mouse&manuscript

Collation Lesson 13 - Dr. Dorrit van Dalen and dr. Peter Webb

The physical production of a new copy of a book in the Manuscript Age could occur in several ways. Lesson 12 explored the most elaborate process involving the oral reading and copying of a book in the presence of a group of scholars, but most manuscripts were produced in less formal settings. An individual could make a new copy himself of an older manuscript, or rent an older manuscript from a professional bookseller (warrāq) and copy it, or, most commonly, an individual or bookseller could commission a new copy to be made by a professional copyist (nāsikh or nassākh).

In all of the above cases, the production of a book involved the transcription of its text from an earlier copy over many hours of writing, hours in which the attention of a copyist could slack. The quality of copied manuscripts accordingly depended on both the quality of the exemplar from which the new book was copied (aṣl) and the care and accuracy of the copyist.

As discussed in Lesson 11, the most valued copies were those which could be shown to have been transcribed from either (i) the original aṣl written by the book's author himself, or (ii) a copy which bore a chain of transcription of copies that trace back to the author's original. In the clear majority of cases, however, copyists did not have access to the author's original or meticulously-copied transcriptions of the author's original, and instead copyists had to work from non-authorised exemplars. Sometimes copyists specify that they produced their version from an 'old copy' (nuskha qadīma), which presumably communicated a degree of authority, but in other cases, the identity of the exemplar aṣl is not mentioned.

Whatever the nature of the exemplar, copyists generally evidence that they double-checked their transcriptions in some way to rectify inadvertent mistakes (see Lesson 14 for more on scribal errors). Sometimes this checking was only performed by the copyist himself in a relatively informal process of double-checking what he wrote (taḥrīr), but better-quality manuscripts were those more carefully double-checked via a process known as collation: Ar. muqābala, muʿāraḍa, or ʿarḍ. Collation involved the checking of the new manuscript copy against the original exemplar aṣl and/or against other exemplars of the same book. This collation process could be undertaken by reading the new copy and the exemplar side-by-side, but in some fields, particularly hadith studies, the collation was ideally performed orally in auditions like those discussed in Lesson 12.

Exhortations to Collate
Collating was the mark of good scholarship, especially amongst hadith specialists. The eighth-century scholar of ḥadīth Yaḥya ibn Abī Kathīr is recorded to have remarked:

مثل الذى يكتب ولا يعارض مثل الذى يقضى حاجته ولم يستنج بلماء
"He who writes and does not collate is like one who uses the toilet and does not clean himself afterwards."

The emphasis on careful collation was more than a matter of jest: even several versions of a hadith (some ascribed inauthentically to the Prophet himself) circulated which remarked that a text written down but unchecked should not have been written. Another early hadith scholar ʿUmar ibn ʿUthmān al-Jinzī composed a short poem on collation:

Collate your writings after you’ve edited them
A text un-collated should be deemed un-written.
If you write, edit and collate,
Even an inexperienced fool can be trusted to read it.

(The above quotations are drawn from al-Samʿānī Adab al-imlāʾ wa-l-istimlāʾ (ed. Weisweiler, Brill 1952) pp.77-79).

The most pressing texts in need of careful collation were collections of hadith, given the symbolic importance of copying the Prophet's words correctly (see Lesson 12). While it is less common to find copies of texts of other disciplines collated and corrected in oral assemblies, muqābala/muʿāraḍa notes can also be found on a good number of careful (and valued) copies of an array of books from other fields too, and the principles of collation were applied in varying degrees beyond the field of hadith. This lesson examines a meticulously collated hadith manuscript to explore how the act of collation was marked and shaped the ways in which manuscripts were produced.

About this manuscript

Title: Al-musnad al-jāmiʿ (also Sunan al-Dārimī)

Classmark: Or. 364

Subject: Hadiths

Author: (Compiler) ʿAbd Allāh b. ʿAbd al-Raḥmān al-Dārimī (d. 255/869)

Copyist: ʿAbdallāh b. Muḥammad Yūsuf... ... b. Musāfir b. Jamīl al-Muqarrī al-Baghdādī

Format: Codex

Extent: 285 folios, leaf height 260 mm, width 175 mm

Date of creation: 634 /1237

Language(s): Arabic

Script: Naskh

See Lesson 11 for background on the al-Musnad al jāmiʿ by al-Dārimī. As noted in Lesson 12, this manuscript was produced from oral reading sessions, and, as the manuscript also evidences, it was re-read after its initial copying. Each of these readings were marked on the manuscript and corrections were made accordingly.

