in tabernaculi latitudine; novam autem hinc, et novem inde in utriusque lateris descensione." The arithmetic is fixed: 28 cubits of width, ten spread over the tabernacle's breadth, nine down each side. novam is therefore certainly novem, and the parallel novem inde two words later is the evidence. But novam is a real Latin word (feminine accusative of novus), so under Pattern 12 it takes no marker and is rendered as ordinary English — "a new one on this hand, and nine on that." A load-bearing number is lost in the literal rendering; this is the second entry that most needs a human's eye. Logged loudly because a reader who does not open these notes will read a nonsensical count where Migne's plate has a one-letter misprint of nine.

0220C — triplicatur, At si contra tres cubitos…. A comma where a full stop is required, followed by capital At. The comma is preserved per Pattern 8; the English reads ", But if…".

0220A — hyacintinarum pellium beside hyacinthinis pellibus at 0216D. The h-less spelling is an ordinary medieval orthographic variant and is exempt under Pattern 9: translated normally, no italics, no marker.

Chunk 0004

0221B — ut firmiter inter sacrificandum acret and ad illud Dominicum praeceptum stacederet — a transposed st. Printed in a single sentence: "Idcirco autem altare ipsum mediotenus terra implebatur, ut firmiter inter sacrificandum acret, et hoc ipsum ad illud Dominicum praeceptum stacederet." Neither acret nor stacederet is a Latin word. The intended reading is staret … accederet: the initial st of staret has migrated forward to the head of accederet, leaving acret behind. One compositor's error producing two non-words, and the transposition is only visible if both are carried. Both carried untranslated in italics per Pattern 7's non-word class, marked [sic: acret] and [sic: stacederet].

0221B — Nec arula … inferebatur, et … imponebatur — a negative that reverses the sentence. Printed: "Nec arula per aditum dictorum ostiorum inferebatur, et super craticulae bases, et super ostium limen imponebatur; terra tamen prius tota illa altaris inferior medietas usque ad situm craticulae implebatur." The whole surrounding argument is that the hearth WAS brought in through the little doors and laid on the bases of the grate — that is the reason the doors exist, and Richard says so explicitly at 0222D ("quod de ostiolorum aditu dictum est"). As printed, the sentence denies it, and then the coordinated imponebatur has to be dragged under the same negative, which produces nonsense. Almost certainly Nam for Nec (or an intruded negative of the class already documented for this author at 11534 col. 1054A and 11537 col. 1069D). Rendered literally as printed per Pattern 7, with no marker, since nec is a real word. Flagged loudly: this is the third and most consequential entry in this batch, because the literal rendering states the opposite of Richard's own conclusion eight lines later.

0222C — induxitque incirculos for induxit … in circulos. Run-together type inside a scripture lemma. Carried untranslated per Pattern 10, marked [sic: incirculos]; because it sits inside an already-italic lemma it takes no second italic marker, per that pattern's closing rule. The correct two-word form stands three lines later ("et ipsos in circulos altaris induxit"), which is the evidence.

0222C — vectes ad immittendum in craticulae annulis fecit. Gerund ad immittendum with an accusative object vectes, and in + ablative where the sense requires accusative. Rendered as closely as the printed words allow ("he made bars for putting into the rings of the grate"); the grammar is not repaired and the reading is not marked, since every word is real.

0221D–0222B — the geometry of grate, hearth and rings is Richard's own reconstruction and is internally consistent. Altar 5 × 5 × 3 cubits, grate at the half-height, hearth beneath the grate but above the bases, rings hanging from the lower ends. The apparent contradiction he wrestles with — Eritque craticula usque ad altaris medium against subter eam in altaris medio arulam — is in the Vulgate, not in the plate, and is left standing as he leaves it.

0221C — ubi subauditur facies. Richard notes that the verb is to be supplied in the lemma. Rendered with straight quotation marks (ours, per Pattern 6) around the supplied verb: where "thou shalt make" is understood. Migne prints no marks here.

Chunk 0005

0225B — ad tringinta ascensionis cubitum for triginta. Printed: "Una via ascensionis ad utrumque serviebat, et ad tringinta ascensionis cubitum introitus ad utrumque patebat." tringinta is not a Latin word. Carried untranslated in italics per Pattern 7 and marked [sic: tringinta] — a load-bearing number, so the crux matters: triginta (thirty) is the height of the lowest storey given five lines earlier at 0223B, but the surrounding clause ("ad … cubitum introitus") does not construe with any reading, so the conjecture stops at the numeral. (Chunk 0009 carries the same misprint; the peer batch marks it identically.)

