| Work / author / dates | Risālat Ḥayy ibn Yaqẓān fī asrār al-ḥikma al-mashriqiyya (رسالة حي بن يقظان) · Abū Bakr Muḥammad ibn ʿAbd al-Malik ibn Ṭufayl al-Qaysī (d. 581/1185-6) |
| Part added | Introduction + the first part of the tale: the two accounts of Ḥayy's origin, his upbringing by the gazelle, her death, the dissection and the burial. Stops before the discovery of fire (see §7). About 30% of the work; 85 passages, 5,947 words. |
| Text id | hayy-tufayl (the id hayy is Ibn Sīnā's Ḥayy); word-ID prefix y1 |
| Reader route | /read/hayy-tufayl (proposed) · apparatus /read/hayy-tufayl/notes (proposed) |
| Status | text of A established (steps 2–6, 8), collated privately with Gauthier 1936, 28 Sept 2026 · reviewed by: — |
Structure of the part added: doxology (passage 0; see §4) · the question and the state beyond description (1–4) · Ibn Bājja and the rank reached by speculation (4–6) · Avicenna's stations of the knowers, quoted from the Ishārāt (6–8) · the parable of the man born blind (8–10) · [lacuna in A, §4] · the two things the question can seek (11–14) · the philosophers of al-Andalus, al-Fārābī, Aristotle and the Shifāʾ, al-Ghazālī (14–27) · the author's own path and the invitation (27–30) · first account: the equatorial island, with the argument on the sun's heat (31–37) · second account: the princess, the chest and the sea (38–42) · first account resumed: the fermented clay and the three chambers (43–54) · the gazelle nurses the child (54–57) · nakedness, leaves, a staff, the eagle's skin (58–63) · the gazelle's death (64–70) · the dissection (71–74) · the heart's empty chamber (75–79) · the ravens and the burial (80–84).
1. Witnesses
| Siglum | Witness | Kind | Vowelled? | Licence | Location / hash |
|---|---|---|---|---|---|
| A | Aḥmad Amīn (ed.), Ḥayy ibn Yaqẓān li-Ibn Sīnā wa-Ibn Ṭufayl wa-l-Suhrawardī, Cairo: Dār al-Maʿārif, 1952 (Dhakhāʾir al-ʿArab 8); the scan is a later free reprint (Dār al-Madā for al-Ittiḥād). Ibn Ṭufayl at printed pp. 55–128 (= PDF pp.); our part pp. 55–77. Source not stated; Amīn's preface names the Egyptian prints and the Damascus print of 1935. His text agrees with the readings Gauthier reports for MSS B, P and the prints αβγ (the vulgate) in nearly every place where they differ from Gauthier's text. | early 20th-c. print (vulgate) | almost none (tanwīn, a few shaddas) | Amīn d. 1954: PD in Egypt; US status uncertain. The medieval text only is used; Amīn's footnotes, headings and numbering are excluded. | (archive.org quran99yaai_gmail_201710), SHA-256 db21ba54e283f36cbc5bbf4a09823ab43e52a138c8d480d15cac6e94462c65c0; OCR text (poor: typing aid only) |
| G | Léon Gauthier (ed. & tr.), Hayy ben Yaqdhân, roman philosophique d'Ibn Thofaïl. Texte arabe avec les variantes des manuscrits et de plusieurs éditions et traduction française, 2nd ed., revised, Beirut: Imprimerie Catholique, 1936 (Institut d'Études Orientales d'Alger III). Critical edition with apparatus: MSS A (Algiers, 1180/1768), B, C, L, P and the prints α β γ; Arabic pp. 1–47 for our part (PDF p. = Arabic p. + 192). | critical ed. | partly | Gauthier d. 1949: PD in France/EU since 2020; in the US probably protected until 2031 (1936 foreign publication, restored under the URAA). Used as a private comparison witness only. | archive.org fr110philo10 (an upload from a personal book site; no licence stated), local copy, SHA-256 647c600c463403cefec2d3d45804a94d4db04324f86427d4b488bc2eaa706399. No Arabic text layer (the djvu OCR is French only). Our consonantal transcription of its main text: kept in the private layer. |
| G¹ | Gauthier, 1st ed., Algiers: P. Fontana, 1900 (text, variants, French translation). Public domain everywhere. | critical ed. | PD | Not found online (archive.org, Gallica, HathiTrust search blocked, Google Books quota): searched 28 Sept 2026. Only Gauthier's 1900 leçon d'ouverture (laphilosophiemus00gaut) and his 1909 Ibn Thofaïl, sa vie, ses œuvres are on archive.org. | |
| O | OpenITI 0581IbnTufayl.HayyIbnYaqzan.Shamela0009734-ara1 (RELEASE v2025.1.9), from al-Shamela, edition not stated; lacks the introduction (starts at «ذكر سلفنا الصالح»). | digital | no | CC BY-NC-SA (OpenITI) | local copy, SHA-256 e91fbf65205f2e6af3407817e5a53f16a8519e039d0ff039a15be800283e94ab. Poor in places (e.g. غرابان … ضار, بعقد for بعلق). Used only as a second check on our reading of A in the narrative. |
