Berkes ————— Greek, Roman, and Byzantine Studies 65 (2025) 80–98 ISSN 2159-3159 Article copyright held by the author(s) and made available under the Creative Commons Attribution License CC-BY https://creativecommons.org/licenses/by/4.0/ Greek Documents and their Scribes in Eighth-Century Egypt Lajos Berkes FTER THE ARAB CONQUEST of Egypt in 642,1 Greek documents continued to be produced in Egypt until the early ninth century.2 The Arabs continued to employ Greek as an administrative language for pragmatic reasons:3 they found a bureaucratic system working in Greek that guar- anteed the smooth collection of taxes. Indeed, official documents constitute the overwhelming majority of our Greek papyri of the early Islamic period. Meanwhile, Coptic documents became more and more numerous, in both the public and the private 1 The phrase “early Islamic Egypt” is common in current scholarship, but on its complex nature see L. Berkes, “Introduction: A Papyrological Perspec- tive on Christians and Muslims in Early Islamic Egypt,” in Christians and Muslims in Early Islamic Egypt (Ann Arbor 2022) 1–9, at 1 and 4, as well as F. Morelli, CPR XXXVI (2022) pp.12–14. I use the word ‘papyri’ here as is common in the discipline, i.e. including texts written on ostraca, wooden tablets, parchment, and other ephemeral writing materials. 2 J. de Jong and A. Delattre, “Greek as a Minority Language in Early Islamic Egypt,” in R. G. Hoyland (ed.), The Late Antique World of Early Islam: Muslims among Christians and Jews in the East Mediterranean (Princeton 2015) 37– 62, provide a useful overview on Greek in Egypt in the early Islamic period. J. H. M. de Jong, “A Changing Position of Greek? Greek Papyri in the Docu- mentary Culture of Early Islamic Egypt,” in J. Bruning et al. (eds.), Egypt and the Eastern Mediterranean World. From Constantinople to Baghdad, 500-1000 CE (Cambridge 2022) 379–407, gives a similar account and connects the disappearance of Greek documents with the language’s loss of status and the increasing use of Coptic (esp. 398–399). 3 See also the discussion of F. Morelli in the introduction of CPR XXX (2010) pp.42–43. A LAJOS BERKES 81 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 spheres.4 However, we find only a few Arabic papyri in the early post-conquest period. Their number increased only in the sec- ond part of the eighth century, but the language becomes dom- inant in the papyrological record only in the ninth century.5 The first decades of Arab rule still produced the same kind of Greek documents that we find in the late Byzantine period, such as letters, legal documents, lists, and receipts. However, in con- trast to the abundance of administrative documents,6 texts from the private sphere in Greek became rare after the conquest. There are only a few examples of private letters in Greek from the later seventh century, but no certain instances from the eighth.7 Similarly, receipts and orders, which were not directly related to tax administration, became increasingly seldom from the late seventh century on. This trend continued in the first half of the eighth century, when documents of a non-fiscal nature rapidly disappeared. The last datable Greek letters from Egypt stem from the governor 4 It seems that the increasing use of Coptic was a trend that started already in the Byzantine period and is not only to be explained by the Arab conquest —although it certainly accelerated the process: see J. van der Vliet, “Coptic Documentary Papyri after the Arab Conquest,” JJurP 43 (2013) 187–208, at 197–198. 5 On the ratio of Arabic, Greek, and Coptic texts in the seventh-eighth centuries see E. Garosi, Projecting a New Empire. Formats, Social Meaning, and Mediality of Imperial Arabic in the Umayyad and Early Abbasid Periods (Berlin 2022) 47–48. 6 It seems that there even is an increase in the number of certain types of fiscal documents after the Islamic conquest, see P. M. Sijpesteijn, Shaping a Muslim State. The World of a Mid-Eighth-Century Egyptian Official (Oxford 2013) 75. 7 The letter CPR XIX 30 was dated paleographically to the eighth century, but it might well come from the late seventh. We do not possess a large enough sample of dated Greek cursive hands from the eighth century to date letters confidently to it. The same applies to legal documents, which were usually dated only by indiction years. There may be several letters or legal documents that have been dated by editors to the seventh century based on their script that were actually written in the eighth. 