0225C — Juxta id autem quod dictum enim de via et ostio medii lateris. enim stands where est is required; there is no finite verb. Rendered with the copula supplied in brackets — "according to that which, indeed, [was] said" — the bracket being our own supplied word in the corpus's existing convention. enim is a real word and takes no marker.

0223D — Has templo adjacentes coaedificationes lateraliter adjacebant. Accusative subject with an intransitive verb; the sentence does not construe. Rendered as closely as the printed words allow ("These buildings lying alongside the temple lay alongside it at the sides"), unmended.

0224A — the boarding lemma ends without a stop: … septem habens cubitos latitudinis Ecce de mensura. Mirrored; no stop supplied (Pattern 8).

0223B — heights of the temple. "triginta cubitos altitudine" (0223B, ablative without in) for the lowest storey, against "centum viginti cubitorum" for the whole house from Paralipomenon. Richard reconciles them by storeys; both figures rendered exactly as printed, and the missing in is left alone as an ordinary ellipsis, not a defect.

0228B — note [n: (Marc. V.)] on the rending of the veil. The verse is Mark 15:38. Migne's citation reproduced verbatim.

0224D — quomodo … clavi illi fortiter affigerentur … loquitur. A question closed with a full stop. Preserved per Pattern 8.

Chunk 0006

0228B — Sed in medium ducimus quod in Paralipomenon legimus citius fortassis elucescet quod quaerimus. Two main clauses in asyndeton, with no conjunction and ducimus indicative where a hortatory subjunctive is expected. Rendered as two clauses joined by a semicolon; the punctuation of the plate (a single full stop at the end) is otherwise untouched.

0230A — Superfluum vero videtur quod expressio materiae reperitur — possible dropped non. Printed with no negative. Richard has just praised the exactness of the description ("Bene autem prius dicitur, quod fecerit ostiola, postea quod fecerit ostia") and at 0230A–B goes on to make an argument that depends entirely on the material being specified twice, olive wood at the oracle against fir at the temple entrance. As printed he calls that specification superfluous — the opposite of what he then does with it. Non superfluum is the likely reading. Rendered literally as printed per Pattern 7. Flagged: this is the fourth possible dropped negative in this stretch of PL 196.

0229B — cujus operatione quaelibet remotior ultra extrinsecus dici poterat. operatione where the sense requires comparatione ("by comparison with which any farther place could be called 'beyond outside'"). Real word, wrong word: rendered literally ("by whose operation…"), no marker, per Pattern 12.

0229B — Sed quantum quomodo id fieri potuit facile perpenditur, legentium conjecturae relinquitur. Two interrogative/relative adverbs stacked with no conjunction and no subject for relinquitur; the sentence does not construe. Rendered as closely as the printed words allow and left unmended.

0230A — quantum ex quatuor ostiolis duo ostia sunt facta, and 0230B–C Sed quantum in hoc templo omnia intrinsecus operiebantur ligno. quantum in a causal/concessive sense where classical Latin wants quoniam or quatenus. This is attested medieval usage rather than a defect: exempt under Pattern 9, translated normally ("in so far as", "in so far as"), no marker. Logged because the two instances sit six lines apart across a chunk boundary and could look like a repeated misprint.

0228C — Tenenda itaque est absque ambiguo quod dictum velum … pendebat. Feminine tenenda est governing a quod-clause. No English exponent for the gender mismatch (Pattern 9); rendered "It is therefore to be held without ambiguity that…", nothing to emend.

0228C — de situ arce for arcae. The ae → e monophthong is ordinary medieval orthography, exempt under Pattern 9.

0228B — Ubi ergo, obsecro, hoc velum pendebat, quod sancta et Sancta sanctorum ab invicem dividebat. A direct question closed with a full stop. Preserved per Pattern 8; no ? supplied.

0227A–0229A — capitalization of Sancta sanctorum. Migne prints Sanctum sanctorum, Sancta sanctorum, and sanctum a Sanctis sanctorum within three columns. The English mirrors the printed capitalization case by case ("Holy of Holies" where the Latin capitalizes, "holy place" for lowercase sanctum); no normalization, since unlike a section label (Pattern 9) this is running text.

Chunk 0007

0232C — atque ad Urium, scilicet majus for atque atrium. Printed: "In eo quod superius dixerat: atque ad Urium, scilicet majus, non satis intelligitur utrum ipsum atrium a praemissa generali sententia excludatur." The sentence itself supplies the correct word four words later (ipsum atrium), and the lemma being recalled is et atrium majus (III Reg. 7:12), quoted correctly in the next line. Urium is not a Latin word; the run is split type under Pattern 10 and the whole run including the real word ad is carried: [sic: ad Urium]. A capitalized nonce-word of this kind is the sort of thing a later reader would take for a proper name if it were silently corrected away.

