Social Networks in Byzantine Egypt

This page intentionally left blank

SOCIAL NETWORKS IN BYZANTINE EGYPT

Social network analysis maps relationships and transactions between people and groups. This is the first book-length application of this method to the ancient world, using the abundant documentary evidence from sixth-century Oxyrhynchos and Aphrodito in Egypt. Professor Ruffini combines a prosopographical survey of both sites with computer analyses of the topographical and social networks in their papyri. He thereby uncovers hierarchical social structures in Oxyrhynchos not present in Aphrodito, and is able for the first time to trace the formation of the famous Apion estate. He can also use quantitative techniques to locate the central players in the Aphrodito social landscape, allowing us to see past the family of Dioskoros to discover the importance of otherwise unknown figures. He argues that the apparent social differences between Oxyrhynchos and Aphrodito in fact represent different levels of geographic scale, both present within the same social model. giovanni ruffini is an Assistant Professor in History at Fairfield University in Connecticut. Publications include Ancient Alexandria between Egypt and Greece (co-edited with William Harris, ) and Ostraka from Trimithis, volume  (co-edited with Roger Bagnall, forthcoming).

SOCIAL NETWORKS IN BYZANTINE EGYPT GIOVANNI ROBERTO RUFFINI

CAMBRIDGE UNIVERSITY PRESS

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521895378 © Giovanni Ruffini 2008 This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2008

ISBN-13

978-0-511-48073-7

eBook (NetLibrary)

ISBN-13

978-0-521-89537-8

hardback

Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Contents

List of figures List of tables Acknowledgements References and abbreviations

page vii viii ix x 

Introduction  The centralized elite of Oxyrhynchos



 The growth of the Apions



 Aphrodito and the strong ties of village society



 Quantifying Aphrodito’s social network



Conclusion



Stemmata



Bibliography Subject index Index locorum

  

v

For Maria Theresa and Leona Jane

Figures

. . . . . . . . . . .

Visualizing networks page  Apionic toponyms and the Oxyrhynchite nome I  A random large estate and the Oxyrhynchite nome I  Apionic toponyms and the Oxyrhynchite nome II  A random large estate and the Oxyrhynchite nome II  Apionic toponyms extracted from the Oxyrhynchite nome  Toponyms new to the Apionic archive in the s  P.Michael. : Kollouthos, Victor and neighbors  P.Michael. : Ioannes and neighbors  Stemma of the Apion family  Abridged stemma of the Dioskoros family 

vii

Tables

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Oxyrhynchite’s raw data Aphrodito’s raw data Data entry for the Oxyrhynchite Data entry for Aphrodito A UCINET network The payments in P.Oxy. . The payments in P.Oxy. . The sizes of the prono¯esiai Calculating the average prono¯esia Attestations of the Apionic archive over time First attestations of Apionic toponyms over time Distribution of Oxyrhynchite ties I Distribution of Oxyrhynchite ties II Distribution of Oxyrhynchite ties III First appearances of each Apionic prono¯esia Measuring topographical centrality I Measuring topographical centrality II Kephalai¯otai in Girgis The Aphrodito Girgis prosopography I The Aphrodito Girgis prosopography II The Aphrodito prosopography’s largest component Closeness and betweenness centrality I Closeness and betweenness centrality II Ties of strength greater than  First attestations of Aphrodito villagers by decade Network characteristics of Aphrodito by decade I Network characteristics of Aphrodito by decade II Cutpoints in Aphrodito’s corrected prosopography

viii

page                            

Acknowledgements

This book began as a dissertation at Columbia University. The following Columbia University professors helped me to propose and defend that dissertation: Roger Bagnall, Peter Bearman, Richard Bulliet, Alan Cameron, William Harris, and Adam Kosto. The following friends, relatives, colleagues, and mentors provided feedback and advice on the dissertation during its preparation: Jason Governale, Shawn Graham, Todd Hickey, Henning Hillmann, Gueorgi Kossinets, Jinyu Liu, William McAllister, Kai Medville, James Moody, Bruce Nielsen, Wouter de Nooy, Julio Ruffini, Sampsa Samila, and Balazs Vedres. Material herein was presented in various forms to and benefited from the feedback of participants in the Columbia University history department prospectus seminar, the Columbia University Late Antique Group and the graduate fellows’ seminar at Columbia University’s Institute for Social and Economic Research Policy. During the process of revision, I have been fortunate to benefit from the extended comments and criticisms of a number of colleagues, including Roger Bagnall, William Bonds, and James Keenan. Jean-Luc Fournet was kind enough not only to comment on my revisions but to provide digital images of the Aphrodito papyri and texts of yet-unpublished papyri from the same archives. The dissertation was completed during a readership at Dumbarton Oaks. The revisions were supported by a postdoctoral fellowship at Columbia University and a visiting research scholarship at New York University’s Institute for the Study of the Ancient World. To all of these institutions I am grateful. Roger Bagnall and William Harris – both mentioned already – have been instrumental not only for the intellectual guidance they have provided and continue to provide me, but for the many opportunities they have presented me as well. I owe them both a debt I cannot repay. ix

References and abbreviations

Abbreviations for journals and other standard works follow the conventions of L’Ann´ee philologique. Papyri and papyrological reference works are cited according to the conventions of John F. Oates, Roger S. Bagnall, Sarah J. Clackson, Alexandra A. O’Brien, Joshua D. Sosin, Terry G. Wilfong, and Klaas A. Worp, Checklist of Greek, Latin, Demotic and Coptic Papyri, Ostraca and Tablets, available online at http://scriptorium.lib.duke.edu/papyrus/ texts/clist.html, December, .

x

Introduction

The feudal model is also by necessity for Egypt an “Oxyrhynchus model.” This is because the mass of evidence for large Egyptian estates and great landowners in the sixth century has Oxyrhynchus as its provenance; and much of that concerns one family, the high-ranking family of the Apiones . . . Nevertheless, for the past fifteen years or so, despite obstacles, there has been a turning toward the evidence of Aphrodito, giving it equal time with that of Oxyrhynchus. Much there runs counter to the Oxyrhynchus model. In its place, or, better, side-by-side with it, the Aphrodito papyri present a picture of a vibrant agricultural community of small landholders, farmers, craftsmen, priests, monks and shepherds . . . where big landowners may be present but do not rule. James Keenan, 

Byzantine Egypt produced social networks of differing shape and size. This book explores two of those networks. The first network in this study is nome-wide, that of the Oxyrhynchite nome’s elite office-holders and families. This study examines the process by which one of those elite families grew its estates and influence to considerable proportions. The evidence available to us reveals this network’s tendency towards hierarchy and social centralization. The second network in this study is that of a single village. This study looks at Aphrodito’s self-styled “small landowners,” farmers, shepherds, craftsmen and others. It then measures the levels of interconnectivity among and social distance between these groups. This picture shows a remarkable degree of social parity and decentralization. These two pictures are not mutually exclusive. A recent monograph on Aphrodito has described a village fast on its way to becoming more like Oxyrhynchos. Constantin Zuckerman’s model is one in which “the Aphroditan evidence . . . would seem to join the Oxyrhynchite in supporting, not contradicting, the traditional ‘large estate model’ of Byzantine 

Keenan , –.



Zuckerman a.





Social networks in Byzantine Egypt

Egypt.” In this vision, Aphrodito’s fiscal independence ultimately collapsed in the face of external aristocratic pressure similar to that presumably felt by individual villages throughout the Oxyrhynchite nome far to the north. All this may be true. Oxyrhynchos and Aphrodito may be credible components in a grand unified theory of Byzantine Egypt, but we will never prove it by trying to dissolve the differences between the two models. The nature of the evidence – nome-wide in one case, village-level in another – does not permit us to propose a seamless whole with any certainty. Instead, we are left with something less satisfying in simplicity but more rich in texture. Byzantine Egypt is in this study a world of both centralizing elites and decentralizing villages, a world in which two types of social structure exist, not side by side but at different levels of scale. The models I build rest on dual foundations. The first foundation, employed in Chapters  and , is traditional prosopography, not yet fully employed on the thousands of available papyri. The second foundation is social network analysis, a theoretical tool with great potential for analysis of the ancient world. In Chapters  and , I first argue that social network analysis, a method imported to the ancient world from anthropology and sociology, provides rigorous quantitative tools for verifying or challenging the results of traditional prosopography. I further argue that the results we derive from social network analysis heighten the impression of profound differences between the social structures of the Oxyrhynchite, a nome, on the one hand, and Aphrodito, a village, on the other. These results, at first glance unsurprising, have profound implications for the study of Byzantine Egypt. Evidence of regional variations in Egyptian society traditionally elicits one of two answers. Scholars can either abandon the search for an overall model of Byzantine Egypt, and by extension abandon hope that Byzantine Egypt can guide generalizations about the rest of the eastern Mediterranean, or they can create a homogenizing model that somehow subsumes apparent regional variations. A regionalist response to these results would accept them as part of a growing body of evidence for social and economic differentiation from region to region within Egypt. A homogenizing response would suggest that any apparent differences represent stages of development or lacunose evidence, in which the evidence from one region simply does not (yet) manifest the characteristics plainly apparent elsewhere. 

Keenan b, . Sarris  does not address Zuckerman’s model, but seems to share a similar vision, in which Aphrodito’s peasant autonomy is exaggerated and its surrounding great estates more pervasive than often thought.

Introduction



This book’s final purpose is to propose a third way, an alternative to both the regionalist and homogenizing approaches. As just suggested, evidence from a nome and evidence from a village represent the proverbial apple and orange. They neither compete for Egypt as a whole nor are complete within spheres of their own. They represent two different levels of scale. Social network analysis provides quantitative tools to explain these differences of scale, and also shows how these differences can co-exist within the same larger model. The two case studies presented in this work, sixth-century Oxyrhynchos and Aphrodito, suggest that where the Oxyrhynchite evidence indicates the presence of a highly centralized aristocratic elite whose economic power grew in relative isolation from social ties, the Aphrodito evidence indicates something quite different. There, the evidence reveals a village society built on strong multiplex ties, a society in which economic action took place on social lines, a decentralized society in which literacy and mobility could give social prominence to men and women of relatively low social standing. Specifically, this study provides the first full-length treatment of sixth-century Egypt with an eye to its social networks and social connectivity. Abundant surviving evidence from Oxyrhynchos and Aphrodito permits a much closer analysis of social relations at and across all levels of society than is possible for other parts of the Roman world. Network analysis gives us the tools to measure the extent of a society’s centralization, to identify topographical patterns in the formation of its large estates, and to locate the most central – and yet frequently unstudied – figures in its social networks. The result is a cross-disciplinary approach, with anthropological and sociological theory being employed in tandem with quantitative approaches to the papyrological evidence. In addition to these introductory remarks, a guide to network analysis, and a concluding essay, this study includes four main chapters, two each for Oxyrhynchos and Aphrodito. In both cases, the first chapter is largely synthetic, the second chapter largely based on network analysis. The synthetic chapters aim to provide a prosopographical and social survey of Oxyrhynchos and Aphrodito. In each case I pay particular attention to previously understudied connections between various actors in each region, and demonstrate how those connections reflect certain structural features of their social networks. The network analysis chapters will be more 

Awareness of the necessity of such a cross-disciplinarity is growing among papyrologists and those who use the evidence of the papyri. For an entertaining discussion of the relationship between history and quantitative sociology from a decidedly different perspective, see Franzosi .



Social networks in Byzantine Egypt

quantitative, and will differ from each other somewhat in approach. The Oxyrhynchos chapter will focus on the Apionic estates, using various estimates from Apionic data to create a hypothetical population pyramid of the Oxyrhynchite nome. That second chapter will use computer analysis solely in respect to the connections between settlements under Apionic jurisdiction. The second Aphrodito chapter, however, will rely almost entirely on computer analysis to determine the Aphrodito network’s rates of centralization, its most central figures, and other structural features. (Technical terms used in these chapters are defined below, pages –.) A set of positive conclusions emerges from each chapter, outlined here in brief. In Chapter , I focus on a prosopographical synthesis of the Oxyrhynchite elite. The leading figures in Oxyrhynchite society often appear in disparate contexts, with no apparent connections to other members of the elite. Yet this may be only the result of the nature of the papyrological evidence. The number of people at the top of the nome’s social pyramid was certainly rather small, and each of these people ought to have been connected to one another in some way. The synthesis presented in this first chapter attempts to connect most of the important figures in the Oxyrhynchite papyrological record. The result is a picture of the Oxyrhynchite elite in our period as a highly centralized group, tying their far-flung holdings together through vertical, hierarchical administrative structures centered around the city of Oxyrhynchos itself. Chapter  presents a more quantitative approach, focused on the rise of the house of Apion. The first part of this approach is demographic: I attempt a census of the portions of the Oxyrhynchite nome under Apionic fiscal responsibility, and through those figures, derive an abstract model of connectivity throughout the nome’s social hierarchy. The second part of this approach is network analytical. Network analysis of the Apionic topographical material contributes to our understanding of large estate formation, and the distribution pattern of agrarian holdings in Egypt more generally. Using a computer technique I have developed elsewhere, I use network analysis to map the connections between the Apionic toponyms and other sites in the Oxyrhynchite nome. This technique helps to determine whether Apionic jurisdiction spread organically from an original rural site, or grew in a more haphazard fashion, directed by landowners and bureaucrats absent in the nome capital. The conclusion I propose therein, that Apionic land acquisition took place at a distance, and that 

See Ruffini , with additional analyses at Chapter  below, pp. , –.

Introduction



the Apionic estates across the Oxyrhynchite nome were physically isolated from one another, suggests that the growth of the Oxyrhynchite great estates did not proceed through rurally based social ties. Chapter  is a prosopographical survey of Aphrodito comparable to the survey of Oxyrhynchos in Chapter , but with somewhat different results. Because of the apparent archival origin of the Aphrodito papyri, it is no surprise that many of the figures in those papyri were socially connected to each other. It is the nature of that connectivity which interests us. I argue that the evidence from Aphrodito stands in contrast to that of Oxyrhynchos. Aphrodito’s social networks were founded on multiplex ties based on relationships between family members, neighbors, guild members, and others. Social ties between neighbors are to be expected, although the consistency with which these ties appear in Aphrodito despite differences in social status is intriguing and can help explain a number of otherwise opaque documents. The communal organizations found in Aphrodito have already been discussed in recent academic literature. My review of that material stresses the degree to which this sort of social connectivity challenges traditional notions of late antique status consciousness and hierarchy. It further provides a direct contrast between landowners in Oxyrhynchos and Aphrodito by showing the extent to which land acquisition in the latter case proceeded directly along social lines, unlike the Apionic expansion surveyed in Chapter . Chapter  is based strictly on network analysis of the prosopography of the Aphrodito papyri published by V. A. Girgis in . Analysis of this network reveals the importance of otherwise overlooked figures. Hundreds of sixth-century texts survive from Aphrodito, attesting to the existence of social connections between several thousand residents of Aphrodito and its environs. Because central portions of the papyrological record from Aphrodito belonged to Dioskoros, the well-known notary/poet, most scholarly attention has centered on him and his family. But analysis of this network with a computer program called UCINET has produced interesting results. We find surprising figures emerging from various connectivity tests: shepherds such as Victor son of Psaios, and relative unknowns such as Ieremias the priest. Much of Chapter  focuses on attempts to explain the structural centrality of these figures, concluding that landowning, literacy and corporate identities played a considerable role in increasing social connectivity, and that even figures of lower social status could benefit from 

See Chapter  below, passim.



Social networks in Byzantine Egypt

these facts. One crucial discovery is that the network analyses presented in this chapter describe network characteristics that are constant over time. Dramatic transformations in Aphrodito’s fiscal status, suggested by at least one modern scholar, appear to have had little effect on Aphrodito’s social networks. In sum, both prosopographical and quantitative analyses of the papyrological record reveal considerable variance in the shape of social networks from one level of scale in Roman Egypt to the next. The evidence from Aphrodito, a mid-sized village in Upper Egypt with one or two lesser satellite settlements, maps a well-integrated social network bound together by myriad lease agreements between small landholders, legal contracts between the community and local guilds, and so forth. The evidence from Oxyrhynchos, a large city dominating a nome of several hundred settlements across a good portion of Middle Egypt, maps a different sort of social network entirely. Oxyrhynchos, a region of large estates, produced an aristocratic network of large central hubs, in the form of wealthy landholders and powerful church institutions, linking a more diffuse and scattered array of less prominent clusters. With these arguments in mind, one final methodological issue will receive extensive treatment throughout this book and its conclusion: whether the archival nature of the evidence from Oxyrhynchos and Aphrodito distorts our conclusions. Might the differences between the finds of Oxyrhynchos and Aphrodito themselves account for the differences I find throughout these chapters? Are my conclusions inevitable, given a nome capital on the one hand, and a lesser village on the other? Would a village archive from (say) the village of Spania in the Oxyrhynchite make the social structures of that nome look more like those of Aphrodito? Would the archive of the former eparch Ioulianos, who Constantin Zuckerman has recently argued must have dominated the landholding regime in Aphrodito, make the village look more like the world of the Apions of Oxyrhynchos? At various points throughout the following chapters, I propose counterfactuals to refute these possibilities. A summary of these points is in order here. The Oxyrhynchite nome as a whole will not appear any less centralized, no matter how much evidence we might one day find from its outlying villages. The ties binding those villages to the city center are exclusive, and are manifest in our evidence precisely because comparable ties from village to village or within the villages themselves were demonstrably absent. Nor will Aphrodito’s social networks one day start to look like the vertical hierarchies of the Oxyrhynchite with the discovery of a hypothetical stash of

Introduction



papyri from the large landowners of Antaiopolis, Aphrodito’s nome capital. The differences of scale I propose between Oxyrhynchos and Aphrodito are real, but more importantly, they are not mutually contradictory; they can co-exist within the same model, without altering or destroying each other.

implications for the study of the later roman empire I suggested above that a regionalizing approach to Byzantine Egypt, in which we expect the Oxyrhynchite nome to have a different socio-economic structure from the Fayum or from nomes in the Thebaid, is an approach that undermines Egypt’s implications for the larger late Roman world. This is a logical reality more often ignored than addressed, still less supported, by the papyrologists implicitly responsible for it. Certainly, the image of late antiquity as a proto-feudal world of coloni bound to their land weakens when some of the best evidence for the phenomenon in Egypt is shown to be atypical even within Egypt. Conversely, the image of Aphrodito as a village of assertive and self-confident middle-class landowners has few implications for a larger stage if Aphrodito is thought to be atypical in a world of ravenous oikoi ready to consume the village at the first opportunity. But the homogenizing alternative is typically too pessimistic. Few people would now subscribe to all of the traits Sir Harold Idris Bell assigned to the Byzantine servile state, but recent arguments by Peter Sarris and Constantin Zuckerman amount to much the same thing: if Aphrodito was not yet like Oxyrhynchos, it would be soon enough. The implications for the late Roman world under this alternative are clear: look for independent small-holders around the Mediterranean to lose their lands to the creeping growth of the large estates. My central claim – that Oxyrhynchos and Aphrodito represent complementary bodies of evidence at different levels of scale – repudiates this pessimistic homogenizing alternative at the same time that it repairs the Egyptian evidence for use on a larger Mediterranean stage. I suggest here that differing forms of social networks resulting from settlements of different scale can be seen throughout the late antique Mediterranean. The papyrological archives at Petra show a city with close ties to the rural world, much like those of Oxyrhynchos. The Petra papyri map a world of wealthy urban landowners. Theodoros, a central figure in the papyri, owned land on his maternal side in the village of Serila, and property on his paternal side in other cities nearby. But we know nothing about Serila and the other



Social networks in Byzantine Egypt

settlements that archaeological remains show to have been nearby. At a different level of scale, Nessana, a Palestinian village of perhaps , people near a local military camp, provides an intriguing contrast. The Nessana papyri show patterns of landholding and acquisition quite similar to those in Aphrodito. Land divisions and cessions in both places show signs of strong social ties between the relevant parties, who are often family members and colleagues in professional capacities as well. Recent work on the inscriptions of Aphrodisias has detailed the active and prosperous lives of both civic officials and private citizens. Yet one searches the inscriptions in vain for any indication of nearby villages or other rural settlements, still less for a network of settlements tied to the city in the way that Oxyrhynchite settlements were tied to Oxyrhynchos. This sort of silence is suggestive; if Petra and Oxyrhynchos both have their villages, can these models instruct our search for other evidence in Aphrodisias and elsewhere? Readers of a specialized study of two sites in Byzantine Egypt may justifiably query that study’s implications for the larger world of the ancient Mediterranean. My final point here, one I will return to at length in this book’s conclusion, is that we should approach Petra, Nessana, Aphrodisias, and other sites in the ancient Mediterranean in much the same way we should approach Oxyrhynchos and Aphrodito. The presence of socio-economic features at Petra and their absence at Aphrodisias, to give a hypothetical example, cannot be used to support universal arguments about socio-economic trends in late antiquity. Further, differences in evidence types and settlement scale prohibit claims of regional differentiation between Palestine on the one hand and Asia Minor on the other. If the third way proposed by this book is correct, that Aphrodito could survive in an Oxyrhynchite environment, a variety of new approaches to other sites around the Mediterranean are possible. It would be wild from this argument to insist that villages like Aphrodito are to be found in the territory of cities like Aphrodisias, but the possibility does exist, and network analysis can provide the quantitative tools to identify such places once we think to look for them. network analysis: a tutorial for ancient historians Readers without the need for a historiographical survey of network analysis or a guide to learning how to perform this sort of analysis can skip this 

See Charlotte Rouech´e’s second edition of Aphrodisias in Late Antiquity, online at http://insaph. kcl.ac.uk/ala/index.html.

Introduction



tutorial and head directly to Oxyrhynchos and Aphrodito. Chapters  and  do not rely on the material that follows, but Chapters  and  will be difficult to follow without familiarity with social network analysis. Three main strands of academic literature have contributed to the current state of the field: mathematical work on graph theory, anthropological work on decision-making and exchange theory, and sociological work on small world theory. The mathematicians led the way. In the late s, one of the twentieth century’s great mathematicians, Paul Erd¨os, made significant breakthroughs in graph theory, the field of mathematics concerned with links and connectivity. His chief contribution to the field was the realization that it was necessary to bridge only a relatively small percentage of a network’s potential links to connect every member in a network. More surprisingly, he discovered that the larger the network, the smaller the necessary percentage actually became. In other words, the larger we estimate the population of an Egyptian village to have been, the lower the social density of that village need have been to connect all of its residents to one another. The published work of the anthropologists started much earlier than that of the sociologists, but in relative isolation from the parallel currents in mathematics. A number of European anthropologists came first: Barnes, Bott, Boissevain, Mitchell, and Epstein all published articles working out some of the basic ideas of social network analysis in the s and s, without really being aware of the new directions towards which their work pointed. In , Jeremy Boissevain and J. Clyde Mitchell coedited Network Analysis: Studies in Human Interaction. The product of proceedings of a symposium on the subject, the articles therein can still serve as a useful introduction to the field. The symposium participants were aware of the youth of their field, and as a result paid close attention to proposing and tightening formal conceptual definitions. This work also remains useful for its historiographical survey of a number of network theory’s more intriguing concepts. 

  

General handbooks on social network analysis: Newman et al. , de Nooy et al. , Degenne and Fors´e , Wasserman and Faust , Scott . For a review of Wasserman and Faust, see Erickson . For an earlier generation’s introduction to the literature and concepts of network analysis, including reachability, degree and density, see Mitchell . For the brief summary that follows, see Buchanan , –. On social density, see below, pp. –. For the evolution of social network analysis, see Boissevain’s preface to Boissevain and Mitchell , and the first four chapters in the same work, exploring various aspects of the field’s theory and methodology. Exchange theory played an important part in contributing to network theory’s attention to individual autonomy: Boissevain and Mitchell , xii.



Social networks in Byzantine Egypt

The first full-length treatment of social network analysis by a single author came the following year, with Jeremy Boissevain’s  Friends of Friends: Networks, Manipulators and Coalitions. The book is a landmark in the field, in which Boissevain opposed the standard structural-functionalist approaches to anthropology and sociology. He argued that people ask what they can get away with and what is good for them as much as they ask typical structuralist questions like what their group wants them to do. Accordingly, actors resolve conflicts based not on analysis of right and wrong, but on analysis of contact strength. According to this argument, structuralism understands the patterns, but misses how those patterns change, by not looking at the level “at which real people interact.” Boissevain as an anthropologist thus saw social network analysis in a different light than would Franzosi and Mohr, a generation later. Writing with an eye towards trends in historical scholarship, they placed social network analysis in the context of the victory of structuralism in modern historiography. Boissevain looked from the opposite direction and saw social network analysis as the reintroduction of the individual into structuralist analysis. These contrasting views highlight the extent to which social network analysis serves as a mediating methodology, bridging the gaps between history, sociology, and anthropology. This work on Byzantine Egypt serves a similar goal, providing a quantitative bridge between recent microhistories that focus on individual agency on the one hand, and the vast amounts of evidence detailing aspects of Egyptian society’s larger social structures on the other. By the mid-s, sociologists had entered the social network arena in force. Stanley Milgram, himself a social psychologist, highly influential in the late s, conducted a unique experiment into what is now called “small world” theory. Milgram, equally well known for his bizarre experiments in pain infliction, devised a series of letter-sending experiments, in which he asked the participants to forward a letter to anyone they thought would be socially closer to the letter’s final intended recipient. These experiments demonstrated the surprising connectivity of the American social landscape, and ultimately gave rise to the popular clich´e that “six degrees of separation” are all that stand between any two people. In , Mark Granovetter thought he had found a way to explain Milgram’s findings. The essence of Granovetter’s proposal was that “strong ties” – those between close friends and family members – are not the most  

 See Franzosi  and Franzosi and Mohr . Boissevain , –. See both Granovetter  and Granovetter .



Milgram .

Introduction



crucial building blocks of social networks. Rather, it is the men and women we do not know well, those at the periphery of our social sphere, that make our larger social networks well connected. Our close friends and family are disproportionately alike to each other, and thus add little in terms of outside connectivity. But the man you see only a few times a year as he travels through your part of the country and brings news from the outside world serves as a crucial bridge between your part of the network and others. This turns out to be a crucial idea for our understanding of late antique Egyptian social networks. For example, we will see in Aphrodito the apparently high social importance of shepherds, marginal, combative, and difficult figures whose physical mobility and multiple social functions made them central figures in the Aphrodito social network. Various other issues have risen to prominence in the historiography of network theory over the last generation. First, scholars have paid attention to exchange theory, the notion that decisions are made based on the value gained and lost in a transaction. Value is used here not simply in an economic sense, but in a social sense: actors in a network make decisions to some degree according to the connections and access those decisions will afford them. This aspect of exchange theory interacts closely with network theory’s pronounced post-structuralist inclination. This approach has tended to see controversies as social rather than issue-based in nature. Thus, the actions and decisions of individual actors in a given moment of crisis are dependent more on their role in a social network than their actual opinions about the nominal issues at stake. (This was the approach Elizabeth Clark took in her network analysis of the Origenist controversy, and one with potential analogies to future work on, for instance, the monophysite schism between the Egyptian and Chalcedonian churches. ) Within the last half-dozen years, the historiography of network theory has come full circle, returning to the mathematics whence it came. Two of the key players have been Steven Strogatz and Duncan Watts, the first mathematicians to formulate exactly why Milgram and Granovetter’s networks worked the way they did. They discovered that the so-called “small world” phenomenon takes place in networks existing in a state somewhere between complete order and complete randomness. With the introduction into an ordered network of a few random links, the average distance connecting each actor in the network begins to plummet dramatically. These   

See Chapter  below, passim, particularly pp. –. For the Origenist controversy, see below, p. . First in Watts and Strogatz , and then in Watts b and Watts .



Social networks in Byzantine Egypt

random links account for the surprising “six degrees of separation” effect. This discovery came only after they removed their networks from the social sphere and retreated into purely abstract mathematical models. The most remarkable aspect of this return to the mathematical has been the extent to which social networks are now seen to share structures identical to those occurring again and again in the realm of the physical and natural sciences. The implications of this work are wide-ranging. It means that in theory, we can describe both the flow of opinion and information in late antique Oxyrhynchos and the spread of AIDS in late twentieth-century America using mathematical models from the physical sciences. In concerning themselves with how small-world networks came into being, Strogatz and Watts ignored how many links belonged to each node in these networks. Using some of the same data-sets employed by Strogatz and Watts, Albert-L´aszl´o Barab´asi and his students began to look at the distribution of node sizes. Put in terms of a social network, this is the equivalent of asking how many friends or connections each person in a network actually has. Other networks are susceptible to this sort of analysis as well. How many web-pages point to any given website on the internet? How many flights leave any given airport to other airports around the country? Barab´asi found that distribution of node sizes in networks of this sort does not follow a bell-curve distribution, but rather follows what is called a power-law distribution. In this sort of distribution a rather large number of nodes have few links, while an increasingly small number of nodes have many more links. In terms of social networks, one or two nodes (the President of the United States, for instance) at the far end of the scale have thousands of social connections, whereas the bulk of the population has only a few dozen or a few hundred connections maintained on a regular basis. A crucial part of this development has been the focus on hubs and connectors. Like Granovetter’s weak ties, hubs can play an important role in analysis of late antique social networks. Like weak ties, they are a crucial building block in a small-world network. In general terms, hubs are members of a network with an inordinately high number of connections  

 

 Barab´ Watts , . asi , –. Plotted on a regular graph, power-law distributions are gradually decreasing curves. Plotted on a so-called “log-log graph” in which each axis proceeds on a logarithmic scale, power-law distributions map a straight line. These numbers are arbitrary. The size of any individual’s personal network is notoriously hard to estimate. See Wasserman and Faust , – for the literature on ego networks generally. See the chapter of the same title in Barab´asi .

Introduction



to other members in the network. If the average Egyptian peasant has fifty acquaintances, a village’s hub is the gregarious or socially important exception with  acquaintances. But hubs are not simply exceptions, they are necessary exceptions, crucial steps decreasing the social distance between many of the network’s members. This work alerts us to the likelihood that certain members of Egyptian society – Dioskoros of Aphrodito comes to mind – will look disproportionately important in our analysis, not simply because of the nature of the evidence, but because people like him served as society’s social hubs. With this work in recent years, social network analysis is poised to enter the mainstream of intellectual thought. The field’s growing trendiness has brought it to the pages of the New York Times. An article on the subject from March  analyzed the implications of network analysis on the National Security Agency’s terrorist surveillance program. An article on the subject from January  featured Duncan Watts sharing his thoughts on network theory with the vice-chairman of the NASDAQ. A morbid article published at slate.com a month after the World Trade Center attacks of September, , brought “small world” theory to bear on the problem of why no one the author knew directly had died in the attack on the two towers. In , social network theory received two full book-length treatments, by Albert-L´aszl´o Barab´asi and Mark Buchanan, both of them deliberately thin on abstract theories and mathematical treatments, and designed for more general consumption. Duncan Watts’ second book on the subject of network theory came out early in the following year. In this wave of thinking about network theory, the focus seems to be on developing a network equivalent of a “Grand Unification Theory,” in discovering why a cell’s chemical reactions, the directors of American corporate boards, and web-links on the Internet all share the same basic network architecture.  





On the problems of distortion potentially caused by Dioskoros in his own archive, see Chapter  below, pp. –. Patrick Radden Keefe, “Can Network Theory Thwart Terrorists?” The New York Times Magazine, March , . See also the brief editorial on the NSA’s telephone number database by Jonathan David Farley, “The N.S.A.’s Math Problem,” The New York Times A, May , . Emily Eakin, “Connect, They Say, Only Connect,” The New York Times, January , , page A. See also William Middleton, “Popular Mechanics,” The New York Times Magazine: Style and Entertaining, November , , analyzing “sociometry” through the hubs and degree centrality of a New York City socialite party. The national security implications of networks are also gaining public attention: see Laura Blumenfeld, “Dissertation Could be Security Threat: Student’s Maps Illustrate Concerns about Public Information,” Washington Post, July , , on Sean Gorman’s dissertation, and its analysis of American communication networks. David Plotz, “Life’s Odds and Sept. : No one I know personally was on the list. Why?” Posted online at slate.com at October , .



Social networks in Byzantine Egypt

The implications are clear: a growing number of people have come to believe that network theory is relevant to nearly every field of intellectual investigation. social network analysis and the ancient world No modern studies of Roman Egypt have employed network analysis on the available data. Jane Rowlandson’s Landowners and Tenants comes close in her approach to the early empire, particularly in her fourth chapter on “The Landowners and their Properties.” But her aim is to highlight social connections only within the context of her own interest in landholdings; it is not her intent to perform a general network analysis. Within the specific context of landholdings, a subset of Oxyrhynchite connections interest her. Namely, Rowlandson hopes “to examine how an urban landowning class derived its wealth from the rural hinterland.” Throughout this study, particularly in the chapters on Oxyrhynchos, I examine how cities form links with their rural surroundings, and how those links contribute to patterns of estate formation, as in the case of the Apionic estates. Rowlandson’s work, although outside of our period, provides some insight into this question. Perhaps unsurprisingly, she notes that the plots of absentee landholders tend to have been more scattered than those of landholders present on site, a pattern we will confirm in our analysis of the expansion of the Apions. As Rowlandson herself notes, equally interesting is the question of how landholders found their tenants: It is . . . probable . . . that most private landlords found their tenants through personal contacts. This need not imply that landlord and tenant were necessarily well-known to each other; the writer of a letter offering to collect rent from a tenant on behalf of her sister (unless ‘sister’ here merely expresses a relation of friendship) evidently thought that the tenant would not know who she was unless the landlord sent a letter of explanation.

These sorts of multiplex ties – in which a person is both a personal contact and an economic one, for instance – have received much attention among network theorists. This is a natural aspect of human behavior: to fulfill certain needs, we are most likely to turn first to those we already know. But     

Rowlandson is reviewed in Keenan . He calls her work a vision of a “gentler and kinder” Roman Egypt (). Rowlandson , . Rowlandson , . For Apionic parallels, see Chapter  below, pp. –. Rowlandson , , citing P.Oxy. .. See e.g. White , with the brief discussion in Chapter  below, pp. –.

Introduction



Rowlandson further suggests that these sorts of “informal and personal ties” between two parties are social structures easily destroyed by any emergent gap of social distinction between landlord and tenant. This proposal, if it could ever be tested, would have interesting implications for the social structures of late antiquity, a period in which scholars have typically believed the social distinctions between rich and poor in Egypt to be growing ever wider. Modern scholarship of the ancient world more generally includes only a few works of social network analysis. The method has yet to receive a general audience among ancient historians, and its practitioners are not taking complementary approaches. The first major work, an article on Cicero’s correspondence networks, appeared in  in Social Networks, a journal read by few ancient historians. Alexander and Danowski tabulated a number of variables relating to Cicero’s correspondents, including their social status and gender, and used the UCINET computer program to run network analysis on the data. The results led them to conclude that senators and knights were not separate social blocs, but in fact shared considerable structural similarity in late Republican society. They emphasized that this result would not surprise a modern Roman historian, pointing out that the value of their work lay in the way it brought a new quantitative method to bear on an old debate. The network theoretical approach has been most common in the field of early Christianity. Here, the landmark work of ancient network analysis is a collective one, the  issue of Semeia guest-edited by L. Michael White. In his preface, White explains that the eight works assembled in that issue serve first “to introduce network theory and analysis; [and] second, to show direct application of this approach to social history research on early Christianity and its Hellenistic-Roman environment.” These contributions are noticeably uneven. Some of them dwell on the theoretical side of network analysis without employing the necessary heavy industry,  



Rowlandson , . To this list of literature on the ancient world should be added a fascinating Byzantine contribution, Chapter  of Margaret Mullett’s  Theophylact of Ochrid. Her interest is in a different type of social network than that which concerns us here. She focuses specifically on one individual’s ego network, rather than a group network. Her interests are also more impressionistic: as she notes (), her analysis does not lend itself to “the statistical techniques used by social anthropologists” such as density and degree. An intriguing part of her discussion is its attention () to the social function of the network itself, the reasons why Theophylact maintained his social connections. See Erickson  for a more general discussion of the implications of social network analysis on the historical discipline, including a survey (–) of examples of historians using network analysis in their work.  White a, vii. Alexander and Danowski .



