Published rights research for a work the Archive cites but does not reproduce.

Rights dossier — Sermones (St. Augustine), citation-only index

Researched: 2026-08-09. Queue item sermones-st-augustine (apps/web/content/cited-sources/manifest.json, category PATRISTICS, citedByCount: 17, citesCount: 0).

Verdict: CLEAR on rights, PARTIAL on coverage — 6 of 17 (35%) of our cited sermons fall inside a public-domain English collection. The other 11 exist in English only in the modern, fully-copyrighted New City Press (Hill/Rotelle) translation. This is a citation-only address index (LS-KG-02 §1/§7 — no source text stored), so nothing here blocks the current entry; the finding matters if/when this work is ever considered for translation-linking or text ingestion.

What we actually cite

The 17 connections are recorded as nodes in apps/api/ingestion/vault/imported/catechism-sources.json with workSlug: "sermones-st-augustine". Each heading is a locus in the form sermon, chapter, section (standard Maurist/Migne PL numbering — plain numbers, no Denis/Dolbeau/Morin/Guelf./Frangipane prefix, so these are definitely the classic 1–396 numbered corpus, not one of the 20th-century "newly discovered" sermon collections that use a separate numbering track):

| Sermon # | Locus (heading) | In PD range (51–147)? | | --: | :-- | :-- | | 18 | 18, 4, 4 | No | | 33 | 33, 2, 2 | No | | 43 | 43, 7, 9 | No | | 52 | 52, 6, 16 | Yes | | 56 | 56, 6, 9 | Yes | | 57 | 57, 7, 7 | Yes | | 58 | 58, 11, 13 | Yes | | 78 | 78, 6 | Yes | | 96 | 96, 7, 9 | Yes | | 169 | 169, 11, 13 | No | | 186 | 186, 1 | No | | 213 | 213, 8 | No | | 214 | 214, 11 | No | | 241 | 241, 2 | No | | 267 | 267, 4 | No | | 272 | 272 | No | | 298 | 298, 4–5 | No |

The numbering trap (worth naming explicitly)

Augustine's Sermones are cited today by Maurist/Migne (PL) numbering — that's what our loci use. The 19th-century English translation (Oxford Library of the Fathers, 1844/45, reprinted with revisions as NPNF1 vol. 6, Buffalo 1888) is not a complete set — it is a selection, titled "Sermons on Selected Lessons of the New Testament," and it renumbers the selection sequentially starting at 1, with the Maurist/PL number given in a bracketed gloss per sermon (e.g. "Sermon 6 on the New Testament [LVI. Ben.]" = Maurist Sermon 56).

I verified the mapping is linear and contiguous, with no gaps, by checking five points across the sequence on the New Advent mirror of the NPNF text (which preserves this bracketed numbering):

| NPNF sequence # | Bracketed Maurist/Ben. # | URL checked | | --: | --: | :-- | | 1 | LI (51) | newadvent.org/fathers/160301.htm | | 6 | LVI (56) | newadvent.org/fathers/160306.htm | | 28 | LXXVIII (78) | newadvent.org/fathers/160328.htm | | 46 | XCVI (96) | newadvent.org/fathers/160346.htm | | 97 | CXLVII (147) | newadvent.org/fathers/160397.htm |

Formula: NPNF sequence # = Maurist # − 50, for Maurist #51–147 inclusive (97 sermons, no skips across the five checkpoints spanning the full range). So: the PD selection covers Maurist Sermons 51 through 147, and nothing outside that band — not sermons 1–50 (mostly Old Testament lessons) and not 148 onward (the de verbis Apostoli/Pauline-lesson sermons continuing past 147, nor any of the de tempore liturgical-season sermons 184–272-ish, nor the de sanctis/various-subjects sermons 273 on).

Candidate editions verified

1. NPNF1 vol. 6 — the recommended PD edition for the 6 in-range sermons

A Select Library of the Nicene and Post-Nicene Fathers of the Christian Church, First Series, Vol. VI — St. Augustin: Sermon on the Mount; Harmony of the Gospels; Homilies on the Gospels (the "Homilies on the Gospels" is the volume's running head for the "Sermons on Selected Lessons of the New Testament" selection). Translated by Rev. R. G. MacMullen, edited by Philip Schaff, preface by E. B. Pusey.