0232D — leguntur hahere cubitos for habere. Non-word; carried per Pattern 7, marked [sic: hahere]. Compositor's h for b.

0232D — decem et octavo cubitorum altitudinis and decimo octavo cubitos altitudinis and cubitos trigesimo quinto — three numerals in ordinal form where cardinals are required. Printed within four lines of one another: the lemma gives the columns' height as decem et octavo (for decem et octo, eighteen: III Reg. 7:15), Richard restates it as decimo octavo, and gives the Paralipomenon figure as trigesimo quinto (for triginta quinque, thirty-five: II Paral. 3:15). All three are real Latin words and are rendered literally, without markers, per Pattern 12 — "of ten and eighth cubits", "eighteenth cubits", "thirty-fifth cubits". These are load-bearing numbers and the whole passage exists to confront the 18-vs-35 discrepancy between Kings and Chronicles; the discrepancy is Richard's subject and is emphatically NOT reconciled here. The ordinal forms are a systematic habit of this compositor, not three independent slips.

0230D — tres ordines lapidum politorum quo quidem superficialium, tertius interjacentium. As printed, the enumeration loses its first member: quo quidem stands where duo quidem is required ("three rows of polished stones, two indeed of the surface ones, the third of those lying between"). quo is a real word; rendered literally, no marker, per Pattern 12. A count is lost — logged with the novam case at 0219D as one of the two places in this batch where the literal rendering destroys a number.

0231A — Expletis quae de templi aedificatione dicturus, erat, repetit. A comma before and after erat, breaking the relative clause. Both commas preserved (Pattern 8) and mirrored in the English clause structure.

0232A — junctura superiorum semper, occurrerit contra medium inferiorum. Stray comma after semper; preserved.

0232B — Sicut in fundamento, licet in superioribus, non potuerit omnes usquequaque esse ejusdem mensurae. Sicut with no answering clause, and nominative omnes where the accusative subject of esse is required. Rendered as closely as the printed words allow, unmended. Possibly Sic ut, possibly a dropped clause.

0233A — Nam epistylium idem sonat quod super. As printed, "epistylium means the same as 'above'" — which is false, and which the very next sentence corrects ("Inde epistylium, quod ponatur supra stylum"). The intended reading is epi, the Greek preposition, whose gloss is super. epistylium is a real word and the sentence construes, so it is rendered literally per Pattern 7 with no marker. The etymological argument is only recoverable because stylus and epistylium are kept in Latin in the English.

0233B — dicam prius quid mihi videtur inter secundum quod potui de hoc libro … colligere. inter secundum quod does not construe; inter has no object and secundum quod wants an antecedent (perhaps juxta id secundum quod). Rendered as closely as the printed words allow, unmended.

0231C — quorum unusquisque quindecim columnas habebat against quadraginta et quinque columnis and columnas decem et quinque. Three ways of saying the same count within eight lines (fifteen per row, forty-five in all, fifteen per row again). All rendered as printed; the arithmetic is consistent and nothing is normalized to a single form.

Garbled, doubtful, or impossible readings in Migne's plate for the second half of the work (cols 0233C–0255), logged here. Chunks 0000–0007 were translated concurrently by another agent and carry their own file; the two are to be merged.

The English renders what the plate prints, per translation-style.md Pattern 7: a printed reading that is construable is rendered literally even where it is odd or contradicts the argument, and type that is not a word at all is carried into the English untranslated, in italics, marked [sic: …] (Pattern 12). Conjectures live here and only here. Note contents ([n: …]) pass through verbatim and untranslated, so citation errors are flagged here only.

This work makes the rule bite unusually hard, because almost every defect is a number. Cubits, boards, rows, orders, pomegranates and regnal years are the argument, not decoration, and Richard is himself doing textual criticism on them (0239C: quadringenta non quadraginta … codices multi vitio scriptorum depravati). Nothing below has been reconciled to Exodus, to 3–4 Kings, to Paralipomenon, or to Richard's own tables. Where two figures in the same sentence disagree, both are printed.

Terminology fixed across chunks 0008–0014

Richard's architectural vocabulary is technical and he distinguishes terms that English blurs; the renderings below are held constant through all seven chunks.

