Share this:

François Déroche, with contributions by Annie Berthier, Marie-Geneviève Guesdon, Bernard Guineau, Francis Richard, Annie Vernay-Nouri, Jean Vezin, Muhammad Isa Waley

Article contents:
The aims and methods of palaeography
Empirical knowledge of scripts
Early ‘typologiesʼ
The palaeographic method
Applications of palaeography
Arabic book hands: preliminary observations*
Difficulties to date*
The working tools: albums of examples in facsimile*
Theoretical research*
Future lines of research
Type NS III script on parchment; datable as fourth/ tenth century. Paris, BNF 5178 d, f. 6.

Apart from some remarks on the most obvious peculiarities of the scripts, I have had to refrain from giving a palaeographical analysis of these specimens. This is due to the fact that in this field even the most basic work still remains to be done. No adequate criteria for the description of Arabic handwriting do as yet exist. The mere fact that all scripts presented here may conveniently be called naskhi proves that this name is hardly of any use and might just as well be discarded. Not even the roughest guidelines as to provenance and dating of Arabic manuscripts have been drawn, and this work can not be done without taking Persian and Turkish palaeography into account.

J.J. Witkam, Seυen Specimens of Arabic Manuscripts (Leiden, ɪ978), p. ɪ8

As briefly mentioned in the introduction, the palaeography of Arabic book hands cannot be divorced from codicology. We have yet to see, however, the development of a serious and coherent body of research in this field. Hence in this chapter we shall simply attempt to familiarise readers with the aims and methods of palaeography, to offer an assessment of existing research, and to indicate potential lines of further investigation.

The aims and methods of palaeography

The word palaeography is roughly contemporaneous with the birth of the field, defined as ‘the science or art of deciphering and determining the date of ancient documents or systems of writingʼ.[1] It was used, in fact, in the title of a book by a French Benedictine scholar, Bernard de Montfaucon, De palaeographia graeca.[2] This work was published in ɪ7o8, some thirty years after Jean Mabillonʼs De re diplomatica,[3] which latter can be considered the foundation stone of palaeography. It was not long before the term was adopted by Arabists; J. G. C. Adler used it in the late eighteenth century when discussing the Kufic coins he published in his Museum cuficum borgianum Velitris.[4]

Empirical knowledge of scripts

Palaeography was nevertheless not a totally unknown practice prior to the publication of the books mentioned above. Reading or copying ancient manuscripts implied a practical knowledge of early scripts on the part of readers and copyists. This deciphering ability is attested to a certain extent, for example, by readersʼ notes added at a date posterior to the manuscript itself. In a somewhat different way, a copy of the Qurʼān (Istanbul TKS H.S. ɪ7) demonstrates the existence of empirical palaeographical knowledge the text has been transcribed in a modern hand in the margin.[5] As Jan Just Witkam has noted, the history of the text of Ibn Ḥazmʼs Ṭa𝓌q al-ḥamāma supposes that at some point a scribe transcribed an original Maghribī version of this work in an Oriental hand.[6] Deciphering old manuscripts did not, on the whole, require particularly advanced technical knowledge, however. In particular, the fact that Arab and Muslim scribes generally made little use of abbreviations in their standard practice (and even then, only at the periphery of the text), obviated many difficulties in reading manuscripts despite the changes in book hands. Reading, in short, normally required nothing more than a good knowledge of the language or languages. Exceptions to this rule include texts written very poorly, whether because of haste or for other reasons (e.g. for the copyistʼs own use), and those written without the dots by which various Arabic letters of similar form are normally differentiated.

Early ‘typologiesʼ

Readers of manuscripts in archaic writing also possessed an empirical typology of early hands. On occasion they might give specific names to these hands, thereby doing the work of the palaeographer, although their meaning may not always be clear. The scribe of MS. Paris BNF arabe ɪ67 specified that the text he copied was originally ‘in Iraqi script (bi-khaṭṭ ‘irāqī)ʼ.[7] The compiler(s) of the mediaeval inventory of the library in Kairouan made a distinction between ‘Kuficʼ, ‘Sicilianʼ, Nabārī and ‘Orientalʼ scripts.[8] Throughout all the various manuscript traditions, the technical nature of handwriting gave rise to treatises on equipment (reed pen, ink, preparation of the writing surface, etc.), and these treatises sometimes touched on the various kinds of hand in use. The special status of writing in Islamic civilisation favoured the emergence of texts that also discussed the practice of ‘calligraphyʼ, including not only the names of scripts but sometimes also examples. Some of these names referred to the hands used by contemporary scribes and calligraphers, and cannot therefore be associated with palaeographical knowledge; others, by contrast, referred to scripts that had fallen into disuse in the more or less distant past, thereby conforming fully to the ‘art of deciphering ancient systems of writing.ʼ