Single volume, contains all 97 sermons (Maurist 51–147) in one book — the simplest citation target for the 6 covered sermons.

2. Library of the Fathers, vols. XVI & XX — the original (pre-NPNF) PD translation

A Library of Fathers of the Holy Catholic Church, Anterior to the Division of the East and West (Oxford, J. H. Parker; general editors Pusey, Keble, Newman). Sermons on Selected Lessons of the New Testament by S. Augustine, Bishop of Hippo, translated by R. G. MacMullen — the same translation NPNF1-06 later reprinted/revised.

| Vol. | Contents | Year | IA id | | :-- | :-- | :-- | :-- | | XVI (Sermons vol. I) | S. Matthew, S. Mark, S. Luke | 1844 | 16ALibraryOfFathersOfTheHolyCatholicV16 | | XX (Sermons vol. II) | S. John, Acts, Romans, 1 Cor., Galatians, Ephesians, Philippians, 1 Thess., 1 Timothy, Titus, James, 1 John | 1845 | 20ALibraryOfFathersOfTheHolyCatholicV20 |

Both pre-1930 (indeed pre-1923 many times over), Oxford-published — PD by any measure. I did not verify the exact Maurist-number split point between vol. XVI and vol. XX from the scans (the AI-summarized title-page read on 16ALibraryOfFathersOfTheHolyCatholicV16 was unreliable — it returned a garbled "1839/1844" date pair, likely conflating the whole-series first-volume date with this volume's own date). This doesn't affect the rights verdict: LF is PD regardless of the split, and NPNF1 vol. 6 (single volume, confirmed range) is the cleaner citation target anyway. If LF is ever preferred over NPNF (e.g. for a translation closer to the original, pre-Schaff-revision wording), the split point should be pinned down from the actual contents pages before citing a specific volume per sermon.

3. New City Press (Hill/Rotelle) — the complete set, NOT public domain

The Works of Saint Augustine: A Translation for the 21st Century, Part III — Sermons, 11 vols., trans. Edmund Hill OP, ed. John E. Rotelle, New City Press, 1990–1997. This is the only complete English translation of the Sermones (all ~590 known sermons including the 20th-century Dolbeau/Denis/Morin discoveries) and covers all 11 of our not-in-NPNF sermons. It is fully in copyright (translation published 1990s; translator/editor deaths well within copyright term). None of it may be used.

Two traps encountered and worth naming in-dossier

earlychurchtexts.com mislabels its Sermon 272 text as "from the NPNF series." I checked this directly (earlychurchtexts.com/public/augustine_sermon_272_eucharist.htm, which states "The English translation below is from the NPNF series"). This is false — Sermon 272 (Maurist #272) is far outside NPNF1 vol. 6's confirmed 51–147 range. The text on that page is almost certainly the Hill/New City Press translation (or some other in-copyright rendering), mislabeled. Do not trust "NPNF" attributions for any sermon numbered above 147 without checking the actual NPNF volume.
wesleyscholar.com hosts freely-downloadable PDFs named Augustine-Sermons-1-19.pdf, Augustine-Sermons-148-183.pdf, Augustine-Sermons-184-229.pdf, Augustine-Sermons-341-400.pdf, etc. These filename ranges match the New City Press 11-volume division exactly (e.g. Goodreads lists "Sermons 184-229Z (Vol. III/6)" as a discrete New City Press volume; archive.org/stream/saint-augustine-sermons/... — an "11 vols" scan — contains sermon 272B, confirming it's the same Hill set). Free and PDF-shaped does not mean public domain — these are pirated/scanned copies of the in-copyright modern translation, not a PD edition. Vatican.va's "Sermons of the Fathers" devotional page for Sermon 241 should be treated the same way unless its translator credit is checked — I did not verify its source and it is not needed for this verdict since 241 is outside the PD band regardless of which modern translation is quoted.

