ClaytBag ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 2015 W. Graham Claytor and Roger S. Bagnall The Beginnings of the Roman Provincial Census: A New Declaration from 3 BCE W. Graham Claytor and Roger S. Bagnall HE PROVINCIAL CENSUS was one of the most durable and pervasive institutions of the Roman Empire.1 Al- though organized at the provincial level and marked by local variation,2 the institution was an emblem of imperial rule. Luke’s famous narrative of the nativity census, while prob- lematic in detail, is important evidence for the provincial im- 1 See generally L. Neesen, Untersuchungen zu den direkten Staatsabgaben der römischen Kaiserzeit (Bonn 1980), especially 33–44; P. Brunt’s review-essay on this work, “Revenues of Rome,” JRS 71 (1981) 161–172 (repr. Roman Im- perial Themes [Oxford 1990] 324–346); C. Nicolet, Space, Geography, and Politics in the Early Roman Empire (Ann Arbor 1991) 133–139; and E. Lo Cascio, “Census provinciale, imposizione fiscale e amministrazioni cittadine nel Prin- cipato,” in W. Eck (ed.), Lokale Autonomie und römische Ordnungsmacht in den kaiserzeitlichen Provinzen (Munich 1999) 197–212. For the census in Roman Egypt see R. S. Bagnall and B. W. Frier, The Demography of Roman Egypt 2 (Cambridge 2006), and A. Jördens, Statthalterliche Verwaltung in der römischen Kaiserzeit: Studien zum praefectus Aegypti (Stuttgart 2009) 62–94. The starting point for study of the Ptolemaic census is W. Clarysse and D. J. Thompson, Counting the People in Hellenistic Egypt I–II (Cambridge 2006). 2 Some provincial censuses focused on registering inhabitants, as evi- denced by the κατ’ οἰκίαν ἀπογραφαί from Egypt, and were thus geared toward the collection of the poll tax (though not exclusively, for women and children were included in the declarations even though not liable to the tax); others followed Ulpian’s forma censualis by including a declaration of prop- erty, thereby facilitating the assessment and collection of property taxes, as in the Arabian census of 127, discussed below. See H. M. Cotton, “The Roman Census in the Papyri from the Judaean Desert and the Egyptian κατ’ οἰκίαν ἀπογραφή,” in L. H. Schiffman (ed.), Semitic Papyrology in Context: A Climate of Creativity (Leiden/Boston 2003) 105–122. T 638 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 pression of the census as universal and stemming from the direct command of the emperor.3 The census reinforced im- perial ideals, strengthening the notion that the emperor could “see everything and hear everything,” even when ruling from the Palace in Rome.4 It also of course aided imperial interests, such as the collection of revenue and the maintenance of social hierarchy. For most provincials, on the other hand, the census and the closely-related poll tax were simply facts of life and burdens from which there was little chance of escape; for some, the imposition of a poll tax and regular censuses could have be- come “a potent symbol of subjection to Roman rule.”5 In short, the census was a common feature of the imperial experience and a key component of Rome’s control over provincial so- ciety. When we move from general considerations to a more de- tailed examination of the mechanics of this system, we find a dearth of evidence. At the very least, it is clear from literary and documentary sources that Augustus played a central role in the extension of the census throughout the empire, begin- ning with his personal reorganization of Gaul (and possibly 3 Lk 2:1: ἐγένετο δὲ ἐν ταῖς ἡµέραις ἐκείναις ἐξῆλθεν δόγµα παρὰ Καί- σαρος Αὐγούστου ἀπογράφεσθαι πᾶσαν τὴν οἰκουµένην, “in those days a decree went out from Emperor Augustus that all the world should be registered” (transl. The New Oxford Annotated Bible 4). The standard translation of ἀπογράφεσθαι as a passive should be reconsidered in light of the middle ἀπογράψασθαι in Lk 2:5 and the consistent use of the middle voice in the papyri. Luke’s impression was shared by later writers (some influenced by Luke): see Nicolet, Space, Geography, and Politics 137. See further B. Palme, “Die ägyptischen κατ᾿ οἰκίαν ἀπογραφαί und Lk 2, 1–5,” Protokolle zur Bibel 2 (1993) 1–24, and “Neues zum ägyptischen Provinzialzensus: ein Nach- trag,” Protokolle zur Bibel 3 (1994) 1–7, as well as Jördens, Statthalterliche Ver- waltung 62–64. 