| — | Fārūq Saʿd (ed.), Ḥayy ibn Yaqẓān li-Ibn Ṭufayl, Beirut: Dār al-Āfāq al-Jadīda, 1992 (private copy) | modern ed. | © | Text layer in presentation forms, badly broken; not used. |
2. Base and why
A (Amīn 1952), as for the Ghurba: it is the only openly usable print on disk, its scans are clear, and the task allows it (Egyptian PD; only the medieval text is transcribed). G¹ (Gauthier 1900), the preferred public-domain base, could not be found. G (1936) is the better text (a critical edition from five manuscripts) but is probably still protected in the US; it is therefore the private check witness, as Corbin is for the Ghurba. No reading of A was replaced by Gauthier's, except obvious misprints (§4), each listed.
3. Collation
- Transcription: A from 250 dpi page images (the OCR was too poor to use as a typing layer: 494 disagreements with our first reading, nearly all OCR noise). Doubtful spots re-read at 450 dpi.
- Proofreading: (1) against OpenITI (narrative only) and the OCR; (2) against G, read page by page from 220 dpi images with its apparatus; every disagreement not explained by Gauthier's apparatus was checked again on A's page image. One typing error of ours was found and fixed (بعاليق → بمعاليق, p. 75); the rest were A's readings.
- Tool:
difflibalignment on a normalised consonantal skeleton (vowels, hamza seats, alif forms, tāʾ marbūṭa neutralised), word by word, our final text against G. - Result: 336 difference units: 54 orthographic, 282 substantive (186 replacements; 67 places where A has words G lacks, 108 words, the doxology included; 29 places where G has words A lacks, 218 words, of which c. 190 are the lacuna below). A 5,947 words, G 5,965 for the same stretch.
- Readings: private layer only.
- Most substantive differences are readings G reports for B, P and αβγ, which A follows (e.g. «من كتابه ذلك», «بل هي أحوال من أحوال السعداء», «خلق مكفوف البصر», «وبها شجر يثمر نساء … جزيرة الواقواق», «الملتفة الشجر», «خرج من كناسه فحمله العقاب», «الأذواق», «حقيقته», «المحمدية»). Where A's reading is attested only in G's MS A (Algiers), e.g. «كتاب الجوهر», «جنات القدس», «شبه حزام», it is noted in the private file.
4. Decisions (substantive)
| Where | Reading adopted | Rejected / A | Reason | By |
|---|---|---|---|---|
| passage 0 (p. 55) | the doxology «الحمد لله العظيم … تسليماً كثيراً», kept as printed | G's text: basmala + «صلى الله على سيدنا محمد وآله وصحبه وسلم» | G relegates this preamble to the apparatus (only in BP αβγ: "peu conforme à la manière sobre d'Ibn Thofaïl"). A is our base; kept, flagged in the label. | AI (Claude), provisional |
| after passage 10 (p. 58) | lacuna recorded, not supplied | G pp. 9–10: «وقد يوجد في النادر … وتصريف وجوه الحيل في اكتسابه» (c. 190 words) | Absent from A; no witness in G's apparatus lacks it, so the omission is A's (or its source print's). Supplying it from G would import G's text; recorded in uncertain and on the passage (lacunaAfter). | AI, provisional |
12 misprints corrected (list in changes, type proof / mechanical) | الآلاء, يعتريه, مما (the verse), كانت, لا على, علم المكاشفة, متكيفة, حاذاها, ويئن, النضيجة, عليها, الذي | الآراء, يعتز به, ممن, كنت, لأعلى, عدم, مكتفية, خاذاها, وينن, النضجية, غليها, الدي | Each a letter-level slip (dot, metathesis, nūn-shaped hamza seat, word division) giving no sense, and each supported by G or by the vulgate reading G reports. | AI, provisional |
| 5 words in Amīn's parentheses | kept, parentheses removed: الإمام, رحمة الله عليه, يرى, أنواع, شبه | — | Amīn's marks of addition; the words are attested (BP αβγ; MS A for شبه). Logged as mechanical. | AI |
| 11 doubtful readings of A | left as printed, listed in uncertain | — | Print clear, sense doubtful; G's reading given in the note (e.g. «مصارفه», «إنما هي كاملة» without غير, «ولا على حالة واحدة», «أو لإلمام», «مع الظبي»). | AI |
5. Quotations verified
- Quran, replaced with the verbatim Tanzil (Uthmani) text (5), each checked as an exact token run of the Tanzil Uthmani text: 96:4–5 (doxology; A continues «ما لا يعلم», Q «ما لم يعلم», left), 4:113 (doxology), 12:111 and 50:37 (end of the introduction; A «ذكرى», Q «لذكرى», left), 76:1 (the mother's prayer).