Social networks in Byzantine Egypt

the quantitative techniques underlying the theory. In other contributions, “social network analysis” seems to serve solely as a trendy substitute label for traditional prosopographical methods. Here, White’s own contributions to the volume provide one of two worthy exceptions. In addition to a survey of network theory and its possible applications to the ancient world, White includes three case studies of varying length in which he uses network theory on ancient topics he thinks useful for the study of early Christianity. The first case is a prosopographical review of the western aristocrats mired in the controversy over Gratian’s removal of the altar of Victory from the Senate in . He concludes that the so-called “pagan revival” was essentially a “family affair” centered around kinship ties and priestly college membership. By logical extension, the controversy over the altar was not solely religious, but “had deeper social roots.” White’s shorter studies, of patronage and the Isis Cult at Pompeii and non-Jewish synagogue supporters in the eastern Mediterranean, point in essentially the same direction. Religious ties are important social mediators, but they must also compete in an individual’s decision-making process with patronage and family ties. Elizabeth Clark’s contribution to this Semeia volume is the other exceptional piece. Her submission is a more focused look at the social networking aspects of the theological debates she treats at full length in The Origenist Controversy. These works follow this intricate theological debate troubling the church in the late fourth and early fifth centuries through a combination of prosopographical and social network analyses. Following the lead of various post-structuralist sociologists whose work we will discuss below, Clark concludes that the real issues underlying the Origenist controversy were not theological at all, but social. In short, social networks – to whom the theologians were connected, and by how many degrees – were “more decisive than issues of motivation or belief in explaining behavior.” The notion of network density – the ratio of actual connections to potential connections within a social network – is crucial here. The Origenist controversy was as heated as it was precisely because of the high degree of social interconnectivity linking all the relevant players to each other. By Clark’s calculations, Jerome and Rufinus, the two major players in the controversy, were members of social networks with a surprisingly high level of density.   

 White b.  White b, .  White b, . White c. See both Clark a, her piece in Semeia, and Clark b, The Origenist Controversy. Her Semeia piece is more directly relevant to the methodology of network analysis.  Clark b, . Clark b, .

Introduction



To my knowledge, John Chow’s  study of  Corinthians and social networks in Corinth represents the first full-length treatment of an aspect of the ancient world informed by network theory. Chow’s goal is “to investigate some of the behavioural problems in the church at Corinth in the light of the phenomenon of patronage.” He appropriately identifies network analysis as a useful tool for understanding the nature of patronage, because of network theory’s ability to handle asymmetrical ties, indirect ties, clusters, and other complex social structures. However, he deliberately forsakes the quantitative analysis typical of most network theory, and also deliberately avoids the structural emphasis most network theorists favor, arguing more in favor of personal agency and initiative. The necessity of such a decision perhaps suggests a poor fit between the subject (Paul’s  Corinthians) and the network analytic technique: in the final analysis, Chow’s study is more a traditional work on patronage than a new application of network theory. A  study by Harold Remus on voluntary associations and networks in antiquity, subtitled “Aelius Aristides at the Asclepieion in Pergamum,” comes much closer to providing a rigorous application of new network techniques. This work focuses on the social connections between incubants in Asclepius sanctuaries, asking how they related to each other, whether and why acquaintanceships begun there grew into friendships, and what social patterns emerge from the answers. Here, the focus is on Aristides’ stay at the Pergamum Asclepieion during two specific periods in the s ad, as described in his Orations and Sacred Tales. A central thesis for Remus is the assertion that the social networks formed at the Asclepieion “played a significant role in whatever healing Aristides experienced . . . and, moreover, contributed to the continuance and ‘plausibility’ of the Asclepius cult in Pergamum.” What separates Remus from most authors who have done social network analysis in the ancient world is his use of tie-strength and the direction of ties, concepts discussed further below. A comparable willingness to tackle network analytical concepts is present in the first major work of network analysis on ancient Judaism. Catherine Hezser’s  work on the rabbinic movement in antiquity employs network theory as one of its theoretical approaches. Lately, students of Jewish antiquity have come to challenge the old orthodoxies placing the Sanhedrin and the patriarch at the peak of a rabbinical hierarchy which ensured   

Discussed in brief at Chow , –. For both points, see Chow , .  See pp. –. Remus , .

 

Chow , . Remus .



Social networks in Byzantine Egypt

the continuity of a unified “Halakhah,” or sum body of Jewish religious thought. With various aspects of that orthodoxy succumbing to deconstruction, Hezser sees various sociological theories, including network analysis, as ideal tools for parsing difficult rabbinical sources. Network theory would be most useful for analyzing the “ties between actors . . . as potential communication channels which can serve to transmit information from one person to another,” and thus as a way to understand the “internal structure of the rabbinic movement in Roman Palestine.” Hezser proposes (Part II, Chapter ) to use rabbinic sources as a way to analyze the contacts between individual rabbis. Using these in concert with the concepts of density, cluster, and centrality, she argues that we can identify which rabbinical leaders were in a position to manipulate others, and which could “monopolize the flow of information and services to and amongst his followers.” Despite opening the door to data so potentially rich in quantitative uses, Hezser, like so many of these authors, does not use network theory in any mathematical way, preferring to keep it as an abstract backdrop to her work. For instance, when she speaks of a “correlation between references to social meetings between small numbers of rabbis and references to agreements amongst small numbers of rabbis,” she admits that this is only a “suspicion . . . supported by network theory which knows of opinion clusters amongst social clusters.” A sociologist working with modern data would demand that such a suspicion be confirmed by a more quantitatively rigorous approach. This avoidance of quantitative approaches makes Hezser uncertain of her conclusions: for instance, whether rabbinic houses are a network theorist’s “cliques,” “action sets,” or a third creature entirely remains an unresolved question. Philip Harland wrote an article called “Connections with Elites in the World of the Early Christians” for a  volume of social science approaches to the history of early Christianity. Network theory is not the only conceptual framework guiding Harland’s work: he concerns himself first with the concepts of “stratification,” “social status,” and “social mobility.” His general aim is to map “the social avenues whereby Christianity made advances within various strata of society, including the preConstantinian elites.” While Harland’s work is informed by the general     

See Hezser  Chapter  for a survey of previous scholarship. Her introduction to network theory comes on pages –, its application in Part II, Chapters – and –.  Hezser , quotation at ; see generally –. Hezser , .  Hezser , –.  Harland . Hezser , . Harland , .

Introduction



surveys of Mitchell, Boissevain, Wellman, and Wasserman and Faust, all authors of seminal works of network analysis, the relative lack of evidence relating to his topic – compared to e.g. Elizabeth Clark’s Christian networks – keeps his survey away from the quantitative side of network analysis. Instead, he works at a high level of abstraction, surveying the sorts of networks through which Christianity spread in specific instances: the family, the military, the familia Caesaris, and so on. The result is not so much an advance in the study of early Christianity as it is a successful demonstration that network theory can help affirm aspects of modern scholarship’s orthodoxy about the spread of Christianity. Compared to most of the authors surveyed so far, Shawn Graham’s  study of brick-makers in imperial Italy is atypically rigorous in its level of quantitative detail. Graham focuses on a particular problem of diffusion and innovation: how did the practice of consular dating in brickstamps spread so rapidly in the early second century ad? He approaches the problem with the epidemiological thrust common to recent studies of information diffusion. A new idea is thought of as a disease, the central network hubs being those actors whose influence contributes to the spread of the disease-idea. In this case, overlapping attestations of names on brickstamps allow Graham to generate a network of figures involved in the brick industry, and to identify the key hubs and bridges in the network. By extension, he argues that these hubs and bridges are themselves the actors whose influence contributed to the spread of consular dating in brick stamps. His unexpected conclusion – that women such as Domitia Lucilla and Plotina were crucial to the structure of the brick-making industry in Italy in this period – is similar to the unexpected importance of minor or tangential figures in other networks: the aunts and in-laws I have identified in previous work on late antique pagan philosophers, and the shepherds of Aphrodito who appear so important in analyses of that village. Katja Mueller’s work is the most striking example of the inroads network analysis and its related disciplines have made in the study of the ancient world. Mueller has written several articles discussing the implications of geographic information systems (GIS) for the study of GrecoRoman Egypt. Her first published article and her dissertation, recently published as Settlements of the Ptolemies (), focus on applying the ranksize rule to recently published Ptolemaic census data to establish theoretical   

See Graham , Graham and Ruffini  and Graham forthcoming. Philosophers: see Ruffini . Shepherds: see below, Chapter , pp. –. Mueller , .



Social networks in Byzantine Egypt

population sizes for settlements in Ptolemaic Fayum. Most significant for the field of ancient network analysis are Mueller’s  articles on multidimensional scaling and Monte Carlo simulations of the data from the Greco-Roman Fayum. These are powerful techniques employed for visualizing the relationships between large groups of data, and for identifying the clusters or groups within that data. Mueller’s aim in these articles was to locate ancient Fayum settlements by employing the power of network analysis on the topographical network of the Fayum as represented in the Prosopographia Ptolemaica Online. Central to Mueller’s work is her assumption, one which I share in my analyses of the Oxyrhynchite nome in Chapter , that settlements “which were geographically closely linked should appear together in the same texts.” Thus Mueller is the first person in the study of antiquity to apply network analysis to places instead of people. A recent conference in Crete in  showed how network theoretical ideas are starting to take root in wider areas of study of the ancient world. Papers at that conference employing network analysis addressed topics as diverse as pagan monotheism, late antique pagan elites, the ThrakoMakedonian monetary system, the relationship between Hellenistic cities and kings, and the relationship between sanctuaries and sacred delegates to those sanctuaries. Published proceedings of this conference may potentially go a long way towards broadening the audience for network analysis among scholars of the ancient world. starting from scratch: how to make and analyze a network data set To trust the results I present based on network analyses in Chapters  and  of this book, readers must be confident in their ability to duplicate those results. That confidence in turn will enable scholars to import the method to questions of more direct relevance to their own research, with data from Egypt or elsewhere in the ancient Mediterranean. The first step   

 

Mueller , , with Ruffini  forthcoming.  See Ruffini  and pp. – below. Mueller a, b. See now Ruffini . Mueller’s techniques have been subject to a challenging critique in Hoffman and Klin . The appearance in the pages of The Journal of Juristic Papyrology of a mathematician and a computer scientist indicates the potential for cross-disciplinarity inherent in these approaches. Publication is anticipated in a special issue of the Mediterranean Historical Review (personal communication, Dr. Christy Constantakopoulou). For those interested in a more in-depth tutorial, de Nooy et al.  is particularly valuable.

Introduction



in network analysis is the creation of a data-set. Unlike modern anthropologists and sociologists, who can interview live subjects before creating networks based on their data, ancient historians are more confined in the options facing them. Nevertheless, ancient data can still supply evidence for an array of connections: () person to person, if we know the exact nature of a social connection between the two people, () person to place, if a person is attested in a text as at, from or going to a specific place, () place to place, if two places are attested in the same text, and () person or place to event, if a person or place is attested in a text with other people or places. The second and fourth examples are two-mode networks, in which evidence of one type (people or places) is linked to evidence of another type (e.g. a social event). In the case of papyrological evidence from Byzantine Egypt, the social events are the papyri themselves – personal letters, loans, rent agreements, receipts – that link people to each other indirectly through their joint involvement in a papyrological event. Ultimately, our interest lies in one-mode networks, in which people are connected directly to other people or places are connected directly to other places. We will discuss conversion from two-mode networks to one-mode networks shortly. It is important to point out that most data-sets a researcher is likely to encounter will begin their lives as two-mode networks. Personal name indices, prosopographical guides, topographical registers, and other standard scholarly tools are two-mode in nature, listing people or places on the one hand and sources attesting them on the other. Common appearance in the same source constitutes an indirect link in data-sets drawn from these tools. It is crucial to remember that the definition of what constitutes a social connection is to a considerable extent up to each researcher. Some people may prefer a more refined sort of connection than mere appearance in the same source. Signatories to a petition – for instance that from Aphrodito to the empress Theodora discussed in Chapter  – may not have been in the same place at the same time, and may conceivably have not even known each other. A more focused approach might create a data-set in which a connection is said to exist only between a lessor and a lessee, or between a rent payer and his or her intermediary. These connections are certainly more precise and easily understood. But these decisions sacrifice sample size, and may leave us with inadequately small networks. Joint appearance in the same text provides data-sets for this book of hundreds of places and thousands of people. 

The data-sets I use for analysis of Oxyrhynchos and Aphrodito are available online in their raw forms at www.grr.net/SNBE/, where they can be downloaded for import directly into UCINET.



Social networks in Byzantine Egypt Table  The Oxyrhynchite’s raw data

The foundation of all analyses on the Oxyrhynchite nome presented in this book is the  Pruneti topographical register. While many Oxyrhynchite papyri have been published since , Pruneti’s register is a large and accurate sample size for this sort of analysis, and additions to our knowledge since  would have little impact on the results. Table  

For a defense of this claim, see Chapter  below, p. , n. .



Introduction Table  Aphrodito’s raw data No.

Name

 

Abr/ Abraam



Abraam



Abraam



Parents

Date

Function

Texts

bo¯ethos gnapheus

P.Flor. . P.Cair. Masp.  P.Cair. Masp.  P.Cair. Masp.  P.Cair. Masp. , ,  P.Cair. Masp.  P.Cair. Masp.  P.Cair. Masp.  P.Lond. .

son of Anouphios  son of B¯esios 



kt¯et¯or

Aurelius Abraam

son of Bikt¯or 

, , 

scribe, witness, kt¯et¯or



Abraam



Abraam



son of Pb¯ekios  son of H¯orbanip¯e 

Abraamios



Abraamios

VI



Abraamios



ge¯orgos apo scholastik¯on, notary lachanop¯ol¯es

P.Cair. Masp. 

shows the first page of this topographical register. While the commentary and dates are valuable, for network analyses of the Oxyrhynchite nome, we are specifically interested in the places and the texts attesting them, which together form our two-mode Oxyrhynchite network. The foundation of all analyses on Aphrodito presented in this book is the  Girgis prosopography. As prosopographies go, this is an imperfect source. Gascou has called it an “instrument de travail des plus m´ediocres.” Nonetheless, for network analysis it provides an adequate starting point, an assertion I defend in greater detail in Chapter . Table  shows the data from the first ten entries in the Girgis prosopography. Again, the various categories all provide useful information, but we are primarily interested in the people and the texts attesting them, which in turn form our two-mode Aphrodito network. The next step is to convert the two-mode networks derived from the Pruneti register and the Girgis prosopography into computer-readable data  

For further discussion of the characteristics of this network, see Ruffini .  See Chapter  below, pp. –. Gascou , .



Social networks in Byzantine Egypt Table  Data entry for the Oxyrhynchite

A. . .k.tiou Abak() Agathammonos Adaiou Aeianou Atha[ Athlitou Athek[ Athuchis Akakiou

poxy pgiss pcol ptheon, pmerton, ptheon, ptheon, poxy, poxy, poxy, poxy, poxy, poxy, poxy, poxy, poxy, poxy, poxy, pprinc, poxy, poxy poxy poxy poxy wo pprinc, poxy, poxy, poxy, poxy, poxy, poxy, pmichinv, psi psi, psi



This table is formatted for readability; it is not necessary to indent the data or place it in strict column form. In the data-set I have provided online (see n.  above) the special characters have been replaced with the letter x to avoid complications with the UCINET computer program.

Table  Data entry for Aphrodito          

pf    , ,     pl 

files. Programs like UCINET accept a number of different files, including raw text and Microsoft Excel spreadsheets. These data-sets are essentially nothing but lists of people and places and the texts in which they appear. Table  shows what the first ten entries in the Pruneti register look like when converted into a two-mode network format known as DL nodelist. Table  shows what the first ten entries in the Girgis prosopography look like when converted into that same format. 

For the data-sets used in this book, I have used raw text files with data in the DL nodelist format, described in the UCINET packaging.

Introduction



In the Aphrodito case, I have substituted numbers for names, to avoid confusion over homonyms. In both cases, I have abbreviated the papyrological references somewhat further than is standard, for ease of dataentry. To make a two-mode network linking people or places and papyri susceptible to the required analyses, we must use UCINET’s “Affiliations” transformation. This transformation creates a direct link between each person or place appearing in the same text, producing a single-mode network as a result. Thus, in the case of the ten Aphrodito entries given in Table  above, Person  links directly to Person , Person  links directly to Person , and so forth. Since it is fairly common for someone from Aphrodito to have appeared in more than one text, a great many of these people overlap to form a network of affiliations which is large even by the standards of social network analysts working on data-sets from the modern era. One weakness to this approach is its implicit definition of connectivity. In this method, two residents of Aphrodito may be affiliated if they appear together as signatories to the same document, as prosecution and defense in records of the same court case, and so on. To some extent, this is reasonable: a connection of some sort no doubt existed between people in these situations. But we must be wary of imagining that each connection in this Girgis data-set can be considered equal. Once a two-mode network is turned into a one-mode network, it is essentially nothing more than a grid in which the members of the network are each assigned a row and a column, and each space in the grid is assigned a  or a  based on whether a connection exists between the two relevant members of the network. Table  below shows a portion of the output generated by UCINET when the Girgis Aphrodito network is displayed in this grid form. NetDraw, a program packaged with UCINET, can create elaborate visualizations of networks of all sizes, in both one-mode and two-mode forms. Figure  shows what a much larger portion of the Aphrodito data from the Girgis prosopography looks like when visualized through NetDraw. Images such as this can be very helpful in both clarifying one’s research agenda and identifying potential problems with one’s dataset. In this case, we see right away that the Aphrodito network is not a unified whole: some large pieces remain separate from the rest of the network. We also see that some portions of the network are uneven or  

Found in UCINET under the menu option Data→Affiliations. See Wasserman and Faust , –, describing such data-sets as e.g.  twentieth-century sociologists,  CEOs and their spouses, and perhaps most famously, Padgett’s  Florentine families.



Social networks in Byzantine Egypt Table  A UCINET network          

         



This grid shows the attested connections between the first  people listed in the Girgis prosopography (rows –) and the first  papyri in my Girgis data-set (each column). A zero represents the absence of a person from that papyrus, a one represents the presence of a person in that papyrus.

lumpy; asking why leads us to the final step in the preparation of our data-set. After all this labor, we must be honest about the fact that our data-sets are full of potentially distorting factors. Both Oxyrhynchos and Aphrodito are useful case studies of the ways in which ancient archives distort our vision. My network analysis of the Oxyrhynchite nome indicates that the Apionic archive is so disproportionately large that it alters our ability to gauge correctly the connectivity levels between settlements in that nome. Only by correcting for this distortion (through a technique I discuss in greater detail in Chapter , p. , n. ) is it possible to analyze the growth of the Apionic estates. My network analysis of the Aphrodito data produces a rather different and unexpected result. Dioskoros and his immediate family might be expected to alter the shape of the Aphrodito social network by virtue of prominence in their own archive. Certainly, our view of Aphrodito is their view, but their presence in the data-set we shape from their archive does little to alter that data-set’s characteristics. Removing Dioskoros and select relatives from the Aphrodito data-set (through a technique I discuss in greater detail in Chapter , pp. –) shows that crucial measurements of network analysis are not affected by the presence of Dioskoros and his family. Generally speaking, this is because the texts we turn into raw data count as social connections between everyone in the text; even when

Introduction





Figure  Visualizing networks For more details on the alterations to the Girgis prosopographical network necessary to create this visualization, see Chapter  below, page . Data-set five, removing large distorting texts, is shown here.

Dioskoros and his relatives are removed from that text, the ties between the other participants therein still remain. Individual texts can sometimes cause distortions in network analysis as well. Here, tax lists and other long documents pose the greatest risk. These are the lumps or abnormally large groups that we can see most easily through NetDraw visualizations such as that in Figure  above. How central a settlement in Oxyrhynchos or a farmer in Aphrodito appears in certain types of network analysis depends to some extent on what types of texts attest to their existence. Unusually large texts can have a considerable distorting effect. In network analyses of both sites, I selectively remove such texts from my data-set, and present the results achieved both with and without these corrections. In a number of these cases, I then proceed to test my results against randomly generated networks of the same size as the networks in question. This method (discussed in greater detail in Chapter , pp. –, , , and Chapter , pp. –) provides comparanda against which to judge whether the network characteristics we discover in our analyses represent real social and topographical features, or mere quirks of the evidence.



Social networks in Byzantine Egypt

While my tests for distortion and my attempts to correct for it receive greater treatment in their respective locations, a brief technical explanation is useful here. UCINET provides a simple procedure for removing a person, place or text from one’s data-set. If one wishes simply to test the distorting effects of the scribe named Abraamios or the village named Takona, this procedure will create a second network allowing one to do just that. However, removing hundreds of items from a data-set of thousands would be time-consuming under the same procedure, and creating an attributes network provides an ideal alternative. An attributes network is a data file in which every member of the network sharing a certain attribute is marked with a “” and every member not sharing that attribute is marked with a “”. A wider range of attributes can also be coded, for example, to isolate members of every type of occupation group. Pajek, another network analysis program packaged with UCINET, and recently the subject of an entire monograph on its analytical features, is the most effective tool for isolating these attributes. In my analyses of the Apionic estates, I load into Pajek both my Oxyrhynchite topographical network and an attributes network listing every Apionic toponym. Pajek can then extract one from the other and allow me to analyze both components in isolation to test for possible distortion. network vocabulary and concepts In the section that follows, I survey some of the main concepts of network analysis and place them in the specific context of network analysis in the ancient world. I provide formal definitions and references to more thorough discussions of various concepts in network theory. In some cases, these concepts are crucial to the work of the following chapters. In other places, I have merely outlined ways in which I think these aspects of network theory can in the future be an aid to answering some intriguing questions about late antique Egypt. Throughout this discussion, I typically provide examples using social (that is, personal) connections. But the concepts outlined in this section apply equally well to networks of topographical (geographic) connections. The most basic component of any network is a node. In simple terms, a node is any point in a graph or network. That   

To be found under UCINET’s menu option Data→Extract. See de Nooy et al. , with its Chapter  on attributes, particularly pp. –. See also de Nooy et al.  with a wide-ranging glossary of terms on pages –.

Introduction



point can represent an actor, such as Flavius Apion of Oxyrhynchos, a text such as Aphrodito’s petition to the empress Theodora or a settlement such as Aphrodito’s satellite village, Phthla. Thus, when we talk about the degree of a node or the number of steps between two nodes, we can be referring to any piece of social, papyrological, or topographical data that the ancient world provides us. A fundamental feature of how the nodes in a network relate to one another is the network’s mode (i.e. one-mode and two-mode graphs). As I have explained above, a one-mode graph establishes connections between comparable types of actors, namely, connections between people, connections between cities and connections between texts. In a one-mode social network, actors connect directly to actors, e.g. Phoibammon son of Triadelphos married a relative of Dioskoros of Aphrodito. Two-mode graphs do not measure such direct connections between actors or individuals, but rather measure connections between actors and events. For our purposes, we can define actors and events quite broadly. We may, for example, assume that if two people or two places appear together in the same papyrus, some sort of connection exists between them, through the social event that that text represents. For instance, in a two-mode social network of the Aphrodito evidence, Flavius Pneis is connected to Aurelius Onnophris by virtue of the fact that both witnessed a sale of land from Aurelius Ioannes to Apollos son of Iosephios, and thus both appeared in the same text. A data-set describing all the links between people (or places) and the papyri in which they appear forms a two-mode graph. Two-mode graphs are at the heart of the analyses I perform in the chapters on Oxyrhynchos and Aphrodito below. Paola Pruneti’s topographic register of the Oxyrhynchite nome and V. A. Girgis’ prosopographical index form the foundations of the data-sets in question. Both networks map the connections between actors (people or places) and events (the texts which name them). But these networks will not help answer certain questions unless they are simplified by turning them into one-mode networks. In both cases, it is necessary to perform an affiliations analysis to turn each two-mode graph into a one-mode graph. The results are affiliation networks, so called because they link social actors through the social events in which they are affiliated; these networks are the fundamental building blocks of network analysis in this study. 

P.Michael. , which I discuss in further detail in Chapter  below, pp. –.



Social networks in Byzantine Egypt

UCINET’s “affiliations” function will do this transformation automatically by making, for instance, the following conversion of the two-mode connection discussed on the previous page: Flavius Pneis → P.Michael.  → Aurelius Onnophris (Flavius Pneis appears in P.Michael 40 with Aurelius Onnophris) becomes Flavius Pneis → Aurelius Onnophris (Flavius Pneis is connected to Aurelius Onnophris) Affiliation networks based on this sort of conversion are risky propositions because we do not really know what sort of connectivity they show. Did Pneis know Onnophris well, or were they in the same place just once, for the purpose of witnessing this land sale? If we leave this concern aside, and grant that joint appearance in a text at least implies connectivity of some kind, affiliation networks will become the bread and butter of our discussion. Constructing, as I do, data-sets linking over  toponyms in Chapter , and nearly , villagers in Chapter , it is not possible to do a close analysis of each text. Do different documentary genres – land leases, personal letters and contracts – generate different sorts of affiliation networks? Does combining evidence from different genres create a distorted affiliation network? Future network analyses of evidence from the ancient world will want to investigate these questions more closely. To my knowledge, the most recent scholarship on turning two-mode networks into single-mode networks has been that of de Nooy et al., in their  book on the Pajek computer program. Their most important contribution is the attention they draw to the “size of an event” and the “rate of participation” of an actor. In our terms, event-size is equivalent to the number of people or places appearing in a text, while participation rate is the number of separate texts in which a person or place appears. These factors can create considerable distortion in attempts to analyze certain features of a network. In my discussion of Oxyrhynchite topographical and Aphrodito social networks, I will point out several such examples. Brief reference to one here will suffice. The text originally published as P.Flor. ., now part of Constantin Zuckerman’s edition of a larger Aphrodito fiscal register, includes people we know from other Aphrodito texts, and thus naturally begs to be included in a graph of that society. Yet, 

de Nooy et al. , . My thanks to de Nooy for providing an advanced copy of this chapter in .

Introduction



the text includes some  names, ten times more than some of the next largest texts from Aphrodito. Here, we have an example of an exceptionally large “event size” – the event in question in this text is simply the act of tax-paying. The Aphrodito fiscal register is nothing but a payment list, which hardly implies social connectivity between the people mentioned therein. But when it is included in network analyses, the people in it appear at the top of our connectivity lists. We should take one lesson away from this: when gathering data from two-mode graphs and converting them into one-mode graphs, we must be on guard for disproportionately large events, which turn into disproportionately large cliques and create disproportionately centralized individuals. We should also consider other sorts of networks available to network analysts, if only to acknowledge some analyses not employed in this work, but available for consideration in the future. Two important analytical categories here are the effective network and the extended network. In the words of an early network theorist, Mary Noble, an effective network consists of “those people known to ego who also are known to each other.” In late antique Egyptian terms, we may expect Dioskoros and Apollos to be a part of Phoibammon’s effective network. But the men and women Dioskoros met during his trip to Constantinople will not be; rather, they are part of his extended network. As Mary Noble put it, an extended network consists of “people who are known to ego but not to other members of ego’s network and who in turn know other people.” Greater in their potential for future papyrological research are the concepts of the signed graph and the directed graph. The absence of both types of graphs is a shortcoming of this study. Directed graphs indicate whether or not a given relationship is reciprocal. For example, a directed graph analyzing revenue collection on the Apionic estates would trace a connection from a tenant farmer to an estate prono¯et¯es (caretaker) to one of the estate’s central managers. In visual terms, these connections can be visualized as one-way arrows in which the central manager would not have a return path or connection to the tenant farmer. Equally, directed graphs can indicate status differential or lines of patronage. For instance, the social network of an Aurelius Phib might have an arrow pointing towards his landlord, Flavius Apollos, to indicate his social subordination; Apollos’ network would have no such reciprocal arrow. I have not created directed versions of the data-sets used in this work: too many of the texts in question  

Mary Noble, in Boissevain and Mitchell , . Wasserman and Faust , –.



Boissevain and Mitchell , .



Social networks in Byzantine Egypt

mention too many pairs of individuals whose status relations are unclear. Later, more focused work on ancient networks could benefit from using directed graphs. A signed graph is merely a graph in which positive or negative values are attached to the links between each node. In other words, a signed graph of the social networks of Aphrodito would include a “+” or “−” value to indicate whether, for example, Dioskoros liked or disliked Flavius Ammonios, the local luminary to whom we know he was connected. To some extent, we must reconcile ourselves to this weakness: the surviving sources simply do not give this sort of information. (Nor should we necessarily let modern emotional constructs intrude on analysis of ancient networks.) Still, using this sort of information to construct signed graphs can be a powerful tool, and may be possible in some future, smaller test cases. Social psychologists and sociologists have used this sort of graph to test for “balance,” to check, in other words, whether networks contain groups of relationships largely pleasing to the actors, or if the networks contain groups riddled with interpersonal tensions. Despite the absence in this study of close attention to signed and directed graphs, Chapters  and  herein argue implicitly for attention to a comparable concept, the valued graph. A valued graph is simply a network “in which the strength or intensity of each tie is recorded.” Valued graphs are thus mere mathematical abstractions for what will really concern us in Chapter , the extent to which social connections between any two people in a village such as Aphrodito are multiplex, comprising several distinct but mutually reinforcing roles. As Jeremy Boissevain wrote in an early discussion of multiplexity, people are “sometimes in touch with the same people in different capacities. A person may know his brother as a neighbour, member of the same religious association, fellow employee, and supporter of the same political party.” The discussion of Aphrodito society in Chapter  will focus on the value of Aphrodito’s social ties. Put in a less abstract fashion, Chapter  argues that the multiplexity or multi-stranded nature of Aphrodito’s social connections explains many of the transactions and conflicts we see in the papyrological record.      

For Ammonios and Dioskoros, see Chapter  below, pp. –. See generally Wasserman and Faust, Chapter , and their discussion of structural balance and transitivity. For a brief introduction to valued graphs, see Wasserman and Faust , –. Wasserman and Faust , –. Boissevain , ; for multiplexity generally, see –, with discussion throughout the volume: consult the index, s.v. multiplexity. See particularly the introductory and concluding remarks at Chapter  below, pp.  and .

Introduction



The only network feature to have received real attention in the historiography of ancient social networks is density, a concept crucial to Elizabeth Clark’s discussion of the Origenist controversy. Rudo Niemeijer has provided an excellent if relatively early treatment of the concept of density. Density is a measure of the ratio of a network’s actual connections to its possible connections. As Niemeijer’s threefold definition puts it: density = the percentage of theoretically possible links actualized = the percentage of the maximal possible degree of a network actualized = the probability of a member chosen at random having a relation with another member of the network similarly chosen at random. The larger the size of the network, the more additional connections are required to increase its density. Some network theorists have drawn attention to weaknesses in density as a descriptive tool. Networks can have high-density “knots,” particularly likely where the same type of link exists between a group of people, which can skew the average density of the network as a whole. (See pp. – below on “cliques” and “clusters.”) Despite these limitations, density provides a generic tool for determining the likelihood that any given pair of names knew each other. To some extent, we can already make this determination at the elite levels. Any two Flavii from the contemporary sixth-century Oxyrhynchite accounts we discuss in Chapter  below probably had some acquaintance with each other, no matter how fleeting. But a more rigorous method is necessary for the lower classes, who were smaller fish in larger pools. For instance, without constructing a series of hypothetical chains of connectivity, to what degree can we guess that (say) Phib and Elizabeth, two debtors from P.Oxy. ., c. ad, knew Apphous, who appears in P.Oxy. ., two years earlier? Niemeijer points out that measuring the probability that any given node is connected to any other given node in a network is essentially the same as measuring that network’s density. If we could arrive at a satisfactory measure of the social density of late antique Oxyrhynchos, we would know how likely it is that Phib and Apphous knew each other. I will argue in Chapter  below that such results are possible for Aphrodito,     

Niemeijer . For another discussion and definition of “density” as a concept in network analysis, see Barnes , –. Niemeijer , . For a concrete example (Barnes , ) evaluated to make this point, see Niemeijer , . Cubitt , – in Boissevain and Mitchell . Niemeijer , .



Social networks in Byzantine Egypt

if not for Oxyrhynchos. Aphrodito is a reasonable consolation prize: measurement of its social density can let us debate whether its numbers approximate those we might find elsewhere in Egypt. Other structural characteristics of social networks can be applied by extension to the characteristics of specific individuals within these networks. First, centrality. Centrality is a crucial test I employ in our discussion of the Aphrodito social network. A “selected unit is central if it has [a] high degree, if it is easily accessible (close to) all other units, [or] if it lies on several geodesics (shortest paths) between other units.” So the actor of highest “degree centrality” is the person with the most links, or direct connections, to other actors. In Aphrodito, a villager can become degree central by appearing in many texts, or by appearing in texts with many other people, or some combination of both. Other centrality measures figure prominently in my discussion of Aphrodito’s social networks. The actor of highest “closeness centrality” is the person with the shortest paths to all the other actors. In other words, an Aphrodito villager will be closeness central if he has a lower average number of social steps between him and everyone else in the village. Measuring closeness centrality is generally more useful than measuring degree centrality; unlike degree centrality, closeness centrality calculates both direct and indirect links in the same measure. “Betweenness centrality” measures who sits on the most routes between two actors; this is considered a good way to find out who influences information flow. (Other measures of centrality, such as prestige, can only be calculated for directed graphs, discussed above. These sorts of graphs will be absent from our analysis, thus limiting the centrality measures we can use. ) A network is highly centralized if one actor or a small number of actors have considerably higher centrality than other actors in the network. This sort of measurement can clarify the hierarchical nature of a society. If late antique Egypt is as hierarchical and status-conscious as much of the traditional historiography has maintained, we would expect a highly centralized network, with some actors exhibiting a hyper-centrality characteristic of the elite, and most others appearing much less central. On the other hand,     

See Chapter  below, pp. –. For recent discussions of centrality, see de Nooy et al. , Chapter , Degenne and Fors´e , Chapter  and Wasserman and Faust , Chapter . It is precisely this weight given to large texts that permits the distorting effects described in Chapter  below, at pp. –, and countered by removal of the texts in question. See discussions of both measures at de Nooy et al. , –. For other centrality measures, see n.  above.

