The ‘Year According to the Arabs’: The Rise of the ‘Hijra’-Era in the Context of the Administrative Structures in the Early Islamic Empire

ABSTRACT

This article offers a survey of the spread and function of what is currently known as the hijrī calendar among different socio-linguistic milieus of the early Islamic empire. In particular, it analyses insider and outsider descriptions of the new imperial calendar as a window into the cultural profile of mediators between the Arabic and Graeco-Egyptian milieus in early Islamic Egypt. I argue that the ways the hijrī calendar was referred to in Greek and Coptic documentary texts diverged depending on the level of the issuing authority in the provincial administration: while documents issued by district officials label the era as the ‘year of the Saracens’ or use it without specifications, documents produced by the gubernatorial office use the designation ‘the year according to the Arabs’ (kata Arabas) instead. The main argument is that the kata Arabas label – as well as other formulaic peculiarities of documents produced in the provincial capital – can be linked to the employment of Hellenized Syro-Aramaean experts among the entourage of Arab governors appointed by Damascus. To flesh out the links between the gubernatorial chancery and a Syro-Aramean milieu, Egyptian evidence will be contrasted with Greek and Syriac texts from Syria-Palestine and Northern Mesopotamia.

KEYWORDS:

• Hijrī era
• calendrical practices in the early Islamic empire
• documentary sources
• Arabic–Greek mediation
• early Islamic imperial governance
• Egypt–Greater Syria relations

Introduction

In spite of the often-lamented dearth of contemporaneous historical narratives, the rise and consolidation of the early Islamic empire in the seventh and eighth centuries are blessed with a relative wealth of documentary evidence. Favourable climatic conditions have meant that the sands of Egypt, in particular, have preserved thousands of papyrological records in Arabic, Greek and Coptic dating as far back as the very inception of Islamic rule. Besides shedding light on several facets of everyday life, these texts offer unmediated insights into the strategies and perceptions that characterized Arab-Muslim imperial governance in the earliest decades after the rise of Islam.

Calendrical practices are one of the arenas in which the cultural otherness of Arab conquerors left some of its earliest visible marks on coeval documents and inscriptions. A clear manifestation of how the new political order left an impact on temporal conventions is the disappearance of the traditional method of dating official documents according to regnal and/or (post-)consular years after 641.¹ At the opposite end of the spectrum, reaction to Muslim rule likely boosted the use of the Diocletian era² among the Christian Egyptian population.³ The appearance of regional eras counting years from the Arab conquest (641–642) in seventh-century Coptic inscriptions from both the Delta and Upper Egypt further indicates that Christian communities perceived the conquest as a disruption in the habitual flow of history.⁴ But the calendrical novelties in post-conquest Egypt reflect not only the processes of reacting to and normalizing the Arab presence,⁵ but also elements of the cultural superstrate introduced by the conquerors. Specifically, papyri, inscriptions and coins indicate that what is currently known as the Islamic (or hijrī) era was already in place in Egypt shortly after the conquest.

The potential of dates and calendars as a doorway into the cultural world of early Islam has not been lost on scholarship. Discussions of temporal norms in early Islamic Egypt have mostly focused on the functional interplay of the different chronological systems employed within administrative and everyday life⁶ – including the several attempts to reform the Islamic calendar to better suit fiscal needs.⁷ Other approaches have emphasized the relevance of documentary findings for reconstructing the chronology of the Islamic calendar.⁸ Yet others have focused on the labels used for eras and calendars to glimpse at notions of groupness, in-group sentiments and political allegiances within the rising Islamic community or among the Christians living under Islamic rule.⁹ Conversely, the actual human agency behind the process of mediation between the new calendar implemented by the conquerors and the parallel chronological systems has largely been neglected.¹⁰

The purpose of this article is to examine the interaction between the Arab-Muslim milieu and not yet Arabicized local elites through the prism of notions attached to the A(nno)H(egirae)-era in seventh- and eighth-century documentary sources. Rather than focusing on the mechanics of synchronizing imperial and local chronological systems, I shall approach the scope of the era in the Egyptian administration and society from the angle of scribal practices. The first part of the article will survey the introduction, different uses and spread of the AH-system among the Arabic-writing and various socio-linguistic milieus of early Islamic Egypt compared with other regions of the empire. I shall then proceed to lay out the various perceptions of the AH-system that can be inferred from the insider and outsider descriptions of the era in documentary sources. Finally, I shall discuss how scribal conventions for describing AH-years can help us reconstruct the cultural profile of officials tasked with bridging the gaps between Arabic and Greek and Coptic writing milieus, thereby shedding some light on the governance system of the early Islamic empire. The scope of my discussion concerns the period before the first quarter of the ninth century, when documentary Greek texts fade out of the historical record.¹¹

The social reach of the new era

Information on the genesis of what is commonly referred to as the Islamic era rests entirely with descriptions made by authors of the Islamic age. The Muslim calendar supposedly originated in the lunisolar calendar in use in pre-Islamic Mecca, from which it differed most notably in the absence of intercalation.¹² The exact workings of pre-Islamic intercalatory practices are difficult to reconstruct but the sources indicate that control over the procedure rested with an individual called the nāsiʾ – also referred to with the obscure title al-qalammas (possibly connected to the Latin calendae).¹³ This office was reportedly passed down in the family of Ḥudhayfa b. ʿAbd b. Fuqaym of the Banū Kināna (ca. fourth century).¹⁴ The precious little information at our disposal indicates that the announcement of an intercalated year influenced the succession of months that were considered sacred by the Meccans and in which warfare was prohibited. Authority over intercalation could thus easily be manipulated for political ends. These political implications of intercalatory practice have been interpreted as a plausible backdrop for Muḥammad’s resolve to abolish intercalation altogether.¹⁵ While Muḥammad is credited with establishing the Muslim calendar already in his own lifetime,¹⁶ the epoch of 622 was only agreed upon – depending on the source – sometime between AH 16 and 18 (637–639) at the instigation of the caliph ʿUmar b. al-Khaṭṭāb (r. 634–644). Administrative paperwork up to that point had supposedly only been dated by month in the archive of the capital (at that point still located in Medina). The ensuing chaos is credited as providing the opportunity for devising a stable system of counting the years. After discarding the possibility of adopting a foreign chronography, the choice for the starting point of the new era fell on Muḥammad’s emigration (or hijra).¹⁷ Despite the epoch-making event having reportedly taken place on a Monday in the month of Rabīʿ I,¹⁸ the beginning of the year was retro-dated to 1 Muḥarram/16 July 622¹⁹ in order to place the start of the year after the pilgrimage in the twelfth month of Dhū al-Ḥijja. To what extent such accounts, composed centuries after the facts, reflect actual historical circumstances is unknown. Anecdotal information, however, is partially complemented by documentary evidence.

The Arabic writing milieu

A mere four–six years after the alleged introduction of the hijrī count, dated Arabic documents enter the documentary record. The earliest example of this is a bilingual Greek/Arabic receipt, currently housed at the Austrian National Library, Vienna, for the delivery of provisions to Arab troopers. The Arabic portion of this document is dated to ‘Jumādā I of the year 22’. The text does not qualify explicitly the chronological system it employs. The Greek part of the receipt, however, is precisely dated to 30 Pharmouthi in the first indiction (25 April 643). A correspondence between Arabic and Greek dates can thus only be obtained if one assumes that the system used in the Arabic text is the lunar count, reckoning years from July 622.²⁰ Another fragmentary Arabic document dated to the ‘year 22’ as well as another from the 20s are also known.²¹ In both cases, the lack of a double date makes it technically impossible to confirm that these papyri were dated according to the epoch of 16 July 622. There is, however, sufficient contextual evidence to assume that this was the case. In the late seventh and early eighth centuries, dated multilingual Arabic-Greek documents are relatively plentiful and allow us to determine that the dating system used in the Arabic parts followed the AH-era. The same considerations apply to dated Arabic monolingual documents, the date of which can be verified against prosopographical evidence (e.g. through the correspondence between dates and known periods in office of governors and other dignitaries) or through synchrony with known historical events. Finally, the few dozen Arabic papyri dated to a precise day and naming the day of the week, match dates calculated using modern chronological tables.²² Therefore, there is no reason to doubt that other less precisely dated Arabic documents used the same era and calendar.²³

Use of the AH-era in the Arabic documentary evidence is not only remarkably early but also virtually exclusive. Between 643 and the end of the eighth century, all Arabic documents from Egypt showing a complete or partially legible date are dated to an AH-year or else refer to one.²⁴ Unique among the territories of the rising Islamic Empire, Egypt is home to rich findings of Arabic epigraphy, some of which reach as far back as the earliest decades of Arab presence in the region.²⁵ Among these, one official decree²⁶ and 20 funerary stelae, most of which originate from the so-called Fatimid Cemetery in Aswān,²⁷ also employ the AH-system.²⁸ Perhaps more significant than the numerical datum per se is the fact that, throughout the timespan surveyed, no Arabic text shows a method for counting years that can be demonstrated to be different from the AH-system. While documents of official character are on average more likely to carry a date, there seems to have been little to no difference in how Arabic documents of either provenance were dated.

Data from Egypt are consistent with the sparser Arabic documentary evidence stemming from different corners of the early Islamic empire. In particular, dates according to the AH-era (i.e. dates that – while not specifying an era – make chronological sense when read as AH-dates) already appear from the 640s in graffiti from the Arabian Peninsula,²⁹ from the 670s in inscriptions related to caliphal patronage in the Hijāz and Syria³⁰ and from the 690s on Arabic all-epigraphic coinage³¹ as well as in Arabic documents on papyrus, paper and leather found in Palestine (seventh–eighth century)³² and Central Asia (eighth century).³³ Except for a handful of Arabian graffiti in which AH-dates are supplemented by an additional reference to salient historical events,³⁴ no pre-800 Arabic document (or Arabic version of bilingual documents) employs a different system for reckoning years. This is particularly significant in view of the fact that material evidence surviving outside Egypt comprises both text types well-attested in the region (such as tax demands and receipts) and written media that are comparatively scarce in the Egyptian theatre, such as coins, building inscriptions and graffiti. This striking consistency across far-flung regions and document types is indicative – if nothing else – of the pervasiveness of a sentiment of ‘mechanical’ solidarity within the early Arab-Muslim (and Arabic writing) community.