82 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 Qurra b. Sharīk (in office 709–715) and were addressed to Basileios, the pagarch of Aphrodito.8 The last dateable cluster of Greek legal documents was produced by notaries in the city of Heracleopolis in the first three decades of the eighth century— the last of them dates to 724/5.9 These papyri are the last examples of a still vivid Byzantine notarial tradition, as the com- petent use of high-level technical vocabulary proves. However, the notaries preparing them were among the very last represen- tatives of this tradition, which survived to the ninth century only in Coptic legal documents.10 After the first quarter of the eighth century, the extant Greek documentation is comprised exclusively of fiscal documents, in- cluding various types of accounts, receipts, tax demand notes (entagia), and orders for delivery. This is also evidenced by Coptic papyri, which often contain Greek sequences usually written in a different style than the rest of the text, such as invocations, dating clauses, totals, and subscriptions.11 After the late 750s, virtually only accounts remain (with some rare exceptions). Other genres disappear even from tax admini- stration: our last dateable Greek tax receipt comes from the monastery of Apa Apollos from the year 758 or later (P.Clackson 46). The last Greek tax demand note, entagion, known so far can be dated to 752 (SPP III 260).12 Our late-eighth-century material 8 T. S. Richter, “ ‘An unseren Herrn, den allberühmten Korra, den herr- lichsten Gouverneur, durch Dich, glorreichster Herr Basilios, Pagarch von Djkow mit seinen Gehöften’. Verwaltung und Verwaltungssprachen Ägyp- tens im 8. Jh. nach den Qurra-Papyri,” in F. Feder et al. (eds.), Ägypten und sein Umfeld in der Spätantike (Wiesbaden 2013) 121–138. 9 S. Kovarik, “Der herakleopolitische Notar Paulos: Ein Kompromiss aus dem umayyadischen Ägypten – der bisher späteste griechische Vertrag,” Tyche 35 (2020) 55–70, for the edition of the latest legal document, P.Vindob. G 19705, see esp. 66–67. 10 See e.g. n.44 below. 11 See e.g. the documents discussed at 85 ff. below. 12 The sheet of uncut entagia, SPP VIII 1199–1200 may date from 744 or LAJOS BERKES 83 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 consists exclusively of accounts and (often bi- or trilingual) writing exercises.13 At present, the last dateable Greek document from the Arab administration is the account CPR XXII 21 which is dated to the hijri year 180, which equals 796/7 A.D. ( fig. 1). Figure 1: CPR XXII 21, A.D. 796/7 © Österreichische Nationalbibliothek, Papyrussammlung. 759 depending on the prosopography of the issuing official, see N. Gonis, “Another Look at Some Officials in Early Abbasid Egypt,” ZPE 149 (2004) 189–195, at 191–192. However, even a dating to 759 would still be com- patible with the following discussion. The genre appears to have been abol- ished as a consequence of administrative reforms, since the last testimony for Coptic and Arabic entagia is the Arabic-Coptic P.Clackson 45 from the year 753. See also 92 ff. below. 13 A rare and therefore important exception will be discussed at the end of this article. 84 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 This suggests that the production of Greek documents ceased in the Arab administration approximately at the turn of the eighth and ninth centuries. The processes of linguistic change manifested themselves in different ways in different regions. The use of Greek was more prevalent in the northern than in the southern regions of the Nile Valley. This is evidenced by the fact that our most recent private legal documents in Greek originate from Heracleopolis, one of the northernmost cities where papyri have been discovered. This can be contrasted with the region of Hermopolis that is situated further to the south. Despite the existence of several Coptic legal documents from the city, there are no examples in Greek from the late seventh or early eighth century belonging to the same genre of texts.14 Scribes trained in church institutions It is unclear whether the individuals responsible for creating the Greek fiscal documents of the Arab administration were merely hired scribes or if they also held other positions in the tax administration. It is important to note that these two options are not necessarily mutually exclusive. As will be demonstrated be- low, the papyrological evidence indicates that these scribes were engaged in the production of other texts as well, including litur- gical manuscripts. This suggests that they were involved in a number of different occupations. It can be assumed that they were compensated for their work, but our sources do not provide information on their salaries. Our primary sources on their activities are their own documents, and thus we must accept a certain degree of ambiguity regarding their exact role and status in the Arab administration. 