Introduction



if our understanding of late antique society has been wrong, and quantitative proof that late antique Egypt was in fact remarkably decentralized were to emerge, that proof would most likely come in the form of a diffuse social network in which every actor was more or less equally central. (As the chapters on Aphrodito will show, the curiously low levels of network centrality found there suggest a relatively low degree of social hierarchy, when compared for instance to data-sets from Renaissance Italy widely used as benchmarks among network analysts. ) Perhaps better known in popular parlance than centrality is the concept of distance. Distance is nothing more complicated than the number of steps it takes to go from Point A to Point B in a network. In social terms, if X knows Y and Y knows Z, a distance of two steps exists between X and Z. The notion that “six degrees of separation” exist between everyone on this planet refers to distance, and asserts that the average distance between any given pair of humans is six. Calculating the degree of separation between a node and the rest of the network – a technique called tracing in earlier literature – does no more than reveal how many steps are necessary for a specific member of the network to connect to everyone else, and how many members of the network are reached with each step. Averaging the results for each member of the network allows us to calculate the average “degrees of separation” or distance that exists between each member of the network. Tracing the social distance (or “degrees of separation”) between two members of a network can show a network’s level of interconnectivity. To use an example I have already employed elsewhere, in the context of Alexandria’s pagan philosophical schools, it can tell us which of the pagan scholars was better connected to the members of his peer group, and thus help explain factional struggles within those schools. Calculating degrees of separation can also help us determine whether Dioskoros of Aphrodito was more socially connected than any one of the hundreds of people the evidence links to him. Given the prominence accorded to Dioskoros by the survival of his own archives, we would expect the evidence to accord him a higher level of connectivity than anyone chosen more or less at random from the records of sixth-century Aphrodito. If calculating social distance shows us that Aurelius Mathias son of Iosephios (hardly a household name among students of Aphrodito ) has an average distance comparable to    

See Chapter  below, p.  with n. . Fararo and Sunshine . Ruffini . Girgis ; P.Cair.Masp. ., P.Cair.Masp. .. (All citations to Girgis herein refer to the number of the individual’s entry in Girgis .)



Social networks in Byzantine Egypt

that of Dioskoros and his equally prominent peers, then we have learned something surprising and new. If calculating social distance reveals wide gaps in levels of connectivity, then we are dealing with a relatively more hierarchical society, and a less dense network. If the gaps are narrow, or non-existent, then network density is high. Remembering that our evidence only shows the minimum connectivity apparent in our texts, we keep in mind that network distance will always be shorter than any results we achieve here. I have already used the term “degree” in a non-technical sense. More technically, network analysts distinguish between the degree of a network and the degree of a person. A network’s degree is the average number of relations between each member of the network. A person’s degree is “the number of relations he has within the network.” Thus, when we say that the Aphrodito network is a network of degree eight, we mean that on average each member of that network has ties to eight other members. But when we say, for instance, that Dioskoros of Aphrodito connects to thirty other people in the Aphrodito prosopography, we are referring to his personal first degree. Thus, when someone employs the clich´e that everyone in the world is separated by “only six degrees,” they are not using the term “degree” in this technical sense, but are referring to distance. One final way to analyze an individual’s place in a larger network is the concept of structural equivalence. According to Burt’s formulation, Structurally equivalent people occupy the same position in social structure, and so are proximate, to the extent that they have the same pattern of relations with occupants of other positions. More specifically, two people are structurally equivalent to the extent that they have identical relations with all other individuals in the study population.

Implicit in this notion of structural equivalence is the presence of larger units within any given network, units such as the clique and the cluster. First, the clique: Wasserman and Faust describe a clique as a subset of nodes, all adjacent to each other, with no other nodes adjacent to all the clique members. Put another way, a clique is “a set of persons with a density of  per cent.” Identification of cliques within social networks such as the Aphrodito data-set discussed in Chapter  is potentially full of difficulty. The average family, in which all members know each other, but not all members would know, say, the father’s tenant-farmers, will  

 Niemeijer , . Niemeijer , –.  Niemeijer , –. Kochen , .

Introduction



appear as a clique in such an analysis. More troublesome is the fact that affiliation networks (defined above, page ) create artificial cliques simply through their use of individual papyri as the standard for a social event. If we treat a papyrus naming ten people as creating an implicit social link between those ten people, those ten people form their own clique by definition. Yet this clique may have no validity outside of the realm of the text. A particularly striking example of this problem, Aphrodito’s fiscal register, is discussed above (in the discussion of “modes” at page ) and in Chapter  on Aphrodito. Another crucial network building block is the cluster. Definitions of a cluster came as early as the s. Niemeijer, writing in the early s, refined an earlier definition by Barnes by describing a cluster as “a set of persons that have a higher personal degree with other set members than with non-members.” Clusters may be the result of political divisions within the network as a whole, or other analogous factors. A more advanced technique for studying a network’s clustering is to measure its clustering coefficient. The clustering coefficient of a particular actor in a network is a measure of how many of that actor’s connections are in turn connected to one another. The clustering coefficient of the network as a whole is in turn a measure of the composite or average of each actor’s clustering coefficient. In general, a network with a high clustering coefficient is a network in which one person’s connections are also connections for others. The formal definition given here is not easy to visualize. Put more casually, the clustering coefficient is a measure of the overlap between social circles. To some degree or another, one’s friends and social connections are likely to know one another. The clustering coefficient of a network measures this phenomenon on a scale of  to . The higher a village’s clustering coefficient, the more likely it is that the men and women directly connected to any given villager are also directly connected to each other. But a high clustering coefficient does not mean that everyone in a network would know everyone else. Rather, it indicates close connectivity  



Niemeijer , . For the most lay-accessible discussion of the clustering coefficient and its import in discovering the nature of so-called “small-world networks,” see Watts . The concept is relatively new, missing from Wasserman and Faust, but appearing in Barab´asi, who largely attributes it to Watts and his circle. The standard piece of technical literature on the subject is Watts a. For a more technical treatment of certain aspects of clustering coefficients, see Newman et al., . See Watts a for the importance of this technical limitation. In his own terms, these concepts only have meaningful applicability in cases where k  n: where the average degree of the network is much less than the number of nodes in the network. And this is a reasonable situation for our



Social networks in Byzantine Egypt

between clusters, and thus measures the importance of membership in such clusters. If clustering measures how a network is put together, cohesion measures how to take it apart. Only recently have theorists made a rigorous attempt to define the cohesion of a network. As Moody and White put it, a “group’s structural cohesion is equal to the minimum number of actors who, if removed from the group, would disconnect [i.e. fragment] the group.” These actors central to the creation of network cohesion are the cutpoints, defined below. In less rigorous forms, cohesion has been put forward as an index of a society’s democratic capacity, the “extent to which a group depends on particular individuals to retain its character as a group,” and a network’s ability to transmit information over distance without significant information degradation. In each case, the higher the number of actors whose removal is needed to disconnect the group, the greater that network’s cohesion, and thus the greater its strength and resistance to fragmentation. A cutpoint is a node crucial to the cohesion of a graph. It functions as a mediating link between two separate sections of a network. Removal of a cutpoint can result in the fragmentation of a network into smaller, disconnected networks. A cutset is a collection of cutpoints the removal of which would break the network into two or more pieces. For instance, in Chapter  on Aphrodito, analysis of that village’s prosopography will identify some twenty-eight cutpoints whose removal from the database destroys that network’s social connectivity. Thus, these twenty-eight people warrant special attention for the nature of their links to various parts of the Aphrodito community. According to Moody and White () there is a direct connection between the number of cutpoints in a network and that network’s cohesion (on which, see above). Networks with a larger cutset – in other words, networks which fragment only with the removal of higher numbers of cutpoints – are considered less “vulnerable to the will and



    

purposes: any given person in Aphrodito will have meaningful social interactions with only a small (if significant) percentage of the total population. See Moody and White  for the latest word on the subject. For an earlier treatment and brief historiography of the concept of cohesion, see Chapter Four in Degenne and Fors´e . Their terminology is somewhat different from that which has developed in the English language: cf. their “firewalls” and the more standard English-language “cutpoints,” discussed here, p. . Moody and White , . Moody and White , passim, the quote coming from . For a more technical definition, see Wasserman and Faust , . Moody and White , . To identify cutpoints in UCINET, use the menu option Network→Regions→Bi-Components; the cutpoints are indicated by the number  at the end of the resulting output file.

Introduction



activities of those who can destroy the group by leaving,” and are less likely to manifest in the hands of an individual or small group the “control of resource flows [which] generate power inequality.” By now it is no doubt obvious that thorough exploration of any data-set with these network features in mind all but requires computer assistance. For this study, I have turned to three programs, UCINET, Pajek, and NetDraw. UCINET is a standard software package used by network analysts to answer all sorts of questions about the structure of a given network. The introduction of computer technology has had a dramatic impact on the field of social network analysis. It has, most importantly, made network analytic tools available to researchers otherwise lacking the necessary mathematical know-how. I count myself in this category. The discussions on “density” and “distance” above pose technical questions which would be impossible to answer without a computer program such as UCINET. Pajek, which comes packaged with UCINET, is harder to use, but exceptionally powerful; its ability to remove links between and within large groups of a network is crucial to my analysis of the Apionic estates in Chapter  below. Also packaged with UCINET is a network visualization program called NetDraw, used to generate the visual representations of the Oxyrhynchite topographical networks in Chapter . A number of caveats are in order about these charts. First, before including NetDraw visualizations in these chapters, I cleaned them by removing isolates (people or places not linked to the network) and isolated groups, and by removing pendants (people or places linked to the network by only one degree). Second, while these graphs do accurately map the distance between each actor in the network in mathematical terms – degrees of separation and centrality, for instance – they do not necessarily represent meaningful physical distance. In the case of Oxyrhynchos, for instance, I use UCINET to generate network graphs of the nome’s topographical register, in hopes of learning more about the connections between the various villages. The resulting chart is not a reliable guide to the physical topography of the nome. Nonetheless, these visualizations remain useful tools: NetDraw’s Oxyrhynchite graphic is, for     

 Borgatti et al. . Moody and White , . For more on Pajek, see de Nooy et al. . UCINET version . comes packaged with NetDraw, a program designed to import UCINET data and produce visual graphs from it. See Chapter  below, pp. –. Although see now the recent work on the topography of the Fayum using just such visualization techniques in Mueller a, with a rebuttal in Hoffman and Klin .



Social networks in Byzantine Egypt

instance, what first led me to investigate whether the Apionic archive itself was distorting the Oxyrhynchite data, a question I answer in the affirmative in Chapter . With these tools in mind, let us turn to the sites themselves. 

Freeman  provides a thorough historiography of network visualization.

chapter 1

The centralized elite of Oxyrhynchos

introduction As they appear in the surviving evidence, the Oxyrhynchite nome’s social networks in the Byzantine period were highly centralized and hierarchical. The chains of social connectivity were generally vertical in nature, formed by unequal social transactions. This chapter’s prosopographical study of the Oxyrhynchite elite uncovers a curious phenomenon. We can rarely prove direct links between members of the Oxyrhynchite elite; however, we can often demonstrate indirect links, in which the local office-holders and landowners hired the same scribes, had contact with the same church officials, and so forth. In network terms, the consequences of this sort of connectivity are clear: rather than documenting a web of mingled horizontal and vertical ties similar to what we see in Aphrodito in Chapters  and , our Oxyrhynchite evidence maps a series of social hubs, in which the landholders are the social centers, and their service personnel are the spokes of the wheel, extending the landholders’ social reach throughout the nome. This pattern holds true not only for the elite families who spanned the mid-fifth to the mid-sixth centuries ad, but also for a number of high-profile examples late in the sixth century. Despite the vast amount of papyrological evidence available, the social history of the Oxyrhynchite nome has long resisted a synthetic book-length treatment. The close prosopographical analysis of the large landholders 

Bowman et al., Oxyrhynchus: A City and its Texts, is a forthcoming collection of essays on various topics. MacLennan’s  Oxyrhynchus has not aged well, and presents a city no longer recognizable in light of the revolution in historiographical attitudes towards late antiquity; for example, he more than once describes Egyptian Christianity as “degenerated” (). Rowlandson’s  Landowners and Tenants in Roman Egypt deals with a specific aspect of Oxyrhynchite society, the social nature of the agrarian regime in the principate. For a fascinating look at Oxyrhynchite topography, including an early attempt to quantify the number of connections between various sites, see Kr¨uger , which nonetheless has not replaced Pruneti’s indispensable topographic register, used extensively in this and the following chapter. See also I.F. Fikhman’s  Oxyrhynchus, City of the Papyri, in Russian with a





Social networks in Byzantine Egypt

and other nome elite I provide here fills a small part of that gap. Papyrologists naturally compensate for the lack of a synthetic treatment by writing micro-histories focused on the archives or dossiers of the nome’s leading landholders. The family of Flavius Eulogios has received such treatment, as have Flavia Anastasia and Flavia Christodote, although given their interest as high-profile women, we do not know nearly as much about them as we would like. But there has been no attempt to connect each prominent individual to a larger whole or to explore the social structures those connections imply. Of course, documenting social ties between members of the Oxyrhynchite elite would be easier if we had evidence for their work in a single institution, such as a civic council. But no such record for this period has yet been published. (P.Oxy. ., a city councilman’s account, is something different, but will appear frequently in these pages.) One recent work includes a list of eleven attested politeuomenoi or city councilmen from Oxyrhynchos dating from  to . Over nine decades, this is not an impressive haul, not enough to put any of the eleven in the same place at the same time. So, a methodological concern emerges from this discussion as well: in many cases, we can demonstrate connections between the elite only via institutions documenting those elites. One might object that such institutional mediation explains why this chapter documents indirect connections, through priests, scribes, and other figures outside of the landholding elite: these intermediate links were not the glue holding the Oxyrhynchite social world together, but they may appear so because of the nature of the surviving evidence. This chapter will therefore conclude with a brief methodological excursus addressing these objections, and affirming the social centralization apparent in our evidence. The nome-wide vertical ties in this chapter are unique, exclusive,

  



French introduction, a full-length treatment of Oxyrhynchos from the fourth to the seventh century. Otherwise, the reader must fall back on over a century of learned articles, starting with Grenfell /. For the Apionic estates, which may well do double-duty for many aspects of sixth-century Oxyrhynchos, see Hickey , Mazza  and Sarris , as well as Hardy , which still remains quite useful. The most extensive bibliographical reference guide to the city itself is the Dizionario of Calderini and Daris, now in its third supplement (): its section on ìOxurÅgcwn p»liv (–) is exhaustive, and can almost substitute for a brief history of the city through the entire Greco-Roman period. See below, pp. –, and n. . Christodote: see below, pp. –, and n. . Anastasia: see below, pp. –, and n. . For evidence from an earlier period, see the references to bouleutic documents compiled in Appendix . of Bowman , and see e.g. the series of reports from late third-century meeting proceedings published in P.Oxy. .–. Laniado , , to which now add the evidence of P.Oxy. , discussed below at p. .

1 The centralized elite of Oxyrhynchos



and not merely illusions based on the nature of the surviving evidence. Future discovery of missing evidence – an Aphrodito-like archive of an Oxyrhynchite village, for instance – will not alter the picture presented herein. the large estates The large estates of Oxyrhynchos were the central hubs of social networks spreading throughout the nome. Although their emergence is still shrouded in obscurity, the fifth-century evidence shows these networks already in place, and they continue unchanged throughout the sixth century. One recent study accurately describes these estates as a microcosm of the world at large: “Above a huge army of subordinate staff were supervisors, stewards, secretaries, stenographers and other clerical staff, senior administrators . . . coin-weighers, bankers, treasurers, storekeepers, rent-collectors, armed personnel, etc.” Despite their importance, estates of this size are more likely to have been the exception than the rule. Even in the Oxyrhynchite nome, where the evidence is undeniable, Gascou has argued that oikoi like those of the Apions should be thought of merely “comme une abstraction comptable, une unite d’assignation ‘fiscalo-liturgique.’” 



 



Thanks in part to the work of Edward Hardy, it was for many years standard to think of late antique Egypt as a world composed solely of large estates. If this characterization is true anywhere in Egypt, it is true in the Oxyrhynchite nome. Keenan has noted that the evidence for the –nap»grafoi gewrgo©, the coloni who stand in for the serfs in the old historiographical model of Egypt as a semifeudal state, comes only from the Oxyrhynchite nome: Keenan b, – n. . P.Mich.inv.  (; Oxyrhynchite in provenance, but referring to two –nap»grafoi gewrgo© from the Herakleopolite, which had close ties to the Oxyrhynchite), provides Fikhman ,  evidence for a more nuanced view, that using enapographos “´etait une caract´eristique de la pratique notariale ´ d’Oxyrhynchus et n’a rien a` voir avec le degr´e de diffusion de l’adscripticiat en Egypte.” For an early statement of the enapographoi ge¯orgoi as feudal serfs, see Hardy , , where the association of those ge¯orgoi with coloni, and the assertion that the “fundamental fact” about the latter “was that they were bound to the soil” are taken almost for granted. For the papyrological silence covering part of the period, see Bagnall and Worp b, and Habermann ; the first treatment was R´emondon b, including Table  for a striking visual reminder of the plummet in the number of attested papyri, from which the author concludes that the decline in documentation () “n’est pas s´eparable de l’´evolution politique, e´conomique, social de l’Egypte a` cette e´poque.” Banaji , . Considering whether late antique Egypt was indeed a world of large estates, Liebeschuetz recently wrote that the “evidence available today seems to be against it”: Liebeschuetz , . Sarris  represents a dissenting view, but see my critique of his search for large estates in Ruffini forthcoming c. Gascou , .



Social networks in Byzantine Egypt

Nevertheless, these book-keeping abstractions had some tangible reality on the ground, in social settings. Powerful landholders managed their estates and their fiscal-liturgical obligations through processes that were inherently social in nature. Thus, mapping connections between members of the local elite and their respective estates can illuminate the social structure of the Oxyrhynchite nome more generally, and help trace those processes from the beginning of the period through its end. The landowners I discuss in this section – Kyria, Alexander, Phib, Anastasia and Apion – share a single feature: their urban activities connect them to rural figures whose village activities in turn rely on the city for financial or institutional support. Flavia Kyria is a fifth-century landowner in whose dossier the characteristics of the sixth century are already clear. She appears in a short order for payment to some donkey drivers in  ad, but her appearance a year earlier in a receipt for part of a water-wheel is more interesting. The text, addressing her as a “most illustrious and most noble landowner,” concerns two registered farmers (–nap»grafoi gewrgo©) named Aurelius Pasoerios son of Kornelios, and Aurelius Ioannes son of Phoibammon from a hamlet named Chaira, a holding belonging to Kyria (lines –: ˆp¼ –poik©ou CairŽ ktžmatov t¦v | s¦v {aumasi»thtov [sc. Kyria]). These farmers needed an axle for the water-wheel named Pso, receipt of which they acknowledge in this text. The high degree of social and economic centralization this text suggests is striking. The resources for replacing the axle were lacking in Chaira. Pasoerios and Ioannes had to go “up to the city” (line : ˆnel{»ntev –pª t¦v p»lewv) of Oxyrhynchos, and find someone there – surely not Kyria herself, but an unnamed middleman – who could secure the replacement 

  



For Kyria, see P.Oxy. . and Gascou a, –, adducing P.Lond. .. Gascou was the first to recognize the existence of a Kyria dossier. See also Fikhman b, ; Banaji ,  (citing Fikhman); and most recently Gonis . See the previous note for Gascou’s identification of Kyria in P.Lond. .. Bell was confused by the reading, and thought that “kur©a does not seem likely to be a name.” P.Oxy. .. Water-wheels, naming them, and getting axles for them: see also P.Oxy. . (), P.Oxy. . (), P.Oxy. . (), P.Oxy. . (), and P.Oxy. . (). These are almost all Apionic; see Gonis b, –. For the most recent published examples of irrigation machine-part receipts, see P.Oxy. . (), P.Oxy. . (), Tacoma  ( = SB .), and P.Oxy. ., with additions to the list in Tacoma, many within that volume itself. See also Mirecki  and P.Col. .’s survey of Byzantine Egyptian axles, the last word on the subject. For irrigation of the large estates more generally, see Bonneau . For which, see Pruneti s.n., citing only this text and SB .. For the latter text, and Chaira’s proposed proximity to Pela, see below, p. , n. .

1 The centralized elite of Oxyrhynchos



for them. Modern scholars have seen this document as a crucial turningpoint, the first text before the Apions to attest to the large landowners, the geouchountes. It also shows us in the fifth century the –nap»grafoi gewrgo© so often described as the serfs of the sixth century. As one author recently put it, “After the slow incubation of the previous decades, owners like Flavia Kyria . . . represented the formation of a genuine provincial aristocracy.” If this text is a turning-point, marking the emergence of a new aristocracy, the long vertical ties we see throughout the sixth century connecting city aristocracy to village peasant are already in place in the s. Ties of this nature are apparent elsewhere in the Kyria dossier. In a text dating to  ad, a boatman of Kyria’s (lines –: ìIw†nnou] n[a]Åtou | t¦v Kur©.a. v lam[prot]†th) received through an assistant named Aphouas payment from the komarchs of Sinaru for transportation of the tax grain to Alexandria. Ioannes is not the only boatman of Kyria’s to survive in the record. In his discussion of Flavia Kyria, Gascou identified two texts dating to / with previously unnoticed connections to Kyria. In the first text, a boat-owner (naÅklhrov) named Phib son of Menas instructed Kyria’s prono¯et¯es or estate manager Apphous to pay five artabae of wheat to Didymos the notarios in exchange for a wine shipment. The second text records a similar exchange between Phib and Apphous, who is this time addressed as prono¯et¯es of the well-known village of Pela. Phib again asked Apphous to make a payment of wheat, this time to Martyrios the assistant. Didymos the notarios and Martyrios the assistant both presumably rendered some service to the estate of Flavia Kyria for which Phib the shipowner acted  



 



An assessment shared by Gascou a,  and Fikhman b, , followed by Banaji , . The LSJ takes –nap»grafov as “registered, esp. of cultivators or serfs.” In P.Oxy. , Grenfell and Hunt rendered it as “labourer” or “cultivator.” For a proposed attestation of the enapographoi ge¯orgoi via restoration from a text dating earlier than P.Oxy. ., see Banaji , . For Hardy’s view of the enapographoi, and recent correctives, see n.  above. Banaji , ; see generally his Chapter , building first upon R´emondon’s term, “les hauts fonctionnaires latifondiaires” (cited at Banaji , ) and discussing an office-holding elite, typically Oxyrhynchite, whose members were actively involved in the running of their own estates, and developed a “characteristic Byzantine habit . . . of emphasizing the subordination of the peasantry” (Banaji , ). SB .; recognized as part of the dossier in Gonis , –, with the text dated at p. . P.Oxy. . and . These texts had been assigned to the “early sixth century” in the minor documents section of the Oxyrhynchus papyri. In P.Oxy. ., as in P.Lond. ., the kur©av lamprot†thv in line  had not been recognized as a name. See Gascou a, . Pela, in the Libos toparchy: see Pruneti s.n., with attestations from the Ptolemaic period through the sixth century, with the plurality coming from the third century. Pela and Kyria’s hamlet of Chaira may have been near each other: both places appear in SB ., a fifth- or sixth-century list of place names, but there is no obvious connection between them, or any of the other places listed in that text.



Social networks in Byzantine Egypt

as a middleman, ensuring that Apphous would pay them both. These texts show Kyria’s estate at work through lower-level functionaries, sometimes outside of the city itself. The boatmen provide a possible connection between Flavia Kyria and the family of Flavius Apion. Hardy long ago suggested that Phib the naukl¯eros may be one of the earliest representatives of four or five known generations of boatmen from the same family. Let us start with the later generations and work backwards. An acknowledgement of an advance addressed to Flavius Apion in  by the council of the headmen (line : t¼ koin¼n tän prwtokwmhtän) of Takona mentions the heirs of a shipowner named Menas son of Asklas, who was captain of a boat belonging to the Apionic estate. This may be the same shipowner named Menas from Koma still alive in P.Oxy. ., which must then predate . His father is presumably the shipowner Asklas from Koma who appears in  taking a shipment of wheat for the embol¯e to Alexandria. Kyria’s Phib, with whom we started this discussion, was himself son of a Menas; Menas son of Asklas may represent a later generation of the family, still carrying his name. If Hardy’s hypothetical reconstruction of this nautical family is on target, then they appear as boatmen in Pela, Koma, and Takona in successive generations. Pela is towards the southern end of the Oxyrhynchite nome, Takona towards the northern. Oxyrhynchite boatmen were not limited to work for one family or in one part of the nome, but probably shipped goods from all ends of the rivers and canals of the Oxyrhynchite, if not beyond. Like Kyria’s enapographoi ge¯orgoi, with whom this section began, the careers of these boatmen suggest a considerable level of centrality to the region’s economic ties. Whether they worked for Kyria, the Apions,  



 

This paragraph largely follows the connections Hardy  proposed at  nn.  and . P.Oxy. .. Koinon of pr¯otok¯om¯etai: this term is discussed in a more general context in Gascou b, –, particularly on  n. . Gascou notes () that “Les membres de ce koin»n se partageaient annuellement les fonctions et les responsabilit´es communales. Les autorit´es municipales et provinciales les consid´eraient comme les r´epondants du village.” I owe this reference to Todd Hickey. For a much earlier discussion of pr¯otok¯om¯etai more generally, see the commentary to P.Oxy. .. at –. P.Oxy. .. Hardy ,  with n.  also suggests that we see an earlier generation of the family in the Asklas appearing in P.Oxy. . and P.Oxy. .. Its editors describe the latter as an “illiterate letter from Asclas to Abonas concerning boats and other matters.” See Pruneti s.n. for references, particularly rich for Pela and Takona, two of the major centers of the nome. For a thorough discussion of Egyptian boatmen from earlier periods, see Vinson , particularly his discussion () of their status, in which he remarks that boat captains were “part of the middle class,” and their crewmen were “not without possibilities for advancement.” For the high status typically enjoyed by boat-owners in later periods, see the recent remarks by Gonis in his introduction to P.Oxy. ..

1 The centralized elite of Oxyrhynchos



or some other large landholder, in Pela, Takona, or somewhere else in the nome, the transactions were directed from and recorded in Oxyrhynchos itself, under the aegis of economic elites living in the city proper. Another of Flavia Kyria’s indirect connections is worth exploring to develop an analogous point. In his initial work on Kyria’s dossier, Gascou identified Phib, the naukl¯eros or boat-owner we have been discussing, with Phib the naut¯es or sailor in P.Oxy. .. That text, now dated to the early s, is a note from Makrobios’ bo¯ethos or assistant, Philoxenos, to the sailor Phib acknowledging his receipt of a military grain delivery. Philoxenos is a common name, but we have only one Philoxenos bo¯ethos from Oxyrhynchos in this immediate period, appearing twice over the next three decades. First, in , a steward of the holy church was ordered to pay four double jars of wine to a Philoxenos bo¯ethos of the exactor or tax-collector. Philoxenos appears as bo¯ethos of the exactor once again, before c.. Then, he received instructions from Flavius Prosdokios to re-register eighteen arourae of land in his name, relieving the nauarch Ioulianos of the responsibility. Philoxenos’ jurisdiction in that text is “the division and estate of Timagenes of illustrious memory.” (On these estates, see the following section. ) Prosdokios is hardly a common name. The Prosopography of the Later Roman Empire lists only one, an exact contemporary, a praetorian prefect and member of Justinian’s commission on the Digest, presumably not the  

 

  

Gascou a,  n. . March/April  by Gascou (see previous note), and originally assigned to the early sixth century; see BL . citing Reiter ,  for the correction to  April . This dating makes difficult, if not impossible, Timothy Teeter’s proposal that the Makrobios lamprotatos in P.Oxy. . is the same as the author of P.Col. . ( ad): see P.Col. , p.  n. . Nonetheless, the coincidence is striking. Could the man in the Columbia text have been the other man’s son or grandson? P.Oxy. .. P.Warr. . The original date of c. was based on references in both P.Warr.  and P.Oxy. . () to the estates of Timagenes, but those estates had considerable longevity and date back to a much earlier period: see below, pp. –. See Keenan a,  for a redating to before c., relying on references in both P.Warr.  and P.Oxy. . to Prosdokios, a member of the clibanarii or “cuirassiers” (certainly attested as such in the former, and possibly attested as such in the latter). P.Oxy. . dates to c. by virtue of the appearance therein of Flavius Hermeias, who also appears in P.Oxy. ., dating to . Prosdokios, alive in P.Warr. , is dead by P.Oxy. .. Keenan’s redating is correct, but with P.Oxy. . () in mind, surely P.Warr.  could date even earlier: Philoxenos the bo¯ethos of the exactor appears in both. (A bo¯ethos of Takona named Philoxenos appears in line  of P.Oxy. ., but this text is from the s, too late to be at issue here. For that text’s date, see below, n. .) For a discussion of this genre, see the introduction to P.Petra .–, with a table of examples at –. See below, pp. –. No others appear in Preisigke’s Namenbuch or through an online search of the Duke Databank of Documentary Papyri.



Social networks in Byzantine Egypt

same man. James Keenan is certainly right to identify this Prosdokios with the one mentioned as deceased in P.Oxy. ., from . In that latter text, Prosdokios is the father of a man named Petros, a member of the “devoted cuirassiers” who appeared before Flavius Hermias, advocate and defensor of Oxyrhynchos, to provide surety for a recently arrested man named Eustochios. An Ammonianus attached to the prefect’s office (line : t¦v ¡gemonik¦v t†xewv) joined Petros in providing surety. To summarize: Phib, the boat-owner we first encountered working for Flavia Kyria, received a grain delivery receipt from a Philoxenos bo¯ethos who may have been the same man re-registering land for Flavius Prosdokios. Prosdokios was probably a military man. His son, who certainly was, provided surety for men in trouble with local officials, and kept company with employees of the governor’s office. These connections, by no means certain, are nonetheless suggestive of the relatively small world of the Oxyrhynchite elite. They provide a miniature social map that included the estates of Flavia Kyria, the estate of Timagenes, members of the local military establishment and bureaucrats whose patronage protected lesser figures from the pressure of Oxyrhynchite government officials. Remaining for a moment in the fifth century, Flavius Alexander filled a social role comparable to that of Flavia Kyria in one key regard. Alexander was a powerful figure in the area, but one of several who, in Hardy’s words, only “come into view for a moment” in the s. Alexander held exceedingly high rank: ex-comes limitis Aegyptii et praefectus augustalis, magister militum, perhaps honorary, and on the receiving end of several entries in the Codex Justinianus. In the papyrological record, he appears only in , in a receipt for a water-wheel part from Aurelius Symphonias of the Piaa kt¯ema; the text also mentions Phoibammon the carpenter, and  

 

PLRE .. The ed. princ. of P.Oxy. . restored the lacuna in line  with tr[apezitän ?; thus the Maximos and Iulianos responsible for the libellus leading to the arrest appeared at page  of Calderini ’s list of banks and bankers. Bogaert ,  does not accept the restoration, thinking peribleptos an inappropriate honorific for a banker, and proposes instead reading tr[akteutän: “fonctionnaires qui appartenaient a` l’officium du pr´efet et qui avaient pour charge de faire rentrer r´eguli`erement toutes les redevances et les arri´er´es de la caisse du pr´efet.” Hardy , . See Fl. Alexander  at PLRE ., with CJ .., .., .. and P.Oxy. .. Banaji ,  does not think the title honorary, calling him “the highest military official for the eastern command, and of a status clearly more exalted than that of” Flavius Apion Theodosios Ioannes, on whom see below, p. . The PLRE entry follows CSBE  in dating the Oxyrhynchos papyrus to  (see BL .). Rea has since presented an argument (see BL .) in P.Rainer Cent. , n. to lines –, favoring the  date of the ed. princ. CSBE has accepted this view ( n. ) and the PLRE entry should thus be changed as well.

1 The centralized elite of Oxyrhynchos



Flavius Ioseph, the property manager of Alexander’s Oxyrhynchite urban holdings (lines –: lam|prot†tou dioikhtoÓ [p]ragm†twn toÓ aÉtoÓ –ndoxo|t†tou ˆndr¼v [sc. Alexander] diakeim”nwn kat‡ tŸn ìOxurugcitän). The implication that Alexander had holdings elsewhere perhaps accounts for his failure to leave a larger trace in the Oxyrhynchite record. It may also explain why his property shows no family continuity: by the second half of the sixth century, Piaa appears as part of the Apionic estates. We have no way of knowing when in the intervening decades the Apions gained control over Piaa, but the connection alerts us to the possibility that large estates in the Oxyrhynchite could expand and contract through transactions between social equals at the highest levels. Further, Flavius Alexander’s sole appearance in the papyri, providing replacement machine parts through several degrees of social separation, is reminiscent of Flavia Kyria’s first appearance with her enapographoi ge¯orgoi. Here too, the suggestion is of considerable centralization in Byzantine Oxyrhynchos, even in the fifth century; as with Kyria, those with ties to Alexander must “come up to the city” (in P.Oxy. ., line : ˆnel{[Ü]n –pª t¦v p»lewv) in order to take care of basic agricultural necessities down on the farm. An account of a city councilman (L»gov –] politeuom”(nou)), which records adaeratio and grain payments from the s, provides clues of a similar degree of centralization at a much later date. One of the payers in that text, Phib son of Matrinos, is surely the same Phib son of Matrinos who appears in a receipt issued in  “either by or to a comes . . . on account of t¦v {e©av dwreŽv (a ‘benevolence’ [perhaps] analogous to the earlier aurum coronarium?).” Phib’s payment at line  of that latter text was through a lamprotatos named Serenos, and the receipt drawn up by Ieremias the notary. At the foot of the same receipt, but written in the opposite direction, we find an account of money payments and gifts to Ioannes the baker for the people of Phthochis, payments to Pamouthios the pork-butcher, and payments to Sergios the horse-breeder and to others,  



 

P.Oxy. .. See P.Oxy. ., a sixth-century undated text typically associated with P.Oxy. ., a similar Apionic document dating from /. The latter text has been re-edited in Mazza , now SB .. P.Oxy. .; for the date of the text, see Gascou , . The ed. princ. dated the text only to the sixth century, and “of about the same date” as P.Oxy. ., in which several of the same contributors appear again. For the dating of the latter to the s, see Gascou b, . P.Oxy. ., following the editors in the text’s introduction. Ieremias: not in Diethart and Worp , nor apparently known from anywhere else. A Duke Databank search finds no other Ieremias notaries in the fifth- and sixth-century papyri.



Social networks in Byzantine Egypt

drawn up in a hand “rather similar . . . but not identical” to that of the text above it. These figures all defy identification, even the lamprotatos Serenos who paid for Phib, although in his case, some identification with the archive of Eulogios and a Serenos therein is tempting. These texts demonstrate an obvious point, that the Oxyrhynchite civic elite had indirect connections to the nome’s lower social strata, through the mediation of the officials handling their accounts. They also demonstrate a less obvious point, that these connections were all centralized in nature. Phthochis, in the nome’s eastern toparchy, was admittedly not too far from Oxyrhynchos. Even so, a city account recording gifts and payments for villagers through their baker suggests a considerable level of village ties to the urban center. SB . is another example of comparable ties in the late sixth century. The text is a receipt from Victor, an attendant at the public bath at Oxyrhynchos, acknowledging a shipment of lead received from count Phoibammon. Phoibammon was the dioik¯et¯es for Flavia Anastasia, whose appearance in the city councilman’s account we just discussed suggests that she was one of the region’s more prominent landholders. Nothing more is known about her dioik¯et¯es, although he may be the same count Phoibammon appearing in a sixth-century account of receipts and expenditures from three Oxyrhynchite villages. (Unpublished Anastasia papyri unavailable to me include at least two other texts concerning the public baths, and one other lead receipt, perhaps connected with the office of the logist¯es, or civic finance officer. ) An employee of Anastasia’s named Anastasios   

  

 See below, p. . P.Oxy. .. For Phthochis, a kÛmh, see Pruneti s.n., with the map in the Tavola. SB . = van Haelst , –, dating the text to / or / CE. For the latter date, see BL .. For Victor’s title, cf. line  and P.Oxy. .. Much of what follows on this document and others relating to lead and lead-workers in the next two paragraphs is based on Bagnall’s introductory remarks to P.Turner –. (Therein, he expresses a preference for the earlier of the two dates for SB .: see page , n. .) This text is comparable to other attestations of the Oxyrhynchos public bath in this period (for which see Kr¨uger , –): we get a glimpse when the services of the metal industry are needed for the maintenance of the bath by the civic elite. See P.Oxy. ., from , in which “Four ironsmiths acknowledge to a senator the receipt of money for expenditure on nails and other materials for a public bath,” in this case, not for the bath’s maintenance, but for its initial construction. For other intersections of the public bath with the civic elite, see below, pp. –. P.Oxy. .. For the Anastasia archive, see the full discussion below at pp. – with bibliography at n. . See also Anastasia’s entry in PLRE .. P.Oxy. ., so proposed at PLRE .; for more on this Phoibammon, see the discussion of Anastasia below, cited in the previous note. This text has not received any BL entries. Todd Hickey, personal communication, concerning the forthcoming P.Anastasia collection, on which see below, n. .

