| Work / author / dates | Rūḥ al-quds fī munāṣaḥat al-nafs (روح القدس في مناصحة النفس, "The Spirit of Holiness in the Counselling of the Soul") · Muḥyī al-Dīn Ibn ʿArabī (560–638 / 1165–1240) · written at Mecca, completed Rabīʿ I 600 / Nov. 1203 (colophon of C), addressed to Abū Muḥammad ʿAbd al-ʿAzīz al-Mahdawī of Tunis |
| Part added | A selection from the second part (the portraits of the saints he met in al-Andalus and the Maghrib): 10 complete portraits (Austin §§ 1, 2, 3, 6, 14, 17, 18, 19, 54, 55), plus Fāṭima bint Ibn al-Muthannā's portrait in the Futūḥāt, ch. 178. 6,492 words (Rūḥ 6,080 + Futūḥāt 412) in 99 reading passages |
| Library id | ruh-quds-teachers · word-id prefix r1 · files data/sources/ruh-quds-teachers-{base,meta}.json → ruh-quds-teachers.json (vowelled) → ruh-quds-teachers-analysis.json |
| Reader route | /read/ruh-quds-teachers (not yet registered, see §12) · apparatus /read/ruh-quds-teachers/notes (proposed) |
| Status | working edition · consonantal text collated against C by AI reading of the page images (28 Sept 2026) · reviewed by: — |
The portraits in this selection:
| # | Portrait | Austin § (p.) | Atlas id · place | Passages | C pp. |
|---|---|---|---|---|---|
| 1 | Abū Jaʿfar al-ʿUraynī, "the first I met on the Way" | 1 (63) | ruh-quds:1 · seville | 1–12 | 44–46 |
| 2 | Abū Yaʿqūb Yūsuf b. Yakhlaf al-Kūmī, "our shaykh and imam" | 2 (69) | ruh-quds:2 · seville | 13–26 | 46–49 |
| 3 | Ṣāliḥ al-ʿAdawī | 3 (73) | ruh-quds:3 · seville | 27–30 | 49–50 |
| 4 | Abu-l-Ḥajjāj Yūsuf al-Shubarbulī | 6 (79) | ruh-quds:6 · aljarafe | 31–39 | 51–52 |
| 5 | Abū Muḥammad ʿAbdallāh al-Mawrūrī (the text: al-Marwazī), with the digression on jurists | 14 (101) | ruh-quds:14 · seville (+ marchena) | 40–68 | 60–65 |
| 6 | ʿAbdallāh b. Jaʿdūn al-Ḥinnāwī, one of the four Pillars | 17 (114) | ruh-quds:17 · fez | 69–71 | 67–68 |
| 7 | Abū ʿAbdallāh Muḥammad b. Ashraf al-Rundī, one of the Substitutes | 18 (116) | ruh-quds:18 · rota (+ ronda) | 72–81 | 68–70 |
| 8 | Mūsā Abū ʿImrān al-Sadrānī, bearer of Abū Madyan's message | 19 (121) | ruh-quds:19 · seville | 82–86 | 70–71 |
| 9 | Shams, Umm al-Fuqarāʾ, of Marchena | 54 (142) | ruh-quds:54 · marchena | 87 | 77–78 |
| 10 | Fāṭima bint Ibn al-Muthannā of Seville (Rūḥ) | 55 (143) | ruh-quds:55 · seville | 88–92 | 78 |
| 11 | Fāṭima bint Ibn al-Muthannā, in the Futūḥāt, ch. 178 (the Fātiḥa story) | 55 | ruh-quds:55 · seville | 93–99 | M 6:26–27 |
Abū Madyan has no portrait of his own in Rūḥ al-quds (Ibn ʿArabī never met him in the body). He is present in portraits 2, 4, 5 and above all 8, where al-Sadrānī brings his message that their meeting would be of spirits, not bodies (passages 82–84). Atlas: the Abū Madyan entry (n = null, place bejaia). "Abū ʿAbdallāh al-Mawrūrī": there is no separate portrait of that name. Abū Madyan addresses ʿAbdallāh al-Mawrūrī as "yā Abā ʿAbdallāh" (passage 41); the portrait itself names him Abū Muḥammad ʿAbdallāh. Fāṭima: in Rūḥ al-quds her portrait is folded into Shams's ("and likewise I met Fāṭima…"); we split it at that sentence. Her fuller portraits are in the Futūḥāt (ch. 178, added here) and in al-Durra al-fākhira (not openly available; Austin translates its extract).
