Added on 5 October 2026 as the page-image pilot (docs/design/PAGE-IMAGES.md) and the start of the Dīvān odes (roadmap 4d). The first poem is the qaṣīda on speech, tanzīl and taʾwīl that the author especially likes: «ای گنبد گردندهٔ بی روزن خضرا». Built as the Rawshanāʾī-nāma (docs/texts/rawshanai-nama.md), with the same Persian-verse reader.
| Work / author / dates | Dīvān, Nāṣir-i Khusraw (b. 394/1004, d. after 465/1073) |
| Part added | Qaṣīda 2 (Mīnuvī–Muḥaqqiq no. 2, p. 5; the first poem of the Taqavī edition and of the 1993 Tehran printing; Ganjoor's no. 1), rhyme in -ā, 54 couplets, complete: Taqavī pp. 2–4 (PDF pp. 118–120), from its first couplet (p. 2, l. 4, under the bismillah and the metre heading) to its last (p. 4, l. 15; the next poem, in mutaqārib, starts below it) |
| Reader route | /read/nasir-divan · about /read/nasir-divan/about · poem #q2 · couplets #v-2001 … #v-2054 (key = 1000 × the poem's Mīnuvī–Muḥaqqiq number + the couplet; the reader shows the couplet's own number) |
| Catalogue kind | ismaili, beside the Rawshanāʾī-nāma: an Ismaili philosophical-didactic ode (taʾwīl as the pearl only the diver brings up, the Prophet's family, Fāṭima at the Judgment). poetry holds the love and wine odes; the kinds are by tradition first. |
| Status | working edition · transcription AI-read from the page images, human proofreading pending · reviewed by: — |
1. Edition, digital source, copyright
- Base: the Taqavī edition. Dīvān-i qaṣāyid u muqaṭṭaʿāt-i Ḥakīm Nāṣir-i Khusraw, bi-ḍamīma-yi Rawshanāʾī-nāma u Saʿādat-nāma u risāla-ʾī bi-nathr, ed. Sayyid Naṣr Allāh Taqavī, introduction by Taqīzāda, appendix by Mīnuvī (Tehran: Majlis Press, for the Tehran library, 1304–07/1925–28). PD-US. archive.org
DivaanNaserXosrowDivanEQasaaidOMuqattaaatEHakeemNasirKhusrauFarsi; localdata/raw/persian/nasir-khusraw/divan_taqavi-tehran-1925-28.pdf. PDF page = printed page + 116; archive.org leaf n = printed page + 115 (checked against the item'spage_numbers.json: leaves are 0-based, one per PDF page). The archive.org OCR is unusable and its offsets do not match: the pages were rendered from the PDF. - Layout. The edition numbers lines per page, every fifth (p. 2 counts its three heading lines, so the poem starts at l. 4). Notes at the foot of the page: numbers (1), (2) … are glosses; Arabic letters (ب), (د) … are variant readings.
- This copy carries a reader's handwritten marginalia (Persian notes in the margins and between the lines, a few words gone over in ink). They are not transcribed; the apparatus says where they touch a reading (vv. 21, 28, 45).
- The introduction (Taqīzāda) and appendix (Mīnuvī) are not reproduced. The edition's notes are short glosses and variants, given in the apparatus with their Persian (PD-US) and our English; the long note on Ghumdān (v. 23) is summarised.
2. Transcription and apparatus
- Source file:
scripts/persian/taqavi-1925-divan.tsv(couplet · hemistichs · page · line · flag), with the poem's section (#S q2) and the edition's heading (#H, the metre). - Method: each page rendered at 300 dpi and read in five bands by AI (Claude); every word follows the image.
- Normalisation: word spacing (the print's ایفلک, ایمادر, درینخانه, همیکرد, اینمرکز split); ی/ي, ک/ك; vowel marks dropped; ـهٔ kept; joined prepositions kept (بسخن, بجوزا); «واشفته» kept as one word (v. 48, the metre reads v-āshufta).
- Apparatus,
scripts/persian/nasir-divan-apparatus.json, shown under the couplets, each line with "See the print" (a crop of the page):- the edition's notes (1)–(7) on pp. 2–4 and its variants (ب) v. 18 فروزان, (د) v. 28 بماند, (ب) v. 36 پر گوهر و یاقوت و پر از; a note (ج) on p. 3, «در نسخهٔ بدل این شعر بنظر آمد», is struck through in the print itself, with its mark near v. 21;
- doubtful readings (flag
?): v. 6 (the last word under an ink blot), v. 18 بروزان, v. 20 (first hemistich short for the metre in both printings), v. 21 (the mark after خوب), v. 28 بگرد, v. 29 (the print's کشتنت read گشتنت), v. 45 (words gone over in ink); - v. 27 «بجوزا»: Gemini, not Thurayyā. The Pleiades are named in v. 18 (the sun alone is better than the seven Pleiades).