Noting Corrections
The aim of collation was to make corrections to the transcribed text and mark points of interest to scholars, and the ways in which marks and corrections were added to manuscripts followed conventions. Annotations were invariably made in the margins (ḥāshiya or hāmish), not in the main text block itself (the matn). The place in the text where something was found to have been omitted or misread is often indicated by a reference mark (ʿaṭfa), which often was written in a curvilinear form, or resembling a v or ٢). Throughout the Arabic manuscript tradition, corrections in the margin were usually accompanied by the word ṣaḥ صح , correct, or simply by ص.

Sahh 95v
صح Correction mark f. 95v

In the correction mark illustrated to the right, the original copyist wrote يحيى عبيد الله (personal names "Yahya 'Ubayd Allah") but the text should have been يحيى عن عبيد الله ("[from] Yahya on the authority of 'Ubayd Allah"). In the process of checking, this omission was noted, and the correct word to be added to the manuscript (عن) was written in the margin alongside the abbreviation صح, with a curvilinear ʿaṭfa mark written in the body text to indicate where the insertion should be made. A similar example is the text below, where the particle يا was left out in the original copy, and inserted in the margin with a note صح to indicate the correction, and its place in the body text is marked by a curvilinear ʿaṭfa

Sahh 98r
صح: f. 98r the corrected text should read يا نبي, the copyist omitted the يا

A range of other symbols might appear in the margins of manuscripts following collation. Symbols include marks noting that a word or claim was wrong ( غ for غلط) or should be inverse (ل for بدل) or to identify a marginal gloss ( ح for حاشية ). An example of the marginal gloss in the present manuscript is a mark ح to clarify the name يعلى which had been written a little unclearly by the copyist in the body text.

Sahh 101v
ح Marginal Clarification f. 101v

Another common marginal annotation produced by collation is خ: i.e. nuskha ukhrā (another version): this note is inscribed where the manuscript was checked against more than one older copy of the book, and via the comparisons, it was noted that a word or phrase was rendered differently between the older copies. By reflecting the variation in the margin, the new copy preserves both versions. In some cases, like the one illustrated below, the variations are extremely minor: the exemplar from which this manuscript was produced rendered the word ‘that’ ذلك, whereas another old copy was found to have rendered it ذاك, an alternative Arabic spelling of the same word. This variation has no effect on the text’s meaning, however meticulous scholars of hadith, who were keen to preserve all possibilities of the original wording, valued these marks.

Sahh 37v
خ 'Another version' f. 37v

Such annotations to marking slight text variations as recorded in different exemplars also appear commonly in poetry collections. In cases of more significant variation, marks of ‘another version’ can be very useful today to explore variations from manuscripts that no longer exist.

Collation Notes
It could take many sessions to read aloud a large manuscript to collate a new copy. Each time the collator stopped his work, he would mark the place, writing a sign to indicate the progress. Conventions usually employed one of the following phrases or abbreviations of them for these marks:

بلغ/بلغت معارضة - buligha(t) muʿāraḍatan - this point was reached in collating
بلغت مقابلة - buligha(t) muqābalatan - this point was reached in comparing
بلغ العرض - buligha al-ʿārḍ the collation reached here
قوبل - qūbila this has been compared
عورض - ūriḍa this has been collated, presented

(note the verb بلغ in the above examples was sometimes vocalised in the active "balagha" or the passive "buligha").

The marks served the copyist or circle of listeners when he/they resumed the work of collation, and the marks also served as a mark of the care in which they text was produced for subsequent readers of the codex.

Sometimes instead of writing the collation formulae, copyists used the symbol Ꙩ to indicate the point reached by a session of collation, but this was not universal practice, and the same symbol Ꙩ was also employed (as in the case of the present manuscript) to indicate any pause or distinction between parts of the text, much as modern English paragraphing functions today).