The Greek-writing milieu

While Arabic administrative texts are attested as early as the 640s, initially they are comparatively few. In Egypt – and elsewhere in the empire – key administrative positions continued to be occupied by Christian personnel drawn from the cadres of the local aristocracy and continuing local administrative practices. Throughout the first two generations or so after the conquest, Arabs, in fact, only rarely appear in the extant documents. Up until the mid-eighth century, the majority of the administrative workload, including outgoing correspondence issued by the office of Arab officials, continued to be carried out in Greek and, increasingly, in Coptic.³⁵ Such texts operating at the juncture between super- and substratal scribal traditions offer glimpses of processes of sociolinguistic boundary-marking and boundary-crossing. For instance, seventh- and eighth-century documents in Greek and Coptic issued in the name of Arab officials to non-Muslim subordinates in Egypt show renditions of Arabic epistolary formulae that are not attested in the coeval correspondence of Christian officials.³⁶ Another signal of the cultural liminality of such texts is that a few of them show the use of AH-years.

In his 1985 survey of AH-years in Greek and bilingual (Greek/Arabic and Greek/Coptic) papyri, Klaas Worp counted 33 texts – with several more similar ones having been edited since.³⁷ A large portion of the documents from Worp’s list are bilingual Arabic/Greek documents featuring AH-years in the Arabic part of the text only. Scriptural and material features of these papyri, however, indicate that the Greek and Arabic parts of bilingual documents were composed independently of each other³⁸ and written by the hands of different individuals.³⁹ The impression that the Arabic and Greek versions of seventh- and early eighth-century bilinguals were the result of separate redactional efforts is reinforced by the fact that the Arabic and Greek portions of demand notes are written about three months apart from each other.⁴⁰ Furthermore, references to notarial personnel in administrative documents of the same period classify scribes according to languages,⁴¹ suggesting that Greek and Arabic scribes worked in separate sections of the administration.⁴² It is only from the second half of the eighth century that biliterate scribes, trained to produce documents in multiple languages, enter the record.⁴³ Given the above, it is preferable not to count bilingual documents in which only the Arabic version shows AH-years as evidence for the use of the era in Greek texts.

In accordance with the criteria just mentioned, there are 39 Greek texts that include a reference to AH-years (see Appendix A). These span a relatively narrow spectrum of document types and invariably belong in the sphere of fiscal administration. The earliest confirmed occurrence of AH-years in Greek is to be found in 18 early eighth-century bilingual (Arabic/Greek) collective tax demands dispatched by the Arab governor Qurra b. Sharīk (in office 709–714) from the provincial capital of al-Fusṭāṭ to Basilios, administrator of the town of Aphrodito in Middle Egypt. From the second to the last quarter of the eighth century, more AH-dates are found in tax demands and tax receipts issued by Arab district governors (called pagarchs): in the Herakleopolite district⁴⁴ in the second quarter and in the Fayyūm⁴⁵ in the third quarter of the eighth century, with further isolated finds from the Hermopolite⁴⁶ and Lykopolite.⁴⁷ Eighth-century fiscal registers and accounts from district chanceries also feature entries organized according to AH-years (Appendix A). From a functional perspective, the use of the AH-year in Greek administrative texts from Egypt is not connected with the dating of the documents themselves. Rather, it concerns references to tax periods while the date proprio sensu is indicated by the indiction year.⁴⁸ It goes without saying that similar caches of texts that have come down to us in a fragmentary state might have originally contained references to AH-era dates. The fragmentary bilingual Arabic/Greek requisition orders issued by the offices of Maslama b. al-Mukhallad⁴⁹ (in office 667–682) and ʿAbd al-ʿAzīz b. Marwān⁵⁰ (in office 685–705), for instance, bear a close resemblance to the better-known demand notes from the eighth century and it is feasible that they originally indicated an AH fiscal year.

The Coptic writing milieu

Towards the second quarter of the eighth century, AH-dates also appear in Greek dating formulae proprio sensu set at the beginning of Coptic documents from Western Thebes. Unlike Greek texts, AH-years in Coptic papyri are not exclusive to administrative business in the strict fiscal sense of the word and rather belong in the legal sphere. Among the large caches of eighth-century Coptic legal texts from the town of Jēme (Medinet Habu) and the surrounding area, are a handful of testaments and donation acts dated by Egyptian day of the month, indiction, Diocletian year as well as AH-year and, occasionally, also by the name of the current pagarch of Hermonthis (by that time an Arab)⁵¹ and that of the local dioikētēs.⁵² In these texts, the inclusion of an AH-date as a (functionally superfluous) supplement to other chronological systems was likely aimed at conferring on the documents notions of officialdom.⁵³ Towards the end of the eighth century, AH-years occasionally appear in Coptic evidence outside officially sanctioned administrative and legal spheres. It is around this time that use of the AH-era for dating documents is attested (albeit sporadically) not only in transactions between Christian parties,⁵⁴ but also in monastic and ecclesiastical contexts. A case in point is a small number of tombstones from the monastery of Apa Jeremias in Saqqāra from the last quarter of the eighth century and early ninth century, which include a date according to the ‘year of the Saracens’ in addition to giving the Diocletian and indiction year.⁵⁵ The presence of AH-years within a monastic framework is indicative of the fact that the ‘Saracene year’ – as the official era of the empire – had some currency across religious divides, irrespective of stringent functional purposes. It is possible that the proximity of the Apa Jeremias monastery to al-Fusṭāṭ facilitated contacts with the Arab milieu and might be the reason behind the swifter spread of a typically Arabic scribal feature such as the use of AH-years as compared to monastic complexes further south, such as Bāwiṭ.⁵⁶ A group of five eighth-century Arabic travel permits issued to residents in the monastery, for instance, suggests frequent exchange with Arab authorities.⁵⁷ A similar example is an early ninth-century Greek receipt (or, possibly, voucher) for the delivery of bread dated to 210 AH (825) and approved by a ‘Theodosios the priest (presbyteros)’.⁵⁸

Putting the results together, clear trajectories emerge along geographical and temporal axes. In regional terms, the point of juncture between the Arabic and Greek writing milieu (and, by extension, the nexus at which AH-years appear in Greek and Coptic writings) moves from al-Fusṭāṭ to the districts of the Nile valley around the second quarter of the eighth century. This shift can be linked to well-known restructuring of the provincial administration taking place around the same time. After 717, all pagarchs attested in the sources bear Arab names, suggesting a policy to tighten the administration’s control over the countryside by installing more dependable Muslim officials without a local power base to supplant local Christian magnates.⁵⁹ More generally, evidence of Muslim landed property outside al-Fusṭāṭ and its surroundings becomes increasingly visible in the sources from the second quarter of the eighth century onwards.⁶⁰ This process of intensification of Arab rule over the Egyptian countryside appears to have been the conduit towards a wider dissemination of the AH-era outside the strictly administrative domain and into the legal realm and, eventually, beyond it. AH-dates in clearly Christian environments indicates that as early as the mid-eighth century the use of the era was not per se confessionally coded, but rather responded to the need (or will) to synchronize local and imperial time in selected official – and, in due time, private – interactions.

The perspective from Greater Syria

The social ecology of the AH-era in Egypt appears to have been somewhat different from that in the neighbouring regions. Several hundreds of seventh- and eighth-century Greek, Arabic and Middle Aramaic papyri have been discovered. As in the case of Egypt, some of the Greek documentary texts from Palestine stem from the milieu of the Arab administration. A group of about 200 Greek, Arabic, Latin and Christian Palestinian Aramaic⁶¹ papyri excavated in the town of Nessana in the Negev desert, in particular, includes eight bilingual Arabic/Greek demand notes for oil and wheat sent to the inhabitants of Nessana⁶² and Sykomazon⁶³ between 674 and 689.⁶⁴ The Greek part of the Nessana bilinguals is dated according to Julian month, indiction number and AH-year. Evidence for an even earlier use of AH-dates in Palestine is provided by a Greek inscription commemorating the renovations of a bathhouse in Hammat Gader sponsored by a certain ʿAbd Allāh b. Hāshim which is dated to ‘5 December, in the 6th indiction, the year 726 of the colony (i.e. the Gazean era), the year 42 according to the Arabs’ (662 CE).⁶⁵ A further mosaic inscription installed in a church in the town of Tamra is dated to June(?) of the 8th indiction, in the year 107 (725 CE).⁶⁶ Further insights into the use of the AH-era among Levantine Christians stem from the Syriac documentary record. In the Nestorian milieu in particular, the adoption of the AH-era appears to have been particularly precocious. As much is suggested by several seventh- and eighth-century colophons of Eastern Syriac manuscripts dated by both Seleucid and (again in a purely supplementary function) AH-eras, the earliest of which (a copy of the New Testament) was completed in 682.⁶⁷ To some extent, the AH-system also made inroads in private writing. A handful of Syriac graffiti found in Kamed el-Loz (Lebanon) are dated to 96 ‘of the rule of the Arabs’.⁶⁸ That the date indeed refers to the era of 622 can be inferred from the fact that the inscriptions explicitly state that they were written under Caliph al-Walīd (I) b. ʿAbd al-Malik (r. 705–715), who died in the same year. From internal references, we know that several of the authors were Christian stoneworkers, natives of northern Mesopotamia and associated, like other specialized workers from different corners of the Empire, with the construction of the nearby madīna in ʿAnjar (ʿAyn al-Jarr).⁶⁹ The engravers appear to have wanted to underscore the caliphal sponsorship of the building project as they go to unusual lengths to emphasize that the works were conducted under the ‘reign of Walīd s. of ʿAbd al-Malik amīr of the Saracens/mhaggrāyē’. Another interesting detail of the Kamed inscriptions is that they are the only instance of eighth-century non-Arabic texts that not only use the AH as a dating system proper but also do so without a supplementary date according to a different era. It is possible that workers far away from their home region and employed within a multiregional workforce preferred to use a super-regional, trans-communal chronological system rather than that of a more parochial era. Given the informal nature of graffiti as a medium, one also speculates whether, in a simpler fashion, the two-digit AH-year might not just have been considered more practical than the corresponding four-digit Seleucid year.⁷⁰

Though the evidence is comparatively slight, documentary and para-documentary evidence from Syria and northern Mesopotamia indicates two major differences in the use of the AH-years when compared to Greek and Coptic documents from coeval Egypt. The first is functional, as AH-years are used in dating formulae proper rather than for referring to fiscal years alone. The second is chronological as – at least as far as complete texts are considered – the earliest attestation of the Arab era in Palestine pre-dates its oldest occurrence in Greek materials from Egypt by more than a generation in the administrative domain, and by a decade in an inner Christian milieu. This chronological diastasis is even more pronounced if we take the Eastern Syriac milieu as a term of comparison, where use of AH-dates in manuscript colophons predates the earliest recorded use of the ‘year of the Saracens’ in Coptic manuscripts by almost 150 years.⁷¹ Stated differently, while in Egypt the AH-era remained in essence a chronological system used for, and by, administrators working in contact with Arab officials until the last quarter of the eighth century, boundaries between the Arab-Muslim and local cultural milieus appear to have been much more porous in regions with a Helleno-Aramean substrate.