14 A comparable difference may be observed at the administrative level as well. While in Hermopolis entagia were written in Coptic from the beginning of the eighth century on, we still find a Greek entagion from Heracleopolis dated to 751/2 (CPR XXII 7); this text contains new Greek formularies and administrative vocabulary, showing that these scribes were well trained in Greek. LAJOS BERKES 85 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 In what follows, I will argue that in the eighth century the circle of scribes writing Greek both for private purposes and for the Arab administration became gradually restricted to those trained at church institutions, often themselves members of the clergy. It seems reasonable to posit that the sustained use of Greek in the liturgy after the conquest guaranteed the continued significance of churches and monasteries in Greek education.15 One of the most evident indications of the participation of clergymen in the Arab administration are tax documents that were drafted and signed by them. For example, the tax receipt SPP VIII 837 was signed by “the most humble priest Georgios”; another, SPP VIII 702, was signed by the deacon Theodorakios. Nevertheless, it is challenging to ascertain the precise role of these clergymen in these administrative procedures. Were they engaged as scribes or also as tax collectors? Typically, the spe- cifics of the collection process, including the location and the process of document issuance, remain elusive. Tax collection fre- quently constituted an internal matter within a tax unit, such as a monastery, and so internal tax documents were not necessarily issued by state officials. Another way to investigate the background of the scribes responsible for writing documents for the Arab administration is to engage with their handwriting. In the second part of the eighth century in Egypt, the minuscule was the sole form of Greek documentary handwriting in use, as the cursive, which had previously been employed for letters and legal documents, ceased to be used concurrently with these documentary genres.16 It is therefore of interest to note that we find several Coptic legal 15 Most of the bilingual Coptic-Greek liturgical manuscripts were copied between the sixth and tenth century, see R. Cribiore, “Greek and Coptic Education in Late Antique Egypt,” in S. Emmel et al. (eds.), Ägypten und Nubien in spätantiker und christlicher Zeit II (Wiesbaden 1999) 279–286, at 283. 16 This does not apply to literary texts, which were still written in other styles too. On the distinction between the minuscule and the cursive in the papyrological sense see the classic discussion of H. I. Bell, “Two Official Letters of the Arab Period,” JEA 12 (1926) 265–281, at 265–266. 86 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 documents that bear notarial subscriptions by clergymen written in the same kind of professional Greek minuscule used in con- temporary tax documents. Figure 2: CPR IV 126, notarial subscription © Österreichische Nationalbibliothek, Papyrussammlung For instance, CPR IV 126 contains the subscription of “the most humble Chael, son of the blessed Damianos, deacon from the church of the archangel Michael” ( fig. 2).17 Similarly, the Theban child donation document P.KRU 96 from the year 775 contains the subscription of “Elisaios, the most humble priest from the city of Hermonthis.”18 The fact that these subscriptions were written in a professional minuscule is even more interesting because notarial subscriptions in Greek documents in the late Byzantine period and early decades of Arab rule were not written in minuscule hands, but in peculiar, regionally different 17 This translation is based on a revised reading of the papyrus: instead of † δι’ ἐµοῦ Χαὴλ ἐλ(α)χ(ίστου) δια(κόνου) τοῦ µακα(ρίου) Δαµιανοῦ νο(µικὸς) διάκο(νος) ἀπὸ Ἀρχ( ) Μαχ( ) ἔγρ(αψα) (καὶ) µαρτ(υρῶ) †, I read † δι’ ἐµοῦ Χαὴλ ἐλ(α)χ(ίστου) δια(κόνου) τοῦ µακα(ρίου) Δαµιανοῦ διακό(νου) ἀπὸ Ἀρχ(αγγέλου) Μιχ(αὴλ) ἔγρα(ψα) (καὶ) µαρτ(υρῶ) †. I assume that the double mention of διακόνου is a scribal error, given that it would be unusual in such a subscription to provide information about the deceased father's occupation. I thank Nikolaos Gonis for suggesting the reading Δαµιανοῦ (Δαµιανονο pap.) instead of Δαµιανοῦ νο(µικός). On plural abbreviations used for singulars see also L. Berkes, “Bemerkungen zu Verwaltungsdokumenten aus dem früh- arabischen Ägypten,” ArchPF 68 (2022) 366–378, at 377–378. A new edition of the whole document is being prepared by Esther Garel. 