by Peter Tarras
Saint Catherine’s Monastery at Mount Sinai has preserved the single most important collection of early Arabic Bible translations and, more generally, of Christian writing in Arabic. One of the most curious manuscripts of this collection bears the shelfmark MS Sinai Ar. 514. It was first described by Aziz Suryal Atiya (1898-1988) in his 1955 Handlist. Notwithstanding the usually dry descriptive tone of a manuscript catalogue, Atiya could hardly hide his excitement about the discovery, describing it as a “Unique Quintuple Palimpsest in three languages – Syriac, Greek and Arabic. Discovered on Tuesday 12 June 1950, towards the close of the Mt. Sinai Expedition, this has been christened by its discoverer and author of this Hand-list, as ‘CODEX ARABICUS’”. This christening was again announced in an article roughly ten years later, in which Atiya also explains, not without pride, that the Codex Arabicus is “the third major treasure” found at Saint Catherine’s, next to the 4th-century Codex Sinaiticus, discovered by Konstantin von Tischendorf (1815-1874), and the 4th-century Codex Syriacus, discovered by Agnes Smith Lewis (1843-1926).
The importance of the Codex Arabicus, however, lay not in its being one of the oldest witnesses to the text of the New Testament, as are its Greek and Syriac siblings. Rather, Atiya saw in the fivefold reuse of parchment “a record unmatched in palimpsest history”. Yet, besides a number of hagiographical texts, the scriptio superior also preserved what is probably the oldest known Christian Arabic translation of the biblical book of Job. According to Atiya, the text was translated from Syriac and copied “in what might be described as early Kufic or perhaps pre-Kufic archaic Naskh by a contemporary of the age of Charlemagne, that is, in the late eighth or early ninth century”. In the same year of Atiya’s article, Būlus ʿAyyād ʿAyyād published an edition of the text, which had been brought to his attention by Atiya. In his short preface, ʿAyyād also judged the Arabic version to be translated from the Syriac. Neither he nor Atiya, however, give any reasons for their estimation. Likewise ʿAyyād dates the translation to the ninth century without justification, specifying, however, that its translator was the Sinaitic monk Thomas of Fustat (Tūmā al-Fusṭāṭī). In 1998, Steven P. Blackburn subjected the translation to further scrutiny in his doctoral dissertation. Again, he simply introduces it as “the product of an Egyptian monk, Ṭōmā al-Fusṭāṭī”. Still, Blackburn was able to narrow down the possible base-texts of the translation as the Peshitta with some influence of the Greek Septuagint, probably mediated through the Syro-Hexapla. In a recent article, Arik Sadan questioned the influence of the Peshitta, which cannot have been the exclusive Vorlage of the text, and pointed to features that more strongly link it to the Septuagint tradition.
The most striking example of the influence of this tradition are the Septuagintal additions to verse 42:17. These begin in the Arabic as follows (42:17b): “This is a volume which was translated from the Syriac, since [Job] used to dwell in the land of Uz on the borders of Damascus and Hawran and his name of old was Jobab” (haḏā muṣḥaf fussira min al-suryānī li-annahu kāna sākin fī arḍ ʿŪḍ fī ḥudūd Dimašq wa-Ḥawrān wa-kāna ismuhu qadīman Abwāb). This Arabic rendition, especially the first few words, must be viewed as an interpretation of the Greek, which starts with the pronoun houtos (translated here as hāḏā). The referent of houtos is somewhat tricky to determine, but is is at least clear that it does not refer to the term biblios (“book”). The text of the Syro-Hexapla, by contrast, is more faithful to the Greek, reading: “This was translated from the Syriac books” (hānā meṯpašaq men ktāḇē sūryānē = houtos hermēneuetai ek tēs Syriakēs bibliou). It seems to me that Atiya’s original postulation of a Syriac Vorlage, followed by later scholars, was based solely on the Arabic translator’s attempt to make sense of an “awkward Greek sentence”. But what about the translator? It turns out that the ascription of the translation to Thomas of Fustat is on similarly shaky ground.
The colophon appended to the Septuagintal additions reads as follows: “With God’s help ends the story of Job the righteous. May God have mercy with the sinful servant who copied it for Mount Sinai, God’s hallowed mountain. He is the sinner Thomas of Fustat.” (tammat bi-ʿawn Allāh qiṣṣat Ayyūb al-ṣiddīq raḥima Allāh al-ʿabd al-ḫāṭiʾ allaḏī katabahā li-Ṭūr Sīnā ǧabal Allāh al-muqaddas wa-huwa Tūmā al-Fusṭāṭī). The use of the verb kataba, commonly employed to designate the activity of copying, makes clear that Thomas identifies himself as the scribe rather than the translator of the text. On the basis of his handwriting, Atiya made him “a contemporary of the age of Charlemagne” (the latter died in 814 CE). The colophon had apparently escaped Atiya’s notice, for had he read Thomas’ name, he could have noticed that a colophon by the very same scribe was published already in 1897.
