M.J De Goeje and the editing of Arabic geographical texts

Share this:

Jan Just Witkam

I wish to thank Dr Arnoud Vrolijk for his help and support during the final stages of the composition of this chapter.

When on Sunday 25th August 1889 two Egyptian gentlemen, ʿAbdallāh Fikrī Pasha and Muḥammad Amīn Fikrī Bey, father and son, visited Leiden in the Netherlands, they were on their way from Egypt, through Italy, France and England, to Stockholm and Christiania (Oslo), to represent Egypt at the Eighth International Congress of Orientalists.[1] The old Pasha, who had previously held the post of Nāẓir al-Maʿārif al-Miṣrīyah (Inspector of Egyptian Education) was an accomplished poet and litérateur. He presented a sample of his own commentary on a qaṣidah by Ḥasān ibn Thābit[2] to the congress while his son Muḥammad Amīn Fikrī presented a paper on problems connected with the use of spoken Arabic rather than standard Arabic.[3] Muḥammad Fikrī subsequently completed and published an account of their travels since his father, who had hoped to do so, died soon after their return to Egypt.[4]

They had come to Leiden - a long-established centre of Oriental Studies - for several reasons. Its university library housed, and still houses, an important and ever-increasing collection of Islamic manuscripts; its Faculty of Letters had in the past counted many famous names in Oriental Studies among its members and last, but not least, Leiden was home to the publishing house of E.J. Brill, which had already published many important edited texts, some of which remain in use today. These facts qualified, and continue to qualify, Leiden as a favourite venue for scholarly tourism, and that was the specific purpose of the Egyptians’ visit. However, once they had settled in their hotel, the Lion d’Or in the Breestraat, they received bad news. They were told that, as it was Sunday, the library and university were closed, as was ‘La Maison Brill’. Seeing the disappointment on their faces, the manager of the hotel told them he would try to find a solution and dispatched an employee on an errand. The man returned with the Professor of Arabic, M.J. de Goeje, who told the relieved Egyptians that he would, in an hour’s time, organise a visit to the library and later take them to Brill’s as well.[5]

Thanks to their hotel manager, they had met the very person who embodied the three points of interest that attracted so many learned tourists to Leiden. Michael Jan de Goeje (1836-1909) had been Professor of Arabic in Leiden since 1869, he was the honorary curator of the Oriental Collection in the University Library,[6] he was the co-author of the Catalogue of Oriental and Arabic Manuscripts in the Dutch collections,[7] while his scholarly editions were largely published by Messrs Brill, whose firm was experiencing unprecedented prosperity, in part because of de Goeje’s editorial projects. De Goeje was the typical exponent of the philological and editorial period in Middle Eastern Studies, when the numerous sources of our knowledge of Arabic and Islamic culture in European libraries were made available to the public for the first time through detailed catalogues, and elucidated in a large number of edited texts that could stand the test of criticism.

It was also the time in which the Lachmannian method of editing texts was gradually adopted. In the academic world of the first half of the nineteenth century, there had been a widespread revival of textual criticism adopting new methodologies, most of which were associated with the work of the German philologist and scholar of classical and Old-Germanic Studies, Karl Lachmann (1793-1851). Lachmann devised a method of textual criticism that was subsequently named after him. It can be summarised as an attempt to reconstruct a text in a version as close as possible to the supposed original of the author by removing the mistakes and interpretations of later periods that had become attached to texts. This is primarily achieved by establishing a stemma, by which is meant a genealogy of manuscript witnesses of a text. With the help of such a stemma, manuscripts having no value for the establishment of the text can be distinguished from valuable ones, and then be eliminated. Ideally, the archetypal version of the author’s text can be rediscovered by application of the stemmatical method.

One of the better-known techniques of textual criticism perfected by Lachmann is the rule of the more difficult reading prevailing over the simpler one. Thus, original readings that have been trivialised by generations of copyists can be reconstructed. Lachmann never wrote a textbook on his method, but he developed and demonstrated it in his impressive production of critical editions, studies on aspects of textual criticism and commentaries on editions. He wrote on, and also edited, important texts not only in Greek and Latin literature, but also in old-German and old-English literature and the New Testament. This choice of texts and fields of scholarship made his influence widely felt. Modern students of textual criticism are still wittingly (but mostly unwittingly) influenced by his findings and are users of his method.