4 Plin. Pan. 80.3. Aristides (Or. 26.33) famously stresses that the emperor could easily govern through letters without leaving Rome. 5 D. W. Rathbone, “Egypt, Augustus and Roman Taxation,” CCG 4 (1993) 81–112, at 86. W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 639 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 Spain) in 27 BCE.6 Actual responsibility for conducting the census generally lay with the emperor’s representatives, the provincial governors, who in turn relied on subordinates and local organs of government to see the process through.7 For example, during the well-known census of Syria and recently- annexed Judaea ca. 6 CE, the assessment of the polis of Apamea was entrusted to a military prefect “on the orders of [P. Sulpicius] Quirinus,” the governor to whom Luke refers at the beginning of his work.8 In Egypt, by comparison, the phrase κατὰ τὰ κελευσθέντα ὑπὸ ΝΝ τοῦ κρατίστου ἡγεµόνος (vel sim.) in census declarations submitted to local officials becomes common from 89 on, although the earlier censuses were probably likewise set in motion by prefectural edict.9 As usual, the papyrological evidence from Egypt provides the most-detailed picture of the workings of this system. Every fourteen years the inhabitants of Egypt were required to submit a declaration to local authorities containing the names, ages, and other identifying information of all co-residents. Many such declarations survive on papyrus, often as lone documents deprived of their original context, but occasionally still glued together as part of administrative rolls or traceable to an ar- chive of family papers. There are now secure witnesses for every census between 33/4 and 257/8 CE, and the extension of this cycle back to 19/20 remains a distinct possibility;10 but 6 Cass. Dio 53.22.5. For a collection of literary, legal, and epigraphic evi- dence on the provincial census see Brunt, JRS 71 (1981) 171–172 (= 345– 346). 7 See Brunt, JRS 71 (1981) 165–167 (= 333–335). 8 CIL III 6687 [ILS 2683]: iussu Quirini censum egi Apamenae civitatis millium homin(um) civium CXVII. 9 Bagnall and Frier, Demography 11. 10 The question hinges on the declaration P.Oxy. II 254, whose date is not preserved, although the addressees are known from P.Oxy. II 252, dated to 19/20. See R. S. Bagnall, “The Beginnings of the Roman Census in Egypt,” GRBS 32 (1991) 255–265, at 259–260, and Bagnall and Frier, Demography 3. 640 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 the relative paucity of early-Roman documentation from Egypt has meant that even in this province the introduction and early development of the institution remain poorly understood, leav- ing the door open to speculation. The declaration published below helps fill this gap. It is the earliest declaration so far identified and comes from an archive that has long been central to discussions of the early-Roman census in Egypt. In 1928 Calderini published nine papyri recently purchased for Milan’s Università Cattolica del Sacro Cuore, which formed the core of what is now known as the Harthotes ar- chive, named after its main figure, a public farmer and minor priest from the village of Theadelphia.11 The text that garnered the most attention in this group was the top half of a census declaration submitted by Harthotes (P.Mil. I 3). Although its precise date was missing, scholars recognized that the oath sworn by Καῖσαρ limited it to the early decades of Roman rule and attempted to fit the papyrus into various schemes for Rome’s introduction of the census into the new province of Egypt.12 Bagnall’s discovery of the document’s lower half in the Columbia collection in 1991 provided both a precise date for the papyrus (12 CE) and a better basis for explaining the scant earlier evidence for the Egyptian census. This evidence hinges on “indications of government attempts to establish population lists around 4/5 and 11/2,” which “acquir[ed] a certain benchmark status” in later declarations.13 Nelson suggested that the first general status examination (ἐπίκρισις) occurred in Augustus’ 34th year (4/5 CE),14 while Hombert and Préaux drew attention to an ἐπίκρισις attested 11 P.Mil. I 4–12, republished by S. Daris in 1967 (P.Mil. I2). For the ar- chive see www.trismegistos.org/archive/99. 