At the same time, the production of forgeries of archaic scripts, as seen in the pseudo-al-Aṣma‘ī (MS. Paris BNF arabe 6726[9]), confirms that such knowledge was disseminated, a fact corroborated by the attribution of archaic Qurʼāns or Qurʼānic fragments to leading figures of the early days of Islam.[10] Despite the highly imprecise nature of this traditional brand of typology, it does represent an early form of empirical palaeography. Care should be taken when applying concepts generated by this practice, if only for reasons related to the transmission of texts, as will be discussed below; but some of them remain to a certain extent functional - the names Maghribī and nasta‘līq, for instance, are still effective labels for scripts.[11] Even a term as vague as naskhī provides an approximate idea of the type of script it relates to. The various names used in the Islamic world, long known to specialists in Arabic, make it possible to discuss these scripts, and can even serve as a typology, however imperfect, for anyone prepared to set aside the niceties of chronological or historical accuracy.

The palaeographic method

Compared to traditional knowledge of this kind, the breakthrough associated with the work of Jean Mabillon and Bernard de Montfaucon involved the introduction of expertise on scripts into the realm of the historical sciences in the form of an auxiliary branch of history. Palaeography, which partially adopted some of the aims of the empirical approach described above, gave a new and much wider scope to the study of ancient writing. Not only did palaeography endeavour to provide a solid basis for deciphering ancient hand-written texts; it also aspired to lay solid foundations for establishing their authenticity, age, and origin. This goal obviously presupposes that the script in question underwent evolution through time and across geographical distance, which is empirically acknowledged to be the case with writing in the Arabic alphabet.

A crucial first step in the methodology of palaeography must therefore be to establish a system of classification. In order to develop a typology of the various types of writing, palaeographers start by assembling a series documents preferably dated or datable, displaying similar graphical features; ideally, these documents should also contain reliable evidence of their geographical origin. Once the documentation has been gathered, it must be critically examined with a view to eliminating any items alien to the group. Since this initial phase (the constitution of series) is conducted empirically, its results must be carefully screened. Once this stage has been completed, a palaeographer can define the characteristics of the script, on the level both of overall appearance and individual letterforms. The next step is to set chronological - and, where possible, geographical - limits for each series documents from clues supplied by colophons, notes (𝓌aqfiyyas or deeds of endowment, reading certificates, etc.), or material details of the production of the manuscript. The scope of a given series has a bearing on its validity: the greater the number of manuscripts written at roughly the same time in the same region, displaying the same graphical features, the less ad hoc the set constituted. Scripts from periods and regions for which only limited documentation exists cannot be identified with the same degree of accuracy as those for which examples abound in various collections.

Applications of palaeography

Once a framework has been set up by this general palaeographic approach, the palaeographer is theoretically in a position to apply his or her expertise to early scripts. The first task is to assign a date and geographical origin to a document, x, that bears no direct information as to date or place of origin. The palaeographer must therefore first determine the type of script under study and assign it to a group, y, whose chronological and geographical parameters are already known; in theory, if document x displays graphical features identical to group y, then it also shares that groupʼs chronological and geographical characteristics. Conducting this comparative study with due care, however, requires that one consider the various factors that might militate against the validity of the theory. A scribe might have travelled from place to place, lived to a great age (thereby preserving a style that had otherwise fallen into disuse), or even revived other styles after a gap of generations (sometimes with fraudulent intent). A thorough comparative study must therefore include other aspects of codicology. To take an extreme - and perhaps overly simple - case, a palaeographer would note that the early ‘Abbāsid script found on ff. 4 and 5 of MS Paris BNF arabe 580 is incompatible, from a chronological standpoint, with the writing material, namely a Western watermarked paper dating from the eighteenth century.[12]