18 The subscription reads: † διʼ ἐµοῦ Ἠλισαῖος ἐλαχ(ίστου) πρε(σβυτέρου) ἀπὸ Ἑρµώνθ(εως) / ἐγένετο †. LAJOS BERKES 87 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 styles.19 These peculiar subscriptions were replaced by the min- uscule in the eighth century, because other documentary styles of writing than the minuscule disappeared and scribes were no longer trained to write them. That the same kind of professional minuscule appears in both private legal documents and papyri of the state administration indicates that the notaries responsible for their production were trained at the same institutions as the scribes employed by the Arab administration. Since many of the notaries are explicitly identified as clergymen, it seems reasonable to infer that they probably received their education at a monastery or a church. Indeed, the Greek minuscule is well attested in the internal administration of church institutions. For instance, in eighth- century documents of the monastery of Apa Apollos in Bawit the professional minuscule is present in the formulaic Greek parts of Coptic papyri.20 Writing exercises from the archives of eighth-century pagar- chies, the administrative centres of cities, further support the hypothesis that scribes of the Arab administration received their training at church institutions or were even connected to one. We find several scribal exercises in Greek (sometimes at the same time in Coptic and Arabic) that were written on the back of liturgical papyri.21 One could easily imagine that these liturgical texts were not needed any longer and so were used by the clergymen working for the Arab administration for practicing documentary formulas or vice versa (see 91 below). Furthermore, regular administrative writing exercizes also often contain 19 See J. M. Diethart and K. A. Worp, Notarsunterschriften im byzantinischen Ägypten (Vienna 1986). 20 See e.g. P.Brux.Bawit 1–27 (all eighth century). 21 E.g. P.Berol.inv. 5476 published in O. Stegmüller, Fragmente frühchristlicher Poesie zum Epiphaniefestkreis (diss. Freiburg 1941) 104–105, and K. Treu, “Christliche Papyri IV,” ArchPF 22/23 (1974) 367–395, at 386 (nr. VII.1). The identification of the verso as a writing exercise from the Arab admini- stration is based on autopsy. 88 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 quotations from the Psalms,22 which constituted the basis of Greek education in the Byzantine and early Islamic period.23 Scribes writing Greek documents in the eighth century After the conquest of Egypt in 642, a Greek-speaking bureaucracy was still in place, and this meant a readily available supply of trained scribes. However, as time progressed, the necessity for individuals to learn Greek diminished, as the language lost its institutional prominence and Coptic became the dominant tongue in the private sphere. Nevertheless, it is evident that Greek retained its significance in the ecclesiastical context. The Greek festal letters from the eighth century written by the Alexandrian patriarch provide evidence of a high level of proficiency in complex Greek.24 It seems plausible to suggest that the first part of the eighth century saw the last generation of scribes trained in Greek who were not necessarily connected to the church. This hypothesis is supported by the evidence pre- sented above, which suggests that the Heracleopolite notaries active in the first quarter of the eighth century may have been the last of their kind. From the second half of the century on- wards, the expertise required to draft Greek documents was likely confined to clergy members, although there may have been exceptions.25 It is important to note that clergymen acting in administrative or notarial roles also existed during the Byzantine period and 22 For a list of Psalm citations in writing exercises from the early Islamic period see L. Berkes, “Schreibübungen mit einem Psalm-Zitat: Neues aus der Verwaltung des Fayums im 8. Jh.,” ZPE 188 (2014) 241–244, at 242, esp. n. 15. 23 M. Mavroudi, “Greek Language and Education under Early Islam,” in B. Sadeghi et al. (eds.), Islamic Cultures, Islamic Contexts: Essays in Honor of Professor Patricia Crone (Leiden 2014) 295–342, at 312. 24 E.g. BKT VI 5 (713 or 719). On the papyrological examples see most recently M. Konstantinidou, “Festal Letters: Fragments of a Genre,” in Pro- ceedings of the 28th International Congress of Papyrology (Barcelona 2019) 144–152. 