Many details of M.J. de Goeje’s life are available to scholars. His father was a Protestant minister in Dronrijp, a village in Frisia, a northern province of the Netherlands, where de Goeje was born on 13th August 1836. Although he was an able and eager pupil in primary school, his choice of profession proved to be somewhat complicated. If everything had gone according to plan, he might have become a pharmacist, either in the Netherlands or in the Dutch East Indies. But these and similar plans failed, and in the end it was decided that he would study theology in Leiden, with the probable intention of becoming a minister, like his father. He began his studies in 1854 and quickly became intrigued by the study of Semitic languages, the knowledge of which was at the time a requirement for students of theology. They were, of course, assumed to be able to read the Scriptures in the original Hebrew, Aramaic and Greek.

In Leiden, de Goeje also received a thorough training in classical philology from his teachers J. Bake (1787-1864) and C.G. Cobet (1813-1889). They may have introduced him to the critical methodology towards texts developed by Lachmann, whose influence at the time was increasingly being felt in Dutch scholarship.[8] The professor whom de Goeje came to


value most as his teacher was, however, the Arabist R.P.A. Dozy (1820-1883). Under the latter’s influence, de Goeje decided that a career as a clergyman was not for him and he choose instead to devote himself to the study of the Semitic languages Hebrew, Syriac and Arabic. He was soon completely under the spell of the latter. In his contacts with Dozy he resolved on the scholarly goal of collecting Arabic texts that were sources for the development of human civilisation in the fields of literature, history and geography,[9] to subject them to critical analysis and then construct an attractive scholarly synthesis of them. In 1859, he was put in the daily charge of the Leiden Oriental Manuscripts Collection.[10] In 1860, he defended his doctoral thesis, which was a partial edition of al-Yaʿqūbī’s Kitāb al-Buldān. He became extra-ordinarius professor in 1866 and, in 1869, this appointment was changed into a full professorship for Arabic in Leiden. In 1906 he retired, internationally acclaimed as one of the greatest philologists of his time. He died on 17th May 1909.

Early in his scholarly career, de Goeje realised that critical editions were of paramount importance and, during his many-sided career from 1856 till his death in 1909, he produced a large number of them on a regular basis. It is not inappropriate to call his office a text-editing factory, for not only did he produce a great number of editions himself, he also had his students work on parts of his projects, proof-reading, indexing and so on. De Goeje’s own editions include several historical and geographical works by al-Yaʿqūbī (1860, to be repeated and augmented in 1892 as volume 7 of the BGA), al-Idrīsī (1866, together with his teacher Dozy), al-Balādhurī (1863-1866), several volumes of the history ofal-Ṭabarī (1879-1901), a project organised by de Goeje himself.[11] Furthermore, he also produced several literary texts, notably the Diwān of Abū al-Walīd Muslim ibn al-Walīd al-Ansārī (1875), and Ibn Qutaybah’s Kitāb al-Shiʿr wa al-Shuʿarāʾ (1904). A torrent of articles on a wide variety of subjects connected to Oriental Studies appeared in international scholarly journals, often combined with spin-offs in Dutch periodicals written for a wider audience. Never again would the Dutch general public be informed about aspects of Middle Eastern culture in such an erudite and entertaining way.

One of the pillars of de Goeje’s lasting fame was his single-handed editing of what he called the ‘Library of Arabic Geographers’(Bibliotheca Geographorum Arabicorum) – a series of eight volumes containing geographical texts - which were published between 1870 and 1894, and commonly referred to as the BGA. The texts in this major project were: Kitāb Maslik al-Mamālik by al-Iṣṭakhrī (1870); Kitāb al-Masālik wa al-Mamālik by Ibn Ḥawqal (1873); Kitāb Aḥsan al-Taqāsīm fī Maʿrifat al-Aqālīm by al-Maqdisī (1877, second edition 1906); Kitāb al-Buldān by Ibn al-Faqīh al-Hamadānī (1885);Kitāb al-Masālik wa al-Mamālik by Ibn Khurdādhdhbih with excerpts from Kitāb al-Kharāj by Qudāmah ibn Jaʿfar (1889); the seventh part of Kitāb al-Aʿlāq al-Nafīsah by Ibn Rustah; Kitāb al-Buldān by al-Yaʿqūbī (1892) and, finally, Kitāb al-Tanbīh wa al-Ishrāf by al-Masʿūdī (1894). All these publications were first-time critical editions of the texts concerned, and all were provided with some information on the manuscripts used,[12] an extensive apparatus of textual criticism containing the variant readings of the major manuscripts known[13] and, finally, indexes and glossaries of technical terms and specialist vocabulary which was not to be found in the usual dictionaries of classical Arabic. In one case - Volume 5 - a French translation was added. The German translation of all texts, which de Goeje had envisaged from the very beginning of the project, was, however, never realised.[14]