Labelling the era

Despite giving insights into the general Sitz im Leben of the AH-era among different groups, documentary sources do not contain explicit information on the people charged with navigating across parallel chronological systems. In some instances, it is implied that the shared temporality of the week could function as the mediating element between the Islamic and local calendars. This is suggested by CPR XXII 23, a fragmentary account from late eighth-century Fayyūm pertaining to ‘1/2 of the third installment of the Saracens’ which contains entries organized according to weekdays converted into the corresponding Islamic and Egyptian days of the month.⁷²

An aspect in which the information derived from literary sources clashes with the extant material evidence pertains to the name of the era. As already referenced, Islamic tradition maintains that the hijra marked the founding event of the calendar as well as a symbolic closing of an earlier, less fortunate age of man’s salvation history, styled as the age of ‘ignorance’.⁷³ The eponymous event of the hijra, however, does not appear in the surviving material sources until the tenth century.⁷⁴ On the contrary, AH-years are almost invariably used without any explicit reference to the chronological system being employed. The exception to this rule is a number of registers of debts on papyrus from the 660s and 670s, possibly produced in the provincial capital of Fusṭāṭ.⁷⁵ In these texts, deadlines for the repayment of debts are indicated as ‘the year (sana) X’ followed by the phrase snh qaḍāʾ al-muʾminīn. This expression is difficult to interpret, with possible translations including ‘the year/calendar (reading snh as sanat) of the reimbursement [or: of the decree/power] of the Believers’⁷⁶ (the collective term by which adherents to the early Islamic community referred to themselves).⁷⁷ It has been suggested, furthermore, that the expression was also used in an abbreviated form simply as sanat (qaḍāʾ al-muʾminīn) in a number of coeval individual acknowledgements and receipts pertaining to debts.⁷⁸ In these instances, the dating clause would accordingly have to be translated as the ‘year X according to the calendar’.⁷⁹ Scholars who have tackled the issue have tended to interpret the expression qaḍāʾ al-muʾminīn as the original name of the AH-era which, following this interpretation, would only later have been connected to the hijra of Muḥammad.⁸⁰ The oddity of the formulation, however, has been explained by various interpretations. Y. Rāġib has argued that the expression denoted a lunisolar calendar, starting with the epoch of 622, used by the early Muslims for legal procedures⁸¹ – an interpretation which, however, is based on an inaccurate reading of the documents.⁸² Naïm Vanthieghem and Mathieu Tillier, who grappled with the issue more recently, maintain that the qaḍāʾ al-muʾminīn was one and the same with the classical Islamic calendar, but point out that contextual and lexical evidence would suggest that the early community considered the foundational event of the calendar to be the Treaty of Ḥudaybiyya rather than the hijra.⁸³ The issue is further complicated by the fact that, as Jelle Bruning points out, the expression could theoretically be read as sunnat qaḍāʾ al-muʾminīn and be interpreted as a legal clause rather than a date (translating roughly to ‘according to the reimbursement custom followed by the Believers’).⁸⁴

We are much better informed on perspectives on the new era from outside the Arab-Muslim community in non-Arabic documentary sources from the seventh and early eighth centuries. Different designations of the era are the subject of several dedicated articles, so a cursory overview will suffice here (see also Appendix B).⁸⁵ Greek and Coptic texts from Egypt occasionally refer to AH-years as the ‘year according to the Arabs’ or as the ‘year of the Saracens’. Only the former of the two expressions is also current in Greek texts from Palestine – a point on which I shall elaborate shortly. It is the Syriac milieu, however, that provides the most diversity when it comes to qualifications of the AH-era. In colophons of early Islamic manuscripts, AH-years are variably dubbed as the year ‘of the mhaggrāyē’ (in all probability a derivative of the Arabic muhājirūn),⁸⁶ ‘of the Arabs’ (ṭayyāyē), ‘of the Ishmaelites’ (ʿīshmʿly) or, alternatively, as the year of the ‘rule (shulṭānā) of the Ishmaelites’ or of the ‘rule of the children of Hagar’ (bnay hagar).⁸⁷ The labels of the Islamic era in Syriac colophons are largely consistent with those found in the graffiti of Kamed, which include the ‘year of the mhaggrāyē’ and ‘year X of/during the rule of the Arabs (ṭayyāyē)’.⁸⁸

Despite regional variations, common trends also emerge. Common to all designations of the era of 622 is a perception of the newly introduced epoch as centred around the establishment of Arab rule. In this regard, designations of the era of 622 are analogous to the non-official regional epochs mentioned at the outset of this article and starting their year-count with the Arab conquest of Egypt.⁸⁹ Furthermore, the calendar’s foundational event is not connected to a charismatic personality but rather to the collective entity of the conquerors – a trait they may share with the Arabic designation qaḍāʾ al-muʾminīn (depending on the interpretation). In terms of formulations, only the Syriac definitions incorporating the reference to the ‘rule of the Arabs’ might have been translations of the Arabic definition qaḍāʾ al-muʾminīn.

Where insider and outsider perspectives on the era differ the most is in the qualification of the conquerors’ group. In Arabic documents, AH-dates are characterized as the year par excellence without mentioning the hijra nor associating the system with a particular group. Conversely, non-Arabic definitions of the AH-era tend to stress its cultural foreignness by underscoring the link of this chronological system to the conquerors’ milieu. The frequent use of biblical and pre-Islamic ethnonyms in relation to the conquerors in Greek, Coptic and Syriac texts further indicates a tendency to assimilate the new rulers into Late Antique notions of alterity – though it is somewhat unclear how the post-conquest environment might have changed the semantic spectrum of previous ethnonyms.⁹⁰ One of the Kamed graffiti, in particular, defines the AH-date as the ‘year of the rule of the flat-nosed ṭayyāyē’, thus adding a layer of ethnic stereotyping to the characterization of the era.⁹¹ Only some of the Syriac graffiti and colophons use the collective identifier mhaggrāyē (Arab. muhājirūn), derived from social notions internal to the nascent Muslim community. Even in those instances, however, Syriac descriptions of the era are characterized by an underlying process of ‘biblicization’. This is most clearly exemplified by the pseudo-etymological understanding of the term mhaggrāyē, which is linked by Syriac scribes to the figure of Hagar and seems to be synonymous with ‘children of Hagar’.⁹² In summary, early Islamic non-Arabic definitions of AH-years share a perception of era of 622 as the cultural marker of a distinct outgroup.

The ‘year according to the Arabs’: a Syrian connection?

While labels of the AH-era have been scrutinized to investigate different notions and perceptions associated with the era itself,⁹³ they also shed some light on the individuals writing the documents. Focusing again on the Egyptian theatre, Greek and Coptic documentary texts from the Umayyad period in particular use the AH-era in three different ways, including the already mentioned ‘year according to the Arabs’ (also attested in Palestine) and the ‘year of the Saracens’. In addition, a number of texts from Egypt use AH-years without further specifications, giving the date simply as the ‘year X’. The chronological and geographical distribution of the evidence indicates that the use of the label ‘the year of the Saracens’ (at least as a label for the epoch of 622)⁹⁴ was initially a localized phenomenon restricted to the Theban region.⁹⁵ Only from the first quarter of the ninth century onward did its use spread to further regions of Egypt⁹⁶ as well as, from the tenth century onward, to nearby Nubia.⁹⁷ To the best of my knowledge, however, there have been no attempts to lay out patterns of use of the other names.

A closer analysis reveals that different designations for AH-years coincide with the different offices issuing the documents. When Greek (or a Greek version of bilingual) papyri produced by Muslim authorities in the provincial capital al-Fusṭāṭ include an AH-date, the latter is invariably labelled as ‘the year according to the Arabs’ (kata Arabas etos). Conversely, this formula is not found in Greek documents produced in Egypt at a district or local level. In tax demands and fiscal registers issued by the chancery of the pagarchies of Herakleopolis,⁹⁸ Fayyūm,⁹⁹ Hermopolis¹⁰⁰ and Lykopolis,¹⁰¹ in particular, dates according to Arab era are simply labelled ‘the year X’ (etos X). This difference is especially noticeable in archives that preserve texts from both gubernatorial and pagarchal milieus: in the archive of Basilios, administrator of Aphrodito, specifically, bilingual Arabic/Greek demand-notes coming from the governor’s office indicate AH-years ‘according to the Arabs’ (kata Arabas). In accounts and tax registers produced in Aphrodito for the pagarchy’s bookkeeping, on the other hand, AH-years feature simply as the ‘year X’ (Appendix A). Adding to the oddity of the kata Arabas label is also the fact that ‘Arab’ and Arab-cognates are not widely used in the seventh- and eighth-century Greek documentary texts from Egypt outside the Qurra dossier.¹⁰² Rather, the designations ‘Saracens’ and Moagaritai (like the Syriac mhaggrāyē, a derivative of arab. muhājirūn) are of more frequent occurrence. Furthermore, this peculiar way to indicate AH-years is not the only distinctive formal feature of Greek correspondence of Arab governors, when compared to texts produced in district-level chanceries. Eight-century Greek documents issued in the name of Arab governors, for instance, employ different opening invocations than those issued by Arab district officials: en onomati tou theou by governors and syn theo by pagarchs.¹⁰³ More generally, Greek letters issued by Arab governors typically feature renditions of Arabic epistolary formulae that are unattested in the Greek correspondence by pagarchs.¹⁰⁴ In other words, it appears that Greek writing personnel in the service of early 8th-century Arab governors had a partially different scribal training than Greek writing clerks in the service of pagarchs.