1. Witnesses
| Siglum | Witness | Kind | Vowelled? | Licence | Location / hash |
|---|---|---|---|---|---|
| O | OpenITI barzakh 0638IbnCarabi.RuhQuds.IAC00030-ara1, from ibnalarabi.com. A typed text of a modern print (not identified; its typography, اللّه and eulogies in parentheses, suggests a Beirut print). Carries no page or section numbers. | digital | no (Quran in a modern vowelling) | no licence stated (OpenITI staging repo; OpenITI releases are CC BY-NC-SA 4.0); the work is PD | local copy, sha256 be939d70…0616; https://github.com/OpenITI/barzakh/blob/6b19bdc83141c0388b134a640b1e75bba0d450ca/0638IbnCarabi.RuhQuds.IAC00030-ara1 |
| C | Risālat Rūḥ al-quds fī munāṣaḥat al-nafs, Cairo: al-Maṭbaʿa al-ḥajariyya al-fāʾiqa al-laṭīfa, early Dhū l-Qaʿda 1281 [March 1865], lithograph, 104 pp. (colophon, p. 104). Printed from MS Yaḥyā Ef. 2605 (Austin, Introduction). Copy: National Library of Israel, SC 56 C 5545 (barcode 000200405926). = "the Cairo print". | early print (lithograph) | sparsely | public domain (1865); scan on archive.org uploaded by a third party (no licence stated on the item) | https://archive.org/details/20230520_20230520_2306 (108 images; image = printed page + 1). Page images read at IA's BookReader scale 1. |
| OF | OpenITI 0638IbnCarabi.FutuhatMakkiyya.JK009237-ara1 (Beirut: Dār Iḥyāʾ al-Turāth al-ʿArabī, 1418/1998), ll. 59440–59469 (page marker V02P341) | digital | no | OpenITI (CC BY-NC-SA 4.0) | https://github.com/OpenITI/0650AH/blob/master/data/0638IbnCarabi/0638IbnCarabi.FutuhatMakkiyya/0638IbnCarabi.FutuhatMakkiyya.JK009237-ara1#L59440 |
| M | al-Futūḥāt al-Makkiyya, ed. ʿAbd al-ʿAzīz Sulṭān al-Manṣūb (12 vols., on the Konya autograph), vol. 6, pp. 26–27 (autograph ff. 14a–15b in his margins) | critical edition | yes | © the editor; released by him free for individual use (archive.org description). Verification only; text and vowels not reproduced | https://archive.org/details/baha-2026-07 |
| — | Not used: (a) Damascus 1964 (ed. Maḥmūd Maḥmūd al-Ghurāb, from the al-Jazāʾirī MS collated with C): not found lawfully online; (b) Maḥmūd Bījū's Damascus edition (c. 2004; based on two prints "one taken from the other", i.e. C and the 1964?) is on archive.org (20230929_20230929_2142) but is a third-party upload of an in-copyright book: not used; (c) the Beirut Majmūʿat rasāʾil Ibn ʿArabī (Dār al-Maḥajja al-Bayḍāʾ) likewise (1_20200702_20200702_1113); (d) an AskZad copy (AAskZad-1147344, ed. Ḥāmid Ṭāhir) is from a commercial database: shadow source, not opened. (e) Asín Palacios printed the Arabic from Escorial MS 741 (Madrid 1939); Asín d. 1944, so PD in Spain since 2025: a possible third witness if a scan exists. (f) The authority is Istanbul University Library MS A 79a (600 H, read to the author, with his marginal corrections and nine samāʿ; Austin's base). | ||||
| — | Translation (private comparison only): R. W. J. Austin, Sufis of Andalusia (1971; Routledge 2013). Notes in the private layer. | translation | © | user's own PDF |
2. Base and why
O was given as the base (it is on disk, in the atlas already, and line-linkable). It is a sound text marred by one systematic fault: a final alif (or final letter) is lost in many words (لا→ل, ما→م, هذا→هذ, كذا→كذ, أبا→أب, tanwīn -an: أبدا→أبد, شيئا→شيئ, and final -nā: عندنا→عندن). Every such form in the selection was restored only where C has the full form, and each restoration is logged. C is public domain, so its adopted readings are published with their page. Futūḥāt section: base OF, corrected against M (weight: M is a critical edition on the autograph; OF is a typed text with systematic hamza errors: أن for إن, ألا for إلا, أمرأة).