• capitellum → capital · epistylium → epistyle · stylus → shaft · basis → base. Richard's whole exposition at 0232D–0233A turns on columna being used now of the shaft alone, now of all four members, so columna is always column and never silently narrowed.
• retiaculum → net · versus → row · ordo → order. The distinction is load-bearing: 0240A opens Nota quod unus dicit versus, alius ordines vocat, and the whole hundred-vs-two reconciliation depends on the two words not collapsing. ordo is also order in the tables' column headings (0249), where Migne uses the same word for a table column.
• funiculus → cord · malogranatum → pomegranate · concatenatio → chaining · catenula → little chain.
• luter → laver · concha → basin · mare → sea. Richard notes at 0237A and 0237B that his two books use different words for one object; the English keeps them different too.
• humerulus → little shoulder (the diminutive is Scripture's, and Richard flags it as Scripture's: quae Scriptura humerulos vocat).
• interrasile → chased · caelatura → engraving · coronula → little crown · plecta → plait · intercolumna → intercolumn · ejectura → projection.
• tabula is plate in the bronze-work of 0237C–0239A and table in the chronological tables of 0247–0255. Same Latin word, two objects.
• conregnare → reign together; conregnantium → those reigning together, in the treatise heads and throughout. (See the congregantium crux below.)
• Biblical names keep Migne's Latin forms (Hieroboam/Jeroboam, Zamri, Ambri, Achab, Josaphat, Ochozias, Joathan, Ozias, Osee, Ezechias…). No conventional English equivalents are substituted, per the scripture policy; where the plate spells one king two ways, both spellings stand and are logged.

Chunk 0008 (cols 0233C–0237A)

• col. 0234C — Utrumque capitelium columnarum fusile erat. capitelium is not a word; capitellum is printed correctly a dozen times in the same paragraph. Carried untranslated: [sic: capitelium]. It sits inside an italic scripture lemma, so per Pattern 10 it takes no second italic marker.

• col. 0234D — Sed non multum curandum quid, quibus verbis dicatur, dummodo sensus auctoris veraciter intelligatur — the paragraph simply stops; the plate prints no terminal point. Pattern 8: none supplied in the English either.

• cols 0234D–0235A — alius autem ordo, septem similiter versuum esset supra centum malogranata, iu uno appensa. iu is not a word (a turned or broken n in in). Carried: hung [sic: iu] one.

• col. 0235A — Bene ergo dicitur: Perfecte duos ordines per circuitum retiaculorum singulorum. The lemma Richard is quoting, printed three lines above at 0234D, reads Et perfecit columnas, et duos ordines…. Perfecte is a real Latin word (the adverb "perfectly"), so under Pattern 12 it is rendered as ordinary English — "Perfectly two orders round about each of the nets" — and not carried. Intended: Perfecit.

• col. 0235A — inferior autem medietas erat super summitatem mologranatorum appensorum. Non-word for malogranatorum, which the same sentence spells correctly twice. Carried: [sic: mologranatorum].

• col. 0236A — primo supra stillim [stylum?] ponebatur epistylium. stillim is not a word, and Migne has already flagged it himself with his own bracketed conjecture [stylum?], which is reproduced verbatim like any other printed matter. Per Pattern 12 (the 10517 ( sic ) precedent) no [sic: …] of ours is added on top — ours marks defects Migne passed over, not defects he caught. The word is carried in italics.

• col. 0234B, II Paral. III lemma — quasi quaedam retiacula quae capita regerent super epistylia. regerent ("should rule") where the Vulgate, and Richard's own next sentence (Ecce de retiaculis dicit, quod tegerent capita), have tegerent ("should cover"). Real word: rendered literally as "should rule," so that the English preserves the very contrast Richard's gloss silently corrects.

• col. 0236D — Richard's own arithmetical crux, and he says so: Et resticula triginta cubitorum cingebat illud per circuitum… omnino impossibile videtur. A cord of thirty cubits round a sea of ten cubits' diameter (π ≈ 3). This is the author's problem, deliberately raised and deferred (haec quaestio melius in sequentibus solvitur), not a misprint; the figures stand.

• cols 0234A / 0235C — the second capital is quinque cubitorum altitudinis and the epistyle tres cubitos altitudinis, while 3 Reg. VI gives the capitals quatuor cubitorum, which Richard reads as breadth not height. Recorded because chunk 0010 (0240D) depends on 5 + 3 = 8 to reach the thirty-five-cubit column of Paralipomenon.

Chunk 0009 (cols 0237B–0240C)

• col. 0237A/B — Et resticula, tringinta cubitorum cingebat illud per circuitum. Non-word for triginta, spelled correctly at 0236D and again within this chunk. Carried inside the italic lemma without a second italic marker: [sic: tringinta].

• col. 0238B — In basibus autem columnare per latera, in tabulis illis quae per quadrum stabant erectae, coronulae erant insculptae. columnare is not a standard noun, and the phrase is very likely intercolumnare / intercolumna — the term Richard defines two sentences later at 0238D (Intercolumna vocat tabulas superius dictas), of exactly these plates. Since a neuter adjectival columnare is at least formable, it is rendered as ordinary English ("the columnar part along the sides") rather than carried; flagged here because the sense depends on it.