As the preceding paragraph makes clear, the palaeographerʼs work is based largely on comparative analysis. The search for dated documents displaying the script under study is therefore of prime importance. Physical familiarity with manuscripts themselves is not to be underestimated, because it often helps to improve the precision of analysis, but any libraryʼs collections has its limits; fortunately, technology that emerged during the nineteenth century has extended the amount of documentation available for the search for parallels and today the scope is even greater. No less than any other field, palaeography has benefited from technical progress, and is perhaps even more dependent than most on such advances, especially in the realm of the reproduction of documents and image processing. The two centuries since the birth of palaeography have witnessed the replacement of hand-made copies - sometimes the only visual record - by photographs, later abetted by photocopies, and subsequently by computer technology (whose future benefit to palaeography remains to be seen).[13] Thus the quality and quantity of documentation available to palaeographers has grown considerably, even though -as will be discussed below - the study of manuscripts in Arabic script is still a long way from having fully exploited this potential. Also worth pursuing are the laboratory techniques that have increased in number and reliability over the years. Palaeographers, then, are specialists whose knowledge and methods should in theory make it possible to identify the date and origin of a given piece of writing. But their attributions are more wide-ranging, because they are also historians of writing. Part of their work involves intensive study of the evolution of scripts and placing them in their historical and geographical context, taking into account the various factors which may have influenced their development. The work of classification and analysis briefly described here thus serves as the basis for an historical overview.[14]

Arabic book hands: preliminary observations...

*The remainder of this article is exclusively available in the printed version of the related book. The book is available in both electronic and printed formats within Our Publications in the following link:

[1] As defined in the Shorter Oxford English Dictionary.

[2] B. de Montfaucon, Palæographia græca, siυe de ortu et progressu literarum græcarum et de υariis omnium sæculorum scriptionis græcæ generibus (Paris, 1708).

[3] J. Mabillon, De Re diplomatica, libri VI, in quibus quidquid ad υeterum instrumentorum antiquitatem, materiam, scripturam et stilum; quidquid ad sigilla, monogrammata, subscriptiones ac notas chronologicas; quidquid inde ad antiquitariam, historicam forensemque disciplinam pertinet explicatur et illustratur… (Paris, 1681). As the title suggests, even at this early date there was a clear concern to discuss not only the scripts but material factors too.

[4] J.G.C. Adler, Museum cuficum borgianum Velitris (Rome, 1782), p. 32.

[5] F.E. Karatay, Topkapı Sarayı Müzesi Kütüphanesi Arapça yazmalar kataloğu (Istanbul, 1962), vol. I, p. 22, no. 64.

[6] J.J. Witkam, ‘Establishing the stemma: fact or fiction?’, MME 3 (1988), p. 90.

[7] FiMMOD 30.

[8] I. Chabbouh, ‘Sijill qadīm li-maktabat jāmi‘ al-Qayra𝓌ān’, RIMA I/2 (1376/1956): see, for example, pp. 346 (sharqī), 347 (nabārī), 366 (ṣiqillī) and passim (kūfī , or kūfī rayḥānī, p. 346).

[9] G. Vajda, Album de paléographie arabe (Paris, 1958), pl. 3; F. Déroche, ‘A propos du manuscrit “arabe 6726”, Bibliothèque nationale, Paris (al-Aṣma‘ī, Ta’rīḫ mulûk  al-‘Arab al-a𝓌𝓌alîn)’, REI 58 (1990), pp. 209-215.

[10] Ṣ. al-Munajjid compiled many attributions of this kind in his book on early scripts, Dirāsāt fi ta’rikh alkhaṭṭ al-‘Arabī (Beirut, 1972), pp. 50–60, 64–76.

[11] It is nevertheless important to remain alert to regional variations in the use of these names; for example, nasta‘līq is generally known in Turkey as ta‘līq (talik), though nestalik is also found.

[12] The folios are facsimiles of the original leaves, held in Copenhagen. See Déroche, Cat. I/1, p. 157, no. 294.

[13] The possibility of automatic analysis of scripts was discussed as early as the 1970s. For one example, see C. Sirat, L’examen des écritures: l’œil et la machine, essai de méthodologie (Paris, 1981). More recently, E. Rezvan and N.S. Kondybaev have pursued developments along these lines: ’New tool for analysis of handwritten script’, Manuscripta Orientalia II/3 (1996), pp. 43–53; ‘The ENTRAP software: test results’, Manuscripta Orientalia V/2 (1999), pp. 58–64.

[14] This seems to be the primary meaning of the term as used by al-Munajjid, op.cit., p. 7: ‘ilm al-bāliyūghrāfiya, a𝓌 ‘ilm taṭa𝓌𝓌ur alkhaṭṭ  al-‘Arabī.

Source note:
This article was published in the following book:
Islamic Codicology: An Introduction to the Study of Manuscripts in Arabic Script _ English version, 2005, Al-Furqan Islamic Heritage Foundation, London, UK, pp. 205-224.
Back to Top