Coverage table (final)

| Sermon # | PD English available? | Edition | Note | | --: | :-- | :-- | :-- | | 18 | No | — | Below range (OT-lesson sermon, not in the "Selected Lessons of the NT" set) | | 33 | No | — | Below range | | 43 | No | — | Below range | | 52 | Yes | NPNF1 vol. 6 (aselectlibrary06unknuoft), seq. #2 [LII. Ben.]; also LF | | | 56 | Yes | NPNF1 vol. 6, seq. #6 [LVI. Ben.] — verified directly; also LF | | | 57 | Yes | NPNF1 vol. 6, seq. #7 [LVII. Ben.]; also LF | | | 58 | Yes | NPNF1 vol. 6, seq. #8 [LVIII. Ben.]; also LF | | | 78 | Yes | NPNF1 vol. 6, seq. #28 [LXXVIII. Ben.] — verified directly; also LF | | | 96 | Yes | NPNF1 vol. 6, seq. #46 [XCVI. Ben.] — verified directly; also LF | | | 169 | No | — | Above range ("de verbis Apostoli" series continues past 147) | | 186 | No | — | Above range (de tempore, Christmas) | | 213 | No | — | Above range | | 214 | No | — | Above range | | 241 | No | — | Above range (Easter) | | 267 | No | — | Above range | | 272 | No | — | Above range (famous Eucharist sermon; watch the earlychurchtexts.com mislabel above) | | 298 | No | — | Above range |

Coverage: 6 of 17 (35%).

What I could NOT establish

The exact Maurist-number split point between LF vol. XVI and vol. XX (not needed — NPNF1 vol. 6 is the recommended single-volume citation target and its range is fully verified).
The translator/source attribution for vatican.va's Sermon 241 excerpt (not needed — 241 is outside the PD band under any translation).
Whether any other obscure pre-1930 anthology (e.g. a 19th-century Anglican "Sermons for the Christian Year" compilation) happens to include one of the 11 out-of-range sermons in full. I ran a targeted search and found nothing; a negative isn't provable, but nothing turned up.

Recommendation

No action needed for the citation-only entry itself — LS-KG-02 §1/§7 means no text is stored regardless. If this work is ever revisited for translation-linking: link the 6 in-range sermons (52, 56, 57, 58, 78, 96) to NPNF1 vol. 6 (IA aselectlibrary06unknuoft) using the NPNF seq = Maurist − 50 mapping above; leave the other 11 unlinked (no PD English exists) rather than substituting a New City Press citation.

Sources

apps/web/content/cited-sources/manifest.json — the sermones-st-augustine entry
apps/api/ingestion/vault/imported/catechism-sources.json — 17 nodes with workSlug: "sermones-st-augustine", giving the exact loci
https://www.ccel.org/ccel/schaff/npnf106.html — NPNF1-06 landing/summary page
https://www.newadvent.org/fathers/1603.htm — index of the 97-sermon selection (sequential #s 1–97)
https://www.newadvent.org/fathers/160301.htm, 160306.htm, 160328.htm, 160346.htm, 160397.htm — bracketed Maurist-numbering checkpoints
https://archive.org/details/aselectlibrary06unknuoft — NPNF1 vol. 6 scan (Buffalo, Christian Literature Co., ed. Schaff)
https://archive.org/details/16ALibraryOfFathersOfTheHolyCatholicV16 — LF vol. XVI (Sermons vol. I, Matt/Mark/Luke)
https://archive.org/details/20ALibraryOfFathersOfTheHolyCatholicV20 — LF vol. XX (Sermons vol. II, John/Acts/Epistles)
https://earlychurchtexts.com/public/augustine_sermon_272_eucharist.htm — mislabeled "NPNF" attribution trap
https://wesleyscholar.com/wp-content/uploads/2019/04/Augustine-Sermons-184-229.pdf (and sibling PDFs) — pirated New City Press scans, in-copyright trap
https://archive.org/stream/saint-augustine-sermons/... — "Saint Augustine - Sermons in 11 vols" (confirms New City Press 11-vol structure, incl. Sermon 272B)
https://www.goodreads.com/work/quotes/16108880 — "Sermons 184-229Z (Vol. III/6)" — confirms New City Press volume-numbering scheme

2026-10-01 (content night) — THE GATE THIS ENTRY CARRIED WAS THE WRONG GATE, AND THE REAL ONE IS NOW SOLVED

Status change: from "CLEAR on rights, PARTIAL coverage, LOW PRIORITY" to BUILDABLE AND SPECIFIED, with a second edition cleared from primary evidence. Nothing in the rights verdict above is withdrawn. What changed is that the build was attempted, and the thing that actually blocks it is not the thing this dossier told a builder to watch.

The gate the registry named was "fingerprint against the dated scan". That is not what stops this build.