• col. 0238C — boundary case, worth a human's eye. adhaerentem sibi teretem quamdam concavitatem in modum fistulae factam cingebat atque conglutinabat: quam liberos luteris vocat, quia per ipsum sibi infusa transigebat. liberos is a real Latin form (accusative plural of liber) but agrees with nothing: the relative quam is feminine singular, and the resumptive per ipsum is masculine singular, so the printed word matches neither side of its own sentence. Rendering it as a real word ("the free-born of the laver") would be actively misleading, which is the case Pattern 12 flags as reopening its boundary. It is therefore carried untranslated — [sic: liberos] — and the conjecture recorded here: labrum or labium, the rim, which is what the whole passage describes and which per ipsum would then resume.

• cols 0238C–D — Hujus autem proceritas semis cubitum tenebat, dimidius vero cubitus in ventrem descendebat, unus autem cubitus exterius apparebat. The total length is half a cubit, yet half a cubit descends inside and a whole cubit shows outside: the parts exceed the whole by a cubit. Rendered as printed. Intended, on the parts, something like sesquicubitum (a cubit and a half) — which is also the measure the lemma at 0238C gives (unum cubitum et dimidium).

• col. 0239C — Sciendum pro certo quod hoc loco legendum sit, quadringenta non quadraginta, secundum quod habent codices multi vitio scriptorum depravati. Not a defect: Richard's own emendation of his Bible, kept intact, along with his appeal per Hebraeos, Hebraicam veritatem consulat.

Chunk 0010 (cols 0240D–0242A)

• Column sequence. This chunk runs [0240D] → [0241A] → [0242A], skipping bands 0241B–D, which appear in chunk 0011. That is the plate's own order: the third treatise opens in the 0241B–D band while the tail of the second treatise is continued at 0242A. Anchors are reproduced in the order the source has them, with no reordering or repair. Verify at merge that chunk 0011's [0241]/[0241B] opening is the twin of this gap.

• cols 0240D–0241A — in uno retiaculo erant centum malogranata in medio ipsius interposita, et centum in uno dependentia. The parallel two sentences later at 0242A reads malogranata in imis dependentia. Both are real words: rendered literally ("a hundred hanging in the one"). Intended: in imo.

• col. 0240D, Jer. LII lemma — malogranata centum retiaculis circumdabantur is syntactically ambiguous, and the Vulgate takes centum with malagranata ("all the pomegranates, a hundred, were surrounded with nets"). Richard reads it the other way — unum retiaculum ex centum retiaculis compositum fuit, and centum secundum considerationem aliam — so the English renders "surrounded by a hundred nets" in all three occurrences, following his construal, since the exposition is unintelligible on the other reading. Flagged because it is an interpretive choice, not a neutral one.

• cols 0240D–0241A — nonaginta et sex / nonaginta sex pomegranates against Richard's own hundred, on which he declares fateor me nescire. A genuine disagreement in the sources, left standing.

Chunk 0011 (cols 0241–0245A)

• col. 0241B — in libro Regum de diuturnitate vevaria successione regnantium. vevaria is not a word: a run-together of vel varia. Carried: [sic: vevaria]. This is in the first sentence of the treatise, the one addressed to Bernard.

• col. 0241D — Ex hoc quidem in Sciptura, quamvis ubique aperte non dicatur. Non-word for Scriptura. Carried: [sic: Sciptura].

• cols 0242A/B — flag loudly. Et notandum quod tempora congregantium quandoque attribuuntur soli patri… The word required by the entire treatise, and printed in both its heads and everywhere else in it, is conregnantium, "of those reigning together." congregantium is a real word ("of those gathering together") and is therefore rendered literally under Pattern 7 — but the sentence it heads is the thesis statement of the whole work, and as printed it appears to be about assembly rather than about joint reign. This is a one-letter defect at the hinge of the argument.

• cols 0242B / 0245D–0246A — Hieroboam at 0242B, Jeroboam from 0245D on, for the same king. Both spellings preserved per occurrence.

• col. 0242B — Ecce Hieroboam rex Israel dicitur regnasse viginti duos annos, et filius ejus Nadab annos duos. Si vero duo isti dicantur viginti sex annos regnando implevisse… 22 + 2 = 24. The printed 26 stands. (It may be deliberate — Richard's whole method is that overlapping years are counted twice — but nothing in the passage says so, and it is not reconciled here.)

• col. 0243A — fecit sibi regem omnis Israel Zamri. 3 Reg. XVI 16 has Amri, and Richard's own argument needs Ambri here (Zamri is the man the army is rebelling against). Printed reading kept.

