# Codex H Evidence Dossier

Date: 2026-06-03

## Question

Does the newly recovered/digitized Codex H material change our understanding of the overlapped New Testament text when compared with the accepted current Greek text? Classify divergences as tangential, minor, major, or critical.

## Bottom Line

No critical textual finding was found in this audit.

The recovered/digitized material is important for manuscript history, page recovery, scribal habits, corrections, marginal apparatus, and early chapter-list/paratactic reading systems. It does not reveal an unknown Pauline passage and does not require a change to core NT theology or to the current critical Greek text on the evidence available here.

There are real material variants. The strongest are known-style additions/omissions and wording differences in places such as 1 Cor 10:28, Col 2:2, Col 2:11, Heb 1:3, Heb 2:9, Heb 4:15, Heb 10:34, and several smaller Pauline variants. These are major in a textual-apparatus sense, but not critical in the sense of overturning doctrine or the accepted Greek text.

## Sources

Primary manuscript data:
- Glasgow Research Data deposit: `https://researchdata.gla.ac.uk/2105/`
- DOI: `10.5525/gla.researchdata.2105`
- Local file: `outputs/audits/2026-06-03-codex-h/source/2105_data.zip`
- Zip SHA256: `c2c04b62d2f88171fd1805a5d4b41cb14da8138a7b0305acd30e118b93165d1f`
- Extracted TEI/XML: `outputs/audits/2026-06-03-codex-h/source/combined-xml.xml`
- XML SHA256: `096b2bc5800033aa70fcadefb1e844a236f6267820c3dd4820ec394e1e2e2be3`
- License shown in the deposit and XML: CC BY 4.0.

Public project/context sources:
- Codex H digital edition: `https://codexh.arts.gla.ac.uk/`
- Glasgow announcement, 2026-04-24: `https://www.gla.ac.uk/news/headline_1263245_en.html`
  - Announces 42 recovered pages.
  - States the manuscript is a sixth-century copy of Paul's letters.
  - States the recovered text contains known portions of Paul's letters.
  - Emphasizes chapter lists, scribal interaction/correction, and medieval reuse.

Mechanical comparator:
- SBLGNT source archive: `outputs/audits/2026-06-03-codex-h/source/SBLGNT-master.zip`
- SBLGNT zip SHA256: `329253319cbe4cb0873a2289c92dcc97e9b4b9307c9f2bbc156027da66137991`
- SBLGNT is CC BY 4.0 and open in electronic form.
- SBLGNT is not NA28/UBS5. It is useful for machine collation, but the report must not pretend it is the accepted standard. SBLGNT itself says the usual standard text is the Nestle-Aland / UBS tradition and that SBLGNT differs from that standard in more than 540 variation units.

Accepted Greek baseline:
- Scholarly shorthand here means NA28/UBS5. I did not mechanically ingest the proprietary NA28/UBS5 apparatus.
- Where the TEI itself uses `supplied source="na28"`, I treat that token as editorial restoration, not independent manuscript evidence.
- Where SBLGNT and NA28/UBS5 differ, SBLGNT-only collation can understate or misstate a Codex H divergence. Heb 1:3 is the clearest example in this dossier.

## Extraction Summary

Derived artifacts:
- `outputs/audits/2026-06-03-codex-h/derived/codex_h_comparison.json`
- `outputs/audits/2026-06-03-codex-h/derived/codex_h_comparison.csv`
- `outputs/audits/2026-06-03-codex-h/derived/codex_h_summary.md`

Observed counts:
- TEI units/pages: 125.
- Unique folio labels: 125.
- Verse references extracted: 338.
- Correction/app readings extracted: 64.
- Books covered: 1 Corinthians, 2 Corinthians, Galatians, Colossians, 1 Thessalonians, 1 Timothy, 2 Timothy, Titus, Philemon, Hebrews.
- Book counts: `1Cor 28`, `2Cor 56`, `Gal 42`, `Col 38`, `1Thess 12`, `1Tim 48`, `2Tim 16`, `Titus 26`, `Phlm 1`, `Heb 71`.
- Markup flags across reconstructed words: `unclear 1042`, `nomSac 206`, `expansion 88`, `supplied 689`, `gap 4`.
- Automated SBLGNT status counts: `exact_reconstructed_match_sbl 55`, `near_match 7`, `variant_candidate 257`, `too_fragmentary_for_full_collation 19`.