3. Collation
- Method: the ten portraits (O lines 996–1863 selectively) read side by side with C's page images, pp. 44–52, 60–65, 67–71, 77–78, by eye (Claude, AI reading of the lithograph, 28 Sept 2026). Close crops were made where a reading was doubtful (e.g. فتقايا p. 47, السبريلي p. 51, the مهاد/أوتاد sequence p. 46). Not a double keyed transcription: a person should re-read the pages.
- C's purely orthographic habits (no hamza, final ى for ي, تا/ة, missing dots) are not recorded as variants.
- Script:
build_base.py(scratchpad of this session; the logic is summarised here, it is not a repo script). - Results: 101 changes to O from C (92 restorations of lost letters/words, 8 orthographic, 1 haplography), 7 readings of C recorded but not adopted (
variants), 9 flagged doubts where O and C agree (uncertain). Futūḥāt: 23 changes to OF from M (20 orthographic, 3 substantive: فماذا, أبعث, اعتنى). - Mechanical clean-up (by rule): 11 ×
بـ"X→بX; 53 quotation marks around names removed; 6 spaces before commas removed.
4. Decisions (substantive)
All 124 are in ruh-quds-teachers-base.json → changes (passage, from, to, context, C page, type, reason). Examples:
| Where | Reading adopted | O had | Reason | By |
|---|---|---|---|---|
| p. 5 (ʿUraynī) | وصاح وطرأ عليه حال | وصار | C; صار leaves the verb without predicate | AI, from C p. 45 |
| p. 7 | فإنك مقر بها معترف لا أنساها الله لك | مقربها … أنساه | word division + lost alif; C | AI, C p. 45 |
| p. 15 (Kūmī) | فتقايأ فيهما لبنا | فتقي | truncated form; C فتقايا (close crop) | AI, C p. 47 |
| p. 49 (Mawrūrī) | الشكاز الباغي | الباعي | "of Bāgha" (Priego), as in his own portrait; C | AI, C p. 61 |
| p. 49 | ومشى بي عليها … الرحا | عليه … الرح | pronoun = المقامات; C | AI, C p. 61 |
| p. 87 (Shams) | وكنت أفرح لها بذلك لها بركات كثيرة | — (لها missing) | haplography; C | AI, C p. 77 |
| p. 91 (Fāṭima) | فقيل لها | فقيل له | said to her; C | AI, C p. 78 |
| p. 98 (Futūḥāt) | حيث اعتنى بي | أعتني | perfect, subject God; M | AI, M 6:26 |
Kept as printed but flagged (both witnesses agree): المروزي (Austin: al-Mawrūrī), السبريلي (the Futūḥāt has الشبربلي … قرية … تسمى شبربل = al-Shubarbulī), السيدراني (Austin: al-Sadrānī), أبو يعقوب الكوفي in the Qāf episode (surely al-Kūmī), فاطمة بنت أبي المتنبي (the Futūḥāt: بنت ابن المثنى القرطبي), فترعني ذلك (sense doubtful), البابين (unidentified). Reader labels use the scholarly forms; the text keeps the witnesses'. Finding for the atlas: C preserves the heading of Ibn Qassūm's portrait («ومنهم … أبو عبد الله محمد بن قسوم», p. 52), which O lacks; the atlas entry n = 7 took the name from Austin and can now cite C.
5. Quotations verified
Against the Tanzil Uthmani text, replaced verbatim: 8 quotations: 78:6–7 and 78:38–39 (passage 11, the imam's recitation of al-Nabaʾ), 3:66 (56), 2:204 (57), 6:116 and 38:24 (65), 51:57 (79), 28:25 (85, "lā takhaf, najawta…": Abū Madyan's greeting to Mūsā). 11 allusions recorded, left as written (5:54, 95:5/83:18, 28:26/18:71, 20:130, 18:65, 12:53, 3:7, 12:56, 6:59, and two stock formulas). 7 hadith identified (the one who passes in front of a person praying; "I am the master of the children of Adam"; "when God loves a servant He tries him"; "the heart of the Qurʾān is Yā Sīn"; the seventy thousand; "whomever God wills good…"; the judge who exercises ijtihād): collection references are AI-supplied, to verify. A skeleton search of all 99 passages against the whole Quran (the quran_refs.py method, run privately without writing to app/) found no further quotation. scripts/quran_refs.py itself has not been run for this text (it needs the text registered, see §12).