• col. 0243B — Achab filius ipsius Zamri. Achab was son of Ambri; the first table at 0247 duly places Ambri. | xii between Zambri and Achab. Printed reading kept.

• Punctuation, Pattern 8 — three interrogatives close with a full stop and are left so: 0242D …regnavit sine patre. (a quomodo … nisi quia question), 0243D An forte idem annus erat filio secundus…, 0244D …dicitur fecisse anno ejusdem regis undecimo. No question marks supplied. The exclamation at 0241B (…recuperari non possit!) is likewise Migne's and is kept.

• col. 0243D — An forte idem annus erat filio secundus, qui erat patri decimus octavus, immediately after the same sentence has given anno decimo nono Josaphat. Both figures preserved.

Chunk 0012 (cols 0245B–0248B)

• col. 0245A/B — the loudest number in the batch. Jehu super Israel decem octo annis regnavit, cujus septimo Joas filius Ahaziae super Judam regnare incoepit. Eighteen is impossible on Richard's own reckoning one clause later: Idem autem annus fuit Joas vigesimus secundus, qui et Jehu ultimus — Joas's 22nd = Jehu's last, and Joas began in Jehu's 7th, so Jehu's last is his 28th (7 + 21). Richard's own first table at 0247 prints Jehu. | xxviii, and 4 Reg. X 36 has twenty-eight. Rendered as printed ("ten and eight years"); intended viginti octo.

• col. 0245B — Joas vero filius Jehu vigesimo tertio anno Joas regnare incoepit, et decem septem annis regnavit. The son of Jehu who began in the twenty-third year of Joas of Juda and reigned seventeen years is Joachas (4 Reg. XIII 1) — as Richard's very next sentences show, twice (Joas filius Joachas, Mortuo autem Joachas filius ejus Joas). Printed Joas kept, producing a sentence in which Joas is his own father.

• col. 0245D — Azarias qui Zozias filius Amasiae. Ozias everywhere else, including both tables (0247 Ozias. | Lii; 0249 i Ozias qui et Azarias). Rendered as a name, as printed.

• col. 0245D — vigesimo septimo anno Jerobam regis Israel. Jerobam is neither a word nor a name-form used anywhere else in the work; carried, [sic: Jerobam]. And in the same sentence, the quotation adduced to prove the point reads Anno vigesimo, inquit, Jeroboam regis Israel — twentieth, not twenty-seventh — while the inference drawn immediately after (Inter primum autem et vigesimum septimum annum invenies annos viginti quinque) requires twenty-seventh. All three figures preserved as printed.

• col. 0246A — Azarias quinquagesimis duobus annis regnavit. An ordinal plural where a cardinal is wanted (quinquaginta duobus). Real form, rendered as ordinary English ("fifty and two years").

• cols 0246C and 0247A/0248A — Migne's own sic, twice. Joathan filius Oziae decimus sextus ( sic ) annos regnasse describitur and alioquin decimus octavus ( sic ) annis potius quam novem regnasse diceretur. In both, an ordinal stands where a cardinal is required, and the nineteenth-century editor caught it. His ( sic ) is reproduced verbatim; no [sic: …] of ours is added, per Pattern 12. The English keeps the ordinal ("sixteenth years", "eighteenth years") so that his mark still has something to point at.

• cols 0246D–0247A — Achab filius Joathan, seven occurrences. The son of Joathan is Achaz (4 Reg. XVI 1), and Richard's own tables print him so at 0247 (Achaz. | xvi) and 0255 (Achaz.). Every prose occurrence is kept as Achab, which in this treatise collides with Achab of Israel and makes the final paragraph read as though one king were two. Not emended.

• cols 0247A / 0248A — the anchor sequence is not monotonic: the first paragraph closes …in regno [0248A] ascriberentur, the italic head then carries [0247], and the second paragraph runs [0247A] [0247B] [0248A] [0248B]. The two tables are set between columns in Migne, so the bands interleave. Reproduced exactly, in order, in place.

Chunks 0013–0014 — the two tables (cols 0247, 0249–0255)

Band-less anchors. The table pages are printed without letter bands: [0247] in 0013, [0249] [0251] [0253] [0255] in 0014 (odd columns only, the tables occupying the full page). Carried in exactly that form. Chunk 0011 also opens with a band-less [0241], and 0012's head carries [0247].

Structure. Migne's rows are rendered 1:1 (35 delimited rows in 0013, 256 in 0014); only the italic head and the italic column-headings are translated, plus the handful of Latin words inside cells (solus, cum patre, cum fratre, in parte, in toto, mortuo patre, mensibus, diebus). Numerals are kept exactly as printed, capitalization and stray points included. No cell was filled in, moved, or aligned: where a row has three cells in Migne it has three in the English, even where the columns are plainly out of register.