The 257 automated candidates are intentionally noisy. They include folio starts/ends, damaged lines, supplied NA28 reconstructions, nomina sacra, movable nu, spelling, and punctuation. The summary file enumerates those raw diff operations; this dossier filters them into evidentiary classes.

## Classification Rules

Tangential:
- Nomina sacra (`κυ`, `θυ`, `χς`, `ις`, etc.) versus expanded forms.
- Accent, breathing, punctuation, capitalization, apostrophe, line break, and word-split differences.
- Itacistic spelling where no lexical/semantic distinction is introduced.
- Paratext, folio titles, chapter numerals/lists, and Euthalian-style apparatus unless they alter the biblical wording.
- Any token supplied from NA28 or another editor without visible manuscript support.

Minor:
- Function-word, article, conjunction, movable-nu, tense/ending, or word-order differences with little or no change in translation.
- Formulaic expansions that do not alter referent or sense.
- Orthographic/grammatical slips that are recoverable from context.

Major:
- Lexical substitution, meaningful pronoun change, addition/omission of a phrase, or a known textual-apparatus variant that would affect translation or exegesis.
- Major does not mean "theologically dangerous"; it means the variant deserves apparatus/commentary attention.

Critical:
- A new or recovered reading that materially changes core doctrine, introduces an unknown NT passage, defeats the accepted critical text in a significant place, or would require changing Biblex's translation base policy.