12 See Bagnall, GRBS 32 (1991) 255–260, for a summary of earlier pro- posals. 13 Bagnall, GRBS 32 (1991) 260. 14 C. A. Nelson, Status Declarations in Roman Egypt (Am.Stud.Pap. 19 [Am- sterdam 1979]) 23. W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 641 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 for year 41 (11/2 CE) in P.Oxy. II 288.35 and proposed that this procedure in fact referred to a census, an interpretation later supported by Braunert.15 Drawing on this observation and on age indications in the Harthotes archive, Montevecchi sug- gested that Harthotes’ then-undated declaration was prompted by a census announced in 11/2 CE.16 Bagnall’s joining of the Milan and Columbia fragments of what is now SB XX 14440 confirmed that a census was indeed conducted in Augustus’ 41st year, 11/2 CE. An interesting feature of Columbia’s fragment of this papyrus allowed Bagnall to further propose a scheme for earlier cen- suses in Roman Egypt, even though direct evidence in the form of actual declarations was lacking. After the declaration proper, Harthotes and his son Harpatothoes were listed with their ages and identifying marks, which are typical features of the official documentation of the province. Before their ages, however, were two notations, κL and ϛL, which Bagnall interpreted as the respective years since each had first been declared in the cen- sus, counting exclusively. Alternatively, if we count inclusively, the two notations could refer to the timespan since the estab- lishment of registers the year after declarations were due, a two-stage process that appears to be attested for the census years 34 (4/5) and 41 (11/2).17 Under either interpretation, the young Harpatothoes would have been first declared in year 34 (4/5), which aligns with the ἐπίκρισις of that year, while his 55- 15 M. Hombert and C. Préaux, Recherches sur le recensement dans l’Egypte romaine (Pap.Lugd.Bat. 5 [Leiden 1952]) 48–49, pointing to the expression κατ’ οἰκίαν ἐπίκρισις in P.Hamb. I 60.8–9 (Hermopolis, 90 CE) (cf. now SB VIII 9869a.6 [Hermopolite, 160 CE] and SB XIV 12110.6–7 [Antaiopolite, 90–91 CE]); H. Braunert, “Zur Terminologie der Volkszählung im frühen römischen Ägypten,” in Symbolae R. Taubenschlag dedicatae III (Eos 48.3 [1956]) 53–66, at 56. 16 O. Montevecchi, “II censimento romano. Precisazioni,” Aevum 50 (1976) 72–84, at 73–74. 17 See Bagnall, GRBS 32 (1991) 264, and D. W. Rathbone, CCG 4 (1993) 89–90. 642 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 year-old father Harthotes would have first been declared in year 20 (11/0 BCE).18 Such an explanation implies both that a seven-year cycle was in place before the later 14-year cycle and that year 20 (11/0) was the first such general census held, since otherwise Harthotes would have been registered earlier. Harthotes now has more to offer to the story. Claytor’s re- search in the University of Michigan papyrus collection has brought to light a dozen more papyri belonging to the archive, including the document published below.19 Although faint and effaced in many places, the type of document and its date are beyond doubt: it is a census declaration submitted by Har- thotes at the end of Augustus’ 27th year (3 BCE) to five registra- tion officials (λαογράφοι) of his home village Theadelphia. Harthotes is described as a 41-year-old public farmer with a house in the village, where he and two other occupants live. His age is in perfect agreement with the later census declara- tion and he is further described as “registered in the village” (λαογραφούµε̣ν̣ος π̣ερ̣ὶ τὴν κώµην̣, 6–7), referring, we presume, to the population register that was first established in 11/0 BCE or the year following. With its date in year 27, this declaration provides the missing piece of evidence for the early seven-year census cycle pro- posed by Bagnall: the census years 20 (11/0 BCE), 27 (4/3), 34 (4/5 CE), and 41 (11/2) are all now attested either directly through declarations or indirectly through the notations in SB XX 14440.20 Bagnall’s reconstruction was incorporated into his and Frier’s 18 Indirect evidence for an initial census around this time “may come from a petition from four priests against a demand for arrears for the years back to, but no further than, 9/8 B.C.,” which Rathbone sees as “another indication that the collection of laographia was put on a new footing after 10/9 B.C.”: CCG 4 (1993) 90. The text in question is BGU IV 1198. 