1 The centralized elite of Oxyrhynchos



delivered the lead, which itself originally came from a lead-worker named Pamouthios. We know this lead-worker from elsewhere. In , Aurelius Pamouthios “became surety to the heirs of Flavius Apion that Aurelius Abraham, a labourer, would remain with his wife and family on an estate [in Tarouthinos] belonging to the heirs.” This deed of surety was executed through Menas, the oiket¯es whose importance in the Apionic archive has long been known to modern scholars. To summarize, Flavia Anastasia paid for lead for the Oxyrhynchos public bath through a lead-worker from the city itself who stood as surety to an enapographos ge¯orgos on the Apionic estates in Tarouthinos. This chain of connectivity (Anastasia, Phoibammon, Pamouthios, Menas, the Apionic heirs) is the maximum case: Anastasia and the Apionic house were no doubt socially closer than this, and did not need lead-workers to connect them. But this case shows still further how large estates created centralizing economic ties between the nome’s urban capital and Oxyrhynchite rural laborers in places like Tarouthinos. With Pamouthios in mind, we should ask how many lead-workers Oxyrhynchos would have had. We know another one from the same decade, a man named Apollos. Apollos was issued five receipts for the supply of various amounts of tin and lead: for soldering a bath pipe, for restoring a cauldron, for caulking cooking equipment, and so forth. Roger Bagnall, who re-edited four of the five texts, was surely right to think the receipts came from the Apionic estates. The references to “our master the lord,” toÓ desp(»tou) ¡män toÓ kur. o. Ó, the Great House, t¦v meg†l(hv) o«k(©av), and the estate of Meskanounis, a known Apionic site, are convincing in conjunction. Bagnall asked whether Apollos the lead-worker was   

  

P.Oxy. ., identified by van Haelst , . Quoting from the introduction to P.Oxy. .. For this text as an exemplar of an innovation in dating only by regnal year, see CSBE . See also an Italian translation at Mazza , –. For the multiple men of the same name in the same position, see P.Oxy. .., n.; Hardy , –; Gascou ,  n. ; Mazza , –; and now P.Oxy. , publishing several documents with a Menas oiket¯es spanning over  years: see index VIII s.n. See Sarris ,  for a novel interpretation of these references, that “Menas was a legal fiction” who is “unlikely” to have actually existed. P.Turner - = P.Oxy. .-, where only partial texts were initially published, dating c., based on the date given in another receipt in the series, P.Oxy. .. See his remarks in P.Turner at –. Our master the lord: P.Turner .. The Great House: P.Turner .. Meskanounis: P.Turner . with n. . Hardy ,  cites P.Oxy. ., one of the receipts in this series, for reference to the Apionic proastion or suburban house. He also cites PSI . in connection with it; baths of the proastion appear at lines  and  therein. See p.  below for a discussion of Apionic bath personnel and their ties to the monastery of Abba Andreas: the Apionic estate “bought mats and ropes from the monks of Abba Andrew’s” (Hardy , ).



Social networks in Byzantine Egypt

“an independent tradesman or an employee of the estate of the Apions? The receipts suggest some formal independence, but there can be no doubt that the Apion estate was his principal customer.” If tradesmen in Oxyrhynchos formed koina or associations such as those well attested in Aphrodito to the south, then the likelihood that Pamouthios and Apollos knew each other is exceedingly high. These texts would then suggest another social connection between workers employed by large estates. They also represent another centralizing link between Oxyrhynchos (their find-spot) and a rural holding, in this case Meskanounis. Social ties between men in the same profession with economic connections to the same large estates are not surprising. More interesting are the men whose social functions connected different estates and men of different social strata. Notaries come to mind as a primary example. To get a sense of the social role of these notaries, consider Anastasios, who executed the deed of surety from Pamouthios the lead-worker to the Apionic heirs (P.Oxy. ., discussed above). As Nikolaos Gonis has pointed out: The notary Anastasius has occurred in a total of eleven texts dating to the period – . . . One of them, P. Warr. , has no connection with the Apion family, while the two Giessen texts are likely to belong to the “archive of Anastasia”; Anastasius was therefore not (exclusively) employed by the Apions.

This Anastasios is thus another example of a bureaucratic link between the Apions and the other large estates. But no doubt more importantly, he is also a link between the estate elites and their social subordinates. The agreement Anastasios executed between Pamouthios the lead-worker and Menas, one of the chief officials of the Apionic estates, is one example. P.Oxy. . is another. That text is an acknowledgement Anastasios wrote to Flavia Praiecta and her son Apion III to supply water to vineyards and arable land (lines –: ï Esw{en toÓ aÉtoÓ ktžmatov ˆntloÓsan e«v Šmpelon | kaª ˆrrÛ. simon g¦n) on behalf of Aurelius Sarmatas, who was himself only a lowly enapographos. P.Oxy. . is a lease agreement between unspecified landowners and an illiterate named Phoibammon, for whom Anastasios wrote. We should not imagine that Anastasios and other    

 

Bagnall’s introductory remarks to P.Turner – on . See Chapters  and  below, passim. See also the discussion of literacy as a social connector in Aphrodito in Chapter  below, p. . Gonis f,  n. : “see J.M. Diethart, K.A. Worp . . . [and] add P.Oxy. LVIII . () and P.bibl.univ.Giss. inv.  and ” as well as P.Oxy. ., with another Apionic attestation now at P.Oxy. . (). On the Apionic family tree, see Chapter  below, pp. –. As they appear in the full re-edition given in the appendix to P.Oxy. .

1 The centralized elite of Oxyrhynchos



notaries were the center of the Oxyrhynchite social universe; that would be an illusion created by the survival of certain types of evidence. But these figures would have been centrally important in another way: they were in some cases the closest the enapographoi and other workers would have gotten to their landlords and employers, serving as the social portals through which the common people might approach the large estates. the estates of theon and timagenes The estates of Theon and Timagenes show how the nome’s social and administrative centralization continued from the fifth century into the sixth. The activities of these two estates share the pattern found in the previous section, in which urban institutions fostered nome-wide rural hierarchies. The identity of Theon and Timagenes has been the subject of considerable debate. Nearly a century ago, Gelzer suggested that Theon had been an exactor whose heirs had to guarantee the taxes he had collected. The early editors of the Oxyrhynchos papyri were agnostic, writing that we cannot be sure “whether Timagenes and Theon should be regarded as large land-owners who had assumed responsibility for the collection of taxes chargeable on their property, or as official exactores whose estates were still accountable, after their death, for the collection in their departments.” Hardy was uncertain whether the Timagenes and Theon documents “refer to private estates or to administrative divisions” in the Oxyrhynchite nome. Such uncertainty began to fade with a later study of the office of exactor in Egypt, which found it likely that the estates of Timagenes and Theon were merely so large that their owners filled an exactorial role as private citizens. Gascou agreed, arguing that the houses 

 



For the Timagenes estate, and characters peripheral to it as well, see P.Oxy. ., ., ., ., ., ., P.Warr. . See also Hardy’s excursus, , –, but with n.  below. For the debate on the nature of the oikoi of Theon and Timagenes more generally, see Gascou , –; Liebeschuetz , ; Syrcou , –, for a brief historiographical summary; and the remarks of Banaji , . Gelzer ,  n. . Commentary to P.Oxy. ., pp. –. Hardy’s belief (Hardy , ) that the estates of Theon and Timagenes were merely “districts into which the Oxyrhynchite nome was divided for certain financial and administrative purposes” strikes me as semantic wrangling. His ultimate conclusion, that “this series of documents may be tentatively removed from those relating to landowners as such,” because they refer to administrative or bureaucratic divisions instead, does not seem plausible. It matters little whether Hardy was right or wrong for my purposes: Theon and Timagenes existed and lent their names to real units of land whose managers had tangible social connections, the real issue at stake here. For recent comments on the office of exactor, see Laniado , .  Thomas , . Hardy , .



Social networks in Byzantine Egypt

of Timagenes and Theon are examples of semi-private institutions in a new, late antique “syst`eme de finances et d’administration municipales,” in which public services were imposed on private estates. These estates had a life of their own, independent of their founders, with a continuity of social relations over a century or more. They were abstract centralizing institutions with bureaucratic responsibilities connecting a wide range of people in the Oxyrhynchite nome. The earliest evidence for the estate of Timagenes playing this role comes from the mid-fifth century. In , the estate’s tax-collecting office received a request for the transferral of payment responsibilities attached to a specific parcel of private land, a request similar to that issued by Flavius Prosdokios, discussed above. Aurelius Mousaios asked that land registered in the name of his father Arion be reassigned to the most eminent Paul. The location of the land is not given, although Mousaios described himself as “from the same city” (lines –: ˆp¼ t¦v aÉt¦v | [p»lew]v, sc. Oxyrhynchos). The estate’s management received a similar request nearly a hundred years later: in , Flavia Euethia, daughter of Apollos, asked “the office of the collection of taxes of the division and estate of Timagenes of noble memory” (a standard phrase in these texts: [t
–xaktorik
t†xe]i m[e]r©dov kaª o­kou toÓ t¦v peribl”ptou mnžmhv Timag”nouv) that land be moved from registration in the name of her mother Theoprepia to that of her husband Iulios. As with Mousaios, nearly a hundred years before, Euethia described herself as from the city of Oxyrhynchos itself. In both cases, the size of the property in question suggests that the estate of Timagenes handled the tax registration of dozens, if not hundreds, of the Oxyrhynchite nome’s smaller landholders. Neither example is particularly large. Aurelius Mousaios transferred . arourae. Flavia Euethia did not give an amount in arourae, but her mention of “. carats for money-taxes of all kinds” leaves the impression that hers was not a large re-registration either. We will examine shortly a re-registration handled by the house of

 

 

Gascou , . P.Oxy. ., where the estate of Timagenes is addressed (lines –) as [mer©]di toÓ o­kou toÓ t¦v peribl”ptou mnžmhv Timag”nouv d. i.‡. | [ . . . .]ou kaª QeodÛrou boh{än –xaktor©av of Oxyrhynchos. Presumably, this represents an earlier phase in the evolution of the estate’s responsibilities; before, it and the estates of Theon are addressed explicitly through the phrase –xaktorikŸ t†xiv mer©dov kaª o­kou. For Prosdokios, see above on P.Warr. , p.  with nn.  and , where recent discussion and examples are cited. P.Oxy. ... The translation herein is that of the editors. P.Oxy. ..: kaª Ëp•r] crusikän panto©wn aÉtän t©tlwn crusoÓ ker†tia ½ktÜ ¤misu Àgdon plžra.

1 The centralized elite of Oxyrhynchos



Theon apparently covering  arourae of land in one entry. No doubt that sort of entry was exceptional, those of Mousaios and Euethia more typical. But the sense of scale is interesting. If the houses of Theon and Timagenes were responsible for the registration of hundreds of arourae of land in parcels ranging from  to  arourae, the officials in charge of that registration process would have had a tremendous range of economic contacts, all converging on their offices in the city. In addition to tax registration, these estates were also responsible for ensuring the assignment of various civic offices in Oxyrhynchos. The estate of Timagenes appears three times over the course of six decades, from  to , acting through intermediaries holding the offices of the pateria, the proedria, and the logisteia. The first office-holders we find working in the name of the estate are the comites Samuel and Phoibammon. (Their possible descent from Timagenes himself we will discuss below, at p. .) They appear together in a  liturgical nomination which documents an interesting connection between the elite civic offices and the most humble levels of society. The nomination is technically addressed to the estate of Timagenes, but via several mediating layers, through the counts Samuel and Phoibammon, and through Serenos the deputy (di†docov), who was probably the matter’s real bureaucratic handler. The nomination itself comes from “the board of the guild of the – and sausage-butchers through us its secretaries and treasurers.” The name of their nominee is lost.   









See below, p. . These comments follow the discussion of Liebeschuetz , . : SB .. : P.Oxy. .. : SB .. Sijpesteijn restored the first text heavily on the left, printing lines – as [di’ Ëmän tän megalopre]p. est†twn kom”twn Foib†mmwnov kaª Samouhl©ou | k. lac»ntwn tŸn lo]gist©an ka. ª p[at]er©a[n] kaª proedr©an taÅthv | [t¦v lam(prŽv) ìOxurugc(itän) p]»lewv. The latter two texts use a different phrasing: to [personal name] lacoÅs tŸn logiste©an kaª proedri†n kaª pater©an taÅthv t¦v lamprŽv ìOxurugcitän p»lewv Ëp•r o­kou toÓ t¦v peribl”ptou mnžmhv Timag”nouv. SB . (for which see R´emondon ) prints l. a. c. »n. t. i. where P.Oxy. . prints lacoÅs, although R´emondon was able to compare the former to the latter. Sijpesteijn ,  n. to line  was wrong to doubt that this is the well-known pair of the same name; his objection that their attestations are six decades after this text disappeared with his own redating of the text in Sijpesteijn . See the following note. SB . = P.Mich.inv.  (Sijpesteijn ) was originally published as Sijpesteijn  text no. , but did not receive an SB entry in that form. Texts – are SB .–, but  is John Rea’s condolence letter earlier in the same volume of ZPE. Sijpesteijn ’s translation at p.  of part of the address at lines –, which reads in full: t¼ koin¼n t¦v | [–rgas©av tän +- kaª tän «]sikiomag©r(wn) di’ ¡män tän par»ntwn kaª —x¦v | [–ggegramm”nwn kefalaiwtän ˆ]nadecom”nwn kaª tŸn gnÛmhn tän Šllwn | suntecnitän. Sijpesteijn took him to be a resident of Oxyrhynchos itself, but only by virtue of his restoration of the lacuna at line . There is room in the lacuna for some other toponym, particularly if the mother’s name were not to be restored therein as well, but since line  makes explicit reference



Social networks in Byzantine Egypt

Despite their assurance that he was “a person of property and qualified” – eÎpo]r. on t¼[n] kaª –pitždion – he was only being nominated to fulfill the duties of a sausage-butcher. Whether or not the sausage-butcher had the abstract estate of Timagenes or the counts Samuel and Phoibammon in his mental landscape (the reverse seems unlikely), Serenos the diadochos no doubt did, and thus formed a social connection between these two extremes of the Oxyrhynchite social ladder. Indirect ties through other aristocrats extended from the estate of Timagenes throughout the province. Flavia Gabrielia, like Samuel and Phoibammon, served on behalf of the estate of Timagenes, appearing in . She is probably the same Gabrielia whose daughter Patricia appears in a civic account with the Apionic estate in the second half of the sixth century. Gabrielia appears to be a remarkable woman: her service as logist¯es, prohedros, and father of the city is eye-catching. But the picture is complex. Gabrielia fulfilled those duties on behalf of the estate of Timagenes, but did not do so herself; a deputy named Christophoros appeared in her place. Perhaps her holdings were close to the original core of the Timagenes oikos, which could explain why she undertook financial obligations for that fossilized institution. Her obligations were presumably part of her patrimony, inherited from previous generations. One modern author has recently proposed that Gabrielia’s role in this situation was to serve solely as a nominator and guarantor, “`a designer un candidat et a` eˆtre ensuite responsable de l’exercice de sa fonction.” In this case, then, Gabrielia had a purely connective social function, being attested only for the social link she created between Christophoros and the institutional abstraction named after Timagenes. Gabrielia’s role as a social nexus did not stop with Christophoros, her diadochos. In the papyrus from , Timothy son of Paul, the water-supplier of the Oxyrhynchite public baths

 

 



to that city’s prison (dh]mos©aƒ –pª t¦sde t¦v p»lewv), there is no reason to think Sijpesteijn was wrong. P.Oxy. .. The note to line  of the ed. princ. of P.Oxy. . states that Gabrielia is not otherwise known, but “may be identical with the late Gabrielia, mother of Patricia . . . [in] .” P.Oxy. .., the civic account, reads: d(i‡) t¦v kÅrav Patrik©av {ugatr(¼v) t¦(v) makar(©av) Gabrihl©av (ˆrt.) nz. On this text, see also Chapter  below, pp. –. P.Oxy. ., lines –: tŸn logiste©an kaª | proedr©an kaª pater©an taÅthv t¦v | lamprŽv ìOcurugcitän p»lewv. See the English synopsis to Fikhman a, . Fikhman describes a situation in which “Not a person but a ‘house’ held office: the offices had become a special sort of munera patrimonii, and probably only a large landowner was in a position to carry out the obligations towards the town which they involved.” Laniado , .

1 The centralized elite of Oxyrhynchos



(lines –: Ëdrop†roco. v. | toÓ dhmos©ou loutroÓ aÉt¦v p»l(ewv)), acknowledged receipt of his salary from Christophoros. The deacon named Ioustos who closed the receipt looks familiar. Rea has identified him, no doubt correctly, with the subdeacon of the same name appearing in three texts of the previous decade. In , he appears in the Apionic settlement case concerning the estates of Diogenes. In , he appears at the end of a millstone-cutting contract addressed to Flavius Apion himself. In , he appears at the end of an acknowledgement of receipt also addressed to Flavius Apion. A network in miniature begins to emerge: the scribal services of a subdeacon who rose to become a deacon provide a direct link between Gabrielia, father of her city, and Flavius Apion, who also held responsibilities for the estates of Timagenes. The estate of Timagenes also appears in the estate account of an Apionic prono¯et¯es or steward from  or later. A single entry in line  of the account’s second column records a payment of  solidi  carats through the heirs of Ioannes, son of Timagenes: d(i‡) tän klhr(on»mwn) ìIw†nnou Timag”. n. o. u. v. Ë(p•r) –. mfu. t. ©av %lex. no(m.) pg ker. h. At some point, then, the family of Timagenes entered into a heritable leasehold agreement with the Apions, confirming connections between these members of the Oxyrhynchite elite. (Ioannes returns below at pages –.) Timagenes and the earliest members of the Apionic dynasty may well have been social equals in the early fifth century, a social proximity that may account for the heritable lease agreement between the families later in the sixth century. It was presumably not an agreement entered into lightly;  solidi is a considerable sum, and must have represented a sizeable holding. The estate of Theon follows the same patterns as that of Timagenes, also creating a web of vertical, centralizing connections for well over a century. The two oikoi appear together, making payments of grain through the same Daniel presbuteros in P.Oxy. . (sixth century). Theon may have been the city councilman and father of his city named in a Christian



     

See P.Oxy. ., n. to lines –, citing Diethart and Worp , , s.n. ..–.., where Ioustos spans the period from  to . See also Diethart and Worp ,  for the suggestion, followed here, that apo diakonon is to be replaced by hupo diakonou.  P.Oxy. .. P.Oxy. .. P.Oxy. ., republished in full in the appendix to P.Oxy. . In SB ., on which see R´emondon , correcting original readings to P.Varsov. . P.Oxy. ., written on the reverse side of P.Oxy. ., which provides the date. These accounts are discussed in more detail in Chapter , below, p. . See Rea’s commentary to lines – and line . For the estates of Theon, see P.Oxy. ., ., ., ., ., and ..



Social networks in Byzantine Egypt

letter from Oxyrhynchos. He may also appear in a shipping declaration for the transport of revenue grain issued in  by an Aurelius Andreas, kubernžthv plo©ou Q”wnov peribl”ptou, “captain of a ship belonging to Theon vir spectabilis.” This may be a glimpse at an earlier period, when the eponymous founder of the estate was still alive. Theon’s high status in Oxyrhynchite society has only recently been appreciated. The crucial evidence is an account in various drafts the first line of which begins: + L»g(ov) çip[a]r(©av) o­k(ou) Q”wnov. Its contents are an index of the heirs of deceased aristocrats providing riparii or law-enforcement officials in the name of Theon’s estate. The dates in the text cover a period of  years up to the s, including seven indictional cycles, and show that the house of Theon maintained responsibility for over  per cent of the riparii of Oxyrhynchos, even after Theon’s death. Names of Oxyrhynchite elites covering those duties for the house of Theon in that account include ¾ ›ndox(ov) ok(ov) (surely the Apions), the heirs of Leontios, those of Philoxenos, and those of Mousaios. Since someone had to be responsible for maintaining this account in the name of the house of Theon, that one person at least had bureaucratic ties to agents providing riparii on behalf of all these heirs.  



 

 

SB . from Keenan d, dated to the sixth century by virtue of a “very hard to place” hand. Azzarello ,  finds the identification only “possibile.” P.Oxy. ., a text which no one to my knowledge has exploited in discussion of Theon. P.Oxy. . does not appear in the BL. Nor, judging from the BL, has anyone yet commented on the later P.Oxy. ., dated , and the daughter of the blessed Theon appearing in that text. Could its Theon be a homonymous descendant a few generations after the establishment of an ancestral large estate? This is purely speculative. The address is from a Flavius Gerontios ˆp¼ t¦v %rsinoitän to (lines –) t a«des©m
wro{” [u¬ toÓ t¦v | makar©av mnžmhv QeodÛrou toÓ kaª %pfouŽ kaª t
[a«]d. e. [s]i.[m]wt†th€ | aÉtoÓ sumb© {ugatrª toÓ makar©ou Q”wnov | ˆp¼ t¦v aÉt¦v p»lewv. At the very least, correspondence of Oxyrhynchite provenance between two Arsinoites is interesting. Cherf  has not looked through the minor documents, and therefore does not list this Gerontios. He is equally missing from PLRE . On this text see Gonis b, –, where the author transcribes the last portion of the text, missing from the ed. princ. The editor does not venture an identification. Pointing to the title of vir spectabilis, he points out that “This is one of the earliest instances of the title. Later it was debased, but at this date Theon, though hardly to be identified, is likely to have been a person of some importance.” Azzarello , . P.Oxy. . with Azzarello  for a re-edition and extensive commentary, particularly at p.  for the chronological proportions involved. See also Liebeschuetz , – and Torallas ,  for earlier discussion. The th indiction in  is presumably /, but not for the reasons put forth by R´emondon via Bonneau ,  n. ; see instead Azzarello , , leaning towards the same date with much less certainty. For riparii financed by the large estates, see the cautions at R´emondon , : “Dans la r´ealit´e, ces riparii priv´es n’existent pas.” Both large estates and villages would nominate candidate riparii for government approval. See Torallas  for the Byzantine ripariate and its relatively high status. Heirs, not oikoi: see Azzarello , , where the o«k() of the ed. princ. is traded for o¬ kl(hron»moi). For discussion of the identities of each of these original landowners, see Azzarello , –.

1 The centralized elite of Oxyrhynchos



Theon’s estate also appears in undated sixth- and seventh-century receipts issued to that estate’s stewards. We see a rare dated reference to the house of Theon in , a request for an alteration in the tax lists. Here, the lady Anna, daughter of the blessed Ioannes and wife (?) of Papphaous, writes to the tax-collection office under Theon’s estate to rid her name (line : {el[žs]ate ˆpokouf©sai t¼ –m¼n Ànoma) and that of (her daughter?) Maria of the taxation responsibilities for the land registered under the names of Dioskoros bo¯ethos, Didymos son of Timotheos, Paoros son of Pambechios, Paulos son of Ioannes, and Phoibammon son of Abraamios, all from Kerkethyris in the Oxyrhynchite. None of these people are known from other texts. The text is fragmentary, so exact personal relations and motives remain elusive. But the catalyst for this particular tax re-registration seems to have been the sale of all the land involved. The text itself refers to  arourae, which caused its editor concern for its sheer size. Perhaps the land, once belonging to Anna’s father Ioannes, had been divided among some halfdozen landowners through purchase, and Anna was taking the last step to finalize the deals. This is the pattern found in other such requests, where the transfers invariably involve a parent or other relative. Kerkethyris, in the Oxyrhynchite’s western toparchy, would not have been very far from Oxyrhynchos, so physical distance was not at issue here. This memo to the tax-collectors of Theon’s estate may reveal what Anna considered too great a social distance: whoever Dioskoros, Didymos, and the others from Kerkethyris were, she could not reasonably be accountable for taxes on land no longer connected to her family in any meaningful way. Anna’s request is not the only tax re-registration text we have from Theon’s estates. P.Oxy. . () is, like several of the Theon and Timagenes documents, addressed to the exactor’s office. In this case, the text is an agreement by Flavia Stephanous, with the consent of her husband Markos, that she would assume some tax duties formerly registered in the   



 

Syrcou  number  ( = SB .), and p.  for complete references to other examples of this genre. For an alternative dating, see P.Oxy. .. Syrcou , . In her n. to line  on p. , Syrcou  proposes the restoration Ëp•r f¯ [tän Âlwn ˆrou]r. än pra{eisän parì aÉtän ‰p†nt[wn], and translates accordingly on p. : “of all arouras sold by all these persons.” See Syrcou’s note on the top of p. : “The size of  arourae of land is too large and it will be a large estate, for example, Theodora’s estate of Hermopolis was not more extensive than  arouras . . . Although Anna is not known from other papyri as an estate owner, it seems to me that f must be interpreted as a figure indicating the amount of arouras owned by Anna.” Judging from the picture published with the article in Tafel XII, the reading of Ëp•r f¯ [ in line  is secure. See P.Oxy. ., ., ., and the other examples cited at Syrcou , . For citations, see Pruneti s.n.



Social networks in Byzantine Egypt

name of her father, because certain of her father’s land had been transferred to her as a dowry. We do not know how much land was in question, but we have reference to a grain tax of  artabae (e«v m•n –m[b]o[l]Ÿn s©tou kan»nov ˆrt†bav —xžkonta tre±v): we are probably somewhere between Flavia Euethia’s more modest transfer of a generation before and Anna’s enormous transfer from the same period. We do not know anyone in this text from elsewhere. Stephanous’ father, Ioannes the scholastikos, was presumably still alive at the time of this transaction. A pair of possible matches come to mind. First, Flavius Ioannes, also called Lampadios, son of Anastasios, a scholastikos and a landowner at Oxyrhynchos, appears in . He is described as sofÛtatov, as is the Ioannes in the re-registration text from . Second, Ioannes son of Hierax, a scholastikos at Oxyrhynchos in the mid-s received a salary from the Apionic estates. The suggestion of links between the Apions and the estates of Theon is tempting, but not solid in this case. If nothing else, these two examples might lead us to ask how many scholastikoi named Ioannes Oxyrhynchos had at any given time. More substantial evidence of ties between the heirs of Theon and the Apions show up in an Apionic account of , the reverse side of which mentions the estates of Timagenes, as noted above. This is the annual account of Theodoros, a prono¯et¯es for the Apions, which we will discuss in detail in Chapter  below. The entries in question are obscure, one recording a payment to the Apionic estates of just over two solidi “from the heirs of Theon from the same holding,” and another recording a payment of just over five solidi “from the same heirs of Theon from the same 

   



Judging from its relative lack of BL entries, this is an understudied text. See Johnson and West ,  n.  for the recovery of a personal name, Kuros, misread in line  of the original. This Kuros may be the same man appearing in another Theon text, P.Oxy. .: Syrcou , . (This would narrow the dating of the latter from “Sixth or seventh century” to some time closer to c.s.) See also Ioannes, father of Markos, at PLRE ., s.n. Ioannes , and Ioannes, father of Stephanous, s.n. Ioannes . With the possible exception of Kuros; see previous note. The text was drafted by a Paulos unknown from any other text: Diethart and Worp , . PSI . as corrected in Keenan . See PLRE ., s.n. Ioannes . BL . cites CSBE for the  date; the ed. princ. gave . PLRE ., s.n. Ioannes , citing P.Oxy. . line . An intriguing experiment: consider the index to P.Oxy. volume . At roughly  pages, with nearly  names per page, the “Personal Names” index (Index V) has not quite , names, and just under  entries for Ioannes. This gives a ratio of not quite  Ioannes for every  Oxyrhynchites. The same volume yields only eleven entries under scholastikos (Index IX), roughly  for every . In an Oxyrhynchite population of c.,, this would give us roughly  scholastikoi named Ioannes, not enough to chance an identification. (Although we should assume that scholastikoi are overrepresented in the written record, and the population of the city may have been lower as well.) See Ruffini  for bibliography on name frequencies. P.Oxy. . and . See Chapter  below, pp. –, with remarks above at p. .

1 The centralized elite of Oxyrhynchos



holding” (p(ar‡) tän aÉtän klhr(on»mwn) Q”wnov ˆp¼ toÓ aÉtoÓ kt[ž]ma(tov)). The name of the kt¯ema in question is lost, but would have been the same for both entries. Here, the house of Theon is not appearing in its capacity as an institution fulfilling civic duties, but in a more modest capacity as a landholder, or more specifically, a leaseholder. So the Apions had much the same relationship with Theon’s heirs as they did with those of Timagenes, perhaps granting a heritable lease. The estates of Theon and Timagenes may have been nothing more than bureaucratic abstractions, particularly in the sixth century, long after the deaths of their founders, but they played an active role in the maintenance of economic and social ties. They appear to some degree responsible for staffing the office of the exactor. One estate worked with the logisteia, proedria, and pateria, while the other was responsible for providing riparii. Both handled written requests for the re-registration of tax responsibilities. Through these duties, the men working for the two estates had contact with guild members (see the sausage-butcher discussed above at pages – ), public bath workers, regional landowners, and many others. This is not merely a series of overlapping social circles, but another example of a phenomenon prevalent in this chapter: throughout the Oxyrhynchite documentation, elite figures or the institutions surviving in their name appear as centralizing agents whose bureaucratic responsibilities tie together lesser regional actors from all walks of life. fl. apion theodosios ioannes, samuel and phoibammon With the recent publication of P.Oxy. volume , we may now tentatively identify the descendants of Timagenes, and show that they too created vertical social hierarchies throughout the Oxyrhynchite nome. These figures have been hidden in plain sight for the last eighty years. We have already seen the counts Samuel and Phoibammon holding the functions of logistai, prohedroi, and fathers of Oxyrhynchos on behalf of the estates of Timagenes. Now it seems they may have acted for the estate because their grandfather was Timagenes himself. Two texts from  and  refer to Samuel and Phoibammon’s father as a peribleptos named Ioannes. A fragmentary contract from  is addressed to Flavius Ioannes, a peribleptos, curialis, comes sacri consistorii, presumably their father, and himself    

Lines  and , with transl. on page . The relevant texts therein are P.Oxy. ., , and . In SB ., discussed above, p. . See Gonis in his introductory remarks to P.Oxy. ., with P.Oxy. ..



Social networks in Byzantine Egypt

the son of Timagenes. We know this Ioannes from elsewhere. In , Flavius Apion Theodosios Ioannes, a peribleptos and a count of the sacred consistory, was also the governor of Arcadia, the larger province of which Oxyrhynchos was a part. That year, his soldiers were paid at the order of count Phoibammon. This Phoibammon is typically assumed to be the same count appearing with Samuel in documents dating into the sixth century. The resulting picture is three generations of a local Oxyrhynchite family functioning at high levels of the elite bureaucracy. Timagenes himself left behind an Oxyrhynchite oikos established in his name. The man we now know to be his son, the comes and provincial governor Ioannes, was an Oxyrhynchite city councilman or politeuomenos. His two putative sons, Samuel and Phoibammon, were also counts and city councilmen. The appointment of both Ioannes and his son Samuel to the rank of comes sacri consistorii, count of the sacred consistory, is one of the striking aspects of this family; these awards were presumably honorific, but they were hardly common. We know of only four other men with Egyptian connections to reach the rank. 





  

See P.Oxy. . for the “likely” identification of the Ioannes therein with that of . and .. Lines – of the text read Flaou© ìIw†nn t peribl”pt k». meti toÓ {e©ou konsistwr©ou | kaª politeuom”n u. ¬ t. o. [Ó] t. ¦. v la. m. p. r“. v mnžmh. v. Timag”nouv. PLRE ., P.Oxy. . and .. Contra Hardy , , we cannot be certain he belongs to the great Apion family: see Gonis a,  n. . Montevecchi ,  labels the Apion Theodosios Ioannes archive number . She lists: P.Oxy. ., ., ., ., ., ., ., ., .. Banaji ,  n.  adds P.Oxy. .. Half of these (P.Oxy. ., ., ., ., .) are texts relating to Samuel and Phoibammon, whom Montevecchi does not give an archive in their own right; in light of the discoveries of P.Oxy. , this archive may perhaps belong not simply to this Apion, but to his descendants as well: see nn.  and  above. P.Oxy. .. Although Gonis is right to point out (P.Oxy. . n. to line ) that this “is not conclusive for identifying the praeses [Ioannes] with Phoebammon’s father,” it is a compelling coincidence. See below, pp. –. The identification began with the ed. princ. of P.Oxy. ..  P.Oxy. ..–. P.Oxy. ... Keenan ,  n. , where Apion Theodosios Ioannes (P.Oxy. .) and Samuel (P.Oxy. .) should be added. The four: Ammonios of Aphrodito (s–s), Serenos (Oxyrhynchos, ), an earlier Ammonios (late fourth/fifth), and Sabinus Antiochus Damonicus (Thebaid, c.– ). Damonicus was a native Antiochene. The second Ammonios, mentioned in the Hermopolite text P.Ryl. ., is a complicated case: PLRE . assumes he signed the text himself, and Keenan assumes him to be Egyptian, but the original editor thought he was acting vicariously, “no doubt, through his local representative.” Ammonios is a distinctively Egyptian name, and Keenan is quite likely correct. For Serenos, the well-known grandson of Eulogios, see below, pp. –. For more on the Aphrodito Ammonios, see Chapter  below, pp. –. For a potential fifth Egyptian count of the sacred consistory, consider the fragmentary text SB . attesting Flavius Ioannes . . . Strategios Ioulianos . . . , addressed as k»meti toÓ {e©ou k . . . . . . ..[.]..[.].[. See Keenan b, –, where the editor resists reading kwnsistwr©ou, while finding it a “strong possibility.”

1 The centralized elite of Oxyrhynchos



Apion Theodosios Ioannes is first attested in the s, and material relating to his career stretches to . Whether Apion Theodosios Ioannes had any connections to the Apion family, as some have argued, he clearly had connections to rural Oxyrhynchos. As Jairus Banaji puts it, his “exalted position in the Byzantine bureaucracy had in no sense distanced him from the immediate life of his estate.” We see this hands-on approach in , when he ordered his butler, Phoibammon, to issue payments of wine to the residents of Sepho and Kesmouchis, and also to a carpenter, a policeman, a fisherman, a porter, and estate guards. Indeed, these toponyms do suggest some connection between this provincial governor and the main Apionic family. Sepho appears in two Apionic accounts, one from the s, and another from the s. Kesmouchis, in a different pagarchy but originally in the same toparchy as Sepho, appears with it in at least five other texts from the Oxyrhynchite nome. If not an Apion himself, his property certainly seems to have been near theirs. As with many of the networks documented in this chapter, Ioannes’ personal ties included those forged both through private economic activities and public or official connections. Ioannes is attested in  with  soldiers in tow. We have an order from Phoibammon comes – perhaps his son, as we have just seen – to “supply to the most noble soldiers of the armigeri who came here with the most magnificent praeses Ioannes.” After taking up his governorship, one of his functions was to hear cases brought to his office by the locals. In c., a bo¯ethos named Pamouthios appealed to Ioannes, complaining that Phib, Elizabeth, Kollouthos, and a priest named Phoibammon, about whom we know nothing else, had failed to discharge their debt to him, and ought to be required to do so.     