Collation 1
The second collation note

Collation notes in al-Dārimī's al-Musnad

According to notes next to the colophon on f283v, this manuscript was collated at least two times (i.e. submitted to muqābala with other copies). It was read once against the original aṣl (the manuscript of Abū al-Waqt): according to the note, this was done “as well as possible” (ʿalā ḥasab al-imkān), perhaps suggesting that part of this manuscript was illegible, missing, or it is a mark of humility. The second collation is noted with a more detailed note:
قابلته جمعاء بنسخة الحافظ ضياء الدين المقدسي فكل ما عليه ض فهو منها كتبه محمد بن أبي الفتح الحنبلي في ذي الحجة سنة ست و[؟]
“I collated all of this with the copy of al-Ḥāfiẓ Ḍiyāʾ al-Dīn al-Maqdisī; any text marked with a letter ḍad is from that copy. Muḥammad ibn Abī al-Fatḥ al-Ḥanbalī wrote this note in the month of Dhū al-Ḥijja in ??6” (the date is difficult to decipher)

Each of these muqābala collations resulted in additions and observations of ‘different versions’ in the text marked by the abbreviation خ, however, there are several different forms in which this marginal note appears in the manuscript. These different forms may indicate separate instances of collation, or the copyist may not have been consistent in how he marked the different versions. See Question 1 for more on these marks.

Collation 2
بلغ العرض

The most frequent marks to indicate the checking of this manuscript are written بلغ العرض. This refers to the public presentation of the manuscript before an assembly (the details of these auditions are discussed in Lesson 12). The spacing of these marginal notes give us an indication of the pace by which pre-modern scholars were able to check manuscripts in one sitting (see Assignment 1).

Other marks refer to muqābala comparison with other copies; these marks sometimes specify whether the collation was from an oral audition or done silently by comparing two manuscripts together, and sometimes add further information about the sources collated.

Collation 3 Collation 3
Collation note f. 19v

For example, the note illustrated to the right reads:
بلغت مقابلة بالأصل المقروء على يد اللتي
“This is a point where al-Latī’s collation read against the original reached”

The note refers to the reading of the early exemplar of al-Dārimī’s al-Musnad in the presence of the Abū al-Munajjā ʿAbd Allāh ibn ʿUmar ibn ʿAlī ibn Zayd ibn al-Latī, and refers to his collation notes (see Lesson 12 for the audition note referring to this scholar).

There are various other detailed marks of collation. Consider another on f. 47v:

Collation 5 Collation 5
f. 47v

سقط من هنا إلى آخر الباب من الأصل المسموع على يد اللتي بجبل قاسيون وقفه الحافظ ضياء الدين

The note refers to a reading done on Mt Qāsiyūn in Damascus, where the version which al-Latī produced in oral audition did not contain a series of hadiths from the point of this note to the end of the chapter; those hadiths were related in the other recension of Ḍiyāʾ al-Dīn al-Maqdisī (the version referred to in the second muqābala note above).

A final collation note to consider is on f. 88r, as it records another common marginal formula in manuscripts:

Collation 4
f. 88r

This note reads: بلغ قراءة رضي الدين وجماعته سماعًا
“Collation read orally from Raḍī al-Dīn and his group”, this refers to the version of al-Dārimī’s al-Musnad narrated by Raḍī al-Dīn al-Muʿāfā, noted in the transmission of the book (see Lesson 11). It specifies that the reading was done before an assembly. See also 98r for another note of this version.

For other collation notes in this manuscript see the Questions and Assignments.

More about this manuscript: lessons 3, 11, 12.

Further reading
A. Gacek, Arabic Manuscripts, a Vademecum for Readers. Brill 2009, 65 ff.
F. Rosenthal, The Technique and Approach of Muslim Scholarship. Analecta Orientalia 24, Rome 1947, 1- 48.
Al-Samʿānī, ʿAbd al-Karīm b. Muḥammad/M. Weismüller, Adab al-imlāʾ waʾl-istimlāʾ. Leiden 1952.

Questions

  1. As noted in the lesson, when readers compared the manuscript against different exemplars and noted variations in old texts, they affixed the symbol خ to indicate the variations in the underlying texts. There are, however, several different kinds of annotations for different versions – the abbreviation خ is written differently or with combinations of letters – these may indicate separate exercises of collation with different manuscripts. Look through the marginal notes in the manuscript, and identify how many different forms of these ‘variation’ notes you can find. What might the differences signify?

  2. Look at f. 59v, what does the lower annotation in the right margin say, and what does this indicate?

  3. Examine the marginal note in the upper-left margin of f. 18r, the two marginal notes in the lower-right margin on f. 26v, and the left margin of f. 27r. What do they say, and why were they added?

Assignments

  1. Leaf through the manuscript and write down where you see that presentation was paused (i.e. marked with the بلغ العرض marginal note). How many pages were presented at a time, on average? What is the longest stretch between the marked pauses?

  2. Examine the marginal notes on f. 25r and v. Transcribe as much of them as possible and explain what they indicate.