Seeing that the use of the kata Arabas formula associates Greek texts from both Egypt and Palestine, documents from the Levant may offer a framework for understanding the terminological discrepancies between the different labels used in the Egyptian context. As already mentioned, evidence for the use of AH-dates in Greek writings from Palestine is provided by the 7th-century bilingual demand notes from Nessana (674–689) as well as the two inscriptions from the bathhouse in Hammat Gader (662) and the church in Tamra (725). Both the Nessana documents and the Hammat Gader inscription introduce AH-dates with the kata Arabas formula. Finally, the AH-era is used without specifications (etos X) only in the later Christian inscription from Tamra.¹⁰⁵ Thus, it appears that in Palestine, unlike in Egypt (where different offices and types of documents used different designations for AH-years), official texts produced in different districts for widely different purposes use one common kata Arabas formula while the diction ‘year X’ only features in texts produced outside the milieu of the Arab administration. After crossing dating clauses with other formulaic features of the documents in question, further correspondences between the Nessana bilinguals and gubernatorial documents from Egypt emerge: both groups of texts begin with the invocation en onomati tou theou. Determining the standing of the Arab officials mentioned in the Hammat Gader inscription and the Nessana demand notes in the administrative hierarchy of Palestine is somewhat difficult. The ʿAbd Allāh b. Hāshim/ʿĀṣim who sponsored the Hammat Gader inscription and the two officials responsible for issuing the demand notes from Nessana are not among the (sub-)governors¹⁰⁶ of the provinces of Urdunn and Filasṭīn known through literary sources. This suggests that we are dealing with district officials.¹⁰⁷ This conclusion is supported by the fact that the operating base Arab officials in the Nessana dossier appears to have been Gaza, a city never known to have been the capital of the (sub-)province of jund Filasṭīn and which is, on the contrary, explicitly identified as the capital of a district (kūra).¹⁰⁸

Returning to calendrical practices, there are some hints that the kata Arabas formula originated within the Levantine scribal idiolect prior to the coming of Islam. A small number of 6^th-century Greek funerary inscriptions from the Beersheba and Hebron regions are dated by indiction year and Eleutherian or Gazean eras while the day of the month is given according to the Macedonian calendar.¹⁰⁹ In these inscriptions, the Macedonian months in use in the Roman province of Arabia are described as ‘according to the Arabs’ (kata Arabas).¹¹⁰ More generally, ‘Arab(-cognates)’ are much more common in Syrian Greek epigraphy in the decades leading up to the Arab conquest than in post-conquest Egypt.¹¹¹ The evidence at our disposal thus suggests that the kata Arabas formula was rooted in Syrian Greek scribal practices. In fact, this is not the only instance of terminology and spellings of seemingly Syrian origin appearing in administrative texts from early Islamic Egypt. A few neologisms and variant spellings supplanting technical terms attested in the region prior to the conquest can be linked to the import of Syrian Greek administrative jargon.¹¹² A similar influx of Syrian technical terminology is in evidence in Arabic administrative documents as well: a surprising number of originally Greek technical terms used in papyri from Egypt were borrowed into Arabic through Middle Aramaic dialects and not directly from the local Greek and Coptic jargon, as one might have expected.¹¹³ A possible explanation is that Arab officials trained in Damascus, at the time capital of the empire, and introduced terminological elements originally rooted in the Syrian administrative idiolect into Egypt.¹¹⁴

Anecdotal literary accounts do in fact suggest that Hellenized Syro-Aramean experts occupied key positions both in the central chancery in Damascus and in the mobile entourage of Umayyad high officials. Muslim historiography provides ample anecdotal evidence for the employment of Hellenized personnel of Syro-Aramean and Christian Arab descent both in the imperial metropole of Greater Syria and in the provinces.¹¹⁵ The most famous example is possibly that of the Manṣūr family, and particularly of Sarjūn b. Manṣūr, who was head of the central dīwān of Damascus under four consecutive caliphs.¹¹⁶ Similar figures occupying key positions in the Umayyad domestic administration were reportedly the Damascene physician-turned-hired-assassin Ibn Uthāl/Athāl, who was put in charge of the dīwān of Ḥimṣ by Muʿāwiya I, the Palestinian secretary of Sulaymān (I) b. ʿAbd al-Malik and overseer of the construction of Ramla, Bitrīq (< gr. Patrikios) b. al-Nakā/Bakā,¹¹⁷ and the secretary to Hishām (I) b. ʿAbd al-Malik, also head of the dīwān of Ḥimṣ Tādhurī b. Usṭīn (< gr. Theodoros s. of Ioustinos)¹¹⁸. Bilingual Arabic/Greek personnel also accompanied Umayyad governors dispatched to the provinces. The chief-governor over the East, Ziyad b. Abihi/Abī Sufyān, for instance employed a certain Stephanos as his personal secretary.¹¹⁹ Turning towards Egypt, personnel from Palestine who came to occupy key positions in the province were, for example, one Yaḥyā b. ʿUmru from ‘the people of Ashqalon’ and his brother ʿIsā, who were appointed as head of the dīwān and of the shurṭa respectively during the second term of Ḥafṣ b. al-Walīd as governor of Egypt (742–745).¹²⁰ The city of Edessa in the Jazīra – which was still governed by the Christian official Athanasios s. of Andreas¹²¹ in the late seventh century – also provided a recruitment pool of mobile multilingual experts, such as Athanasios br. Gumōyē. The latter – a tax official and entrepreneur skilled in Greek and Arabic in addition to his native Syriac – notably served as tutor to the Umayyad prince ʿAbd al-ʿAzīz b. Marwān and became his chief secretary when the latter was appointed governor of Egypt. Contextually, Athanasios’ sons were also granted administrative positions (and generous remuneration).¹²² Furthermore, the figure of Athanasios is one of the fortunate instances in which biographical information derived from literary sources finds confirmation in documentary evidence contemporary to the facts. From a surviving inventory of expenses destined for ʿAbd al-ʿAzīz’s retinue found in Aphrodito (P.Lond. IV 1447), we learn that the ‘endoxotatos chartoularios Athanasios’ had his own complement of scribes, distinct from other retinues of secretaries in the central administration¹²³ and possibly formed by clerks of Syrian stock. When the following governor, ʿAbd Allāh b. ʿAbd al-Malik (in office 705–709), decided to replace Athanasios with an Arab secretary, the choice once again fell on a native of Syria, Ibn Yarbūʿ al-Fazārī from Ḥimṣ.¹²⁴

To reformulate this, lexical and anecdotal evidence suggests that in Egypt the task of mediating between the Greek and Arabic writing milieu in the early Arab administration rested with multilingual experts drawn from the ranks of the Syrian and North Mesopotamian Hellenized Syro-Aramean and Christian Arab officials who were part of the Arab governor’s retinue. Reasonably, the presence of this cohort of transregional experts in Egypt would only have interested the main administrative centres. This, in turn, explains why the correspondence from the capital would have different formulae than later documents issued at a district level – where Greek-writing functionaries would have been drawn from the ranks of the local population. With regard to the training of scribes in particular, there is a direct link between the formulaic structure of Greek documents issued in the chancery of the Sufyanid and Marwanid governors of Egypt and the Syro-Palestinian scribal milieu. Greek and Arabic scribal exercises for high-end correspondence on marble tiles have been excavated at different caliphal estates across Syria-Palestine.¹²⁵ Examples written in Greek in particular, show the same formula en onomati tou theou practised – just like the kata Arabas phrase – in the Fusṭāṭī milieu in Egypt and in the administrative documents from Nessana.¹²⁶ By contrast, the formula en onomati tou theou is – with a single exception¹²⁷ – not attested in known examples of scribal exercises from the coeval Egyptian countryside.¹²⁸

One of the main reasons behind the employment of bureaucratic personnel from Greater Syria in Egypt was probably their unique positioning as mediators between the Arabic and Greek writing milieus. Documentary papyri from sixth-century Petra, for instance, indicate that the local population, while writing in Greek, was versed in both Arabic and Aramaic dialects.¹²⁹ The assimilation of conjecturally Syrian personnel into the Islamic administration might also have been facilitated by the fact that, up to a few generations prior to the coming of Islam, Syria had been home to the Christian Arab phylarchs of the Ghassānid confederation, who resorted to both Greek and Arabic for their public statements.¹³⁰ After the conquest, members of Christian Arab tribes such as the Banū Ghassān in Syria and the Banū Taghlib in northern Mesopotamia went over to the Muslim camp,¹³¹ contributing to the rise of Islamic statecraft,¹³² and occasionally rising to high offices under the Muslim administration.¹³³ A second major factor was the social geography of the Umayyad imperial polity itself. Besides Damascus being the physical centre of Umayyad power, the Syrian army was both the military backbone of Umayyad power¹³⁴ and an important recruiting pool for its administrative class.

Concluding remarks

The early and generalized use of the AH-system in Arabic writings from the first two Islamic centuries is indicative of the prominence of the era of 622 as a distinctive cultural identifier of the Arab conquerors. The limited assimilation of the AH-system into different languages appears to have been mostly due to its status as the official imperial chronology rather than as a response to practical needs. This is suggested by the use of the ‘Arab year’ as a supplement (rather than an alternative) to competing chronological systems in Greek Coptic and Syriac writings of both official and – less frequently – private origin. Documentary evidence further indicates that the pace of the spread of the AH-system outside the Arab-Muslim milieu was proportional to the density of the Arab imperial presence. Most notably, AH-years appear to have made inroads into both administrative routines and non-administrative settings in Umayyad Greater Syria several generations prior to similar developments in Egypt.

The several ways in which non-Muslim ‘outsiders’ labelled the era of 622 speak to calendrical practices being one of the cultural interfaces through which boundaries between the Arab conquerors and local elites were both negotiated and established. The analysis of the different labels for the Arab era used in seventh- and eighth-century Greek texts, in particular, allows us to conclude that the different ways in which the AH-era is described in Egyptian Greek papyri appear to have been dependent on the issuing office, with the gubernatorial chancery using the label kata Arabas and the different pagarchal chanceries using the formula ‘etos X’. I have argued that this difference speaks not only to the existence of distinct clerical practices in different offices of the provincial administration, but also to the different origins of the notarial personnel active at either level. Furthermore, I have argued that the formula kata Arabas probably originated in the Greater Syrian Greek scribal tradition, where it was used to indicate the calendar of the Roman province of Arabia, and was introduced into Egypt from there. Finally, documentary, archaeological and anecdotal evidence converge in indicating that the vector of this development were mobile Hellenized Syro-Aramean experts employed as multilingual mediators in the provincial capital as part of Umayyad governance strategies.

Several crucial aspects of temporal practices in early Islamic Egypt remain hidden behind the wall of silence surrounding the surviving documentary evidence. Who was actually responsible for the arithmetical calculation of dates across calendars as opposed to who was charged with penning the documents? Were they one and the same group? Which processes of mediation were involved? These are some of the questions that remain unanswered. It is possible that new evidence will eventually present itself to draw a clearer picture of the individuals and environment that shaped the reception of the Arab era outside the Arab-Muslim minority. While that is still pending, scribal features provide us with precious glimpses of the otherwise inscrutable human face behind the process of creating a dialogue between the overlapping temporalities in the early Islamic administration and society.

Acknowledgements

Research for this article was carried out under the auspices of the VIDI Project Lived Time and could not have been accomplished without the generous financial support of the Dutch Research Council (NWO). I thank the other project members Sofie Remijsen, Renate Dekker and Elsa Lucassen, as well as Lajos Berkes, for commenting on an early draft of this article, saving me from many mistakes. All remaining mistakes are mine and mine only.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Universiteit van Amsterdam.

Notes

1 Bagnall and Worp, Chronological Systems, 43–54. Cf. also the disappearance of the era of Oxyrhynchus in the late seventh century.

2 Here and in the rest of the article, I use the terms ‘era’ and ‘epoch’ to denote ‘a fixed point in time from which a series of years is reckoned’ and ‘an instant in time chosen as the origin of a particular calendar era’, respectively.