This second colophon is found in MS Strasbourg, Bibliothèque nationale et universitaire, Or. 4.225 (f. 226v) and was published in a German article by the Danish Orientalist Johannes Østrup (1867-1938) together with a description of the contents of this codex and another manuscript. The colophon of the Strasbourg manuscript reads as follows:
This precious volume was completed with the help of our Lord, Christ, in the year 288. It was copied by Thomas, the poor sinner, of Fustat at Mount Sinai, God’s hallowed mountain. He copied it for the gentle-hearted and generous father, distinguished in the faith in God’s true word, Anbā Mūsā son of Ḥakīm the priest of Aḏraḥ [= Darʿā]. When you read in this my book, my brother, understand it and remember me in every hour you read in it. May Christ remember you in his kingdom and elevate you to his right and may the Lady Mary and all the saints plead for him who read and him who copied and him who [had the copy made], Amen, Amen, Amen (kamala hāḏā l-muṣḥaf al-šarīf bi-ʿawn rabbinā <al-Mas>īḥ fī sana ṯamāniyya wa-ṯamānīn wa-miʾatayn <kata>bahu Tūmā al-ḫāṭiʾ al-ḥaqīr al-Fusṭāṭī fī Ṭūr <Sīnā> ǧabal Allāh al-muqaddas katabahu li-l-ab al-wadīʿ <al-karīm> al-šarīf bi-l-īmān bi-kalimat Allāh al-ḥaqq Anbā Mūsā ibn Ḥakīm al-qissīs al-Aḏraḥī lammā anta qaraʾta kitābī hāḏā yā aḫī fa-ifhamhu wa-uḏkurnī fī kull sāʿa taqraʾa fīhi ḏakaraka al-Masīḥ fī mulkihi wa-aqāmaka ʿan yamīnihi wa-li-man qaraʾa wa-li-man kataba wa-li-man <istakataba> bi-šafāʿat Martmaryiam wa-ǧamīʿ al-qiddīsīn āmīn āmīn āmīn).
This colophon allows us to assign a precise date to Thomas’ activity as a copyist, namely 288 AH, which is December 900 through November 901 CE. Thus, Thomas could hardly have been a contemporary of Charlemagne! This also gives us an approximate date for the production of the Codex Arabicus, namely the second half of the ninth or the first half of the tenth century. From the colophon we learn that the Strasbourg manuscript was produced “at Mount Sinai” (fī Ṭūr Sīnā). The very same formulation is used in a colophon preserved in another manuscript, which was also copied by a monk Thomas and known since 1915.
This colophon is found in MS Bryn Mawr, Bryn Mawr College Library, BV 69 (f. 4v [2v]) and reads as follows:
This precious book was copied at Mount Sinai, the hallowed mountain. The self-negligent monk Thomas copied it, he who is poor, weak and very sinful copied it _ _ _. I ask our Lord, Christ, to forgive the sins of him who copied, him who read and him who purchased and had the copy made. May he [sc. the purchaser?] and the scribe be given what [Christ] has given to the pious saints, and may he be elevated to [Christ’s] right and may [Christ] make him hear the voice filled with joy, Amen. May the mother of light, Lady Mary, and all the saints plead [for them] for forgiveness, Amen (kutiba hāḏā l-muṣhāf al-šarīf fī Ṭūr Sīnā al-ǧabal al-muqaddas katabahu Ṯūmā al-rāhib al-ġāfil ʿan nafsihi katabahu _ _ _ al-maskīn al-ḍaʿīf al-kaṯīr al-ḏunūb wa-anā asʾalu sayyidanā al-Masīḥ an yaġfiru ḫaṭāyā man kataba wa-man qaraʾa wa-man iqtanā wa-istakataba wa-yuʿṭī mā ʿaṭā al-qiddīsīn al-abrār wa-yuqīmuhu ʿan yamīnihi wa-yusmiʿuhu al-ṣawt al-mamlūʾ faraǧ wa-li-l-kātib āmīn bi-šafāʿat al-maġfira umm al-nūr Martmaryiam wa-ǧamīʿ al-qiddīsīn āmīn).