Another feature of de Goeje’s work must be mentioned here, namely his sense of the importance of teamwork. Certainly for the two larger projects in which he was involved - the BGA and the Ṭabarī edition - and also the other editions that he produced, he was all too aware of the fact that teamwork and international co-operation between scholars were essential to the success of such vast enterprises. On the one hand, the mere size of a project made joining forces necessary and, on the other hand, it was important to have good connections, in the pre-photography era, to other European manuscript collections. Colleagues and friends often provided de Goeje with notes and quotations from manuscript sources, which they had to copy themselves. This is illustrated by de Goeje’s dedications in each volume of the BGA: these dedications mention both his learned friends and academic colleagues, and for several colleagues it is clear that the dedication was well deserved.[15]

The intricate internal relationship between the different Balkhī texts is an illustration of how these individual texts have fared. Even with the perspective that we now have on the manuscripts, it is virtually impossible to extricate the separate texts from one another. This was demonstrated by the way de Goeje had to handle his manuscripts. His point of departure was the classical, predominantly static, approach to texts. He looked at texts as stable and well-defined entities that had been composed and completed by their authors and then transmitted to later periods by scribes. Such an approach may be feasible and fruitful with texts from classical antiquity,[16] but for a more recent and continuous textual tradition like that of Arabic literature, it is not always possible to maintain such a static view of texts. With the early geographical texts of the Balkhī School such an approach was clearly demonstrated to be too simplistic. Even if de Goeje had wanted to distinguish between the works of al-Iṣṭakhrī and Ibn Ḥawqal he could not have escaped with using his findings with the manuscripts of one text as a textual witness for establishing another. His two editions of the oldest texts of the Balkhī geographers, BGA Volume 1 (al-Iṣṭakhrī) and BGA Volume 2 (Ibn Ḥawqal), are, therefore, so closely interrelated that each edition is used as a textual witness for the other.[17]

To edit purely literary texts the editor needs a sufficient amount of textual evidence, together with a fair knowledge of the literary usage of the language of the text. The term ‘fair knowledge’ is a recurrent theme in discussions about textual criticism. It is sometimes referred to by Middle Eastern scholars as al-dh awq al-salīm, ‘good taste’, meaning or implying that the editor having good taste knows the right level and register of language. He should therefore be able to change the text for the better, as the author might have done. Or even do it ‘better’ than the author, if the editor falls into the trap of identifying himself too much with his object of research. The extent to which this ‘good taste’ is exercised distinguishes a good edition from a bad one, the use of ‘good taste’ being the personal element in textual criticism. Extreme caution is, however, necessary when the editor feels tempted to use his ‘taste’, and most mistakes in applying textual criticism and devising conjectures are caused by its reckless use. Conversely, the editing of texts is not a mechanical procedure, as some followers of Lachmann maintain.[18]

When editing scientific texts the editor’s knowledge of science becomes more important. It is obvious that whoever edits an old mathematical text should know not only about codicology and paleography, but also about the ins and outs of the history of mathematics. Thus, proposals on how to improve corrupt texts should not only be made on the basis of philological arguments but should also take account of mathematical considerations. The same is true for other technical texts. To a certain extent one can consider geographical texts as literary texts, travelogues or descriptions of far-away countries and the manners and customs of their inhabitants. But geography does not confine itself to descriptions of regions. It often contains a wealth of information on economic matters as well. Ibn Ḥawqal’s writings are a case in point since it is likely that he was economically active during his travels as well as acting as a Fatimid missionary; activism for a cause and commercial activities have generally gone hand in hand in the history of Islam.