25 The same probably applies to Coptic legal documents as well. LAJOS BERKES 89 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 were certainly not a post-conquest phenomenon.26 The distinc- tion lies in the fact that from the beginning of the eighth century on, only clergymen were capable of setting up Greek documents and therefore remained the only pool from which scribes could be hired. It is unsurprising that the Arab administration relied extensively on Christian scribes, despite the potential for ideo- logical conflicts that this may have caused. The Abbasid court at Baghdad, for example, is well-documented to have employed numerous Nestorian secretaries (kuttāb) over the centuries, who were trained at monasteries.27 While the upper echelons of Egyptian administration were predominantly Muslim from the eighth century on, the majority of their scribes remained Christian, as evidenced by the crosses that appear on documents issued in their name. It is possible that some of these scribes converted to Islam, but their presence is challenging to discern.28 Greek, Coptic, and Arabic in eighth-century offices The writing exercises pertaining to pagarchical offices under- went a notable shift towards bilingualism or trilingualism during the latter half of the eighth century. Scribes trained in Greek minuscule were no doubt equally capable of setting up docu- ments in Coptic as well. However, it is more difficult to assess whether they were able to write Arabic too. While there are 26 See the evidence collected in G. Schmelz, Kirchliche Amtsträger im spät- antiken Ägypten nach den Aussagen der griechischen und koptischen Papyri und Ostraka (Munich 2002) 241–254. 27 See in general C. Cabrol, “Une étude sur les secrétaires nestoriens sous les abbassides 762–1258 à Bagdad,” Parole de l’Orient 25 (2000) 407–491, and for the training of the scribes esp. 425. 28 In several eighth/ninth-century documents, crosses are replaced by a symbol consisting of two slanting strokes (//). However, we cannot know if the choice of this symbol was always the scribes’ own decision or was perhaps enforced in the administrative offices. On this symbol see T. S. Richter, “Spätkoptische Rechtsurkunden neu bearbeitet (III): P.Lond. Copt. I 487 – arabische Pacht in koptischem Gewand,” JJurP 33 (2003) 213–230, at 223– 230. 90 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 examples that suggest the possibility of scribes proficient in writing in Greek, Coptic, and Arabic, it is not possible to reach a definitive conclusion in these cases, given the limited and cir- cumstantial nature of the available evidence.29 Common sense suggests that there were indeed such scribes: the question is, then, how many of them were to be found in an administrative office in the second part of the eighth century? How shall we imagine the internal work organisation of these bureaus? Figure 3: SPP X 70, list of villages © Österreichische Nationalbibliothek, Papyrussammlung 29 See the case study L. Berkes and Kh. Younes, “A Trilingual Scribe from Abbasid Egypt? A Note on CPR XXII 17,” ArchPF 58 (2012) 97–100. LAJOS BERKES 91 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 A number of documents provide insight into the manner in which eighth-century administrative offices managed their tri- lingual documentation. In the Fayum, we find a list of villages written in Greek accompanied by their Coptic equivalents ( fig. 3).30 Furthermore, there is also a Greek account containing vil- lages to which the Arabic names of the settlements were added.31 Similarly, in an Arabic account of the region of Oxyrhynchos, Greek equivalents of Arabic village names were added supra- linearly.32 This suggests that the data gathered from Coptic and Arabic documents were ultimately summarized in Greek ac- counts by scribes who were proficient in both Greek and Coptic, as well as at least able to read, if not necessarily able to write, Arabic. It seems probable that a process of transition occurred in these offices during the latter part of the eighth century, when Arabic assumed an increasingly prominent role, and a growing number of scribes who had received training in Arabic began to be em- ployed. It is also possible that some scribes may have switched from writing in Greek and Coptic to writing in Arabic in the administrative offices in which they were employed. An indica- tion for this could be liturgical papyri which were written on the back of reused Arabic administrative documents. this may sug- gest that Christian scribes working for the Arab administration reused their waste papers for writing liturgical texts.33 Indeed, as 30 SPP X 70 (Fayum, eighth century). 31 P.Ross.Georg V 73 (Fayum, eighth century). For another Greek account, in which Arabic village names were added, see L. Berkes and N. Van- thieghem, “Notes on the Careers of Nāǧid b. Muslim and ῾Abd al-Malik b. Yazīd,” ChrÉg 95 (2020) 154–161, at 159–161. 32 P.Khalili 1 (mid-eighth century) with the corrections made by N. Gonis, “Some Oxyrhynchite Villages in the Eighth Century,” ChrÉg 75 (2000) 128– 132. 