3 Bagnall and Worp, Chronological Systems, 64; Papaconstantinou, ‘“What Remains Behind”’, 455. For notions of identity attached to the era of Diocletian (eventually rebranded as the ‘era of the Martyrs’) within the Coptic community in particular, see Mikhail, From Byzantine to Islamic Egypt, 131–4.

4 Four Coptic funerary stelae dated from 665 and 701 found at the monastic site of Kellia count years from or refer to ‘the dominion of the Lami’ (probably a rendition of the Arabic tribal nisba Lakhmī) in addition to giving indiction and Diocletian years. For an overview of the dating formulae on the Kellia inscriptions, see Luisier, ‘Années de l’indiction’. By comparing the different dates, it is possible to determine that the engravers of the stones consider Arab rule in the region to have started in November 641; see Bruning, Rise of a Capital, 27–30, who suggests that the pact of Babylon served as the starting date of Arab rule in the Kellia inscriptions. On Lakhmid involvement in the Arab conquest of Egypt, see Sijpesteijn, Shaping a Muslim State, 54–5. The nisba al-Lakhmī is also occasionally attested in Arabic papyri from Egypt from the eighth and ninth centuries, the earliest dated attestation being P.KhanLegalDocument (707; editions of Arabic papyri are quoted according to the abbreviations in Sijpesteijn, Oates, and Kaplony, ‘Checklist’). A slightly posterior inscription commemorating the renovation works in an Upper Egyptian topos of Apa Abraham (SB Kopt. III 1584; editions of Greek and Coptic papyri are quoted according to the abbreviations in Oates, Checklist) is dated to 13 Phaopi in the 11th indiction (= 10 October 697), in the year 414 of Diocletian, ‘while the Saracens rule over the country (for) year 55’. In this case, the engraver of the inscription considered Arab rule to have started in 642, the year of the capitulation of Alexandria. I thank Arietta Papaconstantinou and Renate Dekker for the reference.

5 Cf. in this context P.Paramone 18, a lease of a vineyard dated to the summer of 641, as the Arab conquest was still underway. Federico Morelli has made an interesting case for interpreting the document’s lacunary dating formula as the ‘first year of the regency (dioikēsis) [of ʿAmr and Martina]’, apparently an attempt to normalize the anomalous political situation by framing the Arab presence in terms of a form of vassalage to the Byzantines (Morelli, ‘‘Amr e Martina). For the dating of P. Paramone 18, see Gonis, ‘P. Paramone 18’.

6 See, for example, Grohmann, Arabische Chronologie; Bagnall and Worp, Chronological Systems, with a focus on the Late Antique and early Islamic Egyptian context; for a more general overview, see Taqizadeh, ‘Various Eras’.

7 For a general overview, see Cristoforetti, Izdilāq; for approaches based on documentary sources, see in particular Casson, ‘Tax-Collection’; Grohmann, Arabische Chronologie, 30–5; Thomann, ‘Institution of the Jalālī Calendar’.

8 Grohmann, Arabische Chronologie, 14; Jones, ‘Dotting of a Script’; de Blois, ‘Calendar at Mecca’, 190–1. Cf. also Robin, ‘Die Kalender’.

9 E.g. Hoyland, Seeing Islam, 547–8; Donner, ‘Qur’anization’, 86–7. For early Islamic Egypt, see in particular Rāġib, ‘Ère inconnue’; Mikhail, From Byzantine to Islamic Egypt, 129–35; Tillier and Vanthieghem, ‘Recording Debts’. For a sociological perspective on the role of calendrical practices as group identifiers, see Zerubavel, Hidden Rhythms, esp. 70–100.

10 For a discussion of coexisting temporalities in a pre-modern inner-Muslim context, see Thomann, ‘Tools of Time’; idem, ‘Ritual Time’.

11 To the best of my knowledge, the latest dated documentary text in Greek is SPP III² 577; see Berkes, ‘Latest Identified Greek Documentary Text’. For some consideration of the role of the Greek language and education in the Islamicate world, see Mavroudi, ‘Greek Language’.

12 The names of the months of Rabīʿ (‘vernal rain’), Ṣafar (etymologically connected to aṣfar ‘yellow’), Ramaḍān (etymologically connected to ramaḍ ‘scorching heat’), and Jumādā (etymologically connected to jamd ‘freezing’) hint at a stable synchrony between calendar and seasonal phenomena and have traditionally been regarded as proof of the original lunisolar character of the pre-Islamic Meccan calendar. See e.g. de Blois, ‘Calendar at Mecca’, 194–5.

13 Moberg, An-Nasí’, 53–4.

14 Based on the number of generations of known nāsiʾs, Armand Caussin de Perceval calculated that the practices of intercalation (or at least some form of institutionalization of the nāsiʾ’s office) would have started around 412 CE (Caussin de Perceval, Essai sur l’histoire, 244, 413–4).

15 De Blois, ‘Calendar at Mecca’, 192. For a gallery of case studies on the political prerequisites behind the developments and reformation of calendars in antiquity, see Stern, Calendars in Antiquity.

16 According to Muslim tradition, Muḥammad abolished the practice of intercalation during the so-called Farewell Pilgrimage (ḥajjat al-waḍāʾ) to Mecca in 632. Assuming a synchrony between the beginning of the year in the pre-Islamic Meccan and Jewish calendars and that the synchronism was maintained until Muḥammad’s decision to abandon intercalation, de Blois concludes that the intercalation had already been abandoned in 622 (when there was the last average correspondence between the beginning of Rabīʿ I and II and that of Tishri I and II before the first documented use of the hijrī era in AH 22) (de Blois, ‘Calendar at Mecca’, 204–6). For the reasons for assuming a synchrony between the pre-Islamic intercalated Meccan calendar and the Jewish calendar, see ibid., 193–204.

17 The clearest accounts of the establishment of the AH-era are those by al-Ṭabarī, Tārīkh, 1251–5 (trans. Montgomery Watt and M. V. McDonald, VI, 157–61) and al-Bayrūnī, Al-Āṯār, 29–31.

18 Muslim narratives maintain that the Jews of Medina were fasting for Yōm Kippūr (10 Tishri) when Muḥammad entered the city. Based on the already mentioned supposed synchrony between the pre-Islamic intercalated Meccan calendar and the Jewish calendar (see n. 16), de Blois argues that al-Bayrūnī’s tradition, according to which Muḥammad arrived in Medina on 8 Rabīʿ I (20 September 622) is to be preferred.

19 In Islamic chronography, days start at sunset rather than at dawn or midnight. Therefore, the beginning of 1 Muḥarram 1 AH fell in the evening between Julian 15 and 16 July 622 CE. As the majority of 1 Muḥarram would have overlapped with 16 July (over 19 hours against the little less than five hours of overlap for the 15th), the latter is generally preferred as the conventional conversion for the starting day of the Islamic era. On the notions of day and night in Arabic, see Fischer, ‘“Tag und Nacht”’.

20 P.World p. 113 (= P.GrohmannApercu p. 41 = P.GrohmannMuhadara II p. 12 = PERF 558). Though the document in question has been known since the late nineteenth century, its significance as a corroborative to narratives on the institution of the Islamic calendar was first explicitly remarked upon by Jones, ‘Dotting of a Script’.

21 P.RagibAnn22 and P.TillierDebts 1 (= Chrest.Khoury I 48), respectively.

22 For the scope of this article, AH to CE conversions have been calculated using J. Thomann’s (1996) Converter housed by the University of Zurich website: UZH – Asien-Orient-Institut – Islamischer Kalender. On the very frequent one-day difference between dates on Arabic documents and the actual corresponding day, see Grohmann, Arabische Chronologie, 30.

23 Rāġib’s (‘Ère inconnue’) suggestion that a parallel lunisolar Islamic calendar was in use in Sufyanid Egypt has been convincingly refuted in Bruning, ‘Legal Sunna’; Shaddel, ‘“The Year”’; Tillier and Vanthieghem, ‘Recording Debts’. See also infra p. 12

24 204 texts, 41 of which are bilingual (Arabic/Greek) and two trilingual (Arabic/Greek/Coptic). For this estimate, I have counted the number of texts rather than the number of documents. Consequently, the presence of multiple independent text-units on the same papyrus has resulted in an entry each. The number does not include five additional documents that are dated only by reference to the day of the week.

25 The earliest Arabic epigraphic text from Egypt is the tombstone of ʿAbd al-Raḥmān b. Ḥayr al-Ḥajrī, dated Jumādā II, 31 AH (winter 652). See el-Hawary, ‘Most Ancient Islamic Monument’.

26 Wiet, Catalogue, 1 (= Combe, Sauvaget, and Wiet, Répertoire, 163) (Upper Manūf; 726–727).

27 On the origin of early Islamic tombstones from early Islamic Egypt, see Schmidt, ‘Problem of the Origin’.

28 This list is further compounded by a few no longer extant historical inscriptions known through quotations by later authors, such as Combe, Sauvaget, and Wiet, Répertoire, 8 (Fusṭāṭ; 685), which, however, have not been considered in this contribution.

29 The earliest dated (Islamic) Arabic graffiti are the so-called Salma and Zuhayr inscriptions, found near Yanbuʿ and Madāʾin Ṣāliḥ respectively and dated to 23 AH/643–644 and 24 AH/644–645; see Kawatoko, ‘Archaeological Survey’, 51–2; Ghabbān, ‘Inscription of Zuhayr’.

30 The earliest securely dated Arabic public inscription is Grohmann, Arabic Inscriptions, Z 68, commemorating the erection of a dam on an estate of caliph Muʿāwiya near Ṭāʾif (58 AH/677–678). Possibly older is the so-called ‘Jerusalem 32’ inscription, found at the foot of the Ḥaram al-Sharīf and seemingly commemorating the fall of Jerusalem. The inscription, which is badly damaged, appears to be dated to AH 32 (652–653); see Sharon, ‘Witnessed by Three Disciples’, 100–1.

31 AH-dates already featured in the Pahlavi legends of Arab-Sasanian coinage issued by mints under the authority of the prefecture of Basra in the 650s and an increasing number of mints from the 670s. See e.g. Heidemann, ‘Evolving Representation’, 163–4.

32 P.Ness. and P.Hoyland Dhimma from Nessana and P.Mird from Khirbat al-Mird.

33 P.Khurasan.

34 A graffito left by one al-Rayyān b. ʿAbd Allāh in Ḥumā al-Numūr (near Ṭāʾif), for instance, is dated to ‘the year the Masjid al-Ḥarām was built (i.e.) in the year 78’ (= 697–698 CE). On the pre-Islamic Hijāzī custom of labelling years after salient events, see, for example, Grohmann, Arabische Chronologie, 4–5.