See the references collected above at n. . Hardy ,  thought it “not improbable” that he was identical to Apion I. Banaji ,  argues that he cannot be in the family’s “main line of descent.” Banaji , . P.Oxy. .. For the inclusion of this text with the other Ioannes documents, see Montevecchi, cited above, n. . s: P.Oxy. .. v., which Mazza ,  assigns to the Apionic archive on grounds of prosopography and topography. : P.Oxy. .., which Mazza ,  assigns to the Apionic archive on grounds of prosopography and archaeology. P.Oxy. ., ., ., ., .. The latter four on this list are all large accounts and payment lists, which do not necessarily suggest topographical ties. Nonetheless, in three of those lists (P.Oxy. ., ., and .) Sepho and Kesmouchis appear back to back. P.Oxy. ..: p[a]rasco[Ó] to±v [genna]iwt†t(oiv) strat(iÛtaiv) tän –. rmig”rwn –l{(oÓsin) –ntaÓ{a met‡ toÓ megalopr(epest†tou) Šrc(ontov) ìIw†nnou; transl. in ed. princ. I share the assumption of Grenfell, Hunt, Bell, and the PLRE (see n.  above) that the Ioannes therein is the same named in full in P.Oxy. .. P.Oxy. .. The date of the text is based solely on the appearance of Ioannes as governor therein, in which capacity he also appears in P.Oxy. ., securely dated to .



Social networks in Byzantine Egypt

As for counts Samuel and Phoibammon, the governor’s apparent sons, they appear, alone or together, in texts spanning  years, from the appearance of Ioannes with his soldiers in , to the reign of Justinian. They were thus contemporaries of Flavius Apion I and Flavius Strategios II, of Eulogios and his descendants, whom we will meet below, and of the other great figures of the Oxyrhynchite nome in that period. Most of the texts – several written before Samuel had become a comes – are orders of payment from the two officials to unspecified recipients. One is an order to give wine to some monks on Christmas Day, and to various prisoners a couple of days later. These examples show how many people throughout the Oxyrhynchite nome had the local government and its elite figures in their sphere of social contacts. Monks and criminals received gifts by order of men who worked for Apion Theodosios Ioannes, who held some of the highest ranks and honorifics and owned property near the Apions, the region’s most powerful family. The office of the praeses, which Ioannes himself held, interacted with palatines (Flavius Eulogios, below), low-level bureaucrats (Pamouthios), debtors, priests, soldiers, and many more. Soldiers working for the governor apparently received their rations by order of that governor’s son. Out in the Thmoisepho toparchy, fishermen and fieldguards received payments of wine by order of the same man, a count but apparently no longer governor. The point here is not simply that the Oxyrhynchite nome was a small world, although it certainly appears to have been. The point is that it was a hierarchical, centralized one as well, with all manners of vertical social chains running through the nome towards a small group of elite figures in the city itself.

the eulogios archive The archive of Flavius Eulogios and his descendants is another example of this tendency, in which several generations of Oxyrhynchite landowners 





Samuel and Phoibammon: P.Oxy. . (), . (), . (), . (undated). Phoibammon: P.Oxy. . (). It is also possible that Coles is right to think of this Phoibammon in his commentary to line  of P.Oxy. .. The “church of Phoibammon” in that text is probably named after a prominent patron, “as the name is not accompanied by Œgiov”; this Phoibammon comes is an obvious possibility, if nothing more. P.Oxy. .. But monks where? And prisoners of whom? The text is vague. The paraschou in line  is followed only by e«v p±n [t]än ‰g©(wn) monaz(»ntwn); that in line  only by e«v t‡ desmotžr(ia).  P.Oxy. ., discussed above, p. . P.Oxy. ., discussed above, p. .

1 The centralized elite of Oxyrhynchos



appear at the top of vertical hierarchies spanning the fifth and sixth centuries. Most of this archive, one of the most significant smaller archives to emerge from the Oxyrhynchite nome, appeared in Oxyrhynchus Papyri volume . A number of those pieces have since been republished by James Keenan and Todd Hickey, and three new items have recently appeared in Oxyrhynchus Papyri volume . The texts in this archive are typically leases of house property. Eulogios himself, the head of the family, was the son of a local notable named Horion. Eulogios appears as early as September , already a landowner, but merely an Aurelius. His first attestation as a Flavius is from February , when he leased a room in the Hippeon Parembole (Cavalry Camp) quarter to a woman named Aurelia Pina. Later that same year he leased an entire house to another woman whose name is lost. In , Eulogios was both a Flavius and ka{osiÛmenov palat±nov, or devotissimus palatinus, a civil servant who owned and leased house property in the Pammenes’ Garden district of Oxyrhynchos. His tenant in that lease was Flavius Paulos, who perhaps described himself as a courier in the office of the praeses (koÅrswr t. ¦. v [¡gem]o. [n]i.k. ¦. [v t†xewv]) in the same city. A libellus which Flavius Eulogios and Megas presented to the  

  

  

Bibliography: see Hardy , –, Keenan a, including Plate IX (P.Oxy. .), and Hickey and Keenan – and . For the latest texts, see P.Oxy. . and .–. Doubt remains about the inclusion of one or two items in this archive. P.Lond. . was first published only in description, and Bell thought it may have been Apionic instead. Hardy suggested a Eulogian connection (Hardy, ,  n. ). Keenan a,  n.  confirmed that via photograph he could read the first line as Fl(aou©) %pf. [oÓti rather than Fl(aou©) %pi.[, and notes that Apanakios appears both therein and in P.Oxy. ., a text clearly dealing with Martyrius and Apphous. Hickey and Keenan (–,  n. ) still had reservations about including this text, but in Hickey and Keenan , they have convinced themselves that Apphous is the landowner in question, and that the text does therefore belong with this group. Another text sometimes adduced to the archive is P.Oxy. .; the proposed connection goes back to the editors themselves, who report that the text was found with P.Oxy. ., another piece relating to the family. Hickey and Keenan – publish P.Oxy. . (now = SB .) in full, but still think that it is only a probable member of the archive. Keenan a,  calls these leases the “majority” of the Eulogian texts. P.Oxy. .. originally read Eulogios’ patronymic as the doubtful ì . ri.g. ”. n. [ou]v, but P.Oxy. .. now gives an unassailable correction. P.Oxy. .. As Gonis points out in his introduction to the text (), this is an important point: “if Eulogius was a man of certain means before joining the service, his wealth did not entirely originate in it. This may serve as a warning when studying the staff of the civil service of the Later Empire, as well as the much-discussed links between the bureaucracy of the time and the ‘new’ landowning class.” P.Oxy. .. Hippeon Parembole: see the references in the third supplement () of Calderini, Diz.geogr., , with the earlier Daris , . P.Oxy. .. P.Oxy. .. For the district, see the third supplement () of Calderini, Diz.geogr., , and Hatzilambrou , .



Social networks in Byzantine Egypt

praeses circa  gives a hint of how Eulogios might have acquired some of this property. He and Megas explained that two men named Serenos and Martyrios fled fifteen years before to evade payment to them, thus entitling them to lay claim to a house in Oxyrhynchos the debtors pledged as security. Eulogios is dead by , from which point we can follow the careers of his descendants for six decades, if not more. Martyrios and Apphous, the sons of Eulogios, had a number of connections to lower social strata. In April of , when they were both merely Aurelii, they leased a portion of a house – the same one their father had leased to Flavius Paulos in the previous decade? – in the Pammenes’ Garden district of the city to Aurelia Martyria, from the village of Ibichis in the Herakleopolite nome. The scribe of the text was Serenos, son of a deceased deacon named Philoxenos. As we will see repeatedly in Aphrodito, minor figures serve as indirect social links through the literate tasks they perform. Serenos was one such character. He appears thirteen years later in , as notary in another lease by the two brothers, now Flavii, of a small house (again, the same one?) in the Pammenes’ Garden district to an Aurelius Apa Nakios. Pointing to the fact that Martyrios is named first, and disappears from the documentation first, Hickey and Keenan suppose he was the older of the two brothers. We know little about Martyrios. He may be the lead figure in P.Oxy. . (late fifth / early sixth century), in which a Martyrios skr. (ini†riov) is attested as leasing an ½rbiopwle±on (vetch-seller’s shop) to an ½rbiopÛlhv (vetch-seller). His brother Flavius Apphous appears 



   

 

P.Oxy. .. There has not been any BL entry on the date of this text. We know only that the text must pre-date , by which point Eulogios was dead. In his recent introduction to P.Oxy. ., Gonis implies agnosticism, declining to name a c. attestation for Eulogios. Whether both men do so, or Eulogios himself does alone, is unclear. The proper name Megas is clear enough: see the note to line  in the ed. princ. But his relationship to Eulogios is never explained, and the rest of the text is in the first person singular. P.Oxy. . and P.Oxy. . = SB .; see the summary at Hickey and Keenan –, –.  See Chapter  below, e.g. pp. –. P.Oxy. .. P.Oxy. .. The initial identification is that of the editors, accepted at Diethart and Worp ,  s.n. .. and ... The Apa has no ecclesiastical force here; the name merely commemorates an actual Apa Nakios. See the discussion at Keenan a, . The text is P.Oxy. . = SB ., the full edition from Keenan a, where notes and plate are included. Hickey and Keenan –, . See Hickey b for new readings and a full edition of this text, and (p.  n. ) on the possible but unproven connection between this text and the family of Eulogios. The dating is only by paleography, through comparison to PSI . and ., but is a refinement over ed. princ.’s “sixth century.” Hickey notes (ibid.) that “The meager evidence for Martyrios’ career certainly does not rule out service as a scriniarius at some point.”

1 The centralized elite of Oxyrhynchos



in his own right in five separate property leases and a loan. The first text dates from the end of , the last from August of . His clients appear to be somewhat down-market. The loan in  was of two solidi to a confectioner (pastillŽv) named Aurelius Theon, son of Ioannes and Herakleia. He also leased property to Aurelius Ioannes, a villager of Senokomis, in ; to Aurelius Anoup, a cumin-seller, in the same year; and to a church official, perhaps an oikonomos, named Anoup, six years later. A third generation of the family of Eulogios also appears in the documentation: Flavius Hatres son of Martyrios, who appears in a lease from , and Flavius Serenos, his brother, who appears as late as . The Hatres text is the family’s third lease to a woman in as many generations: Aurelia Nonna leased from Hatres a hall in the Dromou Sarapiou quarter of Oxyrhynchos. The Serenos text from  provides the family with some fascinating and underdeveloped connections. The text is a contract with a horse-trainer or stableman (stabl©thv), Aurelius Serenos, also called  

 

 

 





P.Oxy. ., P.Mich. ., P.Oxy. ., . = SB . (Hickey and Keenan –), and P.Oxy. .: surveyed at Hickey and Keenan –, . P.Oxy. . = SB . (Hickey and Keenan –) may have a later date. It has the same notary as SB . (Sijpesteijn ), dated to March . According to Grenfell and Hunt, it was found with P.Oxy. . and other texts relating to Martyrios and Apphous, but the names of the addressees are lost, and certainty as to its inclusion is not possible: see Hickey and Keenan –, . P.Oxy. .. P.Oxy. .,  ad. This text is exceptional in the archive of Eulogios and his descendants in its concern for property outside of the city of Oxyrhynchos itself: we learn that Apphous at least owned property in Senokomis. The lease describes Ioannes as ˆp¼ kÛmhv SenokÛmewv (line ), and then records his agreement to lease ˆp¼ tän Ëpar | c»ntwn t
s
eÉgene©aƒ [sc. Apphous] diakeim”nwn –pª t¦v aÉt¦v kÛmhv (lines –). P.Mich. ., where Sijpesteijn rightly rejects (p.  n. ) any connection between this Anoup [ku-]|m. inopÛlhv and the Anoup leasing from Apphous in P.Oxy. .. P.Oxy. ., , but with Anoup’s position missing at the start of line : [+-  t¦v] ‰. g©av –k[klh]s©av. The ed. princ. printed “o«kon»mov?” Hickey and Keenan –,  n. to line  remain agnostic, simply listing di†konov, o«kon»mov, and presbÅterov as possibilities. PSI .. Hatres did not earn an entry in the PLRE. See PLRE ., Serenus , which identifies him with PLRE .’s Serenus . Neither entry draws any attention to this man’s connection to a larger archive. From the evidence of P.Oxy. ., the existence of the archive was not yet apparent. But the connection can hardly be doubted. In lines –, he is addressed as u¬ to[Ó] t¦v lamprŽv mnžmhv Martur©ou, geoucoÓnti –[n]taÓ{a | t
lamprŽ€ ìOxurugcitän p»lei. PSI ... Nonna described herself as from the city of Oxyrhynchos in line . For references to the quarter Dromou Sarapidos, see Daris , , citing PSI . et al., and the third supplement () of Calderini, Diz.geogr., . P.Oxy. .: Keenan a,  includes it in his list of the archive’s documents, as do Hickey and Keenan –, followed by BL .. See also BL ., BL . and Hickey ,  for various corrections.



Social networks in Byzantine Egypt

Kortibos, son of Ioustos and Maria, who “undertakes the superintendence of the racing stable belonging to Flavius Serenus, a comes, for one year.” Serenos the stableman appears again, in the records of the monastery of Abba Andreas, which include a receipt given to Serenos only five years later, in . Maybe the landlord who owned the barn from which Serenos carried the hay for the monastery was Flavius Serenos, and the Abba Andreas monastery rented from the family of Eulogios. The monastery had other interesting connections of its own. Another of its receipts, dating from the following year, , was issued by the head of the monastery to Ioustos, an Apionic bath attendant: pericÅth toÓ geoucik(oÓ) loutroÓ t¦v meg†l(hv) o«k(©av), for four mats. We have already encountered the Apionic bath above, where I proposed a connection between the leadworkers employed for work on that bath, and the lead-workers employed for work by other great landowners, such as Flavia Anastasia. These are more indirect, vertical connections: the family of Eulogios to the Abba Andreas monastery through a stableman, the monastery to the Apions through a bath attendant. Finally, in the family’s original contract with the stableman, the scribe of the text, Philoxenos, also demonstrates indirect connections between the family of Eulogios and the family of the Apions. According to Grenfell and Hunt, the scribe of that text also wrote a receipt for a seed advance paid to the council of village headmen of Takona by Flavius Apion, dating to . This is not the only time Philoxenos apears in an Apionic context. The previous year he drafted a receipt for an Šxwn by the Aurelii Papnouthios and Menas from the kt¯ema Nik¯es, which its editors thought was “probably” addressed to Flavius Apion. Nineteen years before, he drafted a similar document addressed to Flavius Strategios from Aurelius Epimachos, also called Apima, from Pagguleeioos, who needed an axle for the water-wheel

 

   

P.Oxy. ., intro. P.Oxy. .. Surely this is a near-certain identification. We may again consider the thoughtexperiment proposed above at n. : how many stablemen named Serenos might Oxyrhynchos have had? Unlike scholastikoi named Ioannes, the indices for P.Oxy. volume  suggest we are in much more rare territory here. P.Oxy. .; for the “great house” as that of the Apions, see Hardy ,  and Mazza . See above, pp. –. Hardy ,  assumes that we are dealing with only one Apionic bath, not two or more, and this seems a reasonable assumption. P.Oxy. .. See now Diethart and Worp , –, listing the other examples discussed herein, with Tafel , Oxy. ... P.Oxy. . = SB ., from Koenen , with Tafel . Koenen accepts the Apionic origin of this text without comment. Mazza ,  includes it via her categories of archaeology and topography.

1 The centralized elite of Oxyrhynchos



called the “Five Arourai.” So, the Philoxenos who drafted these texts provides at least an indirect link between the Apions and the descendants of Eulogios, as well as links to a number of settlements throughout the nome. But direct connections may have existed between the two families as well, in two texts not typically cited in reference to Eulogios and his descendants. The first text, discussed in greater detail in Chapter , is an extremely detailed account of receipts and expenses by the Apionic estates, dating from /. One of the entries records a payment in line  to men under the stewardship of count Serenos (sunecwrž({h) to±v aÉto±v kaª prokeim(”noiv) –x –pitrop(¦v) toÓ k»m(itov) Seržnou). The second text is an account for water-wheel axles, identified as an Apionic text by virtue of its place names, and dating to no later than /. One item of this account (line ) is paid [ˆp¼ t]än –n. e. c{”nt(wn) ˆ. p¼ [%]lex[a]ndr(e©av) –x –p. it. rop(¦v) to. Ú. k»met(ov) Seržnou, from those brought from Alexandria in the service of count Serenos. These entries are ambiguous to us, although presumably clear enough to the accountant. If the count Serenos in these two texts is the same as the grandson of Eulogios, also a count, they show direct connection between his family and the Apionic estates. Other evidence may show that the family of Eulogios had economic ties beyond the Oxyrhynchite nome proper. A fifth- or sixth-century letter from an unnamed author, who had arrived in Oxyrhynchos with someone named Alis, wrote to their master Apphous, a notary, whom the editors consider “possibly” the same as the Apphous son of Eulogios we have already met. The letter’s writer also mentions giving a copy of another letter to a lord Eulogios, “a further point of contact,” as the editors cautiously put it, between this text and the family of Eulogios. The letter is a plea to the addressee to ensure that builders come to prevent the flooding of some vineyards. Keenan and Hickey have argued that the Eulogios and Apphous here are not those of the archive, because their connection to the    



P.Oxy. .. For water-wheels and axles, see above, p.  with n. . P.Oxy. .. For the date, see BL ., citing P.Oxy. ., p.  (“The calculated date of  is unassailable”) where arguments are assembled. See Chapter  below, p. . P.Oxy. .. For the date, BL ., P.Oxy. .. n. and Gonis b, . P.Oxy. .. Its presence in the archive is contested by Hickey and Keenan –,  n.  because of its reference to land in the Herakleopolite village Gessias; Apphous is never attested as being outside of the Oxyrhynchite. The PLRE . entry on Apphous does not use this piece. See also the note to line  in the ed. princ., where the editors point out that this text “was not found with” P.Oxy. ., .–, and ., the main batch of Eulogios family texts. P.Oxy. ., –, and , and discussed above, pp. –.



Social networks in Byzantine Egypt

Herakleopolite village of Gessias would be that family’s only appearance outside of the Oxyrhynchite. (Gessias was probably near Palosis, in the Thmoisepho toparchy. ) However, as we have already seen above, the family did at least have ties to people from the Herakleopolite: remember their female tenant, Aurelia Martyria, from the Herakleopolite village of Ibichis. The Eulogios family disappears from sight in the second half of the sixth century. Grenfell, Hunt, and Bell suggested that the Apphous appearing in P.Oxy. . was also a member of the family, the text having been found with others in the family archive. The text dates to , which would make it the latest certain attestation of the family. The document is a loan receipt from a priest named Anoup, from the kt¯ema of Kame. This loan reflects the pattern the family established through more than a century of documentation. Eulogios and his descendants are not the Oxyrhynchite super-elite, with vast estates and middle managers throughout the nome. They lease property on a small scale, to isolated figures from the countryside, to vetch- and cumin-sellers. If the leases of property in the Pammenes’ Garden district all refer to the same house, their urban holdings may not have been that large to begin with. But the offices they held and the scribes and other contacts they shared with the Apions suggest that the highest levels of the Oxyrhynchite elite were socially quite close. the female elite: fl. euphemia and fl. anastasia The Oxyrhynchite elite appearing in the first half of this chapter spanned several generations from the mid-fifth to the mid-sixth century. Two highprofile women, Flavia Euphemia and Flavia Anastasia, continue this pattern in the mid- to late sixth century. The first, Flavia Euphemia, leaves a paper trail spanning half a century. In , she appears as Euphemia endoxotat¯e,   

 See above, p. . See Falivene , s.n. Gessias. Hickey and Keenan –,  n.  think that the Apphous in P.Oxy. . is probably not the same as that of the rest of the archive, but might perhaps be a grandson. For Flavia Euphemia, see Hardy , –, P.Oxy. ., ., ., ., P.Mich. ., and PLRE .. The latter entry ignores the Michigan text, which is accepted as referring to the same Euphemia in the most recent treatment of her career; see n.  to P.Oxy. .. The Michigan text is her first attestation, in . Her last attestation, P.Oxy. ., dates to , when she is already dead. Her immediately previous attestations are P.Oxy. . (), P.Oxy. . (s) and P.Oxy. . (); she died therefore between  and , and her estate entered the collective hands of her heirs. In his note to line  of the Michigan text, Sijpesteijn also cited P.Oxy. . for “another” Flavia Euphemia, but I am unable to tell why: of the three Euphemias in the text, none appear as Flavia or have any other status indication suggesting they might have held that rank. P.Oxy. . has received a number of entries in the BL, but none relating to this particular

1 The centralized elite of Oxyrhynchos



the nominal issuing party in an order for a payment of money for meat, issued through her steward to Ioustos, a prono¯et¯es. She appears again, described as the daughter of a deceased Mousaios, in a  lease agreement for a ground-floor room in a house in Oxyrhynchos. The transaction is conducted through Flavius Anastasios, her dioik¯et¯es, and Ieremias, her enoikologos (rent-collector), with Aurelios Stephanos, a baker. Assuming that Euphemia lived in Oxyrhynchos proper and was not renting part of her own residence to a baker, we may conclude that she had several pieces of property within the city of Oxyrhynchos itself. We may know the dioik¯et¯es Anastasios from somewhere else. According to the Prosopography of the Later Roman Empire, Flavius Anastasios, whom the Euphemia lease styles “admired” (per©bleptov), is “Possibly identical with the father of Zacharias (oeconomus of the church of the Holy Resurrection at Oxyrhynchus and u¬¼v toÓ t¦v peribl”ptou mnžmhv %nastas©ou, and a native of Oxyrhynchus)” who appears in P.Oxy. .: “If so, he [Anastasios] was dead by late .” If this identification is correct, then Flavia Euphemia had an employee with family ties to the local church. But the text mentioning Zacharias, Anastasios’ son, has nothing to do with the church at all: it is a deed of surety in which Zacharias vouched for Aurelius Pambechios, an Apionic fruit-grower, before Menas, Apion’s assistant, and Pambechios’ landowner, Flavius Apion himself. If the proposed chain of connections – Flavia Euphemia, Anastasios, his son Zacharias, Pambechios, Menas, Flavius Apion – is correct, it is surely the greatest social distance which could have existed between Euphemia and Apion. We have already seen that it was a small world at the social pinnacle of Oxyrhynchos. Yet here, as with Anastasia and the heirs of Apion on page  below, it is striking that such indirect vertical connections are our best proxy evidence for direct horizontal ties.

 



issue. The text from  is a contract between Flavius Ioannes (see PLRE . s.n. Fl. Ioannes ) and Flavius Iulios concerning the pronoetic jurisdiction of the latter. Euphemia having died, Ioannes has inherited her property. Ioannes also appears in P.Oxy. . as the addressee of a receipt for a cogwheel. P.Mich. .. P.Oxy. .. Mousaios may in turn be the man whose heirs – presumably then including Euphemia – are responsible for maintaining riparii through the house of Theon at P.Oxy. ., where line  records a payment o¬ kl(hron»moi) Mo. usa©ou u¬oÓ Stra. . . . [..] –. po©hsan oÌ(twv), although the identification is not likely; see Azzarello , –. PLRE .. At P.Oxy. ..–, Zacharias is described as o«kon»mov | t¦v –kklhs©av t¦v ‰g©av ˆnast†sewv u¬¼v toÓ t¦v peribl”ptou | mnžmhv %nastas©ou. Anastasios was therefore dead by the date of the text, which the ed. princ. gives as  or , unable to resolve a conflict between the regnal and consular years of Maurice. CSBE  n.  takes the date as  without comment.



Social networks in Byzantine Egypt

Flavia Anastasia, another leading Oxyrhynchite female, shows a pattern of connectivity similar to that of Euphemia. Only a few Anastasia texts have been published, and as many as sixty pieces at Giessen with some connection to her await assembly and publication. Still, a picture emerges from the published scraps. She appears with her sister Maria and other Oxyrhynchite landholders in the L»gov . . . .] politeuom”(nou) dating from the s. The two women receive consecutive entries, lines –, and pay identical amounts, . artabae. From this, van Haelst argued that the sisters inherited their wealth, presumably from their father, Menas son of Eudaimon. (Their grandfather Eudaimon in turn may have been the same man prominent as a comes and Oxyrhynchite landowner in the late fifth and early sixth centuries. ) Van Haelst described Anastasia as the owner of “vastes propri´et´es,” but as Banaji has recently pointed out, “the difference in scale between Anastasia and the Apions presumably deprives the term ‘vastes propri´et´es’ of much meaning.” Nonetheless, she appears in impressive company in the civic councilman’s account just mentioned: there, in the s, we see payments from the divine household and the oikos of the Apions (lines –), and the heirs of Phib son of Matrinos (line ), who himself served as a witness for 



  





For Anastasia, see PLRE .; Montevecchi’s Archive  (P.Oxy. ., P.Bibl.Giss.Univ.inv.  = originally in partial transcription at Eger’s brief note at , ; now van Haelst  and SB .); van Haelst  (SB .),  (SB .), and ; Gascou b,  n. ; Keenan ,  for the possible restoration of Anastasia in the gap of P.Oxy. .; Gascou , – for a more detailed discussion of the same point (he does not cite Keenan, and receives credit for the proposal at BL .); Sijpesteijn  for the publication of what is now SB .; Gonis ,  n. , following Keenan and/or Gascou by citing P.Oxy. .; Hickey ; and most recently, P.Oxy. .–. Hickey is now preparing a complete edition of the Anastasia documents, forthcoming as P.Anastasia, by which abbreviation it is cited herein. P.Anastasia, cited in the previous note. This is an improvement from the situation reported by van Haelst , –, when he reported that  of the items at Giessen were inaccessible. See Hickey  n.  for the location of other Anastasia items. Hickey, now preparing P.Anastasia, remains undecided on whether the items were assembled in antiquity: see Hickey , . P.Oxy. .. For Maria’s relationship to Anastasia, see van Haelst , . P.Giss.Univ.Bibl.inv.  ( = van Haelst , SB .) lines –, where the ed. princ. reads toÓ t¦v –n[d»]xou | mnž. [m]hv M..a. EÉda. ©.monov, with the line notes, where the obvious MhnŽ Euda©monov is proposed; see further argument at van Haelst , –. But for a caution against using cases of female office-holding such as Anastasia for evidence of hereditary tenure, see Gascou b,  n. . (For the most recent attestation of the patronymic, although only via restoration, see Hickey , . Gonis ,  n.  also draws attention to PSI ., “a letter addressed to a Eudaemon, [which] refers to ¡ s• (l. sŸ) ˆdelf(Ÿ) %nastas©a.”) See Gonis , –, although the exact connections remain unproven. For Eudaimon see e.g. P.Oxy. . (; line : EÉda©mwn k»mev), P.Oxy. . (; line : EÉda©monov geoucoÓntov –n t
aÉ. t.
lamprŽ€ ìOxurugcitän p»lei). Banaji , –, used throughout this paragraph. He is citing van Haelst , .

1 The centralized elite of Oxyrhynchos



a settlement involving the Apions in the s. But the payment amounts in that account are revealing. Flavia Anastasia’s assessment, . artabae, was less than one-third the assessment on estates belonging to Ioustos and Ptolemaios: their heirs paid ,. and ,. artabae respectively (lines –). In another civic account, Ioustos and his brothers, and the heirs of Ptolemaios, are in turn assessed at a rate hardly half the size of the Apions. Anastasia is only in eighth place in the first civic account, and must have possessed far less land than her contemporary, Flavia Euphemia. It also seems that unlike the Apions, Anastasia had no property outside of the Oxyrhynchite nome. Flavia Anastasia appears through the mediation of her dioik¯et¯es, the count Flavius Phoibammon. Is the fact that Anastasia’s estate management was in the hands of a count a sign of her importance, of social distance from her holdings? Presumably, having an estate manager of such high rank ensured the orderly running of her affairs. In , Anastasia received through Phoibammon an acknowledgement from Aurelius Onnophris, a brick-maker, that Onnophris would pay her an annual rent on land at Maiouma. Van Haelst was troubled by this text, particularly the lack of specificity about payment methods and the absence of the “as long as you want” (–f’ Âson cr»non boÅlei) clause for the lease’s duration, and found in this sloppiness a lax attitude towards the traditional juridical system. This seems a little forced. Perhaps we are dealing with owners and tenants already familiar with each other, people who, to reuse Hardy’s clever phrase, were “quite well aware of what the system was.” In any event, this acknowledgement connects Anastasia to the Apions. The scribe of the text was Papnouthios, who also signed for Onnophris, 



  

 

See P.Oxy. ., and below, pp. –, on the case of Diogenes. For P.Oxy. .’s date, see Gascou , . But Phib and Anastasia were presumably not exact contemporaries: P.Oxy. . merely gives his name, but P.Oxy. ., dated to , refers to him as “of blessed memory.” P.Oxy. .. gives the Apionic total as no(m.) h k(er.) ie. P.Oxy. ..– gives the assessment of the heirs of Ptolemaios (line ) at no(m.) b k(er.) i{d/, and their contribution for a half-share of the pagarchy (line ) at n»(m.) a k(er.) i
; in lines –, Ioustos and his brothers are assessed under the same divided arrangement at no(m.) b k(er.) i for themselves and n»(m.) a k(er.) i
for the other half-share of the pagarchy. Points first made by van Haelst , . Personal communication from Todd Hickey, July . See PLRE . Phoibammon , where he is tentatively identified with the Phoibammon appearing in P.Oxy. .. (see the apparatus) and ; see also his appearance above, p. . See also SB . (/ or /) and SB . (), with van Haelst , van Haelst  and Eger, cited above in n. . For a possible analogy in the Fayumic estate of Appianus, see Chapter  below, pp. –.  Van Haelst , –.  Hardy , . SB . ( = van Haelst ).



Social networks in Byzantine Egypt

described as agrammatos. This is probably the same Papnouthios who drafted a  loan agreement issued by representatives of the heirs of Flavius Apion. This Papnouthios also appears in  drafting a receipt for a wheat donation to a local hospital, a donation made by the same Apionic heirs. This demonstrates at least indirect social contact between the Apionic estates and Anastasia, through use of the same scribe. But the plot thickens: Maiouma, where Onnophris rented land from Anastasia in , is attested in Apionic hands in an account dating to /, and in an Apionic account from over seventy years before, in /. The Apions and Anastasia were not only connected socially, their properties were near each other, perhaps even intertwined. The two estates used the same notaries. And yet even here, we are unable to draw horizontal connections between the two families, or between the farmers of their holdings, but can only trace indirect connections through social unequals. Anastasia appears again in a lead receipt issued through Phoibammon for work by the lead-worker Pamouthios in either / or /. If a restoration at line  is correct, the receipt was issued by Victor, [pericÅth toÓ dhmos©o(u) l]outr(oÓ), an attendant at the public bath, apparently for work on that bath. (Compare Flavia Gabrielia, whose involvement with the Oxyrhynchos public baths came through her office-holding, particularly as logist¯es; unpublished Anastasia texts connect her to the logisteia as well. We may be in the same context here.) We have already noted the lead-worker Pamouthios providing an indirect social connection between the estates of Anastasia and those of the Apions. Pamouthios himself received payment through an intermediary, the chartularius Anastasios.  



     

P.Oxy. .; Van Haelst , . P.Oxy. .. Van Haelst ,  also cites P.Oxy. ., but must mean , where Papnouthios appears again in an Apionic context. For a list of Papnouthios’ abundant attestations, see Diethart and Worp , –. Todd Hickey informs me in personal communication that he is common in P.Anastasia. Respectively, P.Oxy. . = ., where an account addressed to Flavius Apion is issued through Stef†nou pro(nohtoÓ) Paggoulee©ou sÆn to(±v) Šll(oiv) m”r(esi) (kaª) Ma[r]gar©tou kaª %mbioÓtov kaª MaioumŽ ktl., and P.Oxy. . (to be taken for dating with BL .), where a payment is recorded in line  p(a.) Foib†mmwni pron(oh.) MaeioumŽ. Anastasios: see above, p. , with new appearances in Anastasia texts at P.Oxy. . and  (late sixth century). SB . ( = van Haelst ). Van Haelst suggests that “Il n’est pas impossible qu’il s’agisse ici de la construction du nouveau bain public du Nord a` Oxyrhynchus,” known from P.Oxy. ., from /. Personal communication from Todd Hickey, July . See above, pp. –. SB .. refers to payments do{(”nta) t aÉt() Pamou{© molubourg() d(i‡) %nastas©ou toÓ eÉdok(imwt†tou). Van Haelst ,  suggests that “D’apr`es le titre eÉdok/ . . . trapez©thv, not†riov ou mieux cartoul†riov,” and cites as parallel P.Oxy. . (republished

1 The centralized elite of Oxyrhynchos



The same Anastasios appears in another Anastasia text, which in turn mentions two Phoibammons, one a klao. u. eik(oul†riov) (jailer) and another a džmiov (public executioner?). Earlier authors saw the Pamouthios transaction as evidence of a hierarchical and specialized bureaucracy: Pamouthios received work orders from Anastasia’s estate, which then issued a receipt to the lead-worker, who then acknowledged to the estate that he had done the work required. But if the analogy to Flavia Gabrielia is correct, Anastasia is involved with the public bath in an official capacity, perhaps through the logisteia, and this is not a private matter at all. These chains of connectivity are not, as van Haelst thought, evidence of an inefficient economic system, but are the standard traces of connectivity left by complex bureaucracies. What is rather more interesting about them is the evidence they provide for the Oxyrhynchite elite’s role as a centralizing force in the region’s social network. These texts suggest that the large estates of Oxyrhynchos created social connections across various parts of local society. Indeed, the connections I have proposed here are an excellent example of the role the local elites played at the top of the social pyramid: whether or not Pamouthios the lead-worker and Phoibammon the executioner knew one another, they both had the same wealthy landlady in their mental landscape. family tension and economic difficulty: the case of christodote and kometes The case of Christodote and Kometes is our final example of an elite family with a wide array of indirect social contacts to other elites. In this case, however, it is also an example of an elite family at grave financial risk. Attention to the great estates or to office-holders, particularly to females such as



 

in full in the appendix to P.Oxy. ), where a chartularius appears as the intermediary in a payment issued by the heirs of Flavius Apion. The publication of P.Mich.inv.  appears to have confirmed his suspicions: see next note. Sijpesteijn ,  nn. to lines  and  of what is now SB .. The two Phoibammons appear to be receiving something par‡ soÓ | toÓ a«des©mou %nastas©ou | carto. u. l. a. r©ou t¦v | –ndoxot†thv «lloustr©av | %nastas©av (lines –). Van Haelst , –. For Flavia Christodote, see Keenan , his re-edition of PSI ., and now P.Thomas , which Hickey and Keenan take to be a fragment of another text related to the PSI affidavit. For Kometes, not Kometos, see P.Oxy. . n.  to lines –, referring to an unpublished Oxyrhynchos papyrus giving the correct spelling. To this family now perhaps add a sister named Flavia Maria, daughter of a Ioannes patricius in P.Oxy. . (). Gonis is probably right (n. to lines –) to economize on the number of patricii named Ioannes in this period. Unfortunately, the text itself tells us little more about Maria’s activities.