33 For discussion of liturgical papyri that were reused for Arabic documents see Á. T. Mihálykó, The Christian Liturgical Papyri. An Introduction (Tübingen 2019) 82–83, esp. n.42. 92 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 already mentioned, well-trained, Arabic-speaking Christian secretaries often appear, even at the Abbasid court.34 The abandonment of Greek in the caliphal administration Generally, the use of Greek as an administrative language seems to have followed similar patterns in Egypt as in other ex- Byzantine provinces of the Caliphate. Papyri from Nessana in Palestine, which include several administrative documents in Greek and Arabic from the second part of the seventh century, show formularies and structures remarkably similar to Egyptian documents.35 Some remarks in narrative sources may also confirm the impression that at the turn of the seventh-eighth century, Christian officials and the Greek language still played an important role in other ex-Byzantine provinces.36 While there is no Greek documentary evidence comparable to the Nessana papyri from the Abbasid Near East, there are other hints that fit very well into the narrative emerging from the Egyptian papyri. The chronicle of Theophanes offers significant insights in its account of the events that transpired in the year 758/9: In this year the Arabs maliciously expelled the Christians from government chanceries for a short time, but were once again 34 See Cabrol, Parole de l’Orient 25 (2000) 407–491. Moreover, even in the tenth century, Sāwīrus b. al-Muqaffaʿ was employed in the Arab administra- tion prior to becoming a monk: A. Papaconstantinou, “ ‘They Shall Speak the Arabic Language and Take Pride in it’: Reconsidering the Fate of Coptic after the Arab Conquest,” Le Muséon 120 (2007) 273–299, at 289. 35 In P.Ness. III. For a case study pointing out the similarities (and differences) between the entagia from Nessana and Aphrodito see Garosi, Projecting a New Empire 79–81. 36 See A. Papaconstantinou, “ ‘What remains behind’: Hellenism and Romanitas in Christian Egypt after the Arab conquest,” in H. Cotton et al. (eds.), From Hellenism to Islam. Cultural and Linguistic Change in the Roman Near East (Cambridge 2009) 447–466, at 450, esp. n.18. LAJOS BERKES 93 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 obliged to entrust the same duties to them because they were unable to write numbers.37 Theophanes portrays a reform that sought to displace Christians from government offices, which ultimately failed because the administrative apparatus was unable to function without the technical expertise of Christian scribes, particularly in the area of bookkeeping.38 While there is no direct confirmation of this event in Arabic or other Oriental sources, the presence of Christians, and more generally dhimmī-s, in the caliphal admini- 37 Chronogr. 430–431 De Boor, τούτῳ τῷ ἔτει φθόνῳ τοὺς Χριστιανοὺς ἐκώλυσαν Ἄραβες ἐκ τῶν δηµοσίων χαρτοθεσίων πρὸς ὀλίγον χρόνον, αὖθις ἀναγκασθέντες ἐγχειροῦσιν αὐτοῖς τὰ αὐτὰ διὰ τὸ µὴ δύνασθαι αὐτοὺς γράφειν τοὺς ψήφους; transl. C. Mango and R. Scott. 38 This is consistent with Theophanes’ observation that the Arabs “were unable to write numbers,” i.e. cyphers. In discussing a similar episode in his entry concerning the year 706/7 (De Boor 376) he states: “He [the caliph al- Walīd] also forbade that the registers of the public offices should be written in Greek; instead, they were to be expressed in Arabic, except for the num- erals, because it is impossible in their language to write a unit or a pair or a group of three or 8½ or 3 (?). For this reason they have Christian notaries until this day,” καὶ ἐκώλυσε γράφεσθαι Ἑλληνιστὶ τοὺς δηµοσίους τῶν λογο- θεσίων κώδικας, ἀλλ' ἐν Ἀραβίοις αὐτὰ παρασηµαίνεσθαι, χωρὶς τῶν ψήφων, ἐπειδὴ ἀδύνατον τῇ ἐκείνων γλώσσῃ µονάδα ἢ δυάδα ἢ τριάδα ἢ ὀκτὼ ἥµισυ ἢ τρία γράφεσθαι· διὸ καὶ ἕως σήµερόν εἰσι σὺν αὐτοῖς νοτάριοι Χριστιανοί (I thank Kent Rigsby for pointing this passage out to me). On cyphers in these two passages of Theophanes see K. Krumbacher, “Woher stammt das Wort Ziffer (Chiffre)?” in J. Psichari (ed.), Études de philologie néo-grecque. Recherches sur le développement historique du grec (Paris 1892) 346–356, at 351–352. However, it is well documented that Arabic documentary papyri fre- quently utilized Greek numerals well after Greek had been abandoned as an administrative language. Theophanes' statement therefore appears to be either misinformed or deliberately misleading. A complete change of admini- strative personnel would have been challenging, since only a few specialists were able to properly read and write fiscal documents employing a complex system of abbreviations and technical terminology. Similar problems arose according to historiographical sources in the transition from Persian to Arabic administrative terminology: see M. Sprengling, “From Persian to Arabic,” AJSL 56 (1939) 175–224, at 196. 