## Attested/Material Divergences

| Ref | Class | Codex H evidence | Comparator issue | Judgment |
| --- | --- | --- | --- | --- |
| 1Cor 10:23 | Minor | Adds `μοι` twice in "all things are lawful" formulation. | SBLGNT omits. | Translation nuance only; no doctrinal effect. |
| 1Cor 10:28 | Major | App/correction adds the Ps 24:1-style quotation after "conscience": "for the earth is the Lord's..." | SBLGNT/critical text omit in v28. | Known Byzantine/TR-style expansion; major for apparatus, not critical. |
| 1Cor 10:31 | Minor | Repeated `-ται` forms where SBLGNT has imperative `-τε` forms. | SBLGNT has direct imperatives. | Likely scribal/orthographic or grammatical drift; sense remains transparent. |
| 2Cor 4:4 | Major | Reads `καταυγάσαι` where SBLGNT has `αὐγάσαι`. | Prefix changes the verb form but not the broad "shine/illuminate" sense. | Lexical variant worth noting; not theologically critical. |
| 2Cor 11:3 | Major candidate | Visible Codex H line has "simplicity ... toward Christ" without SBLGNT's "and purity." | SBLGNT includes the extra phrase. | Material if the visible line is complete at that point; confidence medium because the verse has multiple unclear/supplied flags. |
| Gal 1:4 | Minor | Uses `ἀπό`/word order and `θέλησιν` where SBLGNT has `ἐκ ... τὸ θέλημα`. | Same broad meaning: deliverance from the present evil age according to God's will. | Minor-to-moderate phrasing difference; no change to theology. |
| Gal 1:6 | Major candidate | Reads `ἡμᾶς` where SBLGNT has `ὑμᾶς` in the calling clause. | Pronoun changes "us" vs "you." | Potentially meaningful, but surrounded by orthographic/unclear flags; not critical. |
| Gal 1:11 | Minor | `δέ` vs `γάρ`. | Conjunction shift. | Minor discourse difference. |
| Gal 5:1 | Minor/Major | Codex order has Christ before "set us free"; SBLGNT's order differs. | The raw diff overstates the issue because `χς` is a nomSac in a different position. | Meaning remains "Christ set us free"; not critical. |
| Col 1:24 | Major | Corrected reading adds `μου` after sufferings and reads `ἀναπληρῶ` rather than SBLGNT `ἀνταναπληρῶ`. | SBLGNT lacks `μου` and has the compound verb. | Major lexical/possessive variant; no doctrinal reversal. |
| Col 2:2 | Major | Ends "mystery of God" with nomSac `θυ`; no visible `Χριστοῦ` after it. | SBLGNT has "of God, Christ." | Major Christological textual variant; important for apparatus, not a new doctrine. |
| Col 2:3 | Major | Reads "the knowledge" (`τῆς ἐπιγνώσεως`) where SBLGNT has "knowledge" (`γνώσεως`). | Lexical/intensifying shift. | Major enough for translation notes; not critical. |
| Col 2:11 | Major | App/correction adds "of sins"; Codex lacks SBLGNT's final "in the circumcision of Christ" phrase in the extracted line. | SBLGNT has "body of flesh" followed by "circumcision of Christ." | Major known-style expansion/omission cluster; no core doctrine is created or erased. |
| Col 3:4 | Major | Reads "our life" where SBLGNT has "your life"; also differs in final verb ending. | Pronoun changes the direct possessive. | Meaningful pronoun variant; not critical. |
| 1Tim 1:17 | Major | App/correction adds `σοφῷ` before God. | SBLGNT lacks "wise." | Doxological expansion known from later tradition; major/minor border, not critical. |
| 1Tim 2:7 | Major | Adds `ἐν Χριστῷ` after "I speak truth." | SBLGNT omits. | Formulaic Pauline expansion; major as phrase addition, not critical. |
| 1Tim 6:10 | Major | `ποικίλαις` ("various") vs SBLGNT `πολλαῖς` ("many") for pains. | Lexical difference. | Affects nuance only. |
| Titus 2:7 | Minor/Major | Adds `ἁγνίαν` ("purity") in the teaching list. | SBLGNT omits. | Meaningful list expansion; not critical. |
| Titus 3:15 | Tangential/Minor | Adds final `ἀμήν`. | SBLGNT omits. | Liturgical ending; not a substantive textual claim. |
| Heb 1:3 | Major | App/corrections include `δι' ἑαυτοῦ`; Codex also has `ἡμῶν` after sins and `τοῦ θρόνου` with majesty. | SBLGNT includes `δι' αὑτοῦ`; NA28-style text normally omits it. SBLGNT omits `ἡμῶν` and `τοῦ θρόνου`. | Major against NA28/UBS5 and partly hidden by SBLGNT; not critical because the purification and exaltation theology stands either way. |
| Heb 2:9 | Major | Reads grace-of-God wording (`χάριτι θεοῦ`) against the famous "apart from God" reading in SBLGNT. | SBLGNT has `χωρὶς θεοῦ`; NA28/UBS5 main text is commonly `χάριτι θεοῦ`. | Important place: Codex H supports the mainstream reading rather than a disruptive alternative. Major historically, not critical. |
| Heb 2:16 | Minor | Does not repeat the final "takes hold" verb in the extracted line. | SBLGNT repeats it. | Translation effect small; possibly stylistic/line-boundary. |
| Heb 4:15 | Major candidate | Extracted Codex line stops after "sympathize with our weaknesses"; SBLGNT continues with "tested in all ways... without sin." | No explicit app entry; surrounding Hebrews folios skip beginnings/ends of verses. | Major if true, but likely fragmentary/page-boundary evidence rather than a secure omission. Needs apparatus confirmation. |
| Heb 10:34 | Major | App/correction adds `ἐν οὐρανοῖς` ("in heaven") to the possession phrase. | SBLGNT omits. | Meaningful eschatological-location expansion; major but not critical. |
| Heb 10:38 | Major | App shows `μου` struck/deleted; extracted Codex lacks final `ἐν αὐτῷ`. | SBLGNT has "my righteous one" and final "in him." | Major in apparatus sense; doctrinally low-risk. |
| Heb 13:25 | Tangential/Minor | Adds final `ἀμήν`. | SBLGNT omits. | Liturgical ending; not substantive. |