The registry's gates for this item were NPNF numbering (Maurist − 50) and a fingerprint against the 1888 scan. Both are real and both are cheap. The blocker is the ADDRESS GRAMMAR, and it is queue item 11's blocker (Barnabas) and Justin Martyr's blocker for the third time.

Our stubs address three tiers; NPNF publishes two. sermones-st-augustine holds 17 DS stub nodes whose headings are sermon, chapter, section (56, 6, 9, 58, 11, 13, 52, 6, 16). Measured tonight in BOTH witnesses:

CCEL's npnf106.xml: the sermons division (div1 id="vii") holds 99 div2 and ZERO div3. Sermon VI (Maurist 56) carries a single continuous arabic section series 1 to 19 and no chapter element of any kind. The full tag inventory for that division is p.c22 body paragraphs plus note/p.endnote apparatus. There is nothing to build a chapter tier out of.
The 1888 print itself (IA aselectlibrary06unknuoft, the scan this dossier names): same answer, so this is not a CCEL transcription loss. The page sets SERMON VI. / [LVI. BEN.] / ON THE LORD'S PRAYER... / 1. THE blessed Apostle, to show that... and then 2. But many ask for what they ought not to ask. No chapter numerals anywhere. Checked on the page, not inferred.

So NPNF1 vol. 6 cannot produce 56, 6, 9. Transcribing at sermon+section would yield 56, 9, the 6 in-band stubs would not rename, and their edges would strand. That is the option the founder REJECTED for Barnabas on 2026-08-07, ruling instead to find an edition carrying the finer numbering. It is not taken unattended here either.

THE EDITION THE FOUNDER'S RULING CALLS FOR EXISTS, IT IS THE SAME TRANSLATION, AND ITS RIGHTS ARE NOW CLEAR FROM PRIMARY EVIDENCE

A Library of Fathers of the Holy Catholic Church, Anterior to the Division of the East and West — Sermons on Selected Lessons of the New Testament by S. Augustine, Bishop of Hippo, VOL. I, S. MATTHEW. S. MARK. S. LUKE. Oxford: John Henry Parker; J. G. F. and J. Rivington, London, MDCCCXLIV (1844). Printer: Baxter, Oxford. IA 16ALibraryOfFathersOfTheHolyCatholicV16.

PD basis, read off the title pages rather than summarised: 1844 UK publication, Oxford/Parker imprint, renewal-independent and expired by any term that has ever applied. This resolves the one thing the 2026-08-09 research could not establish: that session recorded a garbled "1839/1844" date pair from an AI-summarised title-page read and flagged it unreliable. The date is 1844, printed in roman numerals on the series title page AND again on the work's own title page, and corroborated by Pusey's dateline on the Advertisement ("E. B. P., Christ Church, Feast of S. Barnabas, 1844").

Two further facts come off the same front matter, both load-bearing:

Translator confirmed in the edition's own words: "the Editors are indebted for the translation to the Rev. R. G. Macmullen, M.A. Fellow of Corpus Christi College." This is the same MacMullen translation NPNF1 vol. 6 later reprinted with light revision, so the two editions are not rival translations and their section numbering is shared.
The base text is named: "the text used is that of the Benedictines." The Maurist edition is the source of the chapter/section apparatus, which is exactly the numbering our CCC citations use. The chapter tier is therefore the edition's own datum and not an editorial invention.

LF PRINTS BOTH NUMBERS. Verified at three of our own cited addresses.

LF sets the Maurist chapter as a roman numeral beside the arabic section number:

1 . The blessed Apostle, to shew that those times when it   i.
ii. 2. But many ask for what they ought not to ask...
iii. 3. You have read in the Holy Psalms...
vi. 9. Give us this day our daily bread.

That last line is our stub 56, 6, 9, and the content confirms it independently: the CCC cites Sermo 56, 6, 9 on the daily-bread petition, and chapter vi section 9 is where that petition is expounded. Three of the six in-band citations were resolved to an exact printed chapter+section pair:

| Cited stub | LF chapter+section found | Verdict | | :-- | :-- | :-- | | 56, 6, 9 | vi. 9 ("Give us this day our daily bread") | EXACT | | 57, 7, 7 | vii. 7 | EXACT | | 58, 11, 13| xi. 13 | EXACT | | 52, 6, 16 | heading OCR-garbled, not mechanically located | needs a page read | | 78, 6 | no marks recovered in this OCR | needs a page read | | 96, 7, 9 | only sections 1 and 5 recovered of ~10 | needs a page read |

The three failures are OCR loss, not edition loss, and the distinction is the whole finding. The marks sit in the MARGIN, and this is a Google scan that routinely welds marginalia into the prose ("Serm. distinguish the proper times", "[55VB ] ^en ls lea(^n£") and drops it elsewhere. Where the OCR preserves the roman. arabic. adjacency the address resolves every time it was checked.

