Repository path: workshop/translations/kalila-qird/collation.md · rendered 2026-09-09
Page metadata (front matter)
| type | note |
|---|---|
| id | kalila-qird-collation |
| status | frozen |
| created | 2026-08-16 |
| updated | 2026-08-16 |
| links | workshop/translations/kalila-qird/R38-v1/translation.md, workshop/translations/kalila-qird/source-wikisource-qird.txt, workshop/translations/kalila-qird/witness2-amiriyya-1937.txt, workshop/translations/kalila-qird/copy-text-adopted.txt, workshop/translations/kalila-labwa/collation.md |
Copy-text and collation — «باب القرد والغيلم»
1. The witnesses
| witness | provenance | state | |
|---|---|---|---|
| copy-text | Arabic Wikisource, كليلة ودمنة/باب القرد والغيلم |
no edition statement; unvocalised; modern orthography | 1,170 words of body text; clean typography, punctuation broken (§2), eight substantive losses (§3.1) |
| witness 2 | Arabic Wikisource, كليلة ودمنة (الأميرية، 1937)/باب القرد والغيلم |
Cairo: al-Maṭbaʿa al-Amīriyya (Būlāq) 1937, index كليلة ودمنة.pdf, pp. 234–242 (nine pages), page images present |
fully vocalised; optical recognition damaged; pagequality level="1" — not proofread |
The same pair of witnesses as the four earlier chapters (../kalila-saih/, ../kalila-nasik/,
../kalila-ibn-urs/, ../kalila-labwa/). The chapter page in witness 2 is a bare transclusion
({{صفحات|فهرس=كليلة ودمنة.pdf|من=234|إلى=242|إلى الفصل=ق-1}}) and returns an empty extract; the
text was taken from the nine صفحة: pages and cut at the chapter headings.
This collation reverses the polarity of the four earlier ones, and that is its main result. At
«باب السائح والصائغ», «باب الناسك والضيف», «باب الناسك وابن عرس» and «باب اللبؤة والإسوار والشغبر»,
witness 2 was the damaged witness and the copy-text was the clean one, and adoptions from witness 2
were few and cautious (five at kalila-labwa, and the collation there says so). Here witness 2 is
substantively fuller at eight places, and at four of them the copy-text has lost narrative the fable
does not work without. The recognition damage is exactly as bad as before; what differs is that
the damage is in the letters and the copy-text's losses are in the sentences, and a damaged
letter can be seen through where a missing clause cannot.
2. Sentence numbering — and why it does not follow the copy-text
The four earlier chapters number sentences by cutting the copy-text at ., ؟ and !. That
rule fails on this chapter and the failure is not marginal. The copy-text omits the full stop
before a great many قالs, so the cut yields 37 sentences of grossly uneven length — one of them
(ما احتباس الغيلم … through … ما الذي يحبسك؟) runs to 139 words and swallows four separate
speeches and a four-sentence maxim. Numbers cut that way cannot locate a figure.
Sentence division therefore follows witness 2's punctuation, which is a printed 1937 edition's
and is sound, and yields 73 sentences, [1]–[73] in copy-text-adopted.txt. Every citation on
the translation page and in the figure inventory uses these numbers. The divergence from the four
earlier chapters' practice is recorded here rather than left to be discovered: [n] on this
chapter is not commensurable with [n] on the other four, and no count that pools sentence
numbers across chapters is valid.
Body length after adoption: 1,201 words in 73 sentences.
3. The collation
Read whole, both witnesses, sentence by sentence. Three classes of divergence are separated: substantive losses in the copy-text (§3.1), lexical divergences where a reading had to be chosen (§3.2), and damage in witness 2 that is transparent and carries no variant (§3.3).
3.1 Eight substantive readings adopted from witness 2
Each is a place where the copy-text is either incomplete or does not construe, and witness 2 is whole. Adopted.
| # | sent | copy-text | witness 2 (adopted) | what it costs the copy-text |
|---|---|---|---|---|
| 1 | [4], [9] |
«فوثب عليه قرد من بيت المملكة»; «قد خفت أن يكون شاب قد عرض له عارض سوء» | «قِرْدٌ شَابٌّ مِنْ بَيْتِ المَمْلَكَةِ»; «قَدْ خِفْتُ أَنْ يَكُونَ قَدْ عَرَضَ لَهُ عَارِضُ سُوءٍ» | the same word, displaced by thirty lines. Young belongs to the usurping monkey in [4]; in the copy-text it has migrated into the tortoise-wife's fear in [9], where it is ungrammatical (يكون شاب with no predicate) and where it makes her say her husband has been waylaid by a young man. One transposition, two broken sentences |
| 2 | [20]–[21] |
«فإن ساكن في جزيرة طيبة الفاكهة. فركب ظهر الغيلم» | «فَإِنِّى سَاكِنٌ فِي جَزِيرَةٍ طَيِّبَةِ الْفَاكِهَةِ. فَارْكَبْ ظَهْرِى لِأَسْبَحَ بِكَ. فَرَغِبَ الْقِرْدُ فِي ذَلِكَ، وَنَزَلَ فَرَكِبَ ظَهْرَ الْغَيْلَمِ» | the invitation and its acceptance. In the copy-text the monkey is on the tortoise's back with no one having asked him and no one having agreed. Fourteen words |
| 3 | [44] |