The Arabic text of the colophon together with a French translation was published by Michel van Esbroeck in 1982 who conveniently not only offered a reading for the lacuna but also a date for the manuscript after “having examined the eroded passage for more than two hours”. It is unclear to me how he arrived at his conjecture, which reads “in the year 338” (fī sana ṯamān wa-ṯalāṯīn wa-ṯalāṯ miʾa), corresponding to 949/50 CE. Marianne Hansen, Bryn Mawr College Library’s Curator for Rare Books and Manuscripts, replied to my request regarding the erased line: “I have just examined the MS, and the top layers of parchment are abraded; the surface is now rough. […] I am uncertain of why [Michel van Esbroeck] cites that date, unless he saw the manuscript before the erasure (which I frankly cannot imagine anyone here doing, and we have held the book since the nineteen fifties)“. Hansen must be right, since Georg Graf remarked already in 1915 that the lacuna was illegible. We find the colophon at the end of a two-page table of contents, which makes a somewhat sloppy impression. Textual additions and corrections are made in the margins as well as above certain words. Especially a number of red strokes across the page of the colophon make it appear as if this was a draft with which the scribe grew dissatisfied. The hand is not the one of Thomas of Fustat, though it bears some similarity to it. It is my impression that it also differs from the one that copied the rest of the manuscript, which, however, could actually be Thomas’ hand. So, one possible story of what happened here, is that the table of contents as we have it now together with the colophon was reproduced by a different scribe who tried to imitate, not very dexterously, Thomas’ hand in order to replace the lost pinax (which would have been written on the first page of the original codex and thus most liable to detach from the binding). The erased line, which could have contained a date, then would have fallen prey to the same process of drafting.
The table of contents of the Bryn Mawr manuscript further reveals that it used to contain an Arabic version of the book of Job as well (likewise bearing the title qiṣṣat Ayyūb al-ṣiddīq), which is now unfortunately lost. It would have allowed us to see if this was the same version Thomas copied in the Codex Arabicus. If the manuscript was indeed produced by Thomas of Fustat, it would further add to our knowledge about this Sinaitic scribe. It would also render van Esbroeck’s conjecture even more unlikely, for, even if Thomas copied the Strasbourg manuscript early in his career, he would have been probably in his sixties or seventies when copying the Bryn Mawr manuscript. All this, however, needs to be analysed in much more detail. What is needed is a palaeographic profile of Thomas as well as a minute codicological study of his manuscripts, which will then surely allow the identification of further manuscripts. For instance, Grigory Kessel, was able to identify the palimpsest fragment MS Munich, Bayerische Staatsbibliothek, Cod. arab. 1066 as a membrum disjectum of the Codex Arabicus. I believe that the same batch of Christian Arabic manuscripts contains another fragment copied in the hand of Thomas, namely MS Munich, Bayerische Staatsbibliothek, Cod. arab. 1068.
It was my aim here to collect some of the information we have on Thomas of Fustat. The results are partly negative. It turned out that, even though we do not know when or by whom the Arabic Job of the Codex Arabicus was translated, there is at least no evidence supporting the supposition the translation should be ascribed to Thomas. His being “merely” the manuscript’s scribe, however, does not diminish his importance. In fact, he is one of the earliest known Christian Arabic scribes active at Mount Sinai who not only produced manuscripts at the request of others, but was involved in building up Saint Catherine’s Christian Arabic library on the spot. On a practical level, this meant gathering and making use of much older parchment codices. From this practice the Codex Arabicus emerged. Another example is the famous Lewis-Mingana Palimpsest (MS Cambridge, University Library, Or. 1287), which was not only in all likelihood produced at Saint Catherine’s as well, but whose scriptio superior, as shown by Alain F. George, bears great similarity to Thomas’ hand. One important question raised by George concerns the mode of manuscript production: Was Thomas part of a scribal workshop? That is, are there traces of an institutionalised form of manuscript production similar to the scriptoria of the Latin West? Given that practical knowledge of manuscript production was brought to Saint Catherine’s and cultivated there by monks from many different linguistic backgrounds, as we see, for instance, in the Greek or Georgian quire marks of Christian Arabic codices, it is very likely that some sort of workshop existed. Yet, more evidence needs to be collected to make a strong case for this. The information we gather from the colophons of scribes like Thomas of Fustat are key to such questions, which will also help us better understand the practical conditions under which Christian Arabic translations of the Arabic Bible were produced, transmitted and disseminated.
Peter Tarras is currently a research assistant in the DFG-funded project “Arabic and Latin Glossary“ based at the University of Würzburg and teaches at the Institute of Near and Middle Eastern Studies in Munich. He is writing a PhD thesis under the supervision of Prof. Peter Adamson on evil, free will and eschatology in the Muslim philosopher al-Fārābī. As an associated member of the Biblia Arabica project, he has worked on biblical quotations in Christian-Arabic apologetical literature and developed a strong interest in manuscript studies and the history of modern manuscript collections.