Geographical texts are not just literary travelogues, nor are they simply collections of commercial information, nor descriptions of the, often bizarre, habits of far-away peoples and their cosmological fantasies. They contain, in fact, some aspects of all these things. Geographical texts sometimes have the extra dimension of maps, and maps put an extra burden on the shoulders of the editor. The earliest editions of geographical texts hardly ever had maps added to the text. On the one hand, this was a result of contemporary technical limitations in reproducing images and, on the other hand, it can be attributed to the nineteenth-century philologist’s preoccupation with texts, to the exclusion of graphic images. It seems to me that de Goeje had no interest in this extra pictorial dimension as an aid to the interpretation of the geographical texts he was editing: de Goeje’s orientation towards his subject was largely of a literary nature.[19] He wanted the text to be as correct as possible, considering it as a collection of maps with explanatory texts was much less his concern.

Textual criticism works by assuming that each copy of a manuscript contains the scribal errors already contained in the exemplar, plus a set of new ones. By analysing the genealogy of errors it is possible to establish a genealogy of the manuscripts. These rules of textual criticism cannot simply be transposed to a critical analysis of hierarchy in the tradition of maps and their relationship with their captions, legends and accompanying texts in the manuscripts. Analysis of maps is greatly complicated because it must take account of more than just the text or legends. Art-historical considerations, especially with regard to the techniques of painting and the conventions of representation, give extra information and therefore complicate matters further. There always remains, of course, a correlation between the map and the geographical reality: the extent to which a map represents the reality on the ground is always relevant.

A typical example is the dilemma facing the editor of the map of the Caspian Sea. In the al-Iṣṭtakhrī maps two islands are drawn. If one compares the map in the Leiden MS with the one in the Persian translation, as contained in the Oxford MS,[20] one can see that the names, or relative positions of the two islands, have been interchanged. That is not only a problem for textual criticism, but also one for the study of cartographic representation.

De Goeje never wrote a survey of his editorial methods (nor did Lachmann for that matter). Their methods are implicit in the texts they edited and their other works. The obvious source for the method used is the critical apparatus. When one examines de Goeje’s critical apparatus, one sees that he does sometimes use the maps, but just as additional evidence, where the text has failed him or there is a lacuna. Furthermore, de Goeje made no distinction between a critical apparatus, serving to elucidate the selection of readings of the text, and the commentary that he needed to give in order to make his readers understand the meaning of the text. Such an approach is not always feasible in editing geographical texts. The textual difficulties are often about the spelling of names; correct spellings cannot always be reconstructed by textual criticism. The editor of a text must, therefore, have recourse to other supplementary works with a better reading, even though they are textually unrelated to the primary text that the editor is establishing. The numerous references in de Goeje’s apparatus are eloquent witnesses to this procedure, demonstrating the special nature of editing non-fictional texts, and the impossibility of accepting the restrictions of editorial theory in all instances.

And so, when father and son Fikrī sat informally with de Goeje in his house, smoking and talking on that quiet Sunday afternoon in August 1889, was it at all surprising that they posed him a geographical question? Where in the world, so they asked the learned editor of so many geographical texts, would he situate the legendary island of Wāq Wāq? De Goeje did not disappoint and, of course, came forward with the correct answer: Wāq Wāq was Japan. He gave his visitors a full account of the numerous literary sources about Wāq Wāq and the etymology of its name, fully supporting his opinion with decisive arguments.[21]

[1] Held from the 2-13 September 1889.

[2] The event is described in Irshād al-Alibbāʾ ilā Maḥāsin Urubbā (Cairo, al-Muqtataf, 1892), pp. 655-658. The full text of the commentary by ʿAbdallāh Fikrī, entitled ʿUjālat al-Bayān ʿalā Diwān Ḥassān is published in al-Āthār al-Fikrīyah (Cairo. 1315), pp.334-396.

[3] Nubdhah fī Ibṭāl Raʾy al-Qaʾilīn bi-Taʿwīḍ al-Lughah al-ʿArabīyah al-Ṣaḥīḥah bi al-Lughah al-ʿĀmīyah fī al-Kutub wa al-Kitābah. The text is published in Irshād al-Alibbāʾ, ibid., pp.674-701.

[4] Irshād al-Alibbāʾ, ibid.