19 Claytor, N. Litinas, and E. Nabney are preparing a publication of four contracts, which will include an updated list of all known texts. 20 SB XXIV 16011, palaeographically dated to Augustus’ reign, might therefore be attributed to one of these four censuses. W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 643 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 1994 monograph The Demography of Roman Egypt and the existence of the early seven-year cycle has been generally accepted. Yet the idea that the Romans waited some twenty years before taking stock of the population of Egypt has met with skepticism. Bowman and Rathbone reasoned that the col- lection of the poll tax, attested in Egypt already in the 20s BCE,21 must have been preceded by a census, an assumption about Roman imperial rule shared by other historians.22 Mon- son, however, has recently weakened the logical support for this idea through his convincing argument that the Roman poll tax in Egypt was an adaptation of Ptolemaic practice.23 The in- novation was not the tax itself, which was collected both before and after the conquest under the name of σύνταξις (among other terms) and based on local population registers, but rather in the method of collection, namely a switch to silver rather than bronze drachmas and a regularization of collection tend- 21 See Rathbone, CCG 4 (1993) 88. Monson’s re-dating of a group of poll- tax receipts on ostraka from Karanis puts a number of them in the same decade, some as early as Augustus’ first or second year (A. Monson, “Re- ceipts for sitônion, syntaxis, and epistatikon from Karanis: Evidence for Fiscal Reform in Augustan Egypt?” ZPE 191 [2014] 207–230). Otherwise, the earliest references to the poll tax in the Fayyum are the declarations P.Grenf. I 45 and 46 (= W.Chr. 200a and b, 19–18 BCE) and the unpublished poll- tax receipt P.Heid. inv. G 1026 (Theadelphia, 16/5 BCE), being edited by Claytor. 22 A. K. Bowman and D. Rathbone, “Cities and Administration in Roman Egypt,” JRS 82 (1992) 107–127, at 113: “the conduct of a census was normally the prerequisite for raising a poll-tax” (in a further analysis, Rathbone abandoned the idea of an early census in favor of a system of annual declarations: CCG 4 [1993] 90). Cf. D. Kennedy, “Demography, the Population of Syria and the Census of Q. Aemilius Secundus,” Levant 38 (2006) 109–124, at 116–117, “in order to function adequately, the Roman taxation system presupposes a census,” as well as Brunt, JRS 71 (1981) 163 (= 329–330), and L. Capponi, Augustan Egypt: The Creation of a Roman Province (New York 2005) 90. Cf. below with n.29. 23 A. Monson, “Late Ptolemaic Capitation Taxes and the Poll Tax in Roman Egypt,” BASP 51 (2014) 127–160. Cf. already Brunt, JRS 71 (1981) 162 (= 327–328). 644 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 ing towards monthly installments. If the poll tax was simply reformed rather than newly imposed, there would not have been such a pressing need for an immediate province-wide cen- sus. Approaching the problem from a different point of view, Capponi found it “odd” that Augustus would have taken censuses in less-bureaucratized provinces (such as Gaul in 27 BCE), but not in Egypt.24 Analogy is no substitute for evidence, however, particularly during the formative early years of the Principate and in regard to an institution which was never co- ordinated across the empire.25 The further hypothesis “that the seven-year cycle of censuses of the Egyptian population began around 26/5 BC” rests on weak foundations.26 In our opinion, it seems safest to conclude on present evi- dence that the Romans perpetuated Ptolemaic methods of pop- ulation registration, supplemented by ad-hoc arrangements,27 until it was decided to hold a general census through household declarations in 11/0 BCE. Such developments may not have been unique to Egypt. We 24 Capponi, Augustan Egypt 90; cf. L. Capponi, Roman Egypt (London 2011) 19. 