Social networks in Byzantine Egypt

Anastasia, misleads. Christodote reminds us that not every member of the Oxyrhynchite elite was economically secure and that this insecurity threatened the female elite as much as the male. Beyond two texts which date to  or , Christodote’s existence is nearly unknown to us. The two documents are copies of an affidavit (diamartur©a) “drawn up at the instance of one Flavia Christodote for signing by the defensor civitatis of Alexandria . . . to have been delivered to the party against whom Christodote complains, an Alexandrian banker named Flavius Eustathius.” In addition to being a banker, Eustathios was also a count, although as Keenan points out, “The exact type of Eustathius’s countship is not given and can only be guessed at. It is likely to have been an honorary title, perhaps acquired by purchase.” Christodote’s complaint involved quite a large sum of money. Keenan’s commentary to the text explains that “Christodote, hard pressed by her creditors and in danger of losing her landed property as a consequence, sought to recover money owed to her by her brother, Kometos. Kometos, however, had reached agreement with Eustathius, an Alexandrian banker, to have the banker repay his sister in his behalf.” The sum in question, a surprising  pounds of gold, , solidi, is “more than a fifth of the annual revenue, in solidi, which the Apions extracted from their Oxyrhynchite estates.” That sum is “many times in excess of the defensor’s competence” to handle this case under law. It thus seems strange that she brought the case to Alexandria’s defensor at all; she may “have foregone [sic] the right to approach the imperial court . . . because his [sc. the defensor’s] services would be cheaper and less time-consuming.” Indeed, it is clear that the issue of legal expenses weighed on Christodote’s mind, for twice in the 

 



Discussed in detail for the first time at Keenan , and in earlier form, Keenan b, –. See also Bogaert . The dating of the text has been a source of some confusion: Keenan b,  follows the dating in the ed. princ. of /. Keenan  gives  (?) on the basis of Aelia Sophia’s presence in the dating formula and Tiberius’ absence from it; he is cited at BL .. Bagnall and Worp , – give a solid case for  May  or , cited at BL .. Banaji ,  n.  accepts this dating. Keenan ,  n.  thought it possible that the anonymous lady in P.Oxy. . was Christodote, but preferred Fl. Anastasia. This restoration is now accepted as standard: see above, n. , and BL ., citing Gascou , , and CPR . p. . (Chrysi Kotsifou reports a text at the Catholic University in Washington DC concerning a Fl() Christodot(); this may be a male komes, not our Christodote.)  Keenan ,  n. to line .  Keenan , . Keenan , . Banaji , –, citing P.Oxy. . v (c.), and its figure of , solidi. In their introduction to P.Thomas  (at p.  n. ), Hickey and Keenan phrase the matter somewhat differently, noting that the Oxyrhynchite “rural receipts of the famous Fll. Apiones roughly fifteen years later were only a bit over four times this.” They cite Gascou a, –, and therein at , it is the same figure at stake, , solidi.  Keenan b, . Keenan b, .

1 The centralized elite of Oxyrhynchos



span of two lines she threatens Eustathios with the burden of repaying her expenses in the case. Christodote’s approach to the defensor at Alexandria appears to have failed. We still have two copies of the deposition, both incomplete and still attached to each other. The two texts have estate accounts on the reverse, suggesting they had outlived their usefulness. Perhaps Christodote received her money after all, either from Eustathios or from her brother, who was legally still responsible for the sum owed to Christodote. Or perhaps Christodote simply gave up, daunted by potential expenses. But what expenses would one have to incur to make it no longer worth pursuing such a sum? Christodote claimed to be willing to pursue the matter all the way to Constantinople: Þv eÉtrepis{¦na© me loip¼n katalabe±n | tŸn basile©da tän p»lewn. Presumably she considered taking the case to Alexandria because the banker was from there. But what was the origin of the connection between the banker and her family? The Kometes whom Christodote called her brother may be the same man found in the two Oxyrhynchite city accounts we have discussed elsewhere above. According to Banaji’s analysis, Christodote’s brother, Kometes, turns up in P.Oxy. XVI  (from the s) with a substantial payment of  artabas of barley, among private landholders the fourth biggest in the list (. per cent of the total). In P.Oxy. , cash contributions for the public bath, his cash payment is almost precisely half the level assigned to the Apion oikos and this time the second highest in the list ( per cent of the total payment). Thus by the s (the presumed date of P.Oxy. ) Cometes had become the formal head of an oikos inherited, presumably, from the patricius John. 

 

  



PSI ..–. At line : prof†sei xenite©av poižsw dapanžmata, with Keenan’s commentary at , , where he gives parallels, and notes that travel expenses “would only have been an insignificant fraction of the amount in dispute.” At line : –k s¦v Ëper{”sewv dapanžmata. Keenan b, . Hickey and Keenan describe these texts as “being prepared for publication by Rosario Pintaudi and Hermann Harrauer” in their introduction to P.Thomas  p.  n. . Keenan , , reports only that “the verso, to the extent that I have read it, contains no indication that it is concerned expressly with Christodote’s estates.”  PSI ..–. See Keenan , . On Eustathios’ Alexandrian background, see Keenan ,  n. : the fact must be assumed from context. P.Oxy. . and .. Banaji ,  accepts the identification as a given. Keenan , – thinks it only probable, and also cites Maspero’s proposal that the man was the son of the dux Thebaidos in Ed. , on which see below, n. . Banaji , . For the date of P.Oxy. ., see Gascou a and the discussion above at n. . For the patricius John: PSI ., in which Christodote is described as a daughter toÓ t¦v eÉkleoÓv mnžmhv ì Iw†nnou genom”nou patrik©ou.



Social networks in Byzantine Egypt

Who is this John? Maspero took John to be the endoxotatos Ioannes son of Kometes who served as dux of the Thebaid in . Christodote’s brother Kometes by this argument would be named after their grandfather. If these connections are correct, Christodote and her brother were near the social center of Egypt’s office-holding elite. (Apion himself had been dux of the Thebaid a decade after their father. ) The appearance of Christodote’s brother Kometes in the two Oxyrhynchite city payment lists further links their family to other Oxyrhynchite elite. But as with so many of the links in this chapter, these connections are only indirect, through secondary personnel responsible for collecting taxes and compiling tax lists. The Ioustos who appears in both lists may be the Ioustos son of Eudaimon who appears still alive as a comes in /, and deceased by . (This Ioustos might in turn be an uncle of a woman we have already discussed, Flavia Anastasia, the grand-daughter of a Eudaimon. ) As we have seen, the lady Patricia mentioned in P.Oxy. ., the later of these two lists, is probably the daughter of the same Gabrielia who appears in  as a patrician, logist¯es and proedros, and father of the city of Oxyrhynchos. Gabrielia held her positions “on behalf of the estate of Timagenes of noble memory,” in turn connecting this text to the estates of Timagenes, discussed above, pp. –. Kometes’ appearance in the same list also puts him in the company of Phib son of Matrinos, who, as we will see below, had direct ties to the Apions. Can we speculate as to the size of the landholdings of Kometes and Christodote? Banaji approaches these civic accounts as if they record sums proportional to the size of the holdings themselves. But this is problematic: we have no real way of knowing how the burdens were assigned for any given payment, or whether the assignments changed over time. And 

       

An argument summarized by von Druffel, Munch.Beitr. ,  n.  and noted without approval or rejection by Keenan ,  n. to line ; for Ioannes the dux see Ed. .. and Ioannes  in PLRE .. The PLRE entry for Kometes, Cometas  (PLRE .–), thinks the identification between this Ioannes and the father of Kometes possible, but does not commit. However, the identification is implicitly rejected elsewhere in the PLRE, which describes Ioannes  (PLRE .) as the father of Christodote and Cometas .  See Chapter  below, p. . About whom nothing more is known: PLRE .. In P.Oxy. .. For a summary of these proposals, see P.Oxy. . n. to lines –. P.Oxy. .. See Gonis , –, with Anastasia, pp. – and n.  above. For further discussion, see above, pp. –. See the discussion of the Diogenes affair below, pp. –. Banaji , –. He does not to my knowledge provide an extended defense of this approach elsewhere. And as Hickey and Keenan point out in their introduction to P.Thomas , page  n. , “the accounts concern different charges (ˆrkarik† vs. –gkaÅmata for a new public bath), and it is very possible that the bases of assessment diverged.”

1 The centralized elite of Oxyrhynchos



why is Christodote absent from both P.Oxy. . and . if she was as wealthy as she seems? Hickey and Keenan provide one possible explanation. Following Banaji’s recent definition of oikoi as estates “held in joint ownership and thus immune to the devastating fragmentation of partible inheritance,” they explain Christodote’s absence by virtue of the appearance of Kometes in her place. Both sets of payment accounts contain entries “through the house of the endoxotatos Kometes.” These accounts bracket the Christodote affair, the former dating to / and the latter to the s, after the episode described in the original affidavit. In the later of the two accounts, the contribution of their oikos has dropped from . per cent of the total to . per cent. Keenan and Hickey speculate that the decline in percentage “might have something to do with the seizure of Christodot¯e’s land by her creditors or the severance of her property from the oikos.” If Banaji is right that the payment proportions here relate to the size of the holdings in question, we can infer that the oikos of Kometes declined nearly  per cent in size during the period in question. This figure should be taken with a grain of salt, because we have no way to check whether the other estates in these payment lists grew or shrank during the same period. The decline nonetheless suggests that a closer look at Christodote’s estates is warranted. In her affidavit, Christodote makes two separate references to her property. Keenan took them both to refer to the same property, but this does not seem likely. Christodote first mentions property left to her in the province of Arcadia (¡ Ëpoleif{e±s† moi ˆk©nhtov oÉs©a kat‡ tŸn %rk†dwn), and next mentions property left to her in the city of Oxyrhynchos (tŸn Ëpoleif{e±s†n moi kat‡ tŸn ìOxurugcitän ˆk©nhton oÉs©an). If in the latter case she had meant to refer to the nome of Oxyrhynchos – the substitute one would expect for Arcadia – she would have used some variation of ìOxurugc©thv, the normal term for the nome, not ìOxurugcitän. Moreover, Christodote seemed to think that different fates may yet await the two pieces of property. The property “in Arcadia” is about to be handed over to her creditors: m”llei | –k perist†sewv parado{¦nai to±v danista±v (lines –). But she was unsure whether there will be a risk to the property “in the (city of the) Oxyrhynchites”: E­te (line ) kinduneÅsw    

Quoting Banaji , . For what follows, see P.Thomas , –.  P.Thomas , p. . Quoting P.Oxy. ..; cf. P.Oxy. ...  See Keenan , , in his note to line . PSI .a, lines  and  respectively. While the form kat‡ tŸn %rk†dwn is unattested elsewhere, the conclusion that the province is meant seems unavoidable. Every other attestation of %rk†dwn in the Duke Databank is followed either immediately or shortly thereafter by some reference to the –parc©a.



Social networks in Byzantine Egypt

(line ). This grammatical subtlety may represent the economic priorities of a member of the Oxyrhynchite elite. When hounded by creditors, a local aristocrat is sure to conclude that the property out in the provinces ought to be the first to go. But the property in the city, where Christodote herself presumably lived, could be protected unless the situation grew worse still. This lends further credence to an impression we have throughout the Oxyrhynchite evidence, that the local elites were quite distanced from their rural properties, coming together in an existence centered around the city itself. The small world of that aristocratic city life is highlighted in other ways. The joint appearance of various large landholders in the same civic accounts suggests that these elite Oxyrhynchite landholders were all closely connected. One wonders whether the Apions of the world looked with disapproval at the family troubles of their peers, knowing that their share of the public burden would increase if Christodote and Kometes did not sort out their difficulties. Christodote hints at the public nature of her problem: “The fact that I am wrestling with debts and am hourly harassed by my creditors is known to all.” While Christodote’s brother Kometes clearly had some connection to Alexandria, through his banker, her own affidavit was found at Oxyrhynchos, and was drafted by a notary who did not know the name of the defensor in Alexandria. Those making her embarrassment so acute were no doubt her fellow aristocratic elite in Oxyrhynchos itself. With Christodote and Kometes in mind, we should remember the career trajectory of the family of Eulogios. Beginning with an Aurelius of moderate means, it culminated two generations later in a count of the consistory. While elites such as Christodote faced economic crisis from internal family tensions, one or two rungs down the social ladder, other families were ready to move up. the oxyrhynchos church The Oxyrhynchite church created hierarchical ties throughout the nome in much the same way as the local aristocracy. However, the role of the Oxyrhynchite church as a central node in the nome’s social network has not been adequately studied. Even a survey of the church’s agricultural   

Transl. by Keenan ; PSI ., line : í Oti pollo±v cr”esin prospala©w kaª ka{’ —k†sthn ãran –nocloÓmai par‡ tän danistän, toÓto pŽsin d¦lon ka{”sthken. A point Keenan makes (,  n. to line ) while noting that the office’s frequent turnover may have created the same effect in Alexandria itself. Montevecchi’s description of the archive of the Catholic church of Oxyrhynchos (Montevecchi , ) was short, eight texts only, and not intended to summarize every reference to the Oxyrhynchite church and its activities. This archive includes: P.Oxy. ., ., ., ., ., P.Flor. , P.Berl.Zilliacus , PSI .. Her reasons for including P.Oxy. .

1 The centralized elite of Oxyrhynchos



activities alone would be a major project. Hardy once speculated that the property of the church in Oxyrhynchos might have surpassed even the vast holdings of the Apionic estates, through the accumulation of large numbers of pious donations. Non-agricultural activity among Oxyrhynchite religious institutions created ties with some of the most prominent families of the fifth and sixth centuries. We have for instance three papyri in the same hand from the monastery of Abba Andreas, one of which records the delivery of four mats for the use of the landowner’s bath in early March, . Specific reference to the oikos of Apion on the reverse makes it clear which landowner is meant. The church also fulfilled all the social functions of a major landowner. Consider the example of Aurelius Pseeis of Kolotes, an enrolled farmer for the church: in  he wrote to Petros, bishop of Oxyrhynchos, through Phib, a priest and principal steward (ka{olik¼v o«kon»mov) to acknowledge the receipt through a monk named Luke of machine parts provided by the church in its capacity as the owner of the hamlet of Kolotes. Philoxenos the notary signed for the farmer Pseeis. Here, the church is interchangeable with any large estate. Tenant farmers requested replacement parts for irrigation machines repeatedly in the Apionic estates as well. And as with other chains of connectivity in this chapter, this example of connectivity is vertical in nature, from a peasant farmer up to the bishop, with a priest, a monk, and a notary in between. The chief characters in this text formed other centralizing connections as well. Abba Petros, the bishop, provides us with a wide range of social



 

 

are not immediately clear. (That text has received no entries in the BL.) P.Oxy. . is more interesting in the context of Samuel and Phoibammon, on whom, see above, p. . To this list add the fragment P.Oxy. .. See also Antonini , – for an index of Oxyrhynchite churches with references. Hardy , . For the property and economic activities of the Oxyrhynchite church, see Wipszycka  and below, pp. –. For Oxyrhynchite churches in general see P.Oxy. .–, all of which are lists for unknown purposes of churches and monasteries in the Oxyrhynchite, at least one of which () the editor believes may have come from an Apionic context. The lists are relatively rich:  for example lists fifteen churches and a monastery. See also Antonini , –, and Timm , .–, s.n. al-Bahnasa. With the Apions: see below, pp. –. With the family of Flavius Eulogios: see above, p. . Sijpesteijn a = SB .–. SB . is the text cited here. The other two are also receipts similar in form, all three beginning –d»\{/(h) di(‡) tän monaz(»ntwn) monasthr(©ou) ˆbba %ndr”ou. The first is for “the use of the machine of the topos of Elias outside the gate,” and the third is for “the use of the machine of the stable of the cursus velox.” Reference to a place ›xw t¦v pÅlhv in the first text also reminds us of other documents with Apionic ties. See the similar P.Oxy. ., issued by the monastery of Apa Hierax, discussed below at p. . P.Oxy. .. Kolotes: ˆp¼ | –poik©ou KwlÛtou ktžmatov t¦v aÉt¦v ka{[o- | lik¦v –kklhs©av (lines –). For Petros, see Worp , , and Papaconstantinou a, . See e.g. P.Oxy. . (), and the intro. to P.Col. .; see above, n. .



Social networks in Byzantine Egypt

connections. We have a calendar of sun†xeiv or special processionals to churches throughout Oxyrhynchos in –. The sun†xeiv itself took place met‡ t¼ katel{. (e±n) | –n %lexandr(e±aƒ) t¼n p†pa, or “after the Papa descended to Alexandria.” The editors had believed the man in question to have been Alexandrian patriarch Timothy IV, who they suppose “had come to Oxyrhynchus on his way back from a tour of inspection in Upper Egypt, and started homewards a day or two before Oct. ,” when this calendar of services begins. This identification has been challenged twice since it was first proposed; Derda and Wipszycka have shown that “Papa” can refer to priests on occasion. A consensus is emerging that the “Papa” in question was the bishop Petros. Under this interpretation, the text shows Petros not only traveling to Alexandria himself, but also leading a processional which no doubt contributed to the social unity of the Oxyrhynchite church. Just under forty churches are listed in this processional itinerary, which must have represented thousands of Oxyrhynchites who were part of the Christian community but did not necessarily see their bishop on a regular basis. The calendar of / is thus not only a list of spiritual commemorations of individual saints, but also a list of specific social events, 



 



P.Oxy. .; Delehaye  and now Papaconstantinou b for a re-edition and extended discussion, critical of Grenfell and Hunt for restorations based largely on later calendars, and Delehaye for following those restorations as if they were genuine readings. The dating, however, remains secure: Papaconstantinou b,  n. . For the original interpretation, of the P†pav as the patriarch, despite its potential use applying to local bishops, see Grenfell and Hunt’s introduction, –. That introduction also includes a fascinating survey of the churches in the text, and the saints they commemorated. P.Oxy. , p. . Derda and Wipszycka , –, where this text is not discussed; clearly, the absence of an identifying proper name here means that no ordinary priest is meant. Petros is thus attested as both an Abba and a Papa; see Derda and Wipszycka ,  for a case of an “apa Leon papa.” Challenges to the identification with the patriarch: Papaconstantinou b,  n.  thinks this identification to be “sans fondement,” and Delehaye ,  thinks it is simply a reference to Abba Petros, arguing that the list was drawn up in anticipation of his absence in the capital. Papaconstantinou shares () the conclusion that the Papas is Petros himself. Under this interpretation he perhaps went to Alexandria for the election of a patriarch during the crisis following the death of Timothy III earlier in . As early as the ed. princ. (see p.  for an additional note to line ) Crum suggested that Severus of Antioch was meant, to which the editors responded that “whether Egyptians would refer to him as well as to the Alexandrian patriarch by the title p†pav is doubtful.” At least ten other churches we know from the evidence do not appear on this list: Papaconstantinou b, . Possibilities include their absence due to the bishop’s desire to cover the city territory in a certain way, or their appearance for feasts during parts of the year not covered by this calendar. See also Papaconstantinou , – for extended remarks on the topography of Egyptian saint cults. Her remarks at  are particularly relevant: “Les e´vˆeques, qui avaient l’occasion de visiter sinon tous les lieux de culte de leurs dioc`eses, du moins les plus importants d’entre eux, jouaient ici un rˆole essentiel, par l’interm´ediaire de la liturgie stationnale sur laquelle le calendrier d’Oxyrhynchos apport un e´clairage particuli`erement puissant . . . On y voit l’´evˆeque couvrir le plus grand nombre

1 The centralized elite of Oxyrhynchos



at which the presence of the bishop in the company of parish priests less familiar with his company no doubt helped to strengthen the fabric of the Oxyrhynchite church. By implication we have in this calendar further evidence for the centralization of Oxyrhynchite society, with the evidence providing examples of vertical ties between the bishop and local churches. No corresponding horizontal evidence is forthcoming to link the local churches directly. Here, as in so many other cases, the chains of indirect connectivity lead back to the Apions. The second stop on the bishop’s tour, on Phaophi , is at Saint Serenos for a day of penitence. This Saint Serenos is attested in four other Oxyrhynchite texts, three of them accounts from the Apionic estates. In two of these accounts, from  and , an Apionic prono¯et¯es is responsible for donations to Saint Serenos funded by one grant from Apollos and another from Phoibammon, both sons of the bishop. Another donation entry records an Apionic payment “to the martyrion of Saint Serenos according to the donation of Apollos son of the bishop.” (But which bishop? No name is specified, and from sixth-century Oxyrhynchos, only Petros himself is attested, two to three decades earlier. ) The Apionic responsibility for these payments is presumably due to the presence of Saint Serenos on Apionic property. Indirect family ties between the unnamed bishop and the Apionic estate suggest that any Oxyrhynchite bishop could well have been in contact with members of the Apionic family and employees of their estate; he almost certainly would have made such connections on his trip to Saint Serenos on October , .

 





 

possible de lieux de culte, tout en e´tablissant entre eux une hi´erarchie reposant sur la fr´equence des visites et l’importance des fˆetes c´el´ebr´ees.” With Papaconstantinou see Frankfurter’s  review. For the presence of the bishop at these events as “very likely” on the basis of comparison to e.g. the Roman stationes, see Grenfell and Hunt’s introduction, p. , and Papaconstantinou b, . Papaconstantinou b,  n. . The identity of the Saint Serenos in the calendar with that in P.Oxy. . is taken by Rea, editor of the latter text, as possible, but not definite: see p.  note to lines –. See the parallel entries at P.Oxy. ..– and P.Oxy. ..–: the donation by grant is described as kat‡ dwre‡n of PN. Both of these texts are discussed in detail in Chapter  below, pp. –. P.Oxy. ... Papaconstantinou b,  n. : Saint Serenos “se trouvait sans doute sur la propri´et´e des Apions.” But property where? Hardy ,  had assumed the church to be in Tarouthinos: P.Oxy. .. follows with –n ktžma(ti) Tarou{©nou dika©ou toÓ ktžma(tov) Kotulee©ou. But Rea (cited above, n. ) tentatively suggests Oxyrhynchos itself, pointing out that P.Oxy. . indicates more clearly than P.Oxy. . that Tarouthinos is merely the location of property which serves as the revenue source for the donations in question. See Worp , . Other churches on the calendar list appear in an Apionic context: Papaconstantinou b, , n. to lines –, points to the appearance of Saint Michael in a number of Apionic accounts, e.g. P.Oxy. . and P.Oxy. .a.



Social networks in Byzantine Egypt

The Oxyrhynchite church also created and maintained social ties through dispensing food and drink. We have one fifth-century order from “the holy church” at Oxyrhynchos, presumably the central church, in which a priest named Gregorios ordered Menas the wine steward to make a wine payment to an unnamed bed-maker. A sixth-century order from “the honorable house” (no name is given, but this term refers to the Apionic estates) instructed Pamouthios, the archimandrite of the Homoousion monastery, to pay  loaves of bread to the people of Tarouthinos. Why the Apionic estates would have been in a position to give such an order is unclear. Did the Homoousion monastery have a lease agreement with them stipulating work on estate ovens or payment in bread? The village of Tarouthinos is known to have Apionic ties in other contexts. In any case, the connection seems clear: monastic officials acted as the middle-men between a large estate and (festival?) food for local villagers. The previous example, of the wine steward taking orders from the priest and paying the bed-maker, is equally an example of a religious middle-man in a vertical social chain. Oxyrhynchite monastic communities reveal further hierarchies connecting the nome capital to the countryside. A text dating to  shows the arrangement by which the monks of Berku received an annual donation from the heirs of a lawyer named Gerontios. Berku had been in the Hermopolite nome, but is securely attested in the Oxyrhynchite in the sixth century. In that period, Berku appears more than once in Apionic accounts. Indeed, Gerontios himself had been in the employ of the Apionic estates, and appears in an Apionic account which also records payments to the monks at Berku. But the annual donation appears to be a private affair: the text is addressed by the monk Kollouthos to the heirs of Gerontios through the dioik¯et¯es Anoup. Here the Oxyrhynchite monastic community is part of a social triad, connected both to the large Apionic estates and to the (presumably much smaller) private holdings of the lawyer Gerontios, both of which in turn were connected to each other.

     

 P.Oxy. ..–. P.Oxy. .. See e.g. P.Oxy. ., with a man ˆp¼ ktžmatov | meg†lhv Tarou{©nou at lines –. PSI ..–. For the date of this text, see Bagnall and Worp a, . See also Hardy ,  for what follows. See Pruneti s.n. P.Oxy. .., a payment to±v mon†z(ousi) berkÆ and P.Oxy. .., an entry recording a receipt of payment –n kÛm(aiv) berkÆ kaª Tapekl‡m. P.Oxy. .: Gerontios in line  (t –llogimwt†t() scolastik() Geront© l»g() filotim(©av)) and the monks of Berku in line  (see previous note).

1 The centralized elite of Oxyrhynchos



The story would normally end here, but Papnuthios, the scribe of the text, complicates matters considerably. A dozen different texts give us insight into his career. He appears often in the Apionic archive in the s, when he drafted documents relating to the heirs of Flavius Apion: the small loan to Samuel and Areotes in , the pronoetic contract of Serenos the deacon in , the receipt of a water-wheel axle by Aurelius Ptollion in , and the receipt of a large charitable donation to the hospital of Abbot Elias in . In the following decade, he appears as a scribe in the archive of Flavia Anastasia, one of the most prominent female landowners of the Oxyrhynchite. He was still active in the seventh century, when he drafted several texts on behalf of a later Flavius Apion. The Papnuthios texts attest an interesting range of toponyms: Melita, Episemus, Adaes, Neos, and Paggouleeios, to name a few. We lack pagos numbers for most of these, but at least three toparchies are represented, the an¯o, ap¯eli¯otou, and libos respectively. As with other examples elsewhere in this work, an academic audience may reasonably take pleasure at the sight of a scribe so central in a social network. It is doubtful that Papnuthios had to travel through the nome to make these connections. More likely, the people in these documents came to him. To be sure, his apparent importance is a distortion caused by the survival of certain evidence. Nonetheless, these texts provide us with a minimum level of connectivity against which to create a more general model. If we ask about the mental landscape of the average monk from Berku, or the average farmer from Melita, figures like Papnuthios are important. They are the nodes through which Oxyrhynchites not only interacted with their landlords (e.g. the Apions) but also shared indirect vertical ties with other figures both lesser and greater (e.g. the lawyer Gerontios or Flavia Anastasia). Another example of the complex ties the church built as a major landholder involves a crime against church property. Pantarou and two brothers, Phoibammon and Elias, all three from the same estate belonging to the church, . . . nes (ˆp¼ ktžmatov | [ . . . .]n. hv toÓ ìOxurugc©tou nomoÓ diaf”rontov t
aÉt
‰g©aƒ ka{. olik
–kklhs©aƒ), provided a deed of      

 P.Oxy. .. Diethart and Worp , . P.Oxy. ., which figures in Chapter  below, p. . P.Oxy. .; compare the similar Papnuthios text, P.Oxy. ., now republished in full in the appendix to P.Oxy. .  See above, pp. –.  See Pruneti s.n. P.Oxy. .. See particularly the Aphrodito examples in Chapter , below, pp. –, . P.Oxy. ..



Social networks in Byzantine Egypt

surety to Menas, the epikeimenos (overseer) of the church of the city of Oxyrhynchos. They guaranteed to the church that Onnophris, son of Thekla and the late deacon Pamouthios, will return the gold he stole from the house of Menas in early Mesore, . Some geographic mobility is suggested here: either Menas himself had a house on the estate in question, or Onnophris, who is from that estate ([ˆp¼ toÓ aÉ]toÓ ktžmatov), stole the gold from a house in Oxyrhynchos itself. Either way, the church of the city of Oxyrhynchos had many ties to the estate: it drew deacons from it (we may suppose that Pamouthios came from there, as his son did), it presumably drew income from it, and it turned to its residents to serve as guarantors for criminals who came from it. We can trace another set of landholding ties starting in , with P.Oxy. .: “The present contract was made by the ‘prono¨etes of the holy church’, and was for the management of a whole village, evidently included in the domains of the church; ecclesiastical property was parallel in its administration to that of the semi-feudal houses.” In this text, the most admirable (thaumasi¯otatos) Makarios appears as a church administrator or prono¯et¯es, receiving an acknowledgement from Pambechis, who agreed to fill the post of clerk for Makarios’ administrators (lines –: tŸn | cÛran mis{©ou tän pronohtän) for the village of Sarapion Chairemonos. Apparently Makarios himself was socially removed from Sarapion Chairemonos. He was a prono¯et¯es with other prono¯etai below him, unnamed officials to whom Pambechis will presumably report. This is an interesting glimpse of the ecclesiastical economic hierarchy: a village clerk, a sub-prono¯et¯es, the chief prono¯et¯es, and presumably the church oikonomos and then bishop above him. Other receipts show a comparable chain of connectivity in somewhat abbreviated form. Two laborers from the same village, Sarapion Chairemonos, issued a receipt in  to a thaumasi¯otatos Makarios, whom the editors consider “perhaps the same person” as the Makarios in the text just discussed. The two villagers, Aurelius Georgios and Ieremias, wrote directly to Makarios to acknowledge their receipt from him of wages for unspecified work. No one has noticed the direct connection between these two Makarios texts and P.Oxy. ., from –, a decade earlier. In this text, the same Makarios and the same village appear in a receipt of . carats “for the public dues on the property of Akoutos” (line :   

 Intro. to the ed. princ. P.Oxy. ., quoting lines –. Although the Oxyrhynchite bishop in this period is unknown; see Worp . P.Oxy. ..

1 The centralized elite of Oxyrhynchos



Ë(p•r) dhmos©ou ktžmatov %koÅtou). If these identifications are correct, then we have an interesting glimpse of a career in motion. Makarios appears here, in the early s, as an assistant (bo¯ethos), working through a grammateus named Ieremias and a village headman (mizon [sic]) named Phoibammon. Yet, a decade later, he is “the most admirable” prono¯et¯es. Perhaps his social origins were in Sarapion Chairemonos, and the hierarchy of church land management provided opportunity for advancement. The curious case of Diogenes provides a final example of ties between the Oxyrhynchite ecclesiastical economic network and the larger world of the great estates, a drama interweaving Oxyrhynchite elites with connections throughout the nome. The story begins in , with the previously unknown monastery of Abba Hierax. A receipt records payment of rope for an irrigation machine through Ioannes, a deacon and the monastery’s archimandrite. The recipient of the rope is addressed as “Phoibammon, sub-tenant (?) outside the gate” (line : Foib†mmwni katame©n(anti) ›xw t¦v pÅlhv). But sub-tenant of whom? Ultimately, perhaps the Apions themselves. The term “outside the gate” (›xw t¦v pÅlhv) is presumably an abbreviation of the complete pro†stion ›xw t¦v pÅlhv, which has long been recognized as a palatial Apionic compound just outside Oxyrhynchos proper. The “little orchard at Eutrygius” in this receipt may be one of the orchards described as ›xw t¦v pÅlhv in an Apionic estate account c.. John Rea had argued that because Ioannes was only a deacon when he issued the rope coils, the monastery of Abba Hierax was a small one. With 



   

 

But in volume , Hunt translated Makar©ou boh{oÓ k»mhv (sic) Serap©onov Curžmwnov (P.Oxy. .) as “Macarius, assistant of the village of Serapion son of Chaeremon,” while in volume , Grenfell, Hunt, and Bell translate –n. prostas©aƒ t¦v kÛmhv Sarap©wnov Cairžmmwnov (P.Oxy. ..) simply as “in the management of the village of Sarapion Chaeremonis.” Is the second genitive a patronymic? Does the village name refer to some historical property-owner in the region? With attestations of the toponym going back to the second century ad (see Pruneti s.n.) there is no way to tell. Akoutos is known only from this text (see Pruneti s.n.) but was presumably near Sarapion Chairemonos. Hunt called him simply an “official,” a vague rendering. Grenfell, Hunt, and Bell use “village headman” to render me©zwn throughout P.Oxy. , where it appears often in a late antique context. Was Phoibammon as a headman superior in status to Makarios as an assistant? P.Oxy. ., which Mazza takes to be Apionic, includes a payment to±v me©z(osi) kaª t boh{ Tak»na, which is perhaps in order of precedence. P.Oxy. .. P.Oxy. . (Apa Hierax), but Abba Hierax in P.Oxy. .. For katameinas as sub-tenant, Rea follows Fikhman , –. So Rea in his commentary to P.Oxy. ... See also Pruneti s.n. “Proastion.” The third supplement () of Calderini, Diz.geogr. does not include the site. For a discussion of the word “proastion” in a late antique context, see Husson , –. P.Oxy. .. See Rea’s remarks, cited in the previous note. See the introduction to P.Oxy. ., p. .



Social networks in Byzantine Egypt

the publication of a later text, he was forced to change his conclusions. In a contract between Apion II and the same monastery of Abba Hierax, dating twelve years later, to , the “monastery acknowledged the receipt of the last instalment of a sum of one hundred and thirty solidi as a pious donation and declared that it had no claim against the Apion estate in respect of a piece of irrigated agricultural land.” The land in question had belonged to an Oxyrhynchite landholder named Diogenes during the time of Strategios II, and is called “the irrigated area called Mao situated in the territory of - aw Ophis” (line : mhcanŸn kaloum”nhn M¯ ¯ diakeim”nhn –n pedia. d. i.  ì [fe. w. [v). The estates of Diogenes, presumably including Mao, appear several times in Apionic accounts as “the dikaion of Diogenes.” But what connected Diogenes to both the Apions and the monastery of Abba Hierax in the first place? The story began in Constantinople, where Diogenes met an agent of the Hierax monastery named Theophilos, and borrowed  solidi from him against the Mao estate, . arourae in the village of Ophis. He took another loan from the same agent for  solidi later on. But he had already mortgaged the property in loan to Strategios. On his death, the monks had no legal claim to the property because of the prior mortgage, but appealed to the piety of Strategios. While Strategios was still alive, his agents paid  solidi to the monastery. The settlement after his death deals with the payment of the remaining  solidi and the renunciation of any claims by the monastery to the land itself. Chief players in the affair include Ioseph, described as a monk, priest, and monastery provost; Theodoros, the monastery’s steward; and Menas, acting for Apion as his o«k”thv (literally “slave,” although perhaps something closer to “household manager” ). Secondary characters include two siblings of Diogenes, Apphouas and Klematia, who renounced their inheritance from him, Pamouthios son  

  

 

P.Oxy. .. See the introduction to P.Oxy. .. See also Mazza , – and Hickey , – for a discussion of this text to better determine the meaning of the term “dikaion,” which appears in relation to lands “acquired from former owners” (Hardy , ) in Apionic accounts. Compare with P.Oxy. ..–: mhcanŸn kaloum”nhn tø mø aø wø diakeim”nhn | –n pedi†di ï Ofewv. For which, see the introduction to P.Oxy. ., with references therein, and Hardy , . See Hickey, cited above in n. , on the meaning of dikaion in this context. P.Oxy. ..: diì –moÓ ìIwsŸf mon†zontov kaª presbut”rou, | proestätov toÓ aÉtoÓ eÉag•v (sic) koinob©ou. See also lines , , , . For the term proestos, see Rea’s note to line . Hardy ,  takes it simply as “servant.” For more on Klematia, see Rea’s notes on line , with a correction to a potentially corresponding entry in P.Oxy. ..