[5] There is an interesting and rarely seen double perspective of this encounter, first the account by father and son Fikrí in their travelogue, Irshād al-Alibbāʾ , pp. 523 529, and a year later by M.J. de Goeje in his review of that book, ‘Een Egyptenaar in Europa’, De Gids, 1 (1893), pp. 566 577.

[6] With the arcane title Interpres Legati Wameriani (Interpreter of Warner’s Legacy) which is now held by the present author.

[7] Catalogus Codicum Orientalium Bibliothecae Academiae Lugduno-Batavae (6 vols, Leiden, 1851-1877); Catalogus Codicumm Arabicorum Bibliothecae Academiae Lugduno-Batavae Editio secunda, (Leiden, 1888-1907), vols I-II.

[8] Knowledge of Latin (which university students had already acquired at grammar school) had another, very practical, purpose as well. Most scholarly works in large parts of Europe were published in Latin, more or less as they are in English nowadays. Several of de Goeje’s works directed at an international audience were written in Latin, a sad fact, since they are therefore almost entirely inaccessible now that knowledge of Latin has become so rare.

[9] For an extensive survey of de Goeje’s life and work, see C. Snouck Hurgronje, ‘Levensbericht van Michael Jan de Goeje’ [with a list of his writings by Th.W. Juynboll], in Jaarboek der Koninklijke Akademie van Wetenschappen, (Amsterdam, 1909).

[10] Then comprising at least approximately 2,000 volumes of mostly Arabic, Persian, Turkish and Hebrew texts, including the important collections of J.J. Scaliger (1540 1609), Jacobus Golius (1596 1667) and Levinus Warner (1619- 1665).

[11] See J.T.P. de Bruijn, ‘Collective Studies of the Muslim World: Institutions, Projects and Collections’ in W. Otterspeer (ed), Leiden Oriental Connections 1850-1940, pp. 94-114, esp. pp. 99-100. See also Arnoud Vrolijk, ‘The Leiden Edition of al-Ṭabarī’s Annals. The Search for the Istanbul Manuscripts as Reflected in Michael Jan de Goeje’s Correspondence’, Quaderni di Studi Arabi, 19 (2001), pp. 71-86, De Goeje’s vast correspondence with his collaborators in the Ṭabarī edition is kept in Leiden University Library, Or. 5585 e- f.

[12] These are a bit sparse, however, according to present-day requirements.

[13] See the Appendix.

[14] Vol. 1 of the series (al-Iṣṭakhrī, 1870), introduction p. viii.

[15] Most notably to Carlo de Landberg for his help in procuring the manuscript of Ibn Khurdādhdhbih’s Kitāb al-Masālik wa al-Mamālik (now MS Vienna, ÖNB, Mixt. 783).

[16] This is not because of an intrinsic characteristic of classical Greek and Latin texts but because of the length of time that has elapsed between the period in which the classical authors lived and our own period. It is, in fact, our lack of knowledge about the early stages of the period in which classical texts crystalised into a fixed shape that creates the false impression of their being static.

[17] See the Appendix to this chapter, which gives a survey of the manuscripts used by de Goeje in preparing his editions. Gerald R. Tibbetts has drawn a stemma of the texts of the Balkhī School, demonstrating the complications of transmission of the material (‘The Balkhī School of Geographers’ in J.B. Harley and David Woodward, Cartography in the Traditional Islamic and South Asian Societies, II/1, [Chicago, 1992], pp. 108 136, on p. 111).

[18] The work by Paul Maas, Textual Criticism, which has been widely used by generations of classical scholars, is very confusing in this respect. While it may be applicable in practical textual criticism of Greek and Latin, its extreme conclusions are hardly useful for textual criticism of Islamic literature. I have voiced my own doubts about the applicability of Maas’s book in ‘Establishing the Stemma. Fact or Fiction?, in MME 3 (1988), pp. 88 101.

[19] De Goeje never travelled to the Orient.

[20] See the bibliographical references in the Appendix below.

[21] Irshād al-Alibbāʾ, op. cit., p. 527.

Source note:
This was published in:
The Earth and its Sciences in Islamic Manuscripts: proceedings of the fifth conference of Al-Furqān Islamic Heritage Foundation – English version, 2011, Al-Furqan Islamic Heritage Foundation, London, UK., p 141-174.