- Second witness, the 1993 Tehran printing (
divan_archive-DivanENasirKhusroFarsi.pdf, printed pp. 53–55 = PDF pp. 52–54; ed. Jahāngīr Manṣūr, in the Taqavī tradition, copyrighted): collated in full, privately; only its variant words are published. Same 54 couplets; vv. 11–12 in the other order. ACQUISITIONS.md placed جانت بسخن at its v. 28: it is v. 27 there, as here. Variants at vv. 1 (قبّهٔ for گنبد; و با — with it the hemistich scans), 2, 4, 6, 7 (دیبای سخن for دیبای خرد), 11–12 (order), 13, 14, 16, 17, 18 (a different hemistich), 20, 21, 22, 24 (دیوی است), 28 (نماند), 30, 31, 37 (صورت تنزیل), 39 (دارندهٔ دنیا), 46, 47 (کرّهٔ غبرا), 48 (بیخردان, مهیّا), 49, 50, 51: 25 couplets differ in at least a word, besides the order of vv. 11–12. - Third witness, Browne's Persian (JRAS 1905, pp. 337–338, after the Tabriz lithograph of 1280/1864; public domain): 12 couplets, vv. 36–42, 49–50, 52–54. It supports Taqavī's own variant at v. 36 (پر گوهر و یاقوت پر از), and reads ارزندهٔ زیبا (v. 39), خورسند (42), مولا … مولا (50), حول (53).
- To do: a human proofreading against the images; Mīnuvī–Muḥaqqiq collated in full (copyrighted: privately; only v. 27 checked).
3. Layers
| Layer | Source | Status |
|---|---|---|
| Persian | Taqavī 1925–28, our transcription | PD-US base; ours |
| English, per couplet | AI (Claude), scripts/persian/ai/nasir-divan-en.tsv, written from the Taqavī Persian; compared afterwards with Browne and Azadpur (§5) | AI, marked |
| Word glosses | AI (Claude), scripts/persian/ai/nasir-divan-glosses.tsv: 422 written forms; function words reuse the other Persian texts' glosses, the poem's own words glossed afresh in its sense | AI, marked |
| Transliteration | AI (Claude), scripts/persian/ai/nasir-divan-tr.tsv, IJMES Persian (§3a) | AI, marked; 89/108 hemistichs scan |
| Browne's English | JRAS 1905 p. 347, scripts/persian/browne-1905-en.tsv, 12 couplets | public domain; Aa → Browne, off by default |
| Page images | scripts/page_images.py → app/public/img/pages/nasir-divan/, manifest app/public/data/pages/nasir-divan.json | Taqavī pp. 2–4, Browne pp. 337–338; 10 crops |
Build: python3 scripts/persian/build_nasir_divan.py → app/src/data/nasir-divan.json; python3 scripts/page_images.py nasir-divan; python3 scripts/build_connections.py nasir-divan.
3a. Transliteration and metre
The edition names the metre in its heading: hazaj musamman akhrab makfūf maḥdhūf, mafʿūlu mafāʿīlu mafāʿīlu faʿūlun (– – u | u – – u | u – – u | u – –). Added to scripts/persian/scan_metre.py as hazaj-akhrab (14 positions; the scanner's length is now per metre). ʿAṭṭār 549/564, the Rawshanāʾī-nāma 362/368 and the Gulshan 1982/2016 are unchanged. Readings the metre decided: v. 3 tīra tan-i khāmush-i khākī-st; v. 8 na-bīnad-sh (the 1993 printing's sukūn); v. 11–12 ārizū; v. 36 khudāy-ast; v. 48 v-āshufta; v. 54 bid-had.