6. Vowelling
- Source: none in O or C (C sparsely); AI vowelling by
scripts/recitals/prepare.py(Claude Opus, headless, 3 workers), every letter checked unchanged; per-charactervowelMask(t = source, a = AI). Quran from Tanzil. - Comparison with an authoritative vowelled edition: none available for Rūḥ al-quds. Futūḥāt: M is vowelled (private comparison possible, not done).
- Note:
prepare.py's system prompt names "Mehren's edition" and its آ rule is Mehren-specific; the letters check treats آ = ا, so any such change is logged inchangesas "mechanical" and should be reviewed.
7. Segmentation and word IDs
99 passages, n = 1–99 continuous across the selection; each passage carries label ("name (i/k)"), portrait {key, name, ar, austin, place, placeName}, atlasId (ruh-quds:<Austin §>), source, sourceUrl (O line), and C's printed pages. Passages 40–90 words (a few 29–38 at portrait ends; one of 94). Two editorial breaks in the Mawrūrī digression (before «ثم إن في المسألة» and «هلا قال عند»). Word ids: r1:<passage>:<token> (analyze.py).
8. Analysis layer
scripts/fusus/analyze.py ruh-quds-teachers.json ruh-quds-teachers-analysis.json r1 (CAMeL Tools): 6,448 tokens analysed, 4,146 with a root, 734 distinct roots, 262 vowel conflicts between CAMeL and the AI vowelling. The unresolved "roots" are the usual particles (ف.#, ع.ل.# = على, ل.#, #.ل.# = الله …), the same pattern as the Ghurba. Vowelling run (prepare.py): 99/99 passages vowelled (passage 84 failed the letter check 3× in the batch; re-run alone with the same prompt, passed), 434 ambiguities recorded by the model; no آ→ٱ changes.
9. Commentaries aligned
None (the work has no classical commentary). Parallel accounts: the Futūḥāt (Fāṭima, Shams, Abu-l-Ḥajjāj al-Shubarbulī named in ch. 178) and al-Durra al-fākhira (not openly available).
10. Translation
AI (Claude Opus) via scripts/recitals/translate.py ruh-quds-teachers opus, from the Arabic, with the glossary in ruh-quds-teachers-meta.json (walī, abdāl, awtād, quṭb, ḥāl, maqām, karāma, kashf, himma, firāsa, malāmatiyya, futuwwa, futūḥ, places…). Its system prompt names other works' translators; Austin was not given to the model. Austin (1971) is compared privately only.
11. Rights and private layers
Published: O's text as corrected; C's readings (PD); Tanzil. Private: Austin (notes only), M (readings adopted are cited, M's text/vowelling not reproduced).
12. Open questions / to review
- A person to re-read C's pages against the edition (the collation is an AI reading). Priority: the 7 variants and 9 doubts.
- Acquire the Istanbul MS A 79a (autograph-certified) or the Damascus 1964 edition; look for a scan of Asín 1939 (PD in Spain).
- Austin's "I, with two of my companions, built a hut" vs O/C/Futūḥāt "with my own hand" (private note): check the Istanbul MS.
- Wiring (needs changes to shared files, not made):
scripts/register_text.py ruh-quds-teachers r1 "Rūḥ al-quds: portraits"(edits build_data.py + quran_refs.py), an entry inapp/src/lib/library.ts, thenscripts/quran_refs.py ruh-quds-teachers.prepare.pydrops the per-passagelabel/portrait/atlasId/sourcefields: they were merged back intoruh-quds-teachers.jsonafter the run (a rerun would drop them again).translate.pywrites an Ibn Zayla commentary stub intoapp/public/data/commentaries/ruh-quds-teachers.json(count 0): harmless but misleading. - Atlas link: the atlas has no URL state; see the report for the proposed
?layer=people&person=<n>wiring.
13. Change log
- 2026-09-28: record created; witness survey; C located (NLI copy of the Cairo 1281 lithograph); 10 portraits + Futūḥāt ch. 178 established; 124 changes; Quran verified; AI vowelling, analysis and translation run.