1 The centralized elite of Oxyrhynchos



of Philoxenos, who subscribed for Ioseph and Theodoros, and Ioustos the subdeacon and symbolaiographos who completed the document. Ioustos is known from four other documents: we have already seen him drafting a receipt for Flavia Gabrielia in . He also drafted a receipt addressed to Flavius Apion in , an Apionic millstone-cutter contract in , and an Apionic stonemasonry contract in . His presence as the drafting author of the Diogenes settlement thus makes perfect sense: he was involved in this case not because of his ties to the church as a deacon, but because of his ties to the other party, the family and estates of the Apions. The witnesses to the Diogenes settlement are an interesting group as well, providing a number of indirect connections to both Oxyrhynchos and its villages. The first and last witnesses named are deacons of the Oxyrhynchite Catholic church: Samuel son of the late Martyrios, and Theotimos, son of a deceased priest named Alexander. If we were in doubt, these two witnesses are proof of social contacts between the city’s central church and the Hierax monastery, in the city’s western desert (line : –n t libik Àrei taÅthv t¦v ìOx[u]rugcitän p»lewv. ). The other witnesses are more striking: Pamouthios son of Papbaous; a son of Markianos whose name is lost; a Flavius Philoxenos son of Ision, late priest of the city; and Flavius Phib son of Matrinos. In their signatures, each of these people asserted their presence in the process at the request of Ioseph and Theodoros of the Abba Hierax monastery. The witnesses Philoxenos and Phib are known from elsewhere. Although deceased, Phib appears in a fragmentary receipt issued in . He is mentioned again in entries recording payments into the public account made through his heirs in the s. The Apionic estates appear in the same account. Philoxenos appears in P.Bad. ., where he is addressed as a landowner in Oxyrhynchos and Spania in : geoucoÓnti –n t
lampr“ƒ ìOxurugcitän | p»lei kaª –pª kÛmhv Span©av (lines – ). That text is a rent agreement between Philoxenos and his tenant, a Ioannes son of Sarapammon. Philoxenos appears again in , in another rent agreement, between Philoxenos on the one hand and Gerontios of   



See above, p. , and n. . For the other references that follow, see Diethart and Worp , . For Philoxenos, see P.Bad. . (); PSI . (); Gonis b,  for a correction to P.Oxy. .; and Gonis f, – for corrections to PSI .. P.Oxy. .. The receipt is addressed to a count and politeuomenos whose name has been lost. Lines – read, in part, kat”[ba]l. [ev Ëp•r (?) | ? toÓ t¦v ˆr©]s. (thv) mnžmhv fªb Matr©nou d(i‡) toÓ lampro(t†tou) Seržnou. P.Oxy. ..



Social networks in Byzantine Egypt

Spania on the other. A Ioannes son of Sarapammon wrote for Gerontios, who was himself illiterate. Surely the shared connection between Philoxenos and the village of Spania permits us to suppose that the Ioannes son of Sarapammon writing for Gerontios in one lease is the tenant of the same name in the other lease. So Philoxenos had several tenants in Spania, who were in turn connected to one another as well. Philoxenos’ property at Spania appears substantial. At the very least, Ioannes son of Sarapammon described a location there solely by reference to Philoxenos’ great house: the land to be leased was ˆp¼ tän Ëparc»ntwn t
| Ëmän lampr»thti diakeim”nwn –n t
| aÉt
kÛm
–p’ ˆmf»dou t¦v meg†lhv o«ke©av. This story shows the remarkable social and economic reach of Oxyrhynchite religious institutions. First, Rea was correct to revise his opinions about the size of the Abba Hierax monastery. This was an enterprise with enough financial strength to lend substantial sums, and to do so through a business agent as far away as Constantinople. Second, we have traced another example of considerable centralization among the Oxyrhynchite elite. The connections in this story took us from the nome’s western desert to the village of Ophis in the eastern toparchy, and on, indirectly, to Spania, in the lower (that is, northern) toparchy. And yet all of the characters are connected by ties to local elites in the city of Oxyrhynchos itself. Third, part of the importance of the Diogenes settlement is its glimpse into estate formation. This is a clear-cut example of the Apionic estates expanding in or into Ophis, in competition with a monastery attempting to make the same expansion. But this expansion did not happen in a vacuum: for both the Apions and the monastery, the attempt was possible only through 

   



PSI .; for the date, see BL ., citing Bagnall and Worp, CSBE , , for a restoration of the original post-consulate formula. For further corrections to the text, see Gonis f, with n.  above. The Gerontios in question appears to be missing from Cherf . Diethart and Worp ,  s.n. .., not identified with anyone else by that name. P.Bad. ..–. According to Banaji , , Philoxenos “dominated a whole quarter of the village [Spania] with his ‘massive residence.’” P.Oxy. ., p. , commentary to lines –. See Goma`a et al., ,  for the identification of Spania with modern Safaniya; see also Gonis g for a late attestation. Safaniya would have been safely in the lower toparchy: see maps in Rowlandson , xiii–xiv. Spania is not discussed therein. Nor was any identification discussed in Pruneti s.n. Rowlandson forthcoming discusses the as yet unsolved problem of Spania’s prior name; that toponym seems to emerge fully formed in late antiquity. (I would like to thank Jane Rowlandson for providing me a copy of her forthcoming chapter on the Oxyrhynchite.) An expansion perhaps more natural for the Apions than the monastery. Rea points out (commentary to line ) Apionic ties to the village (citing Gascou a,  n. , thus P.Oxy. ., from ) and its presence in the eastern toparchy, “presumably therefore not very near the monastery which was located in the Western desert of Oxyrhynchus.”

1 The centralized elite of Oxyrhynchos



leveraging a prior financial agreement. The root cause of the affair was Diogenes and his ability to get several conflicting loans. We presume he was an Oxyrhynchite, and that his own social connections helped secure these loans and pave the way for the Apionic expansion which followed. Finally, we should note the social connections each side brings to the table. I have speculated that Ioustos the symbolaiographos appears here because of his connections to the Apionic estates. Everyone else was involved because the Abba Hierax monastery wanted them to be: two viri clarissimi, Flavius Philoxenos being one, and Flavius Phib son of Matrinos, himself a comes, witnessed the agreement at the monastery’s request. In other words, when the double-dealings of a financially troubled landholder like Diogenes caused problems for one member of the local elite, others appear on the scene. We can gather that the monks were not intimidated by the Apionic estates, for they went to some lengths to recoup their lost investment, including the recruitment of powerful landowners as social counterweights. Perhaps the monks approached these other Flavii because their social connections to the Apions would provide credibility and assurance to the Apionic estate that the settlement they made would last. conclusion One aim of this chapter was to demonstrate the social proximity of the Oxyrhynchite elite. At some level, it is intuitively obvious that all or most of the great landowners of fifth- and sixth-century Oxyrhynchos would have known or known of one another, but little effort has been made to prove the point in the past. Presumably, one reason is that the connections we can demonstrate are almost all indirect or vertical in nature. Indirect ties between Flavia Anastasia and the Apionic estates passed through the likes of Aurelius Pamouthios, a late sixth-century lead-worker. A deacon named Ioustos provided indirect ties between the Apionic estates and Flavia Gabrielia in the s. In the s, the descendants of Eulogios had indirect ties to the Apions through Serenos the stableman and the Abba Andreas monastery. The point here is not that lead-workers and stablemen connected the Oxyrhynchite elite; these figures surely had a close awareness of each other’s activities already, as the public embarrassment of Flavia Christodote demonstrates. Rather, the point is that our 

See above, pp. –.



See above, p. .



See above, pp. –.



Social networks in Byzantine Egypt

Oxyrhynchite evidence records connections that were considerably vertical and hierarchical in nature. In contrast to Aphrodito, where we might readily imagine the lead-workers or stablemen acting as a group, interacting with other comparable groups, here in Oxyrhynchos they are found alone, with the great landowners everywhere on their mental horizons. Evidence of horizontal connections among the lower social strata of the Oxyrhynchite nome is lacking. Throughout this chapter, I have also stressed that the evidence from Oxyrhynchos suggests a considerable degree of social and economic centralization in our period. The model here is one in which the city of Oxyrhynchos itself was the center of the wheel; individual settlements throughout the countryside were all oriented towards the center like spokes in that wheel. This is symptomatic of a region whose social geography was structured around smaller settlements connected not to one another but to larger villages, which in turn connected to other parts of the nome. This pattern manifests itself in a number of ways in our evidence. In , when farmers of Flavia Kyria needed to replace a water-wheel axle in Chaira, they “went up” to the city to do so. In the sixth century, when the Homoousion monastery received an order to distribute bread to the people of Tarouthinos, the order was received (and presumably issued) in Oxyrhynchos itself. When Apion Theodosios Ioannes ordered his butler to issue wine to the people of Sepho and Kesmouchis in , that order was executed in Oxyrhynchos as well. The skeptic will object that this centralization is likely an artifact of the archaeological provenance of the bulk of our evidence: if all of the papyri we have discussed herein came from the city itself, residents of that city will naturally appear at the center of our picture. But this objection does not get us very far, for many of the connections this chapter documents are inherently exclusive. The very fact that the farmers in Chaira were going to Oxyrhynchos to get their axle fixed means that they did not do so in nearby Pela. Indeed, even estates headquartered at Oxyrhynchos had at times to get their new axles all the way from Alexandria. The fact that Ioannes the baker received money and gifts for the people of Phthochis in Oxyrhynchos itself means that he was not getting those payments and gifts back home    

For these corporate groups in Aphrodito, see below, Chapters  and , passim.  P.Oxy. ., discussed above, p. . See particularly the conclusion to Ruffini .  P.Oxy. ., discussed above, p. . P.Oxy. ., discussed above, p. .  P.Oxy. . (), an Apionic axle receipt. On Pela and Chaira, see above, pp. –.

1 The centralized elite of Oxyrhynchos



in Phthochis. When Philoxenos issued a receipt for a seed advance paid by Flavius Apion to the headmen of Takona, he was in Oxyrhynchos, not Takona. No amount of evidence from undiscovered village archives will change these facts. The picture we have may be incomplete, but what we do have seems clear enough: the Oxyrhynchite nome in our period was economically centralized, with the nome’s countryside bound to the center through socially hierarchical vertical ties. 

P.Oxy. ., discussed above, p. .



P.Oxy. ., discussed above, p. .

chapter 2

The growth of the Apions

introduction The Apions were the largest and most influential of the great Oxyrhynchite families. The growth and spread of their estate through the Oxyrhynchite nome transformed the region’s social geography. Their estate’s economic activities – both land ownership and fiscal responsibility – connected them to a considerable percentage of the nome’s population. I argue below that the Apionic bureaucracy had ties to roughly  per cent of the nome’s population, over , people, at a conservative estimate. How the family gained this level of prominence has never been satisfactorily explained, but the analyses in this chapter provide helpful clues. Network analysis of Apionic toponyms indicates that the Apionic estates did not grow outward from a rural, ancestral core with a specific regional focus. Rather, Apionic places appear everywhere in the Oxyrhynchite topographical network, linked as much to each other as to other places beyond Apionic reach. Moreover, network measurements of centrality show that the Apions were most important in locations that were in some cases relatively obscure. These conclusions, taken in combination, enhance the picture presented in Chapter , that the social and economic ties of the Oxyrhynchite elite created considerable hierarchical centralization throughout the nome, a centralization driven from the city itself. Because the Apionic oikos is the best-documented institution in late antique Oxyrhynchos, if not in all of late antique Egypt, it represents a singular opportunity for exploring the region’s social connectivity. But whether to treat the Apionic oikos as a social unity remains somewhat unclear, because we do not know how the estates came into being. Without network analysis, there is no quantitative way to tell whether Apionic holdings expanded organically from an original center, or were scattered 

2 The growth of the Apions



in little clusters more or less disconnected from each other. Exploring this question of estate formation to some extent continues Jane Rowlandson’s work on the early empire. But there is a bureaucratic twist that Rowlandson did not face in earlier periods: we do not understand the topographical relationship between land the Apions owned and land that simply fell under their fiscal jurisdiction. After sketching a history of the Apionic family, this chapter outlines two ways to analyze the size and growth of their oikos. The first way, and the more traditional of the two, is to measure the social reach of that oikos by taking an approximate population estimate of Apionic territory. The Apionic documents include general accounts listing several prono¯esiai or pronoetic jurisdictions, and more specific accounts listing individual payments within a single pronoetic jurisdiction. By comparing these accounts, it should be possible to extrapolate approximately how many people throughout the Oxyrhynchite nome came in contact with members of the Apionic bureaucracy, which should in turn let us map the relative size and influence of the Apionic house in relation to its peers. This model administrative pyramid will further map the minimum parameters for social connectivity across the nome. The second way is through network analysis, but of a kind quite different from that presented later for Aphrodito. Aphrodito in the sixth century presents a fairly unified prosopographical network, with its largest component including over  interconnected names. With this sort of network, we can identify its most central figures and the cutpoints whose social ties connected different parts of Aphrodito’s network. It might be possible to do something similar for late antique Oxyrhynchos, but not without an Oxyrhynchite prosopography. Such a project must wait for another day. One alternative is to treat topographical network analysis as an analogue for real social analysis. By treating each settlement as a social unit, and by analyzing the attested ties between those settlements, we may use the topographical evidence in Paola Pruneti’s register of the Oxyrhynchite nome as a substitute for the social connectivity of the nome as a whole.   

 

Fikhman and Mazza, discussed at p.  below, argue that Apionic holdings were scattered, but without any quantitative methods to support their conclusions. On Rowlandson’s work, see my Introduction above, pp. –, and below, pp. –. Both pronohs©a and prostas©a are used to refer to these jurisdictions, in some cases within the same document: see P.Oxy. . (), the pronoetic contract of Serenos discussed below, p. . See Mazza ,  for the format of these accounts, with her earlier remarks at Mazza , –, and –, where the “charge and discharge” structure is introduced. See Chapter  below, p. .



Social Networks in Byzantine Egypt

I have outlined elsewhere a network analytical method for using UCINET to analyze Pruneti’s topographical register of the Oxyrhynchite nome. This register represents a data-set linking over  place names through joint attestations in over , papyri. Analysis of this data-set clarifies the hierarchy of settlements in the nome and suggests that each small settlement interacted with its neighbors through the mediation of larger hub villages. Further analysis of this topographical network can establish the extent to which the individual Apionic holdings are connected to one another in the papyrological record. Analyzing the Apionic holdings as a subset of the Oxyrhynchite topographical network shows that the ties between Apionic holdings were no different from the ties between other Oxyrhynchite toponyms. This in turn suggests that the Apionic holdings did not spread outward from some original center, but instead were acquired at a distance, without particular regard to their placement in relation to other holdings. (This conclusion is in marked contrast to the patterns of land acquisition we see in Chapter  for Aphrodito, where land acquisition proceeded through social ties to one’s neighbors.) the family and career of the flavii apiones The Apionic oikos is one of the best-documented economic institutions in Roman Egypt. According to Roberta Mazza’s recent tabulation, the published material from the Apionic archive includes nearly  texts covering over  years, from  to /. The Apionic estates, those of “the noble house,” have long been considered the largest in the Oxyrhynchite nome. The first documented member of the family, Strategios I, was a curialis (town councilman) and curator of Oxyrhynchite estates belonging to the empress Eudocia, particularly in Tampemou, Ptel and Senyris. Despite  





See Ruffini . Now add over  new texts relating to Strategios I, Strategios II, Apion II and Apion III in P.Oxy. , along with full editions of seven texts that appeared in P.Oxy.  merely as descripta. For the earliest attestation of the family, see Azzarello ,  n.  on P.K¨oln inv. . For the full archive: Mazza , –. For related texts outside of the archive: Mazza , –. For the term “archive” instead of “dossier,” see Mazza , . ›ndoxov o²kov: inclusion of this phrase in a document is enough to label it Apionic. Gascou ,  n.  writes that it is the “D´esignation sp´ecifique et exclusive, a` Oxyrhynchus, de la maison des Apions.” Demonstrable counter-examples are non-existent. P.Oxy. .–. See Mazza , , and Hickey , –. The editor of  cites Fl. Strategius I (= ) at PLRE ., but this cannot have been the same person. The latter Strategios is attested in , but that of P.Oxy. .– comes from the middle of the fifth century, a generation or more earlier. Flavia Isis, daughter of Strategios I, helps bridge the gap: see Mazza , , with the stemma in the Appendix, and the stemma included here at Figure , p. .

2 The growth of the Apions



these relatively minor jurisdictions, the account of riparius-duties we discussed in Chapter  (see p. ) suggests that the family was predominant among its peers as early as the middle of the fifth century. That account dates to the s, but records distribution of responsibilities for providing riparii going back to the middle of the fifth century. The Apions handled four of the fifteen years of duty in that account’s first full indictional cycle (–), five in the second cycle (–) and nine in the next two cycles. For the region to be relying on the family’s financial strength at such a relatively early date suggests even earlier origins for the family than we currently imagine. Apion I, perhaps the son-in-law of Strategios I, is the best-attested member of the family in the literary evidence, where he left a considerable trace. A man of consular rank, Apion I served the emperor Anastasius in a high capacity in the Persian campaign of . His son, Strategios II, appears to have had a close relationship with the emperor Justinian. In , Justinian chose him as his personal vicar, presiding over a church council aimed at reconciling the orthodox and monophysite factions of the eastern church. He served as count of the sacred largesses for five years, from  on, appearing in that capacity in several of Justinian’s surviving laws, and earning notice by Procopius as “a man of wisdom and noble birth.” He even participated in the reconstruction of the Hagia Sophia. Apion II, the son of Strategios II, held the consulship in , one of the last private citizens to be granted the honor. There is some indication that the consulship had originally been intended for his father, and that Apion  



  



P.Oxy. ., with Azzarello , , calling P.Oxy. . “cos`ı la pi`u vicina alla nascita della domus gloriosa.” See the indictional chart at Azzarello , . For possible family members from an earlier period, see Gonis’ introduction to P.Oxy. ., where another high-ranking Strategios from the fifth century is discussed. P.Oxy. .’s utility in identifying the family’s early importance was noted at Gascou , . In e.g. Malalas, Marcellinus Comes, Procopius, Theophanes. See the references in Gascou , ; PLRE .–; and Mazza , –, with a full treatment in Mazza , . See also Hardy , , and the treatise dedicated to “Appion” mentioned in John of Beth Aphthonia, Vita Severi, PO .. For Apion I’s relationship to Strategios I, see Gonis a, , expanding on remarks at P.Oxy. . n. . Hickey ,  n. ; see Preger ,  and –, where an anonymous description of the construction of the Hagia Sophia describes Strategios as ¾ toÓ basil”wv ˆdelfopoiht»v. A first-hand account of this meeting, including a speech by Strategios, still survives: Innocent of Maronea, Epist. de coll. cum Sev. – (= ACOec. ., p. ). Procopius BP ..: Strathg© te patrik© ˆndrª kaª tän basilikän qhsaurän Šrconti, Šllwv d• xunet kaª eÉpatr©d
. For the laws, see Hardy , – and Just. Nov. , Just. Nov.  epil., Just. Nov. . Gascou , .



Social Networks in Byzantine Egypt

received it instead out of courtesy when his father died. His life seems to have been the time of greatest growth for the family’s estates. Following Apion II’s death, probably in , the Apionic estates were not owned by any one individual, but functioned simply in the name of “the heirs of Apion” until . These heirs included Apion II’s son Strategios III; the latter’s daughters after his death; his sister Praiecta; and her children Apion III and George. In / those children were old enough to appear in their own right, both addressed as honorary consuls. Apion III appears with his mother Praiecta as joint owner of the family estates in Oxyrhynchos, and then as sole owner from  on to c.. The so-called pseudo-Strategios who married Praiecta, Strategios Paneuphemos, belongs more properly to a discussion of the elite in the Fayum, despite early appearances in the Oxyrhynchite in the s. Well attested between  and  in Herakleopolis and, predominantly, Arsinoe, he appears there as a patrician, honorary consul, pagarch, and strat¯elat¯es. He is one of only two patricii attested from that region. Flavius Strategios IV, son of Apion III and grandson of Strategios Paneuphemos, is the last known member of the family. Modern scholars have noted that he appears at the start of the seventh century in the letters of Pope Gregory the Great. 

 

  







Hardy , , arguing on the basis of the Chronicon Paschale’s reference to “Apion, son of Strategios” under Olympiad , when the other consuls are not given patronymics; and on the basis that Apion does not appear to have held imperial office at the same level as his father. See Hardy , , and p.  below with Table  for the chronological distribution of Apionic texts, growing considerably during the period of Apion II. Here I follow Palme a, –, a new outline of the aftermath of Apion II’s death. Both Hickey , – and Mazza , –, accept his interpretations. For an alternative explanation, see Beaucamp , –. For the dating of Apion II’s death, see the remarks by Gonis in the introduction to P.Oxy. .–. For a survey of the documents concerning these Apionic heirs, see Mazza , , with the remarks on terminology by Gonis, P.Oxy. .-. P.Oxy. .; see also BL . and Mazza , –. With Praiecta: P.Oxy. .. Sole owner in : P.Oxy. .. Dead by : P.Oxy. . (t –nd»x o­k pot• %p©wni | toÓ makar©tou genom”(nou) patrik©ou). Paneuphemos: Palme a and Palme a, with the historiography at Palme a, –; – for his career; and Palme a, – for the relationship between him and the larger Apionic family, particularly as regards the division of property. See also notes in P.Oxy. .; Banaji , ; Mazza , –; Worp , ; R´emondon a, ; and a clear, concise discussion at Gascou , –. For the reconstruction of one text in particular, see Worp , –. See the list at Mazza , –: undated texts include items from the early seventh century and one from –. Only the Oxyrhynchite attestations of this pseudo-Strategios date earlier than . P.Vindob.G  is an unedited outlier attesting to the survival of his house as late as : see Palme a,  n. . On the topography of his holdings, see Palme a and Banaji , –, and . Banaji () calls the “archive of pseudo-Strategius III, potentially one of the most valuable . . . also one of the least unified and most fragmented.” See his Appendix One, Table  for a list. Worp , ; Gascou , ; Mazza , ; Mazza ,  and , with bibliography at n. . See Greg. Reg. ep. VIII  () and Reg. ep. XIII  ().

2 The growth of the Apions



The recent publication of P.Oxy. . and  gives two more texts from the Apionic archives in the s, after which the fate of the estate is unknown. quantifying the apionic population This brief biographical sketch of the family masks the vast social reach of the bureaucracy under their control. By the mid-sixth century, the Apionic “large estate” was starting to reach its fullest form. During Apion II’s life, nearly eighty Oxyrhynchite toponyms make their first appearance in the estate archives. But what was the nature of Apionic control over these places? Since Gascou’s work on “les grandes domaines” over twenty years ago, most scholars have accepted his “fiscal shares” model, in which the Apions were responsible for tax-collection not only on their own property, but on numerous other holdings throughout the nome. With this model in mind, the historiographical image of large estate owners as responsible for most of the nome may represent something close to the truth. If we do not ask how many Oxyrhynchites lived on land the Apions owned, but how many Oxyrhynchites lived on land under the Apionic fiscal share, the number might have been very high indeed. In calculations below, I argue that at least , people, if not more, were directly accountable to the Apionic estates. Since fiscal responsibility in the ancient world implied some social interaction – in contrast to the modern world, in which we typically do not meet those to whom we pay taxes – our population estimates for the Apion’s fiscal share shape our understanding of social connectivity throughout the Oxyrhynchite. The larger the Apionic population, the more steep and narrow the Oxyrhynchite social pyramid would have been at the top. The higher the percentage of the population financially responsible to the Apions, the fewer elite would have been needed for fiscal shares covering the rest of the population. The larger the number we derive here for an “Apionic population,” the greater the nome’s centralization would have 

 

P.Oxy. . () is addressed to t –nd»x o­k pot• %p©wnov in line . P.Oxy. . () is a less certain case, but mentions Partheniados, which was Apionic four years before in P.Oxy. ., and refers in line  to ïExw t¦v PÅlhv, site of the Apionic proastion, on which, see n.  in Chapter  above. See also P.Oxy. . and . See below, p. . Gascou . The major exception is Sarris , against which see Ruffini forthcoming c. While I dispute Sarris’ proposed bipartite nature of the Apionic estate, whether his conclusions are right or wrong would do little to alter my model. As I indicate below, p. , a peasant’s social tie to the Apionic bureaucracy would exist whether the Apions owned the land in question or merely had fiscal responsibility for it.



Social Networks in Byzantine Egypt

been. The estimates we produce below, in which the Apions and the great estates held combined fiscal responsibility for perhaps as many as , people, suggest an extremely steep nome-wide social pyramid and confirm our impressions from Chapter . The surviving accounts from the Apionic estates provide the evidence necessary to estimate the size of the Apionic bureaucratic reach. By size, I mean neither land area nor sheer agricultural output, issues that have been tackled already, by A. H. M. Jones long ago and Todd Hickey more recently. Their conclusions are discussed below. Instead, I am interested in how many men and women in the Oxyrhynchite nome were in some way financially accountable to the Apions. In network analytical terms, how many Oxyrhynchites had an Apionic employee in their ego network, or primary zone of contacts? (Nome-wide land area or agricultural output figures do not provide helpful parallels here: percentages of total land under Apionic jurisdiction might prove quite different from percentages of total population.) This approach makes no distinction between people living on land that the Apions owned directly, and those living on land for which the Apions merely assumed tax-collecting responsibilities. Indeed, I am not sure to what extent the Apionic accounts make such a differentiation possible. For my purposes, it makes little difference: one either paid money and grain to Apionic officials, and had social connections to them, or one did not. As I have indicated, the two most significant discussions of the size of the Apionic estates were A. H. M. Jones’ efforts in his Later Roman Empire, and Todd Hickey’s recent response. Jones was interested only in the physical size of the Apionic holdings, and his method of calculation was simple. Pointing to a sixth-century text putting the rate of the embol¯e at roughly . artabae to the aroura, and noting a text putting the Apionic contribution from their Oxyrhynchite and Cynopolite oikoi at almost , artabae, Jones concluded that their estates covered over , arourae, which amounts to some , acres, or  square miles. This number seems unreasonably high, amounting to roughly  per cent of the total land in both nomes, but Jones did not question it. Hickey, interested instead in the size of Apionic vineyards, estimated that Apionic vineyards amounted to little more than  arourae. Reasonably   



Jones ,  and Hickey , –. See also Sarris , , with a conclusion similar to that of Jones, that the Apions “owned at least a third of the cultivable land around Oxyrhynchus.” Citing P.Cair.Masp. . and P.Oxy. .. Indeed, as Hickey ,  points out, it seemed to Jones a reasonable affirmation of the proto-feudal model still widely accepted at the time. See Hickey , – for the figures, citing Bagnall , – for nome land areas. Hickey , –.

2 The growth of the Apions



doubtful that “a family of such power and wealth would have invested so little in capital-intensive agriculture,” Hickey reasoned that Jones’ figure must have been grossly high of the mark. The error stemmed from the assumption that the Apions paid taxes only on land they owned, a notion Hardy had already challenged years ago. With Gascou’s model of fiscal shares in mind, it seems much more likely that what Jones had calculated, if anything, was a proportional measure of the Apionic estate’s nome-wide collection burden. Hickey’s much lower figure for ownership makes sense: by way of comparison, we should note that even the largest landholding institution in Aphrodito, the monastery of Apa Sourous, owned only a few hundred arourae. Neither Jones nor Hickey concerned themselves with the size of the populations involved. To estimate this Apionic population, I turn to a handful of the chief Apionic accounts. These accounts are difficult documents to read. As Hardy put it, “some Apion accounts show very little trace of any order at all.” The following survey of these texts will focus less on prosopographical connections and more on population figures. Individual pronoetic accounts give a good sense of the size of some of the hamlets in question; other accounts give a sense of how many such settlements were part of the standard prono¯esia; still others tell us how many of these jurisdictions were managed by the Apionic bureaucracy as a whole. From these figures, and comparisons between the Apions and their peers, come the means to construct an outline in broad strokes of the social connectivity of the Oxyrhynchite nome as a whole. I will organize this discussion on the following lines: (a) the nome-wide texts, listing payments from various large estates; (b) the Apionic estate-wide texts; (c) the Apionic pronoetic texts. Thus the discussion will proceed from the largest sphere to the smallest. A. The nome-wide texts For a population estimate of the Apionic estates to be most useful, we must know how the Apionic estates compared in size to the other Oxyrhynchite   



 Hardy , , cited at Hickey , . Hickey , . Cf. the remarks at Mazza , . See Hickey ,  for something different: he calculates the size of the Apionic permanent workforce. P.Oxy. . allows him to calculate the average wine distribution per worker, and P.Oxy. . gives the maximum amount of wine distributed to the workforce. By this calculation, there were  individuals in the permanent workforce. Hardy , . Part of the confusion, discussed further at Hardy , –, is the Apionic accounting “aim of altering the credit side as little as possible from year to year,” leading to inclusion of tax remissions under the expense side of an account, and so forth.



Social Networks in Byzantine Egypt

estates. Here, I follow Hardy and Gascou in arguing that the Apions had fiscal responsibility for one-quarter to one-third of the total collectively shared by all Oxyrhynchite great houses. A standard approach to this problem, going back at least as far as Hardy, begins with Oxyrhynchite accounts listing payments from the Apions and the other major oikoi. One such account describes “a pair of muslins bought and given to the praetorium on account of the Persians.” The entry records the Apionic contribution of . carats out of a total of six carats (ˆp¼ (kerat©wn) m”r(ouv) toÓ –nd»x(ou) o­k(ou) ker(†t.) a
h/). Hardy wrote that it was “perhaps too much to conclude from this that when a fund was raised from all Oxyrhynchus the proper contribution of the Apion estate was about a quarter.” As we shall see, in other cases this ratio is not far from the truth. Two overlapping civic accounts from the second half of the sixth century record payments from a number of prominent Oxyrhynchite elite. P.Oxy. . dates from before  and after P.Oxy. ., which was probably from /. P.Oxy. . is described in its first edition as an account of “sums received as an adaeratio” by an unspecified urban official. The payments were made for specific purposes, in the case of P.Oxy. . for a public bath. These texts do not help us measure the Apionic estates themselves, but they do help us gauge how they rank in relation to other households, by size of payment. With this approach in mind, P.Oxy. . provides the chief exception to the evidence awarding the Apions nearly a one-third responsibility in such contexts. The first five full payments in grain listed in column two of P.Oxy. . are shown in Table . If the noble house is the Apionic estate, their payment is mysteriously small. But what does the reference to t‡\ v/ prokeim”nav (“the abovementioned”?) mean? Presumably the Apions made a number of previously recorded payments into this fund, perhaps divided among their various pronoetic jurisdictions. The Apionic estate had at least sixteen such  



 

Gascou , –. See Sarris  (n.  above) for a dissenting approach. P.Oxy. ..–, with the transl. of the ed. princ.: (Ëp•r) sindon. ©.(wn) z[u]g(oÓ) a ˆgorasq(”ntov) | kaª doq(”ntov) e«v t¼ praitÛr(ion) l»g() tän Persän. See Hardy , ; Liebeschuetz , ; and Laniado , , who uses this text to measure proportions of curial contributions to civic finance in various cities in the eastern empire. The date is debated: see BL .; Gascou ,  n.  citing Carri´e  (/); Mazza ,  with n.  (/). P.Oxy. . and .. Both texts appear throughout my discussion of the Oxyrhynchite elite in Chapter  above, passim. For a neat summary of P.Oxy. .’s payment amounts, payers, and percentages, see Alston , . See also Gascou b,  for that text’s implications for the nature of the pagarchy, on which see Chapter  below, pp. –. Gascou a,  n.  following van Haelst ,  for these dates. Skipping line , a payment of . artabae from the divine house, Ëp•r toÓ ktžm(atov) Mon©mou, a trivial sum in comparison to the others around it.



2 The growth of the Apions Table  The payments in P.Oxy. 16.2020 Line

Payer

Amount

    

The divine house The noble house, t‡\ v/ prokeim”nav The holy church The heirs of the most noble Ptolemaios The heirs of the most noble Ioustos

 artabae . art.  art. . art. . art.

divisions (see below, pp. –). If the payment in line  was representative of the other unrecorded Apionic payments, we would arrive at an Apionic total of approximately  artabae, which would put their total payment in the same order of magnitude as the holy church, and the two payments by the subsequent endoxotatoi. Nonetheless, the divine house made deposits over twice the size of the next largest landholder. If these payments are indeed any indication of the relative size of Oxyrhynchite large estates, the message is a cautionary one: the Apionic estates would then be dramatically overrepresented in the textual evidence when compared to estates like those of Ptolemaios and Ioustos. P.Oxy. . provides the basis for a similar comparison between the respective payments of the large estates, and produces an Apionic proportion close to one-third. The list of contributions for public bath fuel includes some of the Oxyrhynchite’s most prominent landholders, as well as two entries for “a half share of the pagarchy,”
m”r(ouv) t¦v pagarc(©av). Gascou has argued that the shares in these entries do not represent geographic areas for which each pagarch remained accountable, but that this account instead represents “un cas de fractionnement de l’autorit´e pagarchique.” If he is right, in Oxyrhynchos, the institution of the pagarch had become depersonalized and replaced with a college of local elites, perhaps as a way to limit “les risques encourus par chaque maison en cas de d´efaut dans le service.” In Table , the imperial house is missing, and the Apionic estates pay nearly twice as much as the next closest contender. (As Gascou points out, this is just under one-third of the total.) The payments recorded by Ptolemaios and Ioustos are not as close in size as they were in P.Oxy. ., and figures like Kometes appear here despite being missing from  

Gascou b, , emphasis in the original, against the case of van Haelst . Gascou b, . See Sarris ,  for an alternative theory, that the Apions held hereditary claim to the nome-wide pagarchy. For the situation in Aphrodito, see Chapter , pp. –.



Social Networks in Byzantine Egypt Table  The payments in P.Oxy. 16.2040 Line

Payer

Amounts

       

The noble house The holy church The noble Kometes The noble Ptolemaios The noble Ioustos Valerius comes The most magnificent Euphemias The admirable Theodoulos

 solidi and  carats  solidi and . carats  solidi and  carats  solidi and . carats  solidus and . carats . carats  solidus and . carats  solidus and . carats



This table follows Alston’s analysis of that text at Alston , .

that other list. More curious is the text’s note that the account covers the entire city (line : p†sh. v. p»lewv). This and other details led Gascou to think of these landholders as members of “une sorte de coll`ege [qui] se partageant les responsabilit´es fiscales et liturgiques,” here paying for the public bath on behalf of a larger group of otherwise unnamed landholders. The account from the s of riparii maintained by Oxyrhynchite elites in the name of the house of Theon also gives a sense of nome-wide scale. The families therein – the Apions, Leontios, Philoxenos, and Mousaios – do not, with the exception of the Apions themselves, overlap prosopographically with those listed in the two civic payment accounts just discussed. The riparius account covers sixty-five years of responsibility for these parties spread over eight consecutive indictional cycles, and lists the proportion of those years when each party was responsible for the riparii in Theon’s name. (Who was responsible for the remaining portion of each cycle is never stated.) The percentage of each cycle for which the Apions were responsible fluctuates from one indictional cycle to the next, ranging from just over  per cent to  per cent. In the third and sixth cycle of the account, the Apions were responsible for precisely one-third of the years. Here, we are again in the range Gascou proposed of Apionic fiscal percentages coming close to one-third. In each case, these texts do not record direct ratios of wealth, but percentages of fiscal responsibility. We do not know how closely the size of a family’s fiscal share corresponded to the extent of its landholdings. These   

Gascou , . P.Oxy. ., also discussed above, pp. – and in Chapter , p. . Gascou , –.