35 On the situation of social trilinguism in early Islamic Egypt see the overview in Richter, ‘Language Choice’.

36 The same phenomena can be seen in process in the Middle Persian and Sogdian correspondence of Arab officials from Iran and Transoxiana. I have argued elsewhere that adaptations of Arabic epistolary formulae into other languages were characterized by a more inclusive religious outlook than in their Arabic archetypes and showed a systematic tendency towards a softening of the explicitly Islamic character in favour of a more pan-Abrahamic outlook. See Garosi, ‘Cross-Cultural Parameters’; idem, New Empire, 217–60.

37 Worp, ‘Hegira Years’. Updates to Worp’s list can be found in Bagnall and Worp, Chronological Systems, 300 n. 1.

38 Arabic and Greek versions, for instance, are distinguished by various technical features, such as the different thicknesses of reed pens used by the scribes and the use of different calendars; see Sijpesteijn, ‘Arabic Script’, 439, 444–5.

39 When mentioned by name, scribes of the Greek parts of bilingual documents are consistently distinct from those of the Arabic ones. In turn, scribes of the Arabic parts have Arab personal names. See in particular Rāġib, ‘Esclaves publics’, 21–6. NB: the scribal signature read as ‘Basīl’ (< gr. Basileios/copt. Basile) by the editor of P.Qurra 2 should be emended to ‘Muslim’.

40 This difference can possibly be explained by assuming that the original tax assessment was penned in Greek, while at a later point an Arabic version was prepared and the Greek text of the original assessment was copied beneath the Arabic one without changing the original date. See Kaplony, ‘Arabisierung’, 398–9. This difference characterizes most notably tax demand-notes issued by the governor Qurra b. Sharīk to the office of Basilios. A similar difference is not visible in P.World p. 130, a bilingual receipt issued by the administration of Qurra to the paymaster of Ihnās, which (according to the editor) bears a synchronous date to 9 Thot of the 8th indiction in the year 91 (5 September 710) in the Greek part and Dhū al-Qaʿda 91 (31 August–29 September 710) in the Arabic part. After examining an image of the badly damaged Arabic date on a digital image, I strongly suspect that Grohmann’s reading of the Arabic date is largely dependent on the conversion of the better preserved date in the Greek text. The combination 9 Thot of the 8th indiction with the year 91 indicates that the indiction cycle in the administration in Fusṭāṭ in this period began in Pachon rather than in Thot. See Bagnall and Worp, Chronological Systems, 22–35 (with Appendix A). For the start of the indiction in the documents of Qurra b. Sharīk, see Cadell, ‘Nouveaux fragments’, 138–52 (who does not discuss P.World p. 130). The difference between the Arabic and Greek dates of writing is not to be confused with the well-known two-year discrepancy between the overall fiscal year and the overall year of writing of the document. On the latter phenomenon, see Casson, ‘Tax-Collection’.

41 P.Lond. IV 1434 (Aphrodito; 716) ll. 229 and 314 and 1447 (Aphrodito; 685–705) l. 140.

42 Kaplony, ‘Arabisierung’, esp. 398; Sijpesteijn, Shaping a Muslim State, 231; eadem, ‘Arabic Script’, esp. 439,

43 Berkes and Younes, ‘Trilingual Scribe’; Reinfandt, ‘Petosiris the Scribe’, 142–3.

44 SPP VIII 1195 (724), 1184 (728), CPR XXII 8 (729–730), 9 (729), and 7 (751–752).

45 P.Rain.Untterricht 92 (753–754), CPR XXII 22 (771–776), 17v (789–790), 21 (796–797), P.Christ.Musl. 16 = SPP X 172r (758) and 65 (780), and P.Rain.UnterrichtKopt. 119 (779).

46 P.Clackson 45 (753) P.Bal. 287 (725).

47 P.Bal. 130 app. a (724). The date is based on Crum’s reading in P.Lond.Copt. I 440. Kahle’s edition in P.Bal. omits the number of the day. I thank Renate Dekker for the reference.

48 The only exception to this rule is P.World p. 130 = SB XX 1444, a bilingual tax receipt in which the Greek text is dated according to the era of 622 in addition to the indiction year.

49 P.VanthieghemCompanion (Ṭuṭūn; 670–671 or 679–680).

50 DelatterEntagion (Fayyūm; 694), P.KarabacekBemerkungenMerx (= P.MerxDocuments p. 55) (Fayyūm; 697–698), P.DiemFrueheUrkunden 1 (Fayyūm; 685–705), P.Gascou 27b (Hermopolis; 695), P.DiemFrueheUrkunden 2 and P. Vind.inv. G 43234 (both from Herakleopolis; 685–705) and P.CtYBR inv. 71 (Egypt; 693?).

51 The eukleestatos amira Fl. Siker s. of Abied (< arab. Shukayr b. ʿUbayd) in P.KRU 70 and the eukleestatos Mamet (< arab. Muḥammad) in P.KRU 106. On the reading ‘Siker’ (correcting the earlier ‘Iosēf’ of the editio princeps), see Berkes and Cromwell, ‘Papyrologica 65’. On the Arab officials active in eighth-century Jēme and their jurisdiction, see Cromwell, ‘Western Thebes’.

52 The earliest example is P.KRU 106 (735), a contract of donation in the event of death dated to 6 Payni in the third indiction in the Diocletian year 451, in the year of the Saracens 117 (= 31 May 735). For the reading of the AH-date, see the correction to Crum’s original edition in Till, Die koptischen Rechtsurkunden, 190 (n. 106).

53 Cf. the fact that three out of four of the Coptic texts from Jēme to record an AH-year (P.KRU 70, 99 and 106) are headed by an Arabic prōtokollon.

54 BKU III 364 (Fayyūm?), an acknowledgement of debt by one Pasēth in favour of a cleric dated to the 8th indiction in the year 167 AH (= 784–785 CE). The document opens with a rendition of the Arabic basmala into Coptic, possibly betraying a degree of familiarity of the writer with Arabic scribal conventions – which in turn might explain the use of the AH-era.

55 Wietheger, Das Jeremias-Kloster; for dating clauses including an AH-date, see esp. 198.

56 I thank Renate Dekker for calling my attention to this. She also came up with the ingenious suggestion that the use of AH-years in the tombstones from Saqqāra might be due to the will to show allegiance to the Muslim establishment in the aftermath of the increasingly frequent Coptic agrarian revolts in the eighth century. On the monastery of Apa Jeremias in Saqqāra, see the recent overview in Innemée, ‘Monastery of Apa Jeremiah’.

57 P.RagibSauf-conduits 1 and 5–8 and P.PiletteSauf-Conduit (dated between ca. 717–719 and 751).

58 See Berkes, ‘Latest Identified Greek Documentary Text’. I thank Lajos Berkes for pointing out the ecclesiastical background of the document to me.

59 Sijpesteijn, Shaping a Muslim State, 91–111; Papaconstantinou, ‘Rhetoric of Power’. Cf. also Reinfandt, ‘Petosiris the Scribe’.

60 Sijpesteijn, ‘Landholding Patterns’; Legendre, ‘Landowners’.

61 CPA papyri were not included in the original publication of the Nessana papyri, which only mentioned a (misidentified?) Syriac document (Kraemer, Nessana III, 9). The CPA fragments from Nessana are now edited in Hoyland, ‘Christian Palestinian Aramaic Papyri’.

62 P.Ness. 60–63 and 65–67.

63 P.Ness. 64 (Nessana; 674). As this is the only item of the Nessana dossier addressed to the inhabitants of Sykomazon and as there is no apparent reason as to why it would have been sent to Nessana instead, the document must have been delivered to Nessana by mistake.

64 The official in charge of redistributing tax quotas to individual taxpayers was possibly the dioikētēs Georgios who is mentioned in several Nessana documents. See Sijpesteijn, ‘Multilingual Archives’, 110–11.

65 Di Segni, ‘Hammat Gader’, 237–40 (n. 54).

66 Di Segni and Tepper, ‘Greek Inscription’.

67 Brock, ‘Use of Hijra Dating’.

68 Mouterde, ‘Inscriptions en syriaque’, 5, 10, 20, 21, 28. For the labels of the AH-era in Syriac graffiti, see infra p. 13.

69 The author of graffito n. 10 specifies that the workers came from Jazīrat Qurdū (today’s Jazīrat Ibn ʿUmar); another Greek graffito from the same site (n. 29) was written by one Georgios aktionarios from Edessa and workmen and superintendents mentioned in n. 16 come from several towns in the region of Mosul. On the various Christian workforces employed in the construction of ʿAnjar (including the Egyptian worker(s) referred to in P.Lond. IV 1434, l. 26), see Chehab, ‘Identification of ʿAnjar’, 45–6.

70 I owe this suggestion to Sofie Remijsen.

71 By comparison, the earliest Coptic manuscript showing an AH-date is Morgan M588 (see van Lantschoot, Recueil des colophons, III), dated 12 Phamenoth in the 5th indiction/ Diocletian year 558/227 AH (= 8 March 842).

72 The papyrus’s date can be narrowed down to 30 December 787(= 15 Rajab)–4 January 788 (8 Tybi) by looking for periods of general synchrony between 15–19 Rajab, 4–8 Tybi and the weekdays Monday–Friday. There is, however, a one-day difference between the Coptic and Arabic days of the month indicated in the text: in 787, 4 Tybi fell on Monday 31 December (and not, as Morelli indicates, on Sunday 30 December), which, however, would correspond to 16 Rajab and not to 15 Rajab, as stated in the document. This one-day error is possibly due to the difficulty of the writer in harmonizing the different beginnings of the day in the Arabic (where the day begins at sunset) and Egyptian (where the day begins at first light) systems. I thank Sofie Remijsen for calling my attention to this document.

73 On the symbolic connotations of calendars in general, see Zerubavel, Hidden Rhythms, 70–100. For examples from the Ancient World, see in particular Stern, Calendars in Antiquity, with further references. For some considerations on the ideological meaning attached to chronological systems in the early Islamic Egypt specifically, see Mikhail, From Byzantine to Islamic Egypt, 127–35.

74 Rāġib, ‘Ère inconnue’, 187–8.

75 Tillier and Vanthieghem, ‘Recording Debts’, 149 n. 2.

76 Ibid., 157–60.

77 For group identifiers used within the rising Arab-Muslim community, see Donner, ‘From Believers to Muslims’; Lindstedt, ‘Muhājirūn’.