Chunk 0013 (col 0247)

• Head — contrarietatem temporum regum super conregnantium Judam et Israel: super stands before conregnantium instead of after it. Cf. the second table's head at 0249, regum conregnantium super Judam et super Israel, which is right. Rendered by sense; the dislocation is Migne's.
• Ozias. | Lii — capital L in an otherwise lower-case numeral column. Kept.
• Phaceia. | li — fifty-one years for Phaceia, who reigned two (4 Reg. XV 23) and is given i Phaceia … ii in the second table at 0255. Kept.
• Achaz. | xvi — the table names this king Achaz; the prose of chunk 0012 calls him Achab throughout. Both kept.

Chunk 0014 (cols 0249–0255)

• xxiii | iii | ii. Idem solus | i. Nadab cum patre Baasa — Nadab's father was Jeroboam; Baasa was the man who killed him and took the throne (Richard says so himself at 0242D, Baasa qui Nadab occidit). Printed as is.
• xlviii | xxviii | i Ambri, in parte. | i Zambrid. vii. — a run-together of Zambri and d. (= diebus), cf. the first table's Zambri. | diebus vii. Carried whole per Pattern 10: [sic: Zambrid.].
• lx | xl | iii | i. Achas cum patre. — Achas where the same orders elsewhere have Achab (and the prose has Achab for Achaz). Kept.
• clxxviii | xii Ozias, mortuo | xxv / clxxix | xiii patre | xxvi — the ablative absolute mortuo patre is split across two table rows in Migne. The English splits at the same point ("on the death" / "of his father") so that no row boundary moves.
• Abbreviations. S. (rows xxxiii S., xxi. S., viii S.) is left unexpanded, since Migne abbreviates; Solus. and solus. are translated where he writes them out. Men. (= mensibus) is rendered "months"/"month", as the first table writes mensibus out.
• Defective or out-of-sequence numerals, all kept as printed:
• viii for xviii in the first order (the row carrying i. Abia.)
• lxxxxiii for lxxxviii
• cii printed twice in succession (the rows after ci)
• cxxx for cxxxv
• cxcv | xxxix where the second order's sequence requires xxix
• ccxviv for ccxxxiv — not a well-formed Roman numeral at all
• stray full stops on numerals: c., clix., xv. (twice), x., ix.
• Repetitions and back-steps in the regnal columns, kept:
• lxxiii | xii followed by lxxiv | xii (second order)
• cxiii | xvii followed by cxiv | xvii (second order)
• lxxxvi | xxv followed by lxxxvii | iv (second order jumps backward)
• cxlvi | ix | x followed by cxlvii | x | ix (fourth order back-steps)
• ccxxxviii | xx | iv — xx in the Juda column where the running sequence requires iv
• clv | xviii | iv | ii — the fourth order restarts at ii with no new king named
• Osee is named twice in the fourth order (ccxxxvii | iii | Osee and ccxlvi | xii | Osee), which is consistent with the prose crux of 0012 (Regnavit itaque Osee anno Achab quarto. Quid ergo est quod in sequentibus dicitur regnare ejusdem Achab anno duodecimo?) — the table records the same double accession Richard is trying to explain. Not a defect; noted because it looks like one.

PL 196 polarity sweep (2026-07-28)

Sites found by the volume-wide sweep for sense-reversing defects (dropped or intrusive negations, and single letters that invert a claim). All are Migne's; all are rendered literally per Pattern 7. Entries already discussed above are repeated here so the class can be read in one place.

### 0211D — letter-inversion (high confidence) Printed: Sed qui assidue producit novos fructus ad recreationem hominis exterioris, nunquam credendus est nullos de caetero scientiae profectus impertiri, ad innovandos sensus hominis interioris?

Conjecture: nunquid credendus est nullos ... impertiri?

Printed as a question with 'nunquam' plus 'nullos' the sentence double-negates into incoherence and, taken at face value, denies the very progress Richard is defending. The following sentence answers the rhetorical question as if it had been 'nunquid': 'Ubi ergo illud est, quaeso, pertransibunt plurimi, et abundabit scientia?' The whole paragraph is written against 'illos qui nil acceptant nisi quod ab antiquissimis patribus acceperunt'; the argument requires 'Is he to be believed to impart NO further advances of knowledge?' expecting the answer no. 'nunquam' for 'nunquid' is a confirmed defect class.

### 0212B — added-negative (medium confidence) Printed: nunquam credendus est nullos de caetero scientiae profectus impertiri, ad innovandos sensus hominis interioris?