* This text is an extract of ongoing research, which will hopefully be published in the near future under the title: “Building a Christian Arabic Library: Thomas of Fustat and the Scribal Workshop of Saint Catherine’s Monastery”. I am grateful to Miriam L. Hjälm and Nathan Gibson for their valuable comments on an earlier draft of this blog. I would also like to express my gratitude to Arik Sadan, Grigory Kessel and Marianne Hansen for insights shared on various aspects of the topics discussed here. Any errors and shortcomings are, of course, my own responsibility.
 A.S. Atiya, The Arabic Manuscripts of Mount Sinai: A handlist of the Arabic manuscripts and scrolls microfilmed at the library of the Monastery of St. Catherine, Mount Sinai, Baltimore: John Hopkins Press, 1955.
 Ibid. 19.
 A.S. Atiya, “Codex Arabicus (Sinai Arabic Ms. No. 514)”, in: Homage to a Bookman: Essays on Manuscripts, Books and Printing Written for Hans P. Kraus on His 60th Birthday Oct. 12, 1967, Berlin: Gebr. Mann Verlag, 1967, 75-85; 75.
 Ibid. 76. It is, however, not the case that the parchment of the Codex Arabicus as a whole was written on five times, as Atiya’s description may suggest. Rather, its raw material was drawn from different manuscripts some of which were already palimpsests. The Codex Arabicus was included in the Sinai Palimpsest Project for multispectral imaging and catalogued by Grigory Kessel. He kindly informed me that he is currently preparing a study of the reused manuscripts and the undertexts preserved in Codex Arabicus, see G. Kessel, “Codex Arabicus (Sinai ar. 514) Revisited”, in: C. Rapp, J. Grusková, G. Kessel and G. Rossetto (eds), New Light on Old Manuscripts: Recent Advances in Palimpsest Studies at Saint Catherine’s Monastery and Beyond, forthcoming.
 On the textual history of the Arabic versions of Job, see M.L. Hjälm, “[Job] Arabic Translations”, in: A. Lange and E. Tov (eds), Textual History of the Bible, vol. 1 C: Writings, Leiden: Brill, 236-40.
 Atiya, “Codex Arabicus”, 78.
 B.ʿA. ʿAyyād, Sifr Ayyūb: Qāma bi-tarǧamatihi min al-suryāniyya ilā l-ʿarabiyya Tūmā al-Fusṭāṭī fī l-qarn al-tāsiʿ al-milādī, Cairo: Maʿhad al-Dirāsāt al-Qibṭiyya, 1975, 3.
 S.P. Blackburn, The Early Arabic Versions of Job (first millenium C.E.), PhD Dissertation, University of St. Andrews, 1998.
 Ibid. 12.
 A. Sadan, “Differences and Similarities Between Early Christian and Judaeo-Arabic (Rabbinical and Karaite) Translations of the Bible into Arabic: The Case of the Book of Job”, Journal of Eastern Christian Studies 70.1-2 (2018): 7-33. I am grateful to Arik Sadan for sharing with me the pre-print and print versions of his article as well as for providing me with ʿAyyād’s edition of the Arabic text.
 On these, see G.B. Gray, “The Additions in the Ancient Greek Version of Job”, The Expositor 19 (1920): 429-31; A.Y. Reed, “Job as Jobab: The Interpretation of Job in LXX 42:17b-e”, Journal of Biblical Studies 120.1 (2001): 31-55. For the Syro-Hexaplaric version of these additions, see J.D. Meade, A Critical Edition of the Hexaplaric Fragments of Job 22-42, PhD Dissertation, Southern Baptist Theological Seminary, 2012, 440. See also idem, “An Analysis of the Syro-Hexapla of Job and Its Relation to Other Ancient Sources”, Aramaic Studies 14.2 (2016): 212-41.
 MS Sinai Ar. 514, f. 160r, l. 16f. On the peculiarities of this rendering, see Sadan, “Differences”, esp. 12.
 Gray, “Additions”, 429. Cf. also Reed, “Job”, 37.
 MS Sinai Ar. 514, f. 160r, l. 26f.