25 It is also more felicitous to turn the logic around: pre-existing institu- tions would make it easier to tax and control the population without much innovation, a point which Capponi herself makes in the conclusion to the chapter (Augustan Egypt 95). 26 Capponi, Augustan Egypt 91. SB XXVI 16683 (a receipt issued by ὁ πρὸς τῇ λαογραφίᾳ τοῦ νοµοῦ) is evidence only for the poll tax (as are the other tax documents mentioned) and its date (25 BCE or 19 CE) is uncertain; for the back-to-back declarations from 19 and 18 BCE see the next note. 27 Into this category may fall P.Grenf. I 45 and 46 (= W.Chr. 200a and b, 19 and 18 BCE), which are back-to-back declarations from a public farmer residing in Theadelphia, with no mention of co-residents (there is no other evidence for annual declarations, which in any case seem infeasible on a large scale), and the fragmentary reference to “the others who have been registered by us in the 15th year of Caesar (= 16/5 BCE),” καὶ ἄλλων τῶν ὑφʼ ἡµ ̣ῶν ἐπὶ το̣ῦ̣ ιε (ἔτους) Καίσαρος λελαογραφηµένων (P.Oxy. IV 711). Cf. Rathbone, CCG 4 (1993) 90. W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 645 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 may compare the province of Arabia, conquered and incor- porated into the empire in 106 CE, which has produced the only papyrological evidence for the census outside of Egypt. In the best-preserved document, we learn that a woman named Babatha appeared before local officials in Rabbat, an ad- ministrative center of Arabia, because “an evaluation of the province was being conducted” by the governor L. Aninius Sextius Florentinus in the year 127.28 Cotton has argued that this was the first census conducted in the recently-conquered province, based on (1) the appearance of the appellation “new province of Arabia” in the dating formula, (2) the holdover of a pre-Roman monetary unit (the melaina) and the Nabataean royal tax (stephanikon), and (3) an indication that the census was personally ordered by Hadrian.29 In both cases, then, the Roman administration would have been satisfied with con- solidating their rule and modifying pre-existing structures be- fore arranging for a general census after a couple of decades of governance. Census Declaration from 3 BCE P.Mich. inv. 4406a30 25.5 x 9.5 cm 26 Jan.–24 Feb., 3 BCE, Theadelphia 28 P.Babatha 16.11–13: ἀποτιµήσεως Ἀραβίας ἀ̣γ̣οµένης ὑπὸ Τίτου Ἀνει- νίου Σεξστίου Φλωρεντείνου πρεσβευτοῦ Σεβαστοῦ ἀντιστρατήγου … 29 H. M. Cotton, “ Ἡ νέα ἐπαρχεία ᾽Αραβία: The New Province of Arabia in the Papyri from the Judaean Desert,” ZPE 116 (1997) 204–208, at 206. Turning to a wider range of evidence, she further argues that we must “dis- pel … the notion that a provincial census followed immediately upon the annexation of a territory to the Roman empire.” Cf. Jördens, Statthalterliche Verwaltung 63–64 and 69. 30 According to the Michigan Inventory of Papyri, the papyrus was purchased by Arthur Boak from David Askren in 1925 as part of a group of “miscellaneous fragments” (P.Mich. inv. 4400–4471). Harold Idris Bell described this group as follows: “those left undescribed may very well likely include some which are of value, but the majority are not promising.” Digital images can be found in Michigan APIS: http://quod.lib.umich.edu/ a/apis/x-12560/4406AR.TIF 646 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 The papyrus was folded from top to bottom in antiquity, as evidenced by fold lines and a repeating pattern of loss on the left side. Otherwise, the papyrus is mostly complete, although the ink is badly effaced in many areas. These losses affect spe- cific readings rather than the overall structure of the text. Early declarations from Arsinoite villages to which we may compare the