94 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 stration had been an ideological problem since the Umayyad period. ʿ Umar II (717–720) was the first to attempt to ‘purge’ the bureaucracy of non-Muslims, and many of his successors fol- lowed in his footsteps. This is also attested for the early years of Abbasid rule: Theophanes’ testimony is consistent with this trend.39 Whatever was the exact reality behind Theophanes’ probably partisan account, we can reasonably assume that there was an administrative reform in the caliphal bureaucracy ca. 758/9. Indeed, this would be consistent with the Egyptian evidence. As mentioned above (82–83), in Egypt, our last Greek administra- tive documents, which are not accounts, come from the 750s. Tax demand notes, for instance, that were issued mostly in the name of Muslim officials no longer appear in Greek or in Coptic,40 but also not in Arabic after the early 750s.41 This seems to suggest that these genres disappeared because of an admini- strative reorganization that made them superfluous. 39 I. Rochow, Byzanz im 8. Jahrhundert in der Sicht des Theophanes. Quellen- kritisch-historischer Kommentar zu den Jahren 715–813 (Berlin 1991) 175–176, mentions no parallel sources. See however A. Fattal, Le statut légal des non- musulmans en pays d’Islam (Beirut 1958) 248–252, and also the summary by Cabrol, Parole de l'Orient 25 (2000) 428–430. 40 It seems that the fate of Coptic and Greek as administrative languages went hand-in-hand. Cf. van der Vliet, JJurP 43 (2013) 194–195: “I am inclined to see Greek and Coptic within the cultural constellation of late- antique Egypt as Siamese twins rather than as markers of conflicting religious or national identities. As soon as the shared culture of late-antique Egypt started to lose its hold, under the Abbasids, from the year 750 onwards, both halves of the twin couple started to wither and decline.” The number of Coptic administrative documents declined significantly from the ninth cen- tury onwards, with the surviving examples originating from the lowest levels of administration, see e.g. L. Berkes, A. Delattre, and N. Vanthieghem, “A Ninth-century Coptic Tax Refund Document. Reedition of CPR IV 197,” ChrÉg 96 (2021) 164–172, and L. Berkes and N. Vanthieghem, “A Late Coptic Tax Receipt from the Egyptian National Library (P.Cair.Nat.Library Inv. 3890),” Journal of Coptic Studies 21 (2019) 13–17. 41 On tax demand notes see n.12 above. LAJOS BERKES 95 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 Another conclusion we may draw from Theophanes’ account is that Greek might have been used for a similarly long period in other ex-Byzantine provinces for administrative purposes as in the Nile valley. It is worth recalling the story of Ḥunayn b. Ishāq (809–873), highlighted by Maria Mavroudi: he criticized Yaḥya b. al-Biṭrīq, a Syriac scholar working on the translation of Greek texts into Arabic at the turn of the eighth-ninth centuries for knowing “only the language of the Byzantines of today and its writing and these are the connected letters, not the old Greek individual letters.”42 Mavroudi suggested that his “training in the Greek language and script was meant as preparation for an administrative career in which only the cursive43 would have been needed, but the change in the administrative apparatus of the caliphate may have forced him to pursue a career as translator.” It is conspicuous that this occurred precisely in the late eighth to early ninth centuries, when the use of Greek ceased in the Egyptian administration, and the bureaucratic apparatus went on to function primarily in Arabic. This suggests an administrative reform, which eliminated Greek from the ex- Byzantine provinces. The scribes responsible for producing Greek documents for the Arab administration were compelled to seek alternative sources of employment. In Egypt, it is plausible that some of them continued to serve as Coptic notaries.44 It is also noteworthy that the scribes of the Arab admini- stration responsible for writing Greek documents appear to have 42 As translated by Mavroudi, in Islamic Cultures 322. 43 What Mavroudi calls a “cursive,” I have referred to in this paper as “minuscule” in keeping with papyrological usage, see n.16 above. 44 This is illustrated by CPR II 224 (Arsinoite, 812), a fragment of a Coptic legal document. The dating clause of the text is written in a handsome Greek minuscule characteristic of late-eighth-century administrative documents; an image is available at https://data.onb.ac.at/rec/RZ00008173. 