78 Tillier and Vanthieghem, ‘Recording Debts’, 154–60. Cf. however, the recent re-edition of P.Ness. 56 in Hoyland, ‘P. Nessana 56’, 146, where Hoyland suggests that sana should be understood as simply a tautological expression rather than an abbreviation. The main argument for the ‘abbreviation’ interpretation is that both the extended sana so-and-so sanat qaḍāʾ al-muʾminīn formula and its supposed abbreviated form ‘sana X sana’ appear in the recording and acknowledgement of debts. Some corroboration to the ‘tautological’ interpretation is offered by Wiet, Catalogue général 1 (= Combe, Sauvaget, and Wiet, Répertoire, 163) (Upper Manūf; 726–727), an incomplete official inscription concerning a land survey relating to the year 108 AH (fī sanat thamān [wa-miʾat] sana …) in which the expression ‘sana X sana’ appears outside the context of debts.

79 It is worth noting that the only unambiguous evidence that the chronological system labelled as sanat qaḍāʾ al-muʾminīn is in fact identical to the AH-system is P.Ness. 56, a bilingual receipt from Nessana in which, however, the expression would only appear in the abbreviated form sanat (qaḍāʾ al-muʾminīn?). On P.Ness. 56, see the previous footnote.

80 E.g. Donner, ‘Qur’anization’, 86–7. Connected with this issue is also the fact that the term hijra only assumed the meaning of emigration in the Abbasid period. See Crone, ‘First Century Concept’.

81 Rāġib, ‘Ère inconnue’, 193–4.

82 Bruning, ‘Legal Sunna’, 368–9; Shaddel, ‘“The Year”’, 307–8; Tillier and Vanthieghem, ‘Recording Debts’, 154.

83 Tillier and Vanthieghem, ‘Recording Debts’, 164–8. For this interpretation to be true, a review of the traditional chronology of the Treaty would be required.

84 Bruning, ‘Legal Sunna’, with corrections in Tillier and Vanthieghem, ‘Recording Debts’, 154–7.

85 A detailed overview can be found in Rāġib, ‘Ère inconnue’, 187–92. For an earlier overview, see Grohmann, Arabische Chronologie, 12–36.

86 On the term muhājirūn, see Lindstedt, ‘Muhājirūn’. Cf. also Crone, ‘First Century Concept’. While the Syriac mhaggrāyē is likely a loanword from Arabic, the term was handled by Syriac writers as being synonymous with the children of Hagar. See Brock, ‘Syriac Views’, 15.

87 Brock, ‘Use of Hijra Dating’, 283–90.

88 Mouterde, ‘Inscriptions en syriaque’, nn. 5, 10, 20, 21, 28.

89 See supra, p. 2.

90 For the semantic spectrum of the word Arab and its derivatives in pre-Islamic Egypt, see de Jong, ‘Arabia’; on the rise of ‘Arabs’ as an ethnonym more generally, see Webb, Imagining the Arabs; Webb, however, does not discuss papyrological evidence. For Saracen, see in particular MacDonald, Literacy and Identity, VIII, 1–26, with further references. For the possible use of the Greek Araps/Araboi as a surrogate for Muslims, see P.Lond. IV 1375 (Aphrodito; 710), a letter requesting funds to cover for the purchase of ‘articles for the maintenance of us and the officials who are with us, both Arabs (Araboi) and Christians (Christianoi)’ (trans. Bell; emphasis and texts between parentheses are mine). The argument could also be made that, in this case, the religiously coloured word ‘Christians’ is being used as an umbrella term for non-Arabs rather than the non-religiously coloured ‘Arabs’ being used as a stand-in for Muslims.

91 Mouterde, ‘Inscriptions en syriaque’, 79 n. 5.

92 Brock, ‘Syriac Views’, 15. See, in particular, the colophons of BM add. 14,666 fol. 56 (682), in which the mhaggrāyē are styled ‘the children of Ishmael, son of Hagar, son of Abraham’, and of Pierpont Morgan 236, in which the conquerors are styled ‘the children of Hagar’.

93 Cf. Shaddel, ‘“The Year”’, 303–4, who points out that seventh- and eighth-century non-Muslim authors treated the Islamic era essentially as a regnal era. A similar understanding also shines through a passage in al-Bayrūnī, Al-Āṯār, who motivates the choice of the hijra as the beginning of the Islamic era stating that ‘after it (i.e. the hijra) the rule of Islam was consolidated and polytheism died out … and there afterwards followed the (Islamic) conquests’, so that the hijra ‘became for the Prophet what the ascension to the throne and the take-over of absolute power is to kings’.

94 Cf. the Coptic inscription mentioned supra p. 2 n. 4.

95 Rāġib, ‘Ère inconnue’, 189, Mikhail, From Byzantine to Islamic Egypt, 134.

96 To the best of my knowledge, the earliest dated Coptic text outside the Thebaid to feature the locution etos Sarakēnōn is a tombstone from the Jeremias monastery in Saqqara dated 818 CE (Wietheger, Jeremias-Kloster, n. 194) in which the ‘year of the Saracens’ date precedes the date according to the Diocletian era. This contrasts with the three other (earlier) tombstones from the same site (ibid. nn. 85, 86, and 189) featuring a hijra date, in which the AH-year is not further qualified and follows after the Diocletian era date. Cf. the appendices to this article.

97 Ochała, ‘Era of the Saracens’.

98 SPP VIII 1195 (724), 1184 (728), CPR XXII 8 (729–730), 9 (729), and 7 (751–752).

99 P.Rain.Untterricht 92 (753–754), CPR XXII 22 (771–776), 17v (789–790), 21 (796–797), P.Christ.Musl. 16 = SPP X 172r (758) and 65 (780), and P.Rain.UnterrichtKopt. 119 (779).

100 P.Clackson 45 (753) P.Bal. 287 (725?).

101 P.Bal. 130 app. (724)

102 For the use of ‘Arab’ and Arab cognates in Greek documentary texts from Egypt, see de Jong, ‘Arabia, Arabs, and “Arabic”’. For Arab related terms, in particular, see Webb, Imagining the Arabs in tandem with al-Jallad, ‘ʾAʿrāb’.

103 Morelli, Documenti greci, 53–4. Among seventh-century materials, the invocation en onomati tou theou is also found in documents issued in the name of Arab amīrs (e.g. P.Apoll. 7, 1). To the best of my knowledge, the only document featuring the invocation en onomati tou theou produced in a pagarch’s office is P.Rain.Unterricht 106a, a scribal exercise from the chancery of Yaḥyā b. Hilāl (in office 743–759). I thank Lajos Berkes for the reference.

104 On the rendition of Arabic epistolary formulae into Greek and Coptic, see Luiselli, ‘Greek Letters’; Richter, ‘Coptic Letters’, 763; Garosi, ‘Cross-Cultural Parameters’, 75–9; idem, New Empire, 219–25; Berkes, ‘“Peace be Upon You”’.

105 One other peculiarity of the Tamra inscription is that the date, contrary to all other coeval Greek texts, is written out rather than given in numerals. See di Segni and Tepper, ‘Greek Inscription’, 345.

106 For the definition ‘sub-governor’ in the early Islamic Syrian context, see Gundelfinger and Verkinderen, ‘Governors of al-Shām’, 257.

107 Seven out of eight of the Nessana bilinguals are issued by an official by the name of al-Hārith b. ʿAbd, whom Patricia Crone (Slaves on Horses, 277 n. 235) and Moshe Gil (History of Palestine, 76 n. 1) have identified with a figure known in the literary sources variably as al-Hārith b. ʿAbd/Khālid/ʿAmr/ʿAbdallāh al-Azdī, a military commander from Palestine, who served as interim governor in Baṣra in 665. By analogy with known examples of collective bilingual tax demands from Egypt, which are invariably issued by the office of the provincial governor, Kraemer (Nessana III, 32–3), followed by several others (Crone, Slaves on Horses, 277 n. 235; Stroumsa, People at Nessana, 127; Morelli, ‘Consiglieri e comandanti’, 161), assumed that al-Hārith must analogously have been the governor of jund Filasṭīn. This hypothesis, however, collides with information from literary sources which indicate that Ḥassān b. Mālik b. Baḥdal was governor over both Urdunn and Filasṭīn under Muʿāwiya I and bequeathed Filasṭīn to Rawḥ b. Zinbāʿ under Yazīd I (see Gundelfinger and Verkinderen, ‘Governors of al-Shām’, 263 n. 30, 264 and 297). The issuer of the latest of the Nessana demands, dated to 689, Ḥassān b. Mālik has been traditionally identified with the already mentioned homonymous governor Ḥassān b. Mālik b. Baḥdal. While the identification in itself is quite plausible, the document (dated 70 AH) postdates the end of Ibn Baḥdal’s term in office in 64 AH. The jund Filasṭīn appears to have been governed by Abān b. Marwān during ʿAbd al-Malik’s reign (Gundelfinger and Verkinderen, ‘Governors of al-Shām’, 298). It possible that towards the end of his life Ibn Baḥadal had been demoted to a district position (cf. Crone, Slaves on Horses, 93–4). It is also worth mentioning that the reading of the name Ḥassan (gr. asssan) is doubtful and was rejected by Bell, see Kraemer, Nessana III, 197n9. Another element worth considering, is that P.Ness. 67 was written in 70 AH, postdating by one year Ibn Baḥdal’s supposed date of death. Incidentally, I believe that the date of P.Ness. 67 ll. 6–8 should be read as (6) (…) wa-kataba (7) [Khālid?] fī jumādā al-awwal min [san]at (8) sabʿīn instead of Day’s reading (6) (…) wa-zayt (7) fī [shahr rabī]ʿ al-awwal min [san]at (8) sabʿīn which has no parallel in the dating formulae of the Nessana dossier. For the name of the scribe being written directly above the yāʾ of the fī, see P.Ness. 61 and 62.

108 The ʿAbd Allāh b. Hāshim/ʿĀsim mentioned in the Hammat Gader inscription as well as several Arab officials in the Nessana dossier carry the title symboulos most notably used in early Islamic Egypt to designate the provincial governor (see in particular Morelli, ‘Consiglieri e comandanti’). Neither ʿAbd Allāh or the symbouloi mentioned by name Nessana (one Abū Rāshid in P.Ness. 72 and 73 and one Muslim in P.Ness, 58; two further unnamed symboouloi are mentioned in P.Ness. 75 and 158), however, feature in the lists of (sub-)governors of Greater Syria passed down by Muslim chroniclers (see Gundelfinger and Verkinderen, ‘Governors of al-Shām’). Thus, it seems that in southern Palestine the title symboulos was employed to identify a district-level dignity. See di Segni, ‘Hammat Gader’, 239; cf. Isaac Hasson, ‘Remarques’, 100–1. An additional hint that we are not dealing with provincial governors is also the fact that the symbouloi mentioned in the Nessana dossier carry the honorific endoxotatos whereas the Arab governors of Egypt are usually addressed with the higher dignities of paneuphēmos and hyperphyestatos, with endoxotatos being more frequently used for dukes and pagarchs.