formulae are few; the earliest are Harthotes’ later census declaration, SB XX 14440 (12 CE, Theadelphia), and SB I 5661 (34 CE, Philadelphia). Also early but metropolitan is SB X 10759 (Arsinoe, 34 CE). No other Arsinoite declarations before the census of 47 CE are known. Otherwise, there are census declarations before that year only from Oxyrhynchos, but these do not follow the same formula, and some of them are uncertainly dated. The papyrus provides a new ‘snapshot’ of the family of Har- thotes. Since Claytor intends to provide a fresh overview and analysis of the archive in a future publication, we keep the fol- lowing discussion brief. With the agreement of his two census declarations, we can place Harthotes’ birth precisely in 44/3 BCE.31 The earliest document in the archive is an unpublished service contract from 20/19 BCE,32 in which Harthotes and his mother Esersythis send Harthotes’ younger brother Marsi- souchos off to work in another villager’s home. In this doc- ument and elsewhere in the archive Harthotes’ father Marres is conspicuous by his absence and was likely already deceased. Esersythis’ kyrios is Harthotes and it is fair to imagine her living as a widowed mother with her two sons at the time of this con- tract. In the new census declaration from 3 BCE, Harthotes states that he is a public farmer living in his own house and declares two co-residents besides. Neither of these is his mother Eser- sythis, who would later rejoin his household before the next 31 On the basis of inclusive reckoning: see N. Kruit, “Age Reckoning in Hellenistic Egypt,” in A. M. F. W Verhoogt and S. P. Vleeming (eds.), The Two Faces of Graeco-Roman Egypt (Pap.Lugd.Bat. 30 [Leiden 1998]) 37–58. 32 P.Mich. inv. 4299, in preparation (see n.19). W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 647 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 census in 11/2 CE. Her absence in the present declaration could indicate that she remarried and joined the new hus- band’s household, but remarriage for Egyptian women with adult sons was uncommon,33 so we prefer to imagine that Har- thotes simply acquired his own house, probably near his aging mother, who was around 55 years old in 3 BCE. The occasion for this move may have been his marriage to Taanchorimphis, who is perhaps the second household member listed in the declaration (see on lines 7–8). By the next census in 12 CE, Harthotes had become a priest of Tutu and was living within the precinct of a local temple.34 In the meantime, he had welcomed back his mother Esersythis, now about 70 years old, and was raising Harpatothoes, his nine-year-old son by Taanchorimphis. Since Taanchorimphis herself is not listed in the declaration, she had either died or at least was divorced from Harthotes. In her place, the aged Eser- sythis was probably looking after the boy, if indeed she were healthy enough to do so. Harthotes also had an elder daughter named Taphaunes (born ca. 1 CE: P.Mil. I2 7) who might have also been expected to help with Harpatothoes’ upbringing, but Harthotes had sent her away to the neighboring village Phila- gris to work at an imperial oil mill, which explains why she was not declared in his household.35 33 Bagnall and Frier, Demography 123–127. Cf. S. Huebner, The Family in Roman Egypt (Cambridge 2013) 97–106, who points to the possibility that the remarriage of younger widows may be underrepresented by the sources. 34 SB XX 14440.2–6: Ἁ̣ρ̣θ̣ώτου τοῦ Μαρήους δηµ[ό]σις γεωργὸς καὶ εἱερεὺς Τοθοήους θεοῦ̣. ἔχω ἐν Θεαδελφήᾳ οἰκίαν ἐντὸς περιβόλου ε[ἱ]εροῦ, ἐν ἧ (etc.). 35 This information comes from the service contract P.Mich. inv. 931 + P.Col. X 249, in preparation (cf. n.19). The agreement states that Taphau- nes is to work for two-and-a-half years starting in Pachon of Augustus’ 39th year, that is, roughly from May of 10 CE through November of 12 CE. The census declaration SB XX 14440 dates to January of 12 CE. It was previ- ously thought that Taphaunes’ absence was due to her already being mar- ried: Bagnall, GRBS 32 (1991) 257. 