## Fragmentary Or Non-Evidentiary Major-Looking Rows

These raw SBLGNT diffs should not be reported as secure Codex H omissions without full apparatus confirmation:

| Ref | Why It Looks Major | Why I Downgrade It |
| --- | --- | --- |
| 1Cor 11:6, 11:8, 11:9 | SBLGNT has opening words not in Codex H extraction. | Folio/line starts create automated insertions; not secure omissions. |
| 2Cor 4:7 | SBLGNT has a long continuation after an abbreviated visible word. | Fragmentary/damaged line; not enough visible Codex H text. |
| 2Cor 11:9, 11:14, 11:16, 11:23 | Multiple phrase insertions/deletions in the diff. | High unclear/supplied counts and line-boundary movement; likely extraction noise unless checked against full apparatus. |
| Gal 2:7-9, 2:20 | Long missing or displaced clauses. | Folio fragments and supplied text make the raw sequence unreliable as a variant claim. |
| 1Tim 3:2 | Looks like omission of "hospitable, able to teach." | Verse terminates after `κόσμιον`; likely physical/extraction cutoff, not a secure textual omission. |
| Titus 2:5 | Looks like omission of "submitting to their own husbands..." | The TEI has an explicit `gap reason="illegible" unit="line" extent="3"` between the visible list and the final verb. |
| Titus 2:8 | Looks like omission of the final purpose clause. | Record is already marked `too_fragmentary_for_full_collation`, and the TEI contains a gap. |
| Heb 2:10, 2:14, 3:1, 3:5, 3:10, 3:12, 3:15, 4:2, 4:16, 12:9, 12:18, 13:21 | Long SBLGNT phrases absent from extracted Codex line. | These align with visible fragment boundaries, gaps, or page truncation patterns. Treat as non-evidentiary until checked against the full edition/apparatus. |

## Tangential Divergence Inventory

Observed repeatedly across the extraction:
- Nomina sacra: `κυ`, `θυ`, `θς`, `χς`, `χρ`, `ις`, `πνι`, `πνα`, etc. The extraction counted 206 nomina-sacra flags. These are manuscript practice, not textual disagreement by themselves.
- Orthographic and itacistic forms: e.g., `υμειν`/`υμιν`, `εστι`/`εστιν`, `οφιλει`/`οφειλει`, `απιθειας`/`απειθειας`.
- Word division/line-break artifacts: e.g., `πλη ρωμα` vs `πληρωμα`, `με ν` vs `μεν`.
- Punctuation and paratext: punctuation marks, folio/page titles, chapter numerals, lectional headings, and chapter lists.
- Editorial supply: 689 supplied flags in the derived extraction; `source="na28"` supplied text is restoration, not Codex H evidence.

## Critical Findings Check

Critical standard applied:
- Unknown NT passage? No.
- New Pauline doctrine? No.
- Loss/removal of a core Christological or soteriological claim not attested elsewhere? No.
- Compels replacement of NA28/UBS5 as Biblex's current-Greek baseline? No.
- Shows the "42 lost pages" contain new canonical material? No; Glasgow says recovered text contains known portions of Paul's letters.

Result: no critical findings.

## Decision/Tradeoff Notes

1. I count phrase additions like 1Cor 10:28, Col 2:11, Heb 10:34, and Heb 1:3 as "major" even when they are familiar later-tradition readings because they affect translation/apparatus. I do not count them as "critical" because they do not overturn the accepted text or doctrine.
2. I downgrade long apparent omissions at damaged folio boundaries because the extraction has many supplied/unclear/gap flags and the TEI often records physical gaps. Treating those as deliberate Codex H omissions would overclaim.
3. SBLGNT is an open comparator, not the accepted critical text. It catches most overlap quickly but can misclassify places where SBLGNT differs from NA28/UBS5.
4. A final scholarly verdict would require licensed NA28/UBS5/ECM apparatus collation and the forthcoming print Codex H edition. This dossier is sufficient for product/research triage, not a replacement for a peer-reviewed apparatus.

## Subreview Questions

Reviewers should focus on:
1. Did this dossier miss any major or critical variant among the extracted rows?
2. Did it incorrectly downgrade any "fragmentary/non-evidentiary" row that should be treated as a secure major omission?
3. Did it overstate any Codex H evidence by counting supplied NA28/editorial text as manuscript evidence?
4. Did it mishandle SBLGNT versus NA28/UBS5, especially Heb 1:3 and Heb 2:9?
5. Is the conclusion "major variants but no critical findings" supported?