 J. Oestrup, “Über zwei arabische Codices sinaitici der Strassburger Universitäts- und Landesbibliothek”, Zeitschrift der Deutschen Morgenländischen Gesellschaft 51.3 (1897): 453-71. The second manuscript is MS Strasbourg, Bibliothèque nationale et universitaire, Or. 4.226, which was copied by Anthony David of Baghdad at Mar Sabas in 885/6 CE; see S.H. Griffith, “Anthony David of Baghdad, Scribe and Monk at Mar Sabas: Arabic in the Monasteries of Palestine”, Church History 58.1 (1989): 7-19. See also André Binggeli, “Les trois David: Copistes arabes de Palestine aux 9e-10e s.”, in: A. Binggeli, A. Boud’hors and M. Cassin (eds), Manuscripta graeca et orientalia: Mélanges monastiques et patristiques en l’honneur de Paul Géhin, Leuven: Peeters, 2016, 79-117.
 MS Strasbourg, Bibliothèque nationale et universitaire, Or. 4.225, f. 226v, l. 8-17. Cf. Oestrup, “Über zwei arabische Codices”, 454f. The first five lines of the colophon seem to have suffered from water damage on the right side, making the first word of each line barely legible. Note that the name Tūmā is written without diacritics (ىوما) just like in MS Sinai Ar. 514. The word al-karīm is Oestrup’s conjecture. The word istaktaba is my conjecture.
 See G. Graf, “Alte christlich-arabische Fragmente”, Oriens Christianus NS 4 (1915): 338-41. I have briefly dealt with Graf’s engagement with this manuscript in an earlier blog post on this website. The following fragments used to belong to this codex: MSS Birmingham, Selly Oak College, Mingana Christ. Arab. 94; Add. Arab. 130, 148, 149; Leiden, University Libraries, Or. 14.238. See M. van Esbroeck, “Remembrement d’un manuscrit sinaïtique arabe de 950”, in: S.K. Samir (ed.), Actes du premier congrès international d’études arabes chrétiennes (Goslar, septembre 1980) (Orientalia Christiana Analecta 218), Rome: Pontifical Institute of Oriental Studies, 1982, 135-47; as well as J.W. Pollock, “Two Christian Arabic Manuscripts in the Bryn Mawr Library”, Journal of the American Oriental Society 110.2 (1990): 330-31. Tony Burke has recently dealt with the history of these membra disjecta here.
 MS Bryn Mawr, Bryn Mawr College Library, BV 69, f. 4v, l. 8-17. Cf. Graf, “Alte christlich-arabische Fragmente”, 340, who gave the following partial translation: “Dieses ehrwürdige Buch wurde geschrieben auf dem Ṭûr Sinâ, dem heiligen Berg. Geschrieben hat es Ṯûmâ (Thomas), der Mönch al-Ġâfâ. Für sich selbst hat er es geschrieben.”
 Esbroeck, “Remembrement d’un manuscrit”, 141, n. 23; Eng. tr. mine.
 Ibid. Note that van Esbroeck only marks fī sana and ṯalāṯ miʾa as conjectures.
 Graf, “Alte christlich-arabische Fragmente”, 340.
 First presented at the conference “New Light on Old Manuscripts: Recent Advances in Palimpsest Studies”, 24-27 April 2018, Vienna. Another membrum disjectum of the Codex Arabicus is MS Oslo, Schøyen Collection, 579.
 This fragment might actually constitute a part of the missing beginning of MS Strasbourg, Bibliothèque nationale et universitaire, Or. 4.225. Cf. Oestrup, “Über zwei arabische Codices”, 454; 458. On the Munich Christian Arabic manuscripts, see G. Graf, “Christlich-arabische Handschriftenfragmente in der Bayerischen Staatsbibliothek”, Oriens Christianus 38 (1954): 125-32.
 By contrast, we know that other manuscripts were purchased for Saint Catherine’s already in the tenth century. Cf. M.N. Swanson, “Solomon, Bishop of Mount Sinai (Late Tenth Century AD)”, in: R. Ebied and H. Teule (eds), Studies on the Christian Arabic Heritage in Honour of Father Prof. Dr. Samir Khalil Samir S.I. at the Occasion of his Sixty-Fifth Birthday, Leuven: Peeters, 2004, 91-111. On the scribal activity at Saint Catherine’s, see also N.P. Ševčenko, “Manuscript Production on Mount Sinai from the Tenth to the Thirteenth Century”, in: S.E.J. Gerstel and R.S. Nelson (eds), Approaching the Holy Mountain: Art and Liturgy at St. Catherine’s Monastery in the Sinai, Turnhout: Brepols, 2010, 233-58.
 See A.F. George, “Le palimpsest Lewis-Mingana de Cambridge, témoin ancien de l’histoire du Coran”, Comptes rendus de l’Académie des inscriptions et belles-lettres 1 (2011): 377-429; esp. 409.
 Together with Miriam L. Hjälm, I am currently preparing a comparative study of early Christian Arabic colophons.