648 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 P.Mich. inv. 4406a. Image published courtesy of the University of Michigan Papyrology Collection W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 649 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 P.Mich. inv. 4406a, detail Ἀκο̣υ̣σιλ̣ά̣ω̣ι ̣ κ[αὶ ̣ ̣] ̣[ ̣] ι̣ ̣σίωι κ̣[α]ὶ ̣ Τιµοκ̣ρ̣[ά]τ̣η̣ι ̣ καὶ Δ∆ιονυσίωι κα̣ὶ ̣ ̣ ̣ ρ̣̣ ̣ ν̣̣ω̣ι ̣ λαογράφη̣ς̣ Θ̣ε̣α̣δ̣ε̣λ̣φ̣[εία]ς̣ παρὰ Ἁ̣ρθώ̣τ̣ο̣υ̣ τ̣ο̣ῦ̣ Μαρρείους τ̣ῶ̣ν̣ ἀ̣π̣ὸ̣ τ̣ῆ̣ς̣ κώµης δη̣µ ̣ο̣σ̣ί ̣ω̣ν γεωργῶν̣ ὡς̣ (ἐτῶν) µα̣. 5 ὑπάρχι µοι ἐ̣ν̣ τ̣ῆ̣ι ̣ κώµηι ̣ ο̣ἰ ̣κία ἐ̣ν ᾗ αὐτ̣ὸ̣ς̣ ἐγὼ κατ̣[α]γ̣ί ̣ν̣[ο]µ ̣α̣ι ̣ λαογραφούµε̣ν̣ος π̣ερὶ̣ τὴν κώµην̣ ̣ ̣ ̣[ ̣ ̣ ̣] ̣ ̣ ̣ ̣[ ̣ ̣ ̣ ̣ ̣] ̣ ̣[ ̣ ̣ ̣ ̣ ̣] ̣ς̣ Π̣ε̣τ̣ε̣- σούχο̣υ̣ ὡ̣ς̣ (ἐ̣τ̣ῶ̣ν̣) λγ κ̣αὶ ἡ̣ τ̣ υ̣̣τ̣ ̣ ̣ ̣ ̣ ̣ ̣ ̣ ̣ Τα ̣ ̣ ̣ ̣ ̣ ̣[ ̣ ̣ ̣ ̣ ̣] ̣ ̣ ̣ ̣ ̣ 10 εὐ̣τ̣ύχει ̣ κατακεχώ̣ρ[ι]σ̣τ̣[α]ι ̣ (ἔτους) κζ Καίσαρος Μ̣ε̣χε̣ὶ ̣ρ ̣ 2 l. λαογράφοις 5 l. ὑπάρχει To Akousilaos and –isios and Timokrates and Dionysios and NN, laographoi of Theadelphia, from Harthotes son of Marres one of the public farmers of the village, about 41 years old. I own a house in the village, in which I myself live, registered for the poll tax in the village, and … son/daughter of Petesouchos, about 33 years old, and NN … Farewell. Registered in year 27 of Caesar, Mecheir . 650 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 1 [Ἀφ]ρ̣[ο]δ̣ι̣σίωι? The right corner of delta is compatible with the visible traces. The earlier trace is compatible with rho but very scant. 1–2 This is the earliest secure reference to λαογράφοι, who thus seem to have been in place from the beginning of the Roman census in Egypt.36 Their next secure appearance is in census declarations of 33/4: SB I 5661 (Philadelphia, 34 CE), also addressed to five λαογράφοι (and the κωµο- γραµµατεύς), and SB X 10759 (Arsinoe, 33/4). We have not found a name to match the traces of the fifth official in line 2. 2 λαογράφη̣ς̣ Θ̣ε̣α̣δ̣ε̣λ̣φ̣[εία]ς̣. It is difficult to read the ex- pected λαογράφοις, since an eta seems to follow the phi.37 The rest of the line becomes increasingly unclear,38 but the mention of Harthotes’ home village Theadelphia is at least guaranteed by the following references to “the village.” 3–4 τ̣ῶ̣ν̣ ἀ̣π̣ὸ̣ τ̣ῆ̣ς̣ | κώµης δη̣µ ̣ο̣σ̣ί̣ω̣ν γεωργῶν̣ ὡς̣ (ἐτῶν) µα̣. That Harthotes was a public farmer was known from sev- eral other texts of the archive, such as P.Oslo II 32 (1 CE). 36 For an overview of this office see L. Capponi, “P.Oxy. IV 786: Con- clusion of a Census-Return,” ZPE 140 (2002) 177–180, at 179 (on line 18). For other possible early references see M. Resel, “Zahlungsanweisungen für Weizen aus dem Tempel des Soknobkonneus in Bakchias – Das Archiv des Sitologen Akousilaos,” JJurP (forthcoming). We thank the author for providing us a draft in advance. 37 We prefer this reading, but cannot absolutely rule out the possibility that these three strokes represent (1) a narrow omicron (cf. that of λαογρ-, just before); (2) a superfluous ligature (cf. the writing of ὑπάρχι in 5); (3) a short iota. Interchange of η and οι is called “occasional” by F. T. Gignac, A Grammar of the Greek Papyri of the Roman and Byzantine Periods I (Milan 1976) 265; most instances are later, in fact, but he cites one example in the same direction as here and closely contemporary with our papyrus, from 1 CE, in the letter P.Oslo II 47.18, giving µη for µοι. 38 We have considered the suggestion that kappa follows eta, with per- haps a raised omicron indicating an abbreviation. A form such as ⟨λε⟩λαο- γραφηκό(σι), however, requires both the omission of a syllable and the reading of an otherwise unattested active use of the verb λαογραφέω. We have also tried to find κώµης in this space, to no avail. W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 651 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 It is noteworthy that he is not identified as a priest here, in contrast to the later declaration (SB XX 14440, 12 CE) and the petitions SB XX 14098 and 14099 (ca. 1–10 CE). He thus seems to have joined the priesthood later in life and perhaps only for a relatively short time. As for the final element of Harthotes’ description, his age, we note the perfect agreement between the two census declara- tions, while rounding is evident in other documents from the archive (e.g., P.Oslo II 32). 