«ثم رجع بالقرد إلى مكانه فلما أبطأ على الغيلم، ناداه» | «ثُمَّ رَجَعَ بِالْقِرْدِ إِلَى مَكَانِهِ. فَلَمَّا قَارَبَ السَّاحِلَ، وَثَبَ عَنْ ظَهْرِهِ، فَارْتَقَى الشَّجَرَةَ. فَلَمَّا أَبْطَأَ عَلَى الْغَيْلَمِ، نَادَاهُ» | the escape itself. The copy-text loses the leap off the tortoise's back and the climb up the fig tree — the event the whole fable exists to reach. Nine words |
| 4 | [63]–[64] |
«وأخذ طريقه إلى الأسد وأعلمه بمكانه وقال له: استعد له» | «وَأَخَذَ طَرِيقَهُ إِلَى الْأَسَدِ، فَسَبَقَهُ ابْنُ آوَى إِلَى الْأَسَدِ، وَأَعْلَمَهُ بِمَكَانِهِ» | who tells the lion. Without the jackal's overtaking, the copy-text's subject carries through and the ass informs the lion of his own whereabouts and tells him to get ready. Six words |
| 5 | [68] |
«فلا يأكل منه شيئاً، فقال لابن آوى» | «فَلَا يَأْكُلَ مِنْهُ شَيْئًا. ثُمَّ إِنَّ الْأَسَدَ رَجَعَ إِلَى مَكَانِهِ، فَقَالَ لِابْنِ آوَى» | the lion's return; without it the speaker of «أين قلب الحمار» is unidentified |
| 6 | [39] |
«ثم قال الغيلم: وما منعك أن تعلمني» | «هُمَّ قَالَ لِلْغَيْلَمِ: وَمَا مَنَعَكَ أَنْ تُعْلِمَنِي» | one letter, and it reverses the speech. The reproach and the invented custom of the left-behind heart are the monkey's; the copy-text gives them to the tortoise, who then asks himself why he was not warned |
| 7 | [59] |
«وأراد أن يثبت عليه» | «وَأَرَادَ أَنْ يَنبَ [= يَثِبَ] عَلَيْهِ» | to pounce on him, not to be established on him. The lion is springing |
| 8 | [24] |
«وذلك بمنعي من كثير مما أريد أن أبلغه من حرصك على كرامتك وملاطفتك» | «وَذَلِكَ يَمْنَعُنِي مِنْ كَثِيرٍ مِمَّا أُرِيدُ أَنْ أَبْلُغَهُ مِنْ كَرَامَتِكَ وَمُلَاطَفَتِكَ» | the copy-text's بمنعي does not construe, and حرصك على is caught from the next sentence, where it belongs |
Two of these — 1 and 8 — are the same accident: a phrase caught from a neighbouring sentence. Both are of the class an editor produces and a scribe does not.
3.2 Six places where a reading was chosen, and the copy-text kept
Witness 2 diverges but its reading is either optical damage or no better than the copy-text's. The copy-text is kept and the divergence recorded.
| sent | copy-text (kept) | witness 2 | why kept |
|---|---|---|---|
[9] |
«فجزعت عليه» | «فَزِعَتْ عَلَيْهِ» | both are words; grieved for him fits «وطالت غيبته» better than was alarmed, and ج/ز is a live confusion in this scan |
[17] |
«يا أخي» | «يَا أَبِي» | the two are peers who eat and drink together ([10]); my father is a dotting error for أخي |
[60] |
«أعجزت يا سيد السباع» | «أَتَجَرْتَ» | witness 2's form is not a word; the damage is ع→ت, ز→ج |
[72] |
«كالرجل الذي يعثر على الأرض» | «يَعْبُرُ» | stumbles on the ground then rises leaning on it is the figure; crosses is a ث→ب error |
[67] |
«رجاء أن يتطير الأسد منه» | «يَتَعَيَّرَ» | take it as an ill omen — the jackal's stated motive; witness 2's form is damage |
[65] |
«فجأش جأش الأسد» | «فَاشَ جَاشُ» with a marginal gloss «الجأش … من معانيه النفس» | both are damaged; فجاش جأش الأسد adopted as the reading both damaged forms point at — the lion's spirit surged. Witness 2's own footnote is what identifies it |
3.3 What witness 2's damage costs, and where it could be hiding a variant
The recognition profile is the one recorded at the four earlier chapters, with the same substitution classes: ت→د/ن (التين → الدين), غ→ع/ف (الغيلم → الفيلم / العيلم), ح→خ/ج (جرب → حرب), ض→ص (أضمر → أصمر), plus stray marginal footnote digits and dropped word-initial letters at line starts.
Two places where a variant could be hiding and nothing is adopted:
[48]witness 2 adds «وَجَهْدٌ» to «فأصاب الأسد جرب وضعف شديد». It is clean in the scan and it is adopted (§3.1 does not list it because it adds no event) — recorded here because a third witness could show it to be a gloss.[33]witness 2 reads «وَفِي وَقْتِ الحَاجَةِ» against the copy-text's «وفي الحاجة». Adopted; but «في الحاجة» in a list of four objects of expenditure is defensible. The choice does not touch the figure inventory — the member-end is «الحاجة» under either reading — and it is recorded because that had to be checked, not because it changed anything.
One limit stated rather than discovered later. Witness 2 is level="1" — not proofread by a
second Wikisource editor. Every adoption in §3.1 is a place where the underlying Arabic is
unambiguous through the damage and where the copy-text is defective on its own terms, which is the
standard the four earlier collations set. No adoption here rests on witness 2's authority against
a sound copy-text reading.
4. What this collation changes about the earlier four
Nothing about their texts. One thing about how they should be described: the Wikisource «كليلة ودمنة» text is not uniformly reliable across chapters. Four chapters gave it a clean record and this one shows it dropping nine-word and fourteen-word stretches and transposing adjectives between sentences thirty lines apart. The earlier collations' phrase "clean typography" was true of what it described and is not a warrant about the text as a whole.
Recorded here and not in a result page's limits section, because it is a fact about a copy-text, not about the project's instruments.