Volume split, also resolved

LF vol. I covers Maurist 51 to 116 (Sermons I to LXVI, read off the sermon headings; the contents page prints Serm. 1. (Ben. 51.), 2. (52.) and so on, which confirms the Maurist − 50 mapping from the edition's own contents rather than from five sampled New Advent checkpoints). All six of our in-band cited sermons (52, 56, 57, 58, 78, 96) fall inside vol. I, so LF vol. XX is NOT needed to serve a single one of our citations. The 2026-08-09 open question about the split point is closed for our purposes.

A SOURCE DEFECT FOUND FOR FREE BY HAVING TWO WITNESSES

CCEL's npnf106.xml heads its sequence-62 sermon Sermon LXII. [XCII. Ben.]. Sequence 62 maps to Maurist 112, which is CXII; XCII is 92, and 92 is already correctly claimed by sequence 42 (Sermon XLII. [XCII. Ben.]). LF's own heading at the same sermon reads SERMON LXII. [CXII. Ben.], which is right. So CCEL (or the 1888 print behind it) carries a wrong Benedictine number at one sermon. Anything that parses the bracketed gloss as the authoritative Maurist number would mint a duplicate 92 and lose 112 silently, because both are valid-looking numbers. Parse the bracket, then assert it equals sequence + 50, and fail the build on a mismatch rather than trusting either witness.

The recommended build, specified

TEXT from CCEL npnf106.xml (clean ThML, 97 sermons at div1 vii, div2 vii.iii to vii.xcix; vii.i and vii.ii are the title and the Advertisement and are not sermons). Match on the sermon heading, never on div2 position, since the division count is 99 against 97 sermons.
CHAPTER BOUNDARIES from LF 1844, which is the only witness that has them. For the stubs this is needed for SIX sermons only, which is a bounded page read rather than a 97-sermon campaign. Carrying the chapter tier across all 97 is the larger job and can follow.
The precedent for this shape already exists in this repo: the documented hybrid that takes text from one source and reference positions from another (edition-supplied:newadvent, recorded against Irenaeus). This is that pattern with a PD edition on both sides, which is strictly better.
VERIFICATION GATE before relying on the hybrid: prove NPNF section N is LF section N per sermon. Confirmed tonight for Maurist 56 section 1, where the two texts agree word for word apart from spelling modernisation ("to shew"/"to show", "ought not be preached"/"ought not to be preached"). It must be asserted per sermon in the converter, not assumed from one sample.
The CCEL integrity audit (npnf110-transcription-defect.md) is still owed on npnf106.xml before its text is published. It was NOT run tonight because the address finding changed what the source is for, and running it would have measured a file whose role was still unsettled.
LOAD PATH: loadAdditive.ts onto the existing sermones-st-augustine row. Measured tonight on the local database: CITATION_ONLY, 17 nodes, all 17 are DS stubs (SECTION with parentId null), 35 incoming Connection edges. Note the dossier's "17" above is the count of LOCI; the EDGE count is 35, and that is the number a before/after assertion must hold. The six in-band stubs carry 14 of those 35 edges; the eleven out-of-band stubs carry 21.
The eleven out-of-band stubs STAY stubs, and that is correct, not an omission to tidy away. No PD English of those sermons exists (New City Press only, in copyright). The work's provenance note must say so plainly, because a reader who reaches a stub inside a published work should be told the gap is the state of the translations and not a defect in the transcription.

What is still NOT established

The chapter marks for Maurist 52, 78 and 96, which need page images rather than this OCR.
Whether a cleaner scan of LF vol. I exists. Not searched tonight.
Whether Migne PL 38 (PD, 1845) is a cheaper machine-readable source for the chapter boundaries than reading LF's pages. It is the obvious alternative and was not evaluated.