6 κατ̣[α]γ̣ί̣ν̣[ο]µ ̣α̣ι̣. We find καταγίνοµαι (in the participial form καταγινόµενον, because the female declarant is list- ing her son before herself) also in SB I 5661. One may well wonder whether the restored [ἀπογραφό]|µεθα in SB X 10759.5–6 should not be replaced by [καταγινό]|µεθα, which is the same length. 6–7 περί frequently complements λαογραφούµενος to indicate the taxpayer’s legal domicile, but not as a rule.39 7–8 Only exiguous traces remain after κώµην,̣ with space for about 20–25 letters before the letters at the end of line 7. Given the size of the lacuna, we anticipate both the rela- tionship and the name of the second person listed, which would make Π̣ε̣τ̣ε̣|σούχο̣υ̣ a patronymic (the case ending cannot be determined from the faint traces of ink alone). A Petesouchos is so far not known in the family of Harthotes, so this patronymic may belong to Harthotes’ wife Ta- anchorimphis.40 A possible reconstruction of these lines, therefore, runs as follows: καὶ ἡ γυνή µου Ταανχόριµφις Πετε|σούχου ὡς (ἐτῶν) λγ.41 We have also considered 39 Pace R. Duttenhöfer, “Drei Todesanzeigen,” ZPE 79 (1989) 228, note to line 1. Besides ten cases of λαογραφούµενος + περί, DDBDP searches return twelve cases of ἐπί and even one of ἀπό (C.Pap.Gr. II.1 37.5, Arsinoe, 141 CE), which is likely due to confusion with the ἀπὸ τῆς κώµης formula. 40 We owe this observation to Paul Heilporn. Taanchorimphis is attested only in the next declaration, SB XX 14440.10–11, as the mother of Har- thotes’ son Harpatothoes. 41 An eight-year age difference between husband and wife would not be 652 THE BEGINNINGS OF THE PROVINCIAL CENSUS ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 (given the uncertainty of the ending) the possibility that this entry concerns Harthotes’ brother Marsisouchos, for whom the given age would be suitable, but the traces at the end of line 7 are not compatible with Μαρσι-. This hypothesis would therefore require the supposition of an error in recording his name. 8–9 κ̣αὶ ἡ̣ τ̣ ̣υ̣τ̣ ̣ ̣ ̣ ̣ ̣ ̣ ̣ ̣ | Τα ̣ ̣ ̣ ̣ ̣ ̣[ ̣ ̣ ̣ ̣ ̣] ̣ ̣ ̣ ̣ ̣ End of line 8 Line 9 The article in line 8 and the Τα at the beginning of 9 indicate that a female member of the household is being declared here, but this is unfortunately the only clear aspect of the reading. After the article, we suspect either ταύτης or τούτου was written, referring to a relation of one of the preceding persons listed. This relationship should then be spelled out at the end of the line, but we have not lighted upon the solution. 11 κατακεχώ̣ρ[ι]σ̣τ̣[α]ι ̣ (ἔτους) κζ Καίσαρος Μ̣ε̣χε̣ὶρ̣ ̣. The day is perhaps δ or λ. It is striking to find a subscription related to registration here, since Hombert and Préaux remark with respect to Fayyum villages, “Aucune dé- claration à destinataire unique [a category that for them includes multiple laographoi] n’a reçu de souscription. Ainsi, aucune n’a été transmise d’un fonctionnaire à l’autre ni remise à titre de reçu au déclarant. Elles sont donc restées en la possession de l’administration.”42 They noted the ___ unusual: Huebner, The Family in Roman Egypt 94. We cannot, of course, rule out the possibility of an otherwise unknown person being declared here. 42 Recherches 89. W. GRAHAM CLAYTOR AND ROGER S. BAGNALL 653 ————— Greek, Roman, and Byzantine Studies 55 (2015) 637–653 contrast between this situation and that found in Arsinoe itself, where such marks of registration were routinely found. Their view was modified (but without comment) in Bagnall and Frier: “Once again, however, we find that they do not (as far as preserved) contain signatures, but that they do have official subscriptions.”43 In fact, SB I 5661, cited above in the note to lines 1–2, has both κατακ(εχώρισται) (BL 9.244) and a following date and signature. Hombert and Préaux do not refer to SB I 5661 in this context, nor on p.95 in their table.44 May, 2015 Basel graham.claytor@unibas.ch New York roger.bagnall@nyu.edu 43 Demography 21. 44 The authors thank Rodney Ast, Dieter Hagedorn, Paul Heilporn, and Sabine Huebner for their comments.