109 Meimaris, ‘Arab Era’, 178–80 (nn. I–V) and 184; idem, Chronological Systems, 120–1 with n. 124 and 307 with n. 16; Bingen, ‘Review’, 195.

110 On the interpretation of the kata Arabas formula in this context, see Bingen, ‘Review’, 195.

111 E.g. SEG XLI 1573 (Pella; 521–522), an epitaph of two soldiers named Johannes, who are said to have been of Arabōn ethnos; on the interpretation of the term ethnos in this context as an administrative unit, see Denis Feissel apud Gauthier et al., ‘Bulletin épigraphique’, 590 (n. 632).

112 The fiscal unit of a village (called kōmē prior to the Arab conquest) begins to be defined by the word chōrion in Greek documents from early Islamic Egypt, in keeping with acceptation of the term in Late Antique Syrian evidence (see Gascou, ‘Arabic Taxation’, 672). A similar instance is the spelling gonachion for ‘blanket’ (< Aramaic gwnk), which in fiscal documents from the Arab period supplants the pre-conquest spellings kanaukēs/gaunakēs (see Morelli ‘Gonachia e kaunakai’, 55–81). Other documentary neologisms in the administrative language of early Islamic Egypt, such as the words sakēlla/sakēllarios to indicate the state ‘treasury’/‘treasurer’, are also believed to be imported Syrian terms (see Legendre, ‘Neither Byzantine nor Islamic?’, 13).

113 Garosi, New Empire, 311–14.

114 For general remarks on the role of Aramaic speakers in mediating between Greek and Arabic milieus, see Hoyland ‘Language and Identity’, 194–8.

115 For a survey of the Christian officials in the employment of the Islamic empire, see Cheiko, Vizirs et secrétaires. Cf. also Fattal, Statut légal, 240–63. For the Abbasid period in particular, see Fiey, Chrétiens syriaques; Cabrol, Secrétaires nestoriens. On the Islamic discourse surrounding the employment of non-Muslim secretaries, see Yarbrough, Friends of the Emir.

116 Cheiko, Vizirs et secrétaires, 73–5 (nn. 57–9); on the Manṣūr family, see Griffith, ‘Manṣūr Family’; Anthony, ‘Fixing John Damascene’s Biography’.

117 Also referred to in the sources as Ibn Bitrīq. See Cheiko, Vizirs et secrétaires, 57 (n. 26).

118 Ibid., 152 (n. 226).

119 Ibid., 134 (n. 179).

120 Al-Kindī, Kitāb al-Wulāt, 82. Ḥafṣ b. al-Walīd’s connection with the Palestinian milieu was likely formed during his service as a commander in the Syrian fleet in the late 730s/early 740s (see Ibn Yūnus, Taʾrīkh I, 132).

121 Cheiko, Vizirs et secrétaires, 136 (n. 196).

122 Ibid., 126–7 (nn. 171 and 172); Mikhail, From Byzantine to Islamic Egypt, 43–4; Debié, ‘Christians’, 55–6; see also infra n. 123.

123 P.Lond. IV 1447 l. 139 and 141 mentioning expenses for 20 and 44 notarioi of the endoxotatos chartoularios Athanasios respectively. They clearly form a different corps from both the notaries of the other (Egyptian) endoxotatos chartoularios Isak (ibid. l. 137) and the personal Arab secretaries of the governor (ibid., l. 140). On Isak, see Cheiko, Vizirs et secrétaires, 129 (n. 179). Another Stephanos secretary of Athanasios ‘in charge of requisitioning in Upper Egypt’ mentioned at l. 144 is in all likelihood identical with Athanasios’ homonymous son. See Mikhail, From Byzantine to Islamic Egypt, 43 n. 34 quoting Morimoto’s (Fiscal Administration, 363) translation of the passage who, however, omits the word n(o)t(arios) in the Greek text.

124 Mikhail, From Byzantine to Islamic Egypt, 116.

125 Garosi, New Empire, 187–8 with further references.

126 Schwabe, ‘Khirbat Mafjar’, 21–2 (i–ii) (Khirbat al-Mafjar).

127 P.Rain.Unterricht 106a (Fayyūm; ca. 743–759); cf. supra n. 103.

128 On Greek scribal exercises from early Islamic Egyptian pagarchies, see Bucking, ‘Training of Documentary Scribes’; Berkes and Younes, ‘Trilingual Scribe’. Cf., conversely, P.Christ.Musl. 3 r/v a seventh/eighth-century Greek scribal exercise in which the formula en onomati tou theou is repeated several times over. This latter document mentions the ‘catholicoi of the barns of Babylon’ (= al-Fusṭāṭ) strongly suggesting that the exercise originated in the Fusṭāṭī milieu.

129 Al-Ghul, ‘Preliminary Notes’; al-Jallad, al-Ghul, and Daniel, ‘Arabic Toponyms’; Vierros, ‘Greek of the Petra Papyri’; al-Jallad, ‘Arabic of the Petra Papyri’.

130 Epigraphic testimonies connected to the Ghassānids are collected by Gatier, ‘Les Jafnides’, 193–222.

131 Al-Qāḍī, ‘Non-Muslims’, 88–93.

132 It has been argued that the use of Arabic in the chancery of Christian Arab phylarchs might have been one of the formative influences behind the development of the Arabic script (see e.g. Hoyland, ‘Arab Kings’, 391–2). Umayyad building projects at the site of prior Ghassānid facilities across Greater Syria were also possibly aimed at taking over Ghassānid tribal networks (e.g. Shahīd, ‘Ghassānid and Umayyad Structures’, 299–307; Tohme, ‘Spaces of Convergence’).

133 See, for instance, the case of Ḥassān ibn Nuʿmān al-Ghassānī, who rose to the position of governor of Ifrīqiya. On the history of the Banū Ghassān after the battle of Yarmūk, see Shahīd, ‘Ghassān post Ghassān’, 321–8; cf. also idem, ‘Sigillography’, discussing a seal possibly belonging to the Ghassānid Jabala b. Ayhām († 640–641).

134 On the composition of Umayyad armies, see Kennedy, Armies of the Caliphs, 7–51.

Appendices

Appendix A: Labels of the Arab era in Greek and Coptic sources from early Islamic Egypt (mid-seventh–early ninth century)

Greek

Table

Text	Origin	Level	Type	Label
P.Cair.Arab. 160 (709–710) P.Cair.Arab. 161 (709–710) [P.Cair.Arab. 162] (709–710) [P.Cair.Arab. 163] (709–710) P.Heid.Arab. I 6 = Chr.Wilck. 256 (709–710) P.Heid.Arab. I 5  = SB I 5638 (709–710) P.Heid.Arab. I a = SB I 5644 (709–710) P.Heid.Arab. I b = SB I 5645 (709–710) P.Heid.Arab. I c = SB I 5646 (709–710) P.Heid.Arab. I d = SB I 5647 (709–710) P.Heid.Arab. I e = SB I 5648 (709–710) P.Heid.Arab. I f = SB I 5649 (709–710) P.Heid.Arab. I g = SB I 5650 (709–710) P.Heid.Arab. I h = SB I 5651 (709–710) P.Heid.Arab. I i = SB I 5652 (709–710) P.Heid.Arab. I k = SB I 5653 (709–710) P.Heid.Arab. I l = SB I 5654 (709–710) P.Heid.Arab. I m = SB I 5655 (709–710)	al-Fusṭāṭ	capital	Tax demands (incl. templates)	Etos Kata Arabas
SPP VIII 1195 (724) SPP VIII 1184 (728) CPR XXII 9 (729) CPR XXII 8 (729–730) CPR XXII 7 (751–752)	Herakleopolite	district		Etos X
P.Rain.Unterricht 92 (753–754) P.Christ.Musl. 16 = SPP X 172r (758) P.Rain.UnterrichtKopt. 119 (779) P.Christ.Musl. 65 (780) CPR XXII 17v (789–790) CPR XXII 21 (796–797)	Fayyūm
P.Bal. 130 app. a (724)	Thebais
P.Bal. 287 (725)	Lykopolite
P.Lond.IV 1434 (716) P.Lond.IV 1435 (716) P.Lond.IV 1436 (719) P.Lond.IV 1437 (718)	Aphrodito		Registers/accounts
CPR XXII 22 (774–776) SPP X 74 (780)^a	Fayyūm
P.Cair.Masp. ΙΙΙ 67359 (715–716)	Antinoopolis
P.World p. 130 = SB XX 1444 (709–710)	al-Fusṭāṭ	Capital	Tax receipts
P.Prag. I 27 (741)	Fayyūm?	District

^a Based on a new reading by Lajos Berkes. I thank him for bringing the document to my attention and for sending me a draft of his forthcoming revised edition.

Coptic

Table

Texts	Origin	Level	Type	Label
P.KRU 106 (735) P.KRU 70 (750)	Thebais	Local	Testaments	Etos Sarakēnōn
SB I 5606 = P.KRU 99 (780) P.KRU 91 (781)*			Donations
SB Kopt. I 783 (799)^a Wietheger 194 (818)	Lykopolite		Funerary stelae
Wietheger 189 = SB Kopt. II 1143 (781?)^b Wietheger 85 = SB Kopt. IV 1995 (787) Wietheger 86 (806)	Memphites			Etos X
BKU III 364 (ca. 784)^c	Fayyūm?		Acknowledgement of debt
P.Clackson 45 (753)	Hermopolite	District	Tax demand

^a The Diocletian era date given in the inscription, 9 Mesore 515 ( =  2 August 799), does not synchronize with the AH-date, 189 ( =  804–805).

^b The Diocletian era date given in the inscription, 1 Mesore 499 ( =  25 July 783 CE), does not synchronize with the inscription’s other chronological indications: 4th indiction in the year 164 (of the Arab era) corresponding instead to the 6th indiction/166 AH. The 4th indiction ran either from 29 August 780 to 28 August 781, if counted from Thoth 1, or from early May 780 to 25 May 781 if, as more often the case in the Theban region, counted from Pachon. The Arab year 164 ran from 6 September 780 to 25 August 781. Assuming the indication of the day to be exact and the year 499 AM to be a scribal error, 1 Mesore 497 ( =  25 July 781 CE) is the only possible date that would fit both indiction and Arab-era dates. On the beginning of the indiction in the Thebais, see Bagnall and Worp, Chronological Systems, 30.

^c The document is dated to the harvest of the 8th indiction which ran from 29 August 784 to 28 August 785 as well as to the year 167 AH which ran from 5 August 783 to 23 July 784. The editor of the text posited that the hijra date might refer to the actual writing of the document while the indiction date might have referred to the foreseeable date of the repayment of the debt. An alternative, simpler explanation is that we are dealing with an indiction starting in Pachon (early May 784), leaving a window of about three–four months in which the 8th indiction and the year 167 overlapped.

Appendix B: Qualifying labels of the Arab era