- Allusions, left as printed (11): 17:43, 39:6, 14:15, 12:64, 20:39 (the chest cast into the sea, Moses), 18:17 (the sun inclining from the thicket, the Cave), 17:85, 76:2, 7:22 (covering with leaves), 20:18 (the staff), 5:31 (the raven that "scratched the ground": the three words match the verse; the episode is retold).
- Hadith (1): «إن الله خلق آدم على صورته» (al-Bukhārī, Istiʾdhān 1; Muslim, Birr 115).
- Poetry and sayings: the shaṭaḥāt of al-Bisṭāmī / al-Ḥallāj; the line «فكان ما كان مما لست أذكره» (commonly ascribed to Ibn al-Muʿtazz; to verify) and «خذ ما تراه ودع شيئاً سمعت به» (commonly ascribed to al-Mutanabbī; to verify); two anonymous verses («برح بي أن علوم الورى»).
- Prose quotations: Ibn Sīnā, al-Ishārāt, namaṭ 9 (abridged; G compares Mehren); Ibn Bājja, Risālat al-ittiṣāl; al-Ghazālī, Mīzān al-ʿamal and al-Munqidh.
- Still to do:
scripts/quran_refs.pycannot be run for this text until it is added to that script'sTEXTStable (a shared script; not edited here).
6. Vowelling
A is almost unvowelled. scripts/recitals/prepare.py hayy-tufayl (28 Sept 2026) supplied AI vowels to all 85 passages (headless Claude, opus, 3 workers; 2 retries), letters verified unchanged, per-character vowelMask ('t' print, 'a' AI); 374 ambiguous readings flagged by the model (ambiguous per passage). The model reordered combining marks in 8 words of the Tanzil quotations; these were restored verbatim afterwards (logged in changes). prepare.py drops printedPage, label and lacunaAfter from base passages; they were merged back into hayy-tufayl.json from the base file. Comparison with a vowelled authority: G is partly vowelled; not yet compared (TODO).
7. Segmentation and word IDs
85 passages (0–84), 27–116 words (mean 70), cut at A's paragraph and sentence breaks; section labels ours. Word IDs y1:<n>:<token> (analyze.py numbering). Stopped at the end of A p. 77 (G p. 47, l. 5), after «فيعتقد أنه ليس في الوجود سوى جزيرته تلك», before «واتفق في بعض الأحيان أن انقدحت نار في أجمة قلخ» (the discovery of fire). Next part starts on A p. 77, last lines.
8. Analysis layer
scripts/fusus/analyze.py hayy-tufayl.json hayy-tufayl-analysis.json y1 (CAMeL Tools): 5,946 words analysed, 3,459 with a root, 728 distinct roots, 158 vowel conflicts (AI vowels vs the analyser). Unresolved weak roots are the usual #-radicals (ف.#, #.ن.ن, ذ.#, ع.ل.#).
9. Commentaries aligned
None. No classical Arabic commentary on Ibn Ṭufayl is known to us; Gauthier's notes (French) and Pococke's Latin (1671) are candidates for later.
10. Translation
AI (scripts/recitals/translate.py hayy-tufayl opus) from the Arabic, with the glossary in hayy-tufayl-meta.json. Gauthier's French (1936) is not published: it is the revised 1936 version, probably protected in the US until 2031; the 1900 translation (PD) was not found. Candidate PD English: Ockley 1708/1731 (archive.org bim_eighteenth-century_hayy-ibn-yaqdhan-engli_abu-bakr-ibn-al-tufail-_1731), Ashwell/Fulton 1929 (historyofhayyibn00ibnu, US status to check).