Conjecture: nunquam → nunquid (a double negative, nunquam + nullos, stacks to assert the opposite of the prologue's whole thesis that God still grants new understanding)

The entire prologue argues FOR fresh understanding beyond the Fathers ('Non est ergo magnum, vel mirum si in uno aliquo aliquid possumus addere'); as printed, the rhetorical question doubly negates itself into a flat denial that any new advance is ever to be believed, contradicting the very case Richard is making in the sentences immediately before and after.

### 0214D — letter-inversion (medium confidence) Printed: Ecce dicitur quod ansulas habeant in omni parte; nec tamen dicitur quod sint in summitatibus, sed tantum quod habeant in utroque latere.

Conjecture: nec tamen dicitur quot sint in summitatibus, sed tantum quot habeant in utroque latere

The lemma Richard has just quoted one line above says precisely that the loops ARE on the tops: 'Ansulas hyacinthinas in lateribus, ac summitatibus facies cortinarum.' The printed gloss denies what its own quoted plate asserts. The missing datum is the NUMBER, and the gloss immediately supplies it for the sides only: 'Quinquagenas, inquit, ansulas cortina habebit in utraque parte.' Richard's conclusion also depends on the tops being conceded and only their measure left silent: 'Sed tacendo de uno, quod determinat, de alio dat auctor intelligere, ut secundum modum laterum fiant et ansulae summitatum.'

### 0221B — letter-inversion (high confidence) Printed: Nec arula per aditum dictorum ostiorum inferebatur, et super craticulae bases, et super ostium limen imponebatur; terra tamen prius tota illa altaris inferior medietas usque ad situm craticulae implebatur.

Conjecture: Nam arula per aditum dictorum ostiorum inferebatur

The very next sentence presupposes the approach was just affirmed as the way things were brought in: 'Per hunc aditum ligna ingerebantur, unde ignis nutriebatur, per eumdem aditum cineres extrahebantur.' The concessive 'terra tamen prius ... implebatur' ('yet first the lower half was filled with earth') only works after an affirmative statement that the hearth WAS brought in and laid down. And at 0222A Richard argues the arula did rest on the bases: 'Recte ergo de arula dicitur, quod subter craticulam esset, quamvis super bases jaceret.' The printed 'Nec' denies exactly what he goes on to prove.

### 0229B — dropped-non (high confidence) Printed: Ex his facile est perpendere, quod oportuit esse, quod vectes exciperet ab interiori parte. Sed quantum quomodo id fieri potuit facile perpenditur, legentium conjecturae relinquitur.

Conjecture: Sed quantum [ad hoc] quomodo id fieri potuit, non facile perpenditur, legentium conjecturae relinquitur.

The sentence contradicts itself and its own neighbour: something 'easily weighed' would not be 'left to the conjecture of readers' (legentium conjecturae relinquitur). The adversative 'Sed' is set against the immediately preceding 'Ex his facile est perpendere' — the contrast requires that this second point be NOT easily weighed. Without 'non' the 'Sed' cancels itself.

### 0230A — dropped-non (high confidence) Printed: Superfluum vero videtur quod expressio materiae reperitur.

Conjecture: Non superfluum ("it does not seem superfluous")

Richard immediately goes on to build an argument out of exactly this specification — the material named twice, olive wood at the oracle against fir at the temple entrance — treating it as significant, not superfluous; as printed he first dismisses as pointless the very detail he then argues from.

### 0244D — letter-inversion (high confidence) Printed: Non enim viginti annis pater regnare potuit, si filius ejus qui regnavit pro eo octavo patris sui anno, vel modico post, ab Jehu occisus fuit.

Conjecture: filius for pater — "Non enim viginti annis FILIUS regnare potuit"

The opinion being refuted is stated in the immediately preceding clause as a claim about the SON, not the father: "quam falsa sit quorumdam opinio ... qui ... dicunt FILIUM JOSAPHAT viginti annis regnasse." Richard's refutation must therefore deny twenty years to the son. Printed as "pater," the sentence denies twenty years to Josaphat — whom Richard himself has just credited with twenty-five, at 0244A: "Josaphat qui viginti quinque annis regnavit." So the plate makes Richard assert the opposite of what he wrote a column earlier. The conditional clause requires "filius" as its subject too: "filius ejus qui regnavit pro eo ... ab Jehu occisus" is Ahazias, son of Joram of Juda, and "octavo patris sui anno" is Joram's eighth year — exactly what the next paragraph states, 0244C: "idem annus Joram regi Judae fuit octavus" and "Ahazias et rex Israel ab Jehu interfectus est." With "pater" as subject the possessive chain collapses; with "filius" it is exact. The gloss supplies the right word one clause earlier.