Result: 89 of 108 hemistichs scan. Of the 19 that do not, 15 scan with taskīn (one long for the two shorts between longs, the licence of the akhrab feet in the rubāʿī); it is tried only as a diagnostic (hazaj-akhrab-taskin), never as the check. Whether Nāṣir allows it in this metre, or the lines are to be read otherwise, is for a prosodist to say: AI observation, to review. Marked † in the reader:
| Hemistich | As read | Note |
|---|---|---|
| 1.2 | bā qāmat-i fartūtī bā quvvat-i burnā | taskīn; with the 1993 printing's «و» (fartūtī u bā) it scans strictly |
| 7.2 | hargiz na-shavad ay pisar az dībā zībā | taskīn (in the last feet) |
| 13.1 | āzār magīr az kas bar-khīra u ma-āzār | does not scan |
| 14.1 | pur-kīna ma-bāsh az hamagān dāyim chūn khār | taskīn |
| 15.1 | k-az gand futād ast ba-chāh andar sargīn | taskīn |
| 16.1 | bā har kas ma-nshīn u ma-bur az hamagān nīz | taskīn |
| 17.1, 17.2 | chūn yār-i muvāfiq na-buvad tanhā bihtar / tanhā bih ṣad bār chu nādān-at hamtā | taskīn |
| 20.1 | aḥvāl-i jahān guzaranda -st | does not scan: too short in both printings |
| 20.2 | sarmā sipas-i garmā sarrā pas-i ḍarrā | taskīn |
| 28.1, 30.1, 33.2, 35.2 | … ba-sukhan bāyad / gardad … | taskīn, all at the same place (ba-sukhan bāyad, ba-sukhan gardad) |
| 29.1 | zinda ba-sukhan bāyad gashtan-at az-īrā-k | does not scan |
| 30.2 | dar ʿālam kas bī sukhan paydā paydā | does not scan |
| 45.1, 45.2 | dar zuhd na-ʾī bīnā … / bar khvānī dar chāh … | taskīn |
| 52.2 | ham ẓālim u ham ʿādil bī hīch muḥābā | taskīn |
4. Notes on the text
- The sea of speech (vv. 36–42). God's word is a sea; its letter (tanzīl) is salt, its inner meaning (taʾwīl) the pearl; seek the diver (ghavvāṣ), not the shore; the Prophet was told to give the taʾwīl to the wise and the tanzīl to the crowd; the diver gives brine to one who shows him enmity. In Ismaili teaching the diver is the Imam of the time, the keeper of taʾwīl (Browne 1905, p. 335, on the Dīvān at large).
- Speech (vv. 27–35): the soul is purified and raised "from the well to Gemini" by speech; Jesus raised the dead by speech (v. 29).
- vv. 1–8: the sphere as a loveless mother of the body only; the speaking soul (jawhar-i gūyā) is not her child.
- vv. 49–54: Darius leaves everything; the Day of Judgment; the poet holds the hem of al-Zahrāʾ (Fāṭima) against the enemies of the Prophet's family.
- Quran (allusions, checked by hand; the poem is Persian, so
scripts/quran_refs.pywas not run): v. 11 Adam and Eve (Q 2:35, 7:19, 20:120); v. 20 sarrā … ḍarrā (Q 3:134); v. 29 Jesus raising the dead (Q 3:49, 5:110); v. 43 the Night of Power (Q 97:1); v. 51 maljā (Q 42:47). Shown in the connections panel.
5. Triangulation (§17)
By hand, app/src/content/triangulation/nasir-divan.json. Translations held: Browne (12 couplets, public domain) and Azadpur (vv. 27, 36–40; in Pearls of Persia, 2013, pp. 82–83, and the preprint (private copy); copyrighted: read privately, described in our words, at most four words quoted). They agree with each other and with our English on the sense of every couplet they cover, except:
| v. | Cause | Readings | Note |
|---|---|---|---|
| 27 | term | ours: Gemini · Azadpur: "to the stars" (the 2013 printing: the constellations) | the text names Gemini (Jawzā), not the Pleiades |
| 40 | syntax | ours and Azadpur: God tells the Prophet · Browne: the Prophet himself makes the inner sense the portion of the wise | «ورا گفت» has no subject |
| 40 | interpretation | ours: to the crowd · Azadpur: to the people · Browne: the letter a "Rock of Offence" | ghawghā |
AI-English check (aiCheck): no contradiction. Doubtful, for review: v. 4 (Taqavī's «خانه و» against v. 3), v. 28 (the doubtful word).
6. Open questions
- Human proofreading; the seven doubtful readings; Mīnuvī–Muḥaqqiq in full.
- The metre: taskīn in this qaṣīda, and the four hemistichs that do not scan at all (13.1, 20.1, 29.1, 30.2).
- The next qaṣīdas (Browne's nos. 2–5 are Mīnuvī–Muḥaqqiq's 3–6?: to check); Persian lemmatisation for the word panel.
7. Change log
- 2026-10-05: qaṣīda 2 transcribed from the Taqavī page images (54 couplets, pp. 2–4) with the edition's notes and variants; collated with the 1993 printing (variant words only) and Browne's Persian; the
hazaj-akhrabmetre in the scanner; transliteration (89/108 scan), glosses (422 forms), AI English; Browne's English; triangulation; page images (Taqavī pp. 2–4, Browne pp. 337–338, 10 crops); reader/read/nasir-divan, catalogue kindismaili.