96 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 received their training at church institutions also in other ex- Byzantine provinces of the Caliphate besides Egypt. The Greek grammars from the Near East, which were widely read in in- tellectual circles of the eighth- and ninth-century Caliphate, indicate that the scribes employed by the Arab administration were drawn from the same pool as those serving in the church. These texts were “frequently written by and for people em- ployed in the administration of the church”; their popularity can be attributed to the utility of Greek for both state and church administration, as well as for translations.45 This is consistent with the Egyptian evidence. As argued above, scribes writing Greek documents for the Arab administration in Egypt appear to have been primarily clergymen or at least trained at ecclesi- astical institutions. Indeed, the latest Greek document from Egypt seems to stem from an ecclesiastical context.46 SPP III2 577 ( fig. 4) is a brief formulaic document, possibly a voucher, concerning a modest quantity of bread. The text is signed by Theodosios, priest, and written in the so-called sloping majuscule, a common script for Greek and Coptic liturgical texts, with the exception of the amounts and the date, which are in minuscule. All of this points to a church or monastic context, where Greek would still have been employed as an administrative language, even if in a fos- silized manner. This text is striking not only because of its Greek language at such a late date, but also because it attests to new realities: the Greek document signed by a priest is dated by the hijri year. The later fate of Greek language and education diverged con- siderably from that observed in other regions of the Near East. 45 Mavroudi, in Islamic Cultures 328. 46 See L. Berkes, “The Latest Identified Greek Documentary Text from Egypt: A Papyrus from 825 AD (SPP III2 577 Reconsidered),” ZPE 209 (2019) 242–244. LAJOS BERKES 97 ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 Figure 4: SPP III2 577, A.D. 825 © Österreichische Nationalbibliothek, Papyrussammlung The Levant seems to have already turned to Arabic in the middle of the eighth century—much earlier than the Nile valley.47 Theodore Abū Qurrah, bishop of Harrān (ca. 750–823) is the earliest Christian author writing in Arabic, while the first Copt known to have produced theological works in Arabic is Sāwīrus b. al-Muqaffaʿ (ca. 915–987). Still, Greek education re- mained much stronger in the Levant than in Egypt, also on account of the Greek-into-Arabic translation movement.48 In contrast to this, Egypt does not seem to have been affected significantly by this movement. In the Nile valley, Greek sur- vived reduced to the church sphere and was less and less under- stood. Hymns and other liturgical texts in Greek from medieval Egypt are often written in a spelling and grammar which is 47 Interestingly, Greek epigraphical data also points to similar tendencies: “the remains of churches at several rural sites west of the Jordan indicate continuous patronage of religious buildings by a Christian elite commem- orated in Greek inscriptions that begin in the late seventh and span the eighth century. The latest among them may have been written in the year 785 AD”: Mavroudi, in Islamic Cultures 309. 48 Mavroudi, in Islamic Cultures 317. 98 GREEK DOCUMENTS AND THEIR SCRIBES ————— Greek, Roman, and Byzantine Studies 64 (2024) 80–98 barely understandable. In essence, as the great papyrologist H. I. Bell articulated in 1922, after the Arab conquest “Greek was now, in Egypt, a doomed language.”49 October, 2024 Humboldt-Universität zu Berlin Theologische Fakultät Seminar: Neues Testament Burgstraße 26 10099 Berlin, Germany lajos.berkes@hu-berlin.de 49 “Hellenic Culture in Egypt,” JEA 8 (1922) 139–155, at 155. This is a revised version of a paper originally entitled “Greek as an Admini- strative Language in Islamic Egypt and the Caliphate,” to which I have referred previously in other publications. I would like to express my gratitude to the two anonymous reviewers for their constructive criticism, which has helped to refine the focus of this paper and improve the quality of my argu- ment. They directed me to C. Palombo, The Christian Clergy’s Islamic Local Government in Late Marwanid and Abbasid Egypt (diss. Princeton 2020), which makes several references to previous versions of this article (e.g. p.51, esp. n. 100 or p.164, esp. n.67); in anticipation of the publication of a fuller study in book form, I will not engage with its argument in detail here.