Translation triangulation (docs/PROTOCOLS.md §17; 2 Oct 2026, scripts/triangulate.py hayy-tufayl). Translations compared: Ockley–Fulton (Ockley 1708 as revised by A. S. Fulton, New York/London 1929; public domain in the US; Fulton omits Ibn Ṭufayl's introduction, so ¶32–84 only), cut to our passages in data/sources/hayy-tufayl-tr-ockley.json; and Gauthier 1936 (French, ©, private layer (private copy), ¶0–84). Both cut by scripts/recitals/tufayl_translations.py from the OCR (Claude marks only where each passage begins, which OCR units are footnotes, and where the OCR has displaced a block; the text is sliced by the script, so every word is the translator's own; alignment logs in (private copy)). No commentaries. Result (app/src/content/triangulation/hayy-tufayl.json): 85 passages, 45 divergences, 8 serious (shown as "Translators differ"): ¶34 what is credited to Avicenna; ¶40 Yaqẓān a kinsman (Ockley) or a neighbour (Gauthier), قريب; ¶46 the spirit "from God's command"; ¶49 the self-reflecting mirror; ¶52 which organs rule the later ones; ¶70 «يعترض سائر أعضائه» (stopping the organs, or reviewing them in the mind); ¶73 the membrane reached (diaphragm or pleura); ¶81 «شوق» (desire to inquire, or longing). Gauthier is quoted at most four words. AI-English check: no contradiction; 12 places flagged doubtful. Fixed in our own words:
| ¶ | was | now | why |
|---|---|---|---|
| 1 | must seek it and strive earnestly to acquire it | must seek these secrets and strive earnestly to acquire them | the pronoun of «بطلبها … اقتنائها» is feminine: the secrets (of the Oriental wisdom), not «الحق» |
| 6 | recalls something of what pertains to them | recalls something of the matter of that holiness | «من أمره» is masculine singular: holiness (القدس), not the plural gardens |
| 34 | or contact with hot bodies, and illumination | or contact with hot bodies, or illumination | three separate causes, any one sufficient |
| 81 | to the body's owner (its mover) and to what moved it | to the body's owner, its mover | «صاحب الجسد ومحركه» is one being described twice |
Left for a person (doubtful, not changed): ¶7 «النبل» (or al-nayl, as in the Ishārāt) and «مصارفه» (the Ishārāt has a form of maʿārafa); ¶32 "supernal light" for «النور الأعلى» in a climatic argument; ¶37 «الذي ما أضاء موقعه من الأرض قط» (ours keeps the negation the Arabic has; the translators paraphrase); ¶47 «يُستضاء به» (ours: "take on illumination"); ¶50 «كمالها»; ¶62 the relative «ما» after «يئس من»; ¶68 «لا يغني عنه شيء» (dependence).
11. Rights and private layers
Public: our working text of A (medieval text only) with our corrections, quotations and notes. Private (not published): Gauthier 1936 PDF, our transcription of its main text, the word-level collation. Nothing of Gauthier's text or apparatus is in the public JSON beyond single cited readings in the notes.
12. Open questions / to review
- Review the 12 misprint corrections and 11 doubtful readings against a second print (Cairo prints; Ṣalībā–ʿAyyād, Damascus 1935).
- The lacuna after passage 10: supply from a public-domain witness (Gauthier 1900, Cairo prints) if one is found.
- Find Gauthier 1900 (BnF has it: catalogue.bnf.fr ark:/12148/cb347277098; not digitised on Gallica as far as searched).
- Verify the verse attributions (Ibn al-Muʿtazz, al-Mutanabbī).
- Compare AI vowelling with G's partial vowelling.
13. Change log
- 2026-10-02: translation triangulation (Ockley–Fulton PD, Gauthier 1936 private): 8 serious divergences shown in the reader; 4 AI-English fixes (¶1, ¶6, ¶34, ¶81), 8 doubtful places listed (§10).
- 2026-09-28: record created; A pp. 55–77 transcribed, proofread twice (OpenITI/OCR; Gauthier 1936), 12 misprints and 5 parenthesis marks logged, 5 Quran quotations replaced; 85 passages written to
data/sources/hayy-tufayl-base.json+-meta.json; private collation with Gauthier (336 units). - 2026-09-28: prepare.py (85/85 vowelled), Tanzil restored in 8 words, base fields merged back; analyze.py (y1) and translate.py (opus, 85/85) run. translate.py also wrote
app/public/data/commentaries/hayy-tufayl.jsonwith its hard-coded Ibn Zayla entry (count 0), which does not apply to this text.