2 The growth of the Apions



lists provide a snapshot of the Oxyrhynchite elite, but since the names change from text to text, we see only one portion of that elite at a time, and thus have no way of knowing how many others there may have been. Moreover, in Chapter  (pp. –) we discussed the case of Christodote and Kometes, in which a major family seems to have taken a considerable blow to its wealth. Thankfully, for our purposes, the relationship between the size of a family’s wealth and the extent of its fiscal responsibility is a question we do not need to answer. The following sections, which outline our population estimate of the Apionic fiscal share, need rely on only one fundamental number. If the more specific proportions in P.Oxy. ., ., and . are at all in keeping with the Apionic fiscal share more generally, that share accounted for between one-quarter and one-third of the entire amount shared by all the Oxyrhynchite large estates. B. The Apionic estate-wide texts These nome-wide texts show the scale of the Apionic fiscal share relative to the other major oikoi. But how large was the estate itself? In absolute terms, the Apionic estate income was impressive in its own right. The gross Apionic receipts in gold recorded in  were a fraction over , solidi. The same text also recorded the amount the estates paid to the bankers in Alexandria in taxes, just over , solidi. The net profit of the Apionic estates in this text, the second figure subtracted from the first figure, is surprising: the surplus of income over expenses was , solidi. As Hardy pointed out, this was “over four times the supposedly generous salary of the augustal duke of Egypt.” A receipt of money to be taken to Alexandria forty years later () provides further figures for Apionic payments to the imperial government in tax revenue. The author of the receipt acknowledged that Ioannes the banker had given him , gold solidi to take to Alexandria, and that he would pay it to Ioannes and Simeonios, money-changers, and bring a receipt from the agent (ˆpokrisi†riov) Theodoros. Discussing this text,     

P.Oxy. . verso col.  line . See Maresch , –, with remarks on dating at –; Gascou a, ; BL .. P.Oxy. ., verso col. , line . Hardy , . See also Hickey’s comment (, ) that the Apions took over  per cent of their after-tax net from lease-income. P.Oxy. .. Mazza ,  assigns the text to the Apionic archive by virtue of prosopography and archaeology. See also Maresch , –. PLRE . s.n. Ioannes ,  s.n. Symeonius , and  s.n. Theodorus .



Social Networks in Byzantine Egypt

Hardy noted that the base sum (, solidi) is exactly thirty pounds of gold. Since there were three annual collection periods, this suggests a flat fee for the Apionic estates of  pounds or , solidi for the full year, a figure remarkably consistent with the , solidi the Apionic estates had paid forty years before. The Apions relied on a system of prono¯etai to collect these sums. A prono¯et¯es was an administrator or steward who contracted to collect crops and money payments from estates and villages under the Apionic jurisdiction. There seems to have been no particular uniformity to the number of settlements each prono¯et¯es was responsible for; presumably the size of each settlement affected the level of burden. A deacon named Serenos, who agreed to serve as a prono¯et¯es in , managed the estate of Matreou and the villages of Episemou and Adaiou. Theodoros, an Apionic prono¯et¯es in the s, collected revenue for seven separate hamlets. In that capacity, he made an annual payment to the Apions of  gold solidi, a collection much larger than those of other prono¯etai. He did not make this payment all at once, but in three separate installments over the course of the year. An account from / suggests that Theodoros was engaging in a common practice. This account, a master list of sorts, records revenue collected by a significant number of Apionic prono¯etai, bo¯ethoi, nomikarioi, and others, in some cases detailing their geographic area of operation. The payments each prono¯et¯es made to the Apionic bankers show considerable variation in scale, ranging from a handful of gold solidi (e.g. at the low end, Ioannes in line ) to those of forty solidi or more (e.g. Victor in line ). Some prono¯etai appear only once, while others make payments on behalf of many different settlements. The “account seems to have been written 



  



Hardy , –, where the author adds that “we do not know how much of the estate this covers and cannot say whether the tax was light or heavy.” Can it really have been for only a portion of the estate? This consistency is all the more striking in light of the high number of new Apionic toponyms attested in this forty-year period; see below, pp. –. It is unclear why the Apionic assessment should have remained unchanged in this period of apparent growth. No doubt new toponyms entered the oikos while others left (see below, p. ), while others simply took some time to appear for the first time in our records.  P.Oxy. .. P.Oxy. .. P.Oxy. .; see the ed. princ.’s commentary to lines –. P.Oxy. .. For the date, see Gascou a, –. The relationship between  and P.Oxy. . is complex. P.Oxy. . begins with a summary of P.Oxy. ., starting in column two and finishing in column one of the recto, and including a similar summary in two columns on the verso. Payments such as those recorded at P.Oxy. .. (no(m.) nakd/ mh/ v/) are exceptions, not being through an individual but through a group of preceding individuals: p(a.) to±v —x¦v –ggegramm(”noiv).

2 The growth of the Apions



at different times; numerous entries have been struck out . . . [with] many alterations in figures.” This in conjunction with the many payments may suggest that it was a running summary compiled over the whole year. When Theodoros made several annual payments to the Apionic bankers, he was not at all alone in this regard. Hardy claimed to count twenty prono¯etai in this text, but this is a high number. By my count, the account mentions twenty-five payments from prono¯etai, six of whom are specifically labeled as identical to others (t aÉt()); three others are evident homonyms, leaving us with sixteen prono¯etai working for the Apions. Hardy may well be right to argue that there were “probably many more.” A closer look at the accounts kept by individual prono¯etai will tell us how each of them operated. First a summary of what we have seen so far: Oxyrhynchite accounts in Section A above suggest that the Apionic oikos held fiscal responsibility for roughly one-third of the total share of the Oxyrhynchite great houses. Apionic accounts in Section B herein suggest that the Apions collected that onethird through the collective effort of at least sixteen prono¯etai, if not more. Those prono¯etai in turn paid the oikos by installment, perhaps for a single settlement at a time, or spread out over the course of the year. With some   



Intro. to the ed. princ., P.Oxy. .. Hardy , , followed by Hickey , . I count from P.Oxy. .: .ri, Anoup, Ioustos, Serenos, Papnouthios, Ioannes, Ioulianos, Abraamios, Philoxenos, ]ro, Theodoros, Victor, Kuriakos, Phoibammon, Pamouthios, and Paulos. Hardy appears to have assumed that the Anoup in lines  and , the Pamouthios in lines  and , the Ioustos in lines  and , and the Kuriakos in lines  and  were all distinct. This seems to me somewhat unlikely. In any case, the figure of sixteen prono¯etai I use here is intended to be a conservative figure. If Hardy was right, and twenty or more is closer to being correct, my argument for a high degree of bureaucratic centralization throughout the nome (summarized below, p. ) is only strengthened. P.Oxy. . is a similar but highly fragmentary account, organized into two districts or dio©khsiv. Although the lacunae prevent certainty, this account appears to list eight prono¯etai in one district, and at least seven, although probably more, prono¯etai in the other district. The similarity in total prono¯etai between this text and P.Oxy. . is striking, and lends confidence in the figure. My argument here assumes that the account’s peculiar form, with entries beginning with para followed by personal names in the dative, denotes payments from individuals, not to them. This is ambiguous. In the ed. princ. to P.Oxy. ., the editors comment on the unusual usage, and cite other examples. Despite the obvious lžmmata heading in that text, the editors claim that “it is evident that the amounts so preceded are to be regarded as expenditure, being a reduction from receipts.” Nothing is quite so evident. To the best of my calculations, the lines – using this construction do not represent the difference between the total in line  and the sub-total in line , nor do they represent the remainder paid by Ioannes in line . The total lžmm(ata) given can only be reached (and even then, not quite) by including these figures as receipts, not payments. As other accounts amply attest, Apionic prono¯etai show themselves abundantly capable of paying their expenses out of their own accounts, without central input. Indeed, it is hard to see what payments other than salary the prono¯etai would have received, and the payments in P.Oxy. . cannot be for salary, as the varying number and size of the payments makes clear. Hardy , . See below, p.  for the implications of this possibility.



Social Networks in Byzantine Egypt

sixteen prono¯etai responsible for one-third of the fiscal responsibility of the great houses, this leaves each prono¯et¯es by crude calculation in charge of an Oxyrhynchite fiscal share of roughly /th or just over  per cent of the total of the great houses. C. The Apionic pronoetic texts These pronoetic jurisdictions were variable in size, by my estimates ranging from a few hundred people to over a thousand. Roberta Mazza’s recent tabulation of the Apionic prono¯esiai gives ten known jurisdictions and the texts attesting them. Mazza assigned letters, used in Table  below, to each of these regions. They proceed generally speaking from the southern end of the nome to its northern end. This list will form the basis of my estimates of the population under Apionic fiscal jurisdiction. But this is by no means a complete list. As I have mentioned, the Apions had as many as sixteen prono¯etai working for them at once, if not more. Hickey has written that “the average prostasia had five epoikia.” A pronoetic jurisdiction might have other holdings as well: kt¯emata, portions of a dikaion or foundation, and so forth. To map the bureaucratic reach of the Apionic house and make an Apionic population estimate, we can use Mazza’s tables to get a sense of the average size of the ten prono¯esiai she records. In the sections that follow, I will survey each of the known Apionic prono¯esiai, and count the settlements and people enumerated therein. I will use the same method with each text: () count how many individual names appear therein, () count how many heirs and partners appear as well, () count how many generically labeled groups of “farmers” and “communities” or koina appear, and () use sample communities enumerated in a pair of Apionic texts to provide an average population multiplier for the others.     

Mazza , . See Mazza , Appendix  for the corresponding map, and Mazza , – for a survey of the toponyms in each jurisdiction, detailing the evidence for each geographic assignment. P.Oxy. ., discussed in detail at n.  above. See Mazza ,  and below, pp. – for these other prono¯esiai. Hickey , , although I think this figure is probably too low. I make a series of assumptions: () I assume that the homonyms remaining undifferentiated in each settlement are in fact different people. () I exclude patronymics. () I count each mention of heirs (p(ar‡) klhr(on»mwn)) or partners (p(ar‡) koin(wnän)) as two people, by virtue of the plural forms, although they may in fact be more. () I count entries recording both a payer and an intermediary as two people. () Generally speaking, I count each payment from a community or group (koinon) of farmers as referring to a separate community or group, unless the second entry



2 The growth of the Apions Table  The sizes of the prono¯esiai A (., .)

B (., .)

C (.)

D (.)

Sites

 epoikia

 kt¯ema,  villages, adjoining places

 epoikia, adjoining places

People

–

NA

[, extrapolating ]

 ( kt¯ema “with other parts,”  epoikion, all with “adjoining places”) NA

E (.a)

F (.)

G (.)

Sites

?

 or more epoikia,  village

People

c.

c.

 villages,  epoikion or kt¯ema,  kt¯ema,  foundation [, extrapolating ]

H (PSI .)

I (.)

L (.)

Sites

 epoikia, one in only one-third share

 epoikia,  foundations

People

c. 

c. 

 and adjoining places ( epoikia,  village ) NA

 Unless noted, parenthetical numbers are P.Oxy. citations. I have retained the labels given each prono¯esia by Mazza (see n.  above); writing in Italian, she did not use the letters J and K.  P.Oxy. ., not using columns two and three, which name other places, but are too badly damaged to yield a complete name count.  See Azzarello ,  for all three sites described as epoikia. Teruthis: Pruneti s.n., where it is both a village and an epoikion through the sixth century. Theagenes: an epoikion in P.Wisc. . (). Eutychias: both an epoikion and a kt¯ema in P.Oxy. . ().  See below, pp. –.  Counting the fragmentary heading kaª tän –po. [ik(©wn) ]u. in line  as two.  Thaesios is not described therein, but is attested as both types of settlements in late antiquity; see Pruneti s.n.  See below, p. .  Aspida is not given an exact description either here or in P.Oxy. .. It appears in P.Oxy. . () as an epoikion, but one in the possession of Flavius Ioannes, son of the Euphemia discussed below, p. .



Social Networks in Byzantine Egypt

Pronoesia A. P.Oxy. . and P.Oxy. . are two of the most important Apionic accounts yet published. They describe the southernmost known Apionic prono¯esia, and one of the largest. P.Oxy. . was compiled by a prono¯et¯es named Ioannes in /. P.Oxy. . was compiled nine years later, in /, by a prono¯et¯es named Theodoros, responsible for the same prono¯esia as his predecessor, which included seven epoikia or hamlets: Apele, Pakiak, Kissonos, Trigeou, Loukiou, Tarousebt and Kotuleeiou. The two accounts are closely related; Roberta Mazza made partial restorations to one by using the text of the other. Both texts are in two parts, the first listing receipts from individual farmers and communities, the second listing expenses in and payments to the same communities. The entries are mixed, listing grain, cash, and other measures throughout. Only at the end of the account is the grain balanced out, leaving a final cash amount which the prono¯etai paid to the Apionic banker Anastasios in lump sums. Because the prono¯etai in both accounts listed each payer, intermediary, and community, these accounts are complete enough to permit an estimate of the size of the tax-paying population of Pronoesia A. Theodoros seems clear and consistent in his use of patronymics, and often specifies when mentioning someone more than once; both clues are necessary to separate one Phoibammon from another. Theodoros’ accounts in P.Oxy. . include by my count receipts from some  people and  distinct “communities” (or koin† ) from his seven hamlets. Later in the account, when he recorded payments back to those communities, he listed the individual members of a specific community or koinon, recording







 

specifies that it refers to the first entry. For instance, I count P.Oxy. ..– as one group, but P.Oxy. .. and  as two groups. For more on the term koinon, see nn.  and  below. Mazza , . See also Mazza , – for partial translations of the two texts and for discussion of the lands under autourgia therein. For P.Oxy. . and its reverse, , see Maresch , – and . For Ioannes, see Mazza , , based on Hickey’s autopsy of the original. There is no reason to think this is the same Ioannes prono¯et¯es appearing in P.Oxy. .. The prono¯esia in question is different, halfway across the nome, and the name is common enough. See above, p.  and P.Oxy. volume , p.  for the close relationship between this account and P.Oxy. .. For the portions of P.Oxy. . not published in the ed. princ., see Mazza , now SB .. P.Oxy. ..: Ioannes pays just over  solidi. P.Oxy. ..: Theodoros pays just over . For a list of attestations of these koina in the Apionic archive, see Mazza , –. See also Gagos and van Minnen , –, where the authors seem to suggest a private, professional nuance to the term. They reject a rendering of the term that would imply the koina were “responsible for the taxes on the land” (), arguing that interpreting the term as a mere fiscal organization “underestimate[s] the resilience of the villagers.”

2 The growth of the Apions



payments to at least twelve different people. An agreement addressed to Flavius Apion in the following decade includes a list of eight signatories from the epoikion of Leo who described themselves in line  as the abovenamed community, t¼ koin¼[n t]än proge[gr]amm(”nwn) ½nom†twn. If we assume that these two examples provide an average size of ten men for a “community” through the rest of the Apionic archive, this means that Theodoros was responsible for collecting Apionic revenues from something like  adult males ( individuals plus  communities of roughly  people each). In terms of total population, using a reasonable multiplier of ., this means that Theodoros had about , men, women, and children in his pronoetic jurisdiction. P.Oxy. ., the second account from this prono¯esia, is equally detailed, with over  lines. A similar population estimate is possible for this text, although it is much harder to distinguish homonyms than it is in the Theodoros account; Ioannes does not seem to have needed notes as meticulous as his successor. The receipts section of P.Oxy. . records approximately  individuals and  communities, although it is possible that a number of these community payments refer to the same community. By the same principles employed for P.Oxy. ., where we suggested that communities had roughly ten men, this gives us a population for P.Oxy. . on the order of  tax-paying males, compared to P.Oxy. .’s approximately  tax-paying males. Again using a multiplier of ., this suggests that just over  men, women, and children fell under Ioannes’ pronoetic jurisdiction. For the year on record, Ioannes paid . solidi to Anastasios, the Apionic banker. In the following decade, Theodoros paid an annual total of  solidi to the Apionic banker. When individual entries record payments from communities, those individual payments cannot give us an exact person-to-payment ratio. However, the figures from Ioannes – payments of  solidi for just over  residents – suggest payments of 

   

P.Oxy. ..–: at least twelve adult males, counting so-called “partners” as two. Under the heading for entries from Kotuleeiou (line ) is included (line ) p(ar‡) toÓ koin(oÓ) tän gewr(gän). Lines – record payments to gewr(go±v) ˆp¼ ktžma(tov) Kotulee©ou, which I take to refer to the same group. This is by no means certain: this second group lists Petros, Esaias, Iob, Esais, Phileas and partners (kaª koin(wno±v)), Ioseph, Phib and partners, and Petros. So at least a dozen people are meant when “farmers” are mentioned, and possibly more. P.Oxy. .. See Bagnall and Frier , , for the average village household size of . for conjugal families, the plurality type. Contrast Rathbone , , where a multiplier of . is used. For an extended treatment of P.Oxy. ., see Mazza , cited in n.  above. For a brief discussion of the potters appearing therein, see Mayerson . The payment totals of both prono¯etai are cited above at n. .



Social Networks in Byzantine Egypt

just over half a solidus for every resident of his portion of the Apionic share. The figures from Theodoros – payments of  solidi for roughly , residents – suggest payments of just over a third of a solidus for every resident of his portion. There is no telling why, at a later date, Theodoros was responsible for more people, yet made a smaller total payment. Perhaps in this period of Apionic growth, individual plots were added to this prono¯esia, hidden under the names of larger settlements. Certainly, these two cases confirm the suspicion voiced above that the proportion of land in a fiscal share may not correspond directly to the proportion of population in that share. Pronoesia B. P.Oxy. ., from , documents the pay and sphere of one prono¯et¯es, but does not help us calculate population figures. The text is a contract between the heirs of Flavius Apion and Serenos, a deacon, who agreed to become prono¯et¯es for a year for certain holdings under Apionic jurisdiction. He seems to have made a distinction between collecting for Matreou (line : –m prostas©aƒ ktžmatov Matr”ou), and for “Your Magnificence’s property in the villages of Episemou and Adaiou and adjacent places” (lines –: kaª tän –n ta±v kÛmaiv ìEpisžmou kaª %da©ou kaª tän –xwtikän aÉtän t»pwn tän diafer»ntwn | t
Ëmän Ëperfue©aƒ). Perhaps this language masks an explicit distinction between Apionic property and Apionic fiscal jurisdiction. The same prono¯et¯es named Serenos from this text appears in P.Oxy. . (?) as well. But that account appears to have been only a draft, and does not give any itemized payments or receipts. From these two documents alone, it does not appear possible to take a population estimate for Pronoesia B. Pronoesia C. Column one of P.Oxy. . (sixth century) is a lacunose pronoetic account drawn up by Ioannes, prono¯et¯es for Teruthis, Theagenes,  



 

See above, pp. –. See Mazza , , and , –, for Italian translations. For Serenos: Gascou a,  thinks that the Serenos in P.Oxy. . is the same Serenos as in P.Oxy. .. He is, however, presumably not the same Serenos who appears as an Apionic prono¯et¯es in P.Oxy. .. The long span of time separating the two texts poses enough of a problem. The Serenos in P.Oxy. . also makes payments for Megales Parorios, not listed in his contract from . Both places are documented from the second to the sixth century, listed together in P.Oxy. ., this text, and P.Oxy. ., from . Adaiou appears in P.Oxy. ., and both appear alone in several texts from the period, e.g. P.Oxy. ., ., .. Mazza ,  places the jurisdiction of Serenos in a region straddling the Upper and Eastern Toparchies, “dove si trovavano Episemou (III pago) e Adaiou (IV pago).” The identification has been standard since Gascou a, , where no argument is given, and is followed by Mazza , . Draft: see intro. to the ed. princ. (“probably only a draft”) and Hickey , , citing it as a possible example of Apionic scribal practice.

2 The growth of the Apions



Eutychias, “and other adjoining places.” All three of these places are described as epoikia or hamlets, although Teruthis is elsewhere attested as a larger village. Columns two and three of this pronoetic account are unpublished but for a few lines. Lacunae and inconsistent usage of toÓ aÉt(oÓ) to indicate duplicate payers make a population estimate of column one difficult, but a conservative count (plural phrontistai and ampelourgoi counting only as two, for instance) yields roughly  payers in this column. If the other columns are roughly proportional, this would give  payers in this account, or some  men, women, and children. But this is only a crude extrapolation, and should not be given the weight of figures for other prono¯esiai. Pronoesia D. P.Oxy. . is a partially published pronoetic account drafted by Stephanos in –. The first lines preserve a heading in which the prono¯esia includes Paggouleeios and “other parts,” Margaritou, Ambioutos, Maiouma, and “other adjoining places.” From the description, the account includes  lines in one column, “probably followed by another which is lost.” This is not enough material to estimate this prono¯esia’s population. Pronoesia E. P.Oxy. .a is an account dating to ; it records the receipts and payments of a prono¯et¯es named Philoxenos, who was responsible for collectors from Notinou Paroriou, the village of Senokomis, the village of Tampemou, the epoikion of Satyrou, and a number of various kt¯emata and m¯echanai. The receipts list payments by roughly  individual payers and  collective groups. If we use the same figure of ten payers per collective group derived in my discussion of Pronoesia A, the result is a population figure for this prono¯esia of  males. With the same standard multiplier of . used for the other prono¯esiai, this figure represents a total population of roughly  men, women, and children. Pronoesia F. P.Oxy. . (late sixth century) is a pronoetic account assigned to the Apionic estates by virtue of its similarity to P.Oxy. .– , and the appearance therein of Papsau, a known Apionic site. Other toponyms it names include Theou, Chenetorios, Samakionos, Graeidos, and Petne, which locate this jurisdiction in the Middle Toparchy, eighth pagus. All of these sites are described as hamlets therein except Petne, a   

Discussed at Mazza , –. See BL . and .; no re-edition has been made, but see proposed corrections at Azzarello , . First described as P.Oxy. ., then partially edited in P.Oxy. ., and according to the BL, not done in full since. Discussed at Mazza , –. Discussed at Mazza , –, and to be associated () with modern Shulkam and Tambu.



Social Networks in Byzantine Egypt

village which survives to the modern day. Like others of its genre, this account is obscure in places. When it lists payments “from the same farmers,” p(ar‡) tän aÉ(tän), it is often unclear which group is meant, and never clear how many people that group includes. Still, some  individual names appear in the receipts portion of this text, representing individual payers and their intermediaries. If the groups of farmers appearing throughout this text mask significant handfuls rather than isolated individuals, the manager of this account dealt with  men or more in the process of collecting his receipts. This represents a total of perhaps as many as  men, women, and children in this prono¯esia. Pronoesia G. This prono¯esia, in the seventh pagus straddling the Thmoisepho and Middle Toparchies, appears in P.Oxy. ., an account of the prono¯et¯es Ioannes’ income and expenses in grain, dating to the late sixth or early seventh century. Three settlements appear therein, Leonidou, Limeniados and Herakleias; all three are attested elsewhere as kt¯emata, with Leonidou also attested as an epoikion. The first line of the text records a total grain receipt of , “cancellus” artabae. Hardy noted that this was the second highest individual grain receipt recorded, a distant second behind an Oxyrhynchite account in which the expenses in grain of one Christodora of Cynopolis exceeded , artabae in one year. Christodora would have needed eight men like Ioannes to recoup that sort of expense; as we have seen, the Apions had at least sixteen. This is an interesting comparison, but more important for our purposes is determining how many payers the , artabae represented. This account does not include individual payments, so the method used for other texts will not work here. But other accounts may provide a crude comparison. In P.Oxy. ., Theodoros recorded (line ) total grain receipts of just over , “cancellus” artabae, and may have had some , men, women, and children in his jurisdiction. If Ioannes had grain receipts of over three times the size as Theodoros, he may then have had some , men, women, and children in his jurisdiction. But using ratios of artabae to count people is crude, and gives us a prono¯esia much larger than the others we survey herein. As with Pronoesia C, this figure should not be given the weight of figures for other prono¯esiai.    

Discussed at Mazza , , with Petne to be associated with modern Itnih. Discussed at Mazza , . To the best of my knowledge, no one has proposed a more exact dating since the publication of this text. On Ioannes, see also n.  above. See Pruneti s.n.  See above, p. . Hardy , , citing P.Oxy. .; see line  therein.

2 The growth of the Apions



Pronoesia H. The prono¯esia appearing in PSI . (sixth century) is located in the eighth pagus. The prono¯et¯es, whose name is lost, detailed receipts from the epoikia of Skutalitidos, Ptolemaios, Mikras Parorios, a one-third share of Chrusochoos, and an epoikion whose name is lost. This prono¯et¯es did not always make clear distinctions between the different koina under his jurisdiction. Line  refers to a payment “from the same aforementioned farmers” after an entry recording payment “from the community of farmers.” I have counted this koinon only once. It is less obvious what to do with such cases as lines –, where payment from the koinon of farmers of Skutalitidos and Krarios is followed by payment from the koinon of farmers and ampelourgoi of Skutalitidos and Krarios. I have counted such entries as two koina. Using the same techniques outlined for Pronoesia A above, I count  individual payers and  koina, or roughly  males, which gives a total of roughly  men, women, and children. Pronoesia I. An Apionic account from  details the northernmost of the known Oxyrhynchite Apionic prono¯esiai. The prono¯et¯es, Philoxenos, kept clear track of his income, identifying repeat entries (e.g. line : p(ar‡) toÓ aÉto(Ó) %nouq©ou) and clearly indicating the geographic origin of each payer. He even indicates when someone is paying in one settlement but comes from another. If we use the same methods employed above, Philoxenos’ clear accounting practices make a population estimate of his prono¯esia relatively easy. The account from Pronoesia I records income from six epoikia and three dikaia broken down into entries for  individuals and nine collective groups, giving a conservative total of  payers. This gives a total population under his jurisdiction of roughly  men, women, and children. This text also allows us to calculate the size of a typical dikaion or foundation. Philoxenos recorded income from  payers and  koinon from the  







Mazza , . P.Oxy. . with Mazza , . BL . gives the date as / (a th indiction), but see now BL . for , not yet noted at Mazza , . The correction is at Bogaert ,  n.  where attention is drawn to a reference to  Tubi in an th indiction, presumably . This was not an unusual occurrence: e.g. Petros son of Menas pays for land in Tillon (line ) but comes from Sinaru (ˆp¼ kÛmhv SinarÆ). A group from Tillon (tän ˆp¼ –poik(©ou)) paid for holdings in Eros (line ). The two men paying for holdings from the Antiochou dikaion or foundation (lines –) were both from elsewhere, one from Tagchis (ˆp¼ Tagc”w), the other from Suphis (ˆp¼ SÅfewv). In places where he records payments only “from the farmers” (e.g. line ) I count the entry as something analogous to a koinon. This may be generous, but is countered by my omission of various entries “from the koinon of farmers” followed by lacunae which may have provided distinguishing characteristics. The account as a whole names nine epoikia, as listed in Table  above, but only six of those appear in the section for receipts.



Social Networks in Byzantine Egypt

foundation of Kephalas, which amounts to roughly  individual payers, or  men, women, and children. (These payments were in a separate section of his account, but included in our previous totals.) This was by no means the full extent of that foundation. Another Apionic account shows the holdings of Kephalas scattered throughout seven different prono¯esiai. One payment “in the name of Kephalas” went through the prono¯et¯es of Adaiou, which was part of Pronoesia B above. Another payment in the name of Kephalas went through the prono¯et¯es of Tampeti, which was part of Pronoesia F above. Since the account detailing Pronoesia F makes no mention of any holding of Kephalas, we must assume our population totals for that jurisdiction fall short. Indeed, these holdings are counted in detail only in Pronoesia I, which suggests shortfalls for at least six of the seven prono¯esiai in charge of holdings of Kephalas. If the other holdings of Kephalas under Apionic jurisdiction were comparable in size to those here, where we count some  people, this may suggest a shortfall in the six other districts of as much as  men, women, and children in our nome-wide population totals, tabulated below. Pronoesia L. A fragmentary Apionic account from / records a prono¯et¯es named Paulos responsible for Aspida, Kuamon, Phna, Spania, “and other adjoining places.” Of these four toponyms, only Spania is a village; the other three are known epoikia. No other information is available on this prono¯esia, but it was near the northern end of the nome, in the vicinity of modern Safaniya, with which Spania has recently been identified. Spania must have been a town of some size: a sixth-century text details  landowners (ktžtorev) from Spania “obliged to make good the damage” incurred when a priest and meiz¯on of Spania suffered a massive theft from his property. With this many landowners, the village must have had several times that many tenant farmers, and a population comfortably over ,. If Apionic jurisdiction included even a sizeable fraction of Spania, not its entirety, this prono¯esia would have been large indeed. The   

   

 See P.Oxy. . col. . P.Oxy. ..–.  P.Oxy. ... P.Oxy. ... It is not clear whether Anoup, named at P.Oxy. .., could be prono¯et¯es for Pronoesia I at another period, as no toponyms are given for his jurisdiction in the account. The other six prono¯esiai are on the basis of the toponyms clearly not Pronoesia I. For the payments in this text, see Maresch , –. Again assuming a demographic multiplier of .; see above, n. . P.Oxy. .. For new readings of the toponyms, see BL .. See the discussion at Mazza , –, with n.  on Goma`a et al. P.Oxy. ..–: gnäsi(v) tän ktht»r(wn) Span©av ½feil»ntwn plhräsai | tŸn genam”nhn bl†bhn –n t o­k() toÓ me©z(onov) KuriakoÓ pre(sbut”rou). For Spania’s size see Mazza , .

2 The growth of the Apions



list of  landowners is not from the Apionic archive, and we thus have no real way to provide population figures for this jurisdiction, unless more of Paulos’ account is one day recovered. Unknown prono¯esiai. The Apionic oikos clearly had at least six more prono¯esiai not documented in full accounts. Obviously, we cannot supply evidence for these jurisdictions where none exists. Takona, for instance, is amply attested in Apionic documents, and is even described as being under the Apionic pagarchy in its entirety. But it cannot be assigned to any of Mazza’s specific jurisdictions. The Apions also collected rent within the city of Oxyrhynchos itself: a work contract from  signed by a psalmist named Ioannes includes  agreement to pay Flavius Apion  solidi for rents Ioannes would collect from their urban properties for the second indiction. The figure is eye-catching. A series of sixth-century urban leases for workshops, houses, and parts thereof, names individual figures ranging from a few carats to two solidi a year. If these ratios are any indication, Ioannes the psalmist could have collected rent on behalf of hundreds of Apionic properties in the city of Oxyrhynchos itself. An Apionic account likely from / details collections from the village of Ibion and the epoikia of Nokle and Ostrakinou. We do not know modern equivalents or the toparchy or pagus of these sites, nor are they attested as belonging to any of the known prono¯esiai discussed above. This may well be an account for one of the unknown prono¯esiai, or for a jurisdiction created a generation or two later. If so, this jurisdiction’s scale is in keeping with other examples we have at hand. By my count, it records payments from approximately  individual payers and  koina, which would put its prono¯et¯es in charge of Apionic fiscal responsibilities for roughly  people. Taking an average of the figures we have calculated, including this unknown prono¯esia, may help fill the gaps. Table  takes , as the average population for Pronoesia A, where we estimated Theodoros and Ioannes 



 

See above, n.  with P.Oxy. .. Other Apionic accounts not generated by prono¯etai also provide interesting topographical and prosopographical material, particularly P.Oxy. ., an account of brick payments, and P.Oxy. ., an account of water-wheel axles (for both texts see Gonis b, –). P.Oxy. . ( or later), on the reverse of a pronoetic account published as P.Oxy. ., is the rough draft of an Apionic account with nome-wide scope, but like P.Oxy. ., it does not show any pronoetic organization to it. P.Oxy. ..–, where we read about a “guild of village headmen of Takona in the pagarchic share under the house of your nobility,” koin¼n tän prwtokwmhtän t¦v kÛmhv T†kona . . . | . . . pagarcoum”nh[v Ë]p¼ toÓ o­kou t¦v Ëmän [sc. Apion] –ndox»thtov. (The translation of the ed. princ. describes the village only as “dependent upon” the Apionic oikos.)  See the sequence P.Oxy. .–. See Mazza ,  for references. P.Oxy. ., dated to the early sixth century in the ed. princ.; see P.Oxy. . n. .



Social Networks in Byzantine Egypt Table  Calculating the average prono¯esia Prono¯esia

Population estimate

A E F H I Unknown (P.Oxy. .) Average

,      

to have had fiscal responsibility for  and , people respectively. It then uses the population figures I calculate for Pronoesiai E, F, H, I, and the unknown jurisdiction above. The result is an average of roughly  men, women, and children per prono¯esia, or some , men, women, and children under Apionic fiscal responsibility nome-wide ( prono¯esiai times  = ,). If the rural population of the nome was roughly , in the Roman imperial period, and we assume decline by roughly a third due to the effects of the plague in the s, this population figure places the house of Apion on top of a social pyramid of around  per cent of the entire Oxyrhynchite. This figure should be checked against others in this chapter. We have noted Gascou’s argument that the Apions often appear responsible for roughly one-third of a given payment’s total. Although population ratios may not correlate to jurisdiction ratios, an Apionic population of only  per cent of the nome total may seem low in light of Gascou’s argument. Other figures help. The documents from Pronoesia A suggest that Ioannes and Theodoros collected payments of between one-third and just over one-half a solidus for every resident of their portion of the Apionic share. The Apionic estate’s gross receipts in gold were a fraction over , solidi in /, before the apparent mid-century growth of the estates. If the collection ratios from Pronoesia A mean anything, this gross figure suggests that our Apionic population estimate of roughly , people was only a fraction of the true total. 



Rowlandson ,  for the imperial period, although see Kr¨uger , –, repeated at , for a much higher estimate of the rural population. See Allen  for the extent of population loss due to the plague and Little  for a series of more recent studies on the plague in general.  See above, p. .  See above, p. . See above, p. .

2 The growth of the Apions



Indeed, that population figure is a conservative estimate. It assumes there are no missing prono¯etai, while Hardy at least believed there were. It ignores one outlier, Pronoesia G, where payment ratios compared to other accounts suggest a population over , for one prono¯esia alone. It counts generic plurals (e.g. “heirs” or “partners”) as only two individuals, even when a large koinon may have been meant. It also ignores the possibility that foundations such as that of Kephalas may have had populations nearing , people scattered uncounted throughout many prono¯esiai. With these limitations in mind, our figures affirm the picture in Chapter  of a considerable degree of centralization to the Oxyrhynchite nome. It means that the entire population of the Oxyrhynchite would have been fiscally accountable to, and therefore only a few degrees of separation distant from, one of a handful of elites, no more than a dozen at most. One final caveat should be added. It is possible that some settlements collected their own taxes, and thus fell outside of the system of fiscal jurisdiction traced here. The ten prono¯esiai described above attest to eight k¯omai or villages. Increasing this number by  per cent to account for Apionic prono¯esiai we know existed, but cannot describe, gives us some dozen villages partially or completely under Apionic fiscal jurisdiction. Assuming, with Gascou, that the Apionic fiscal share was roughly onethird of the total, that would mean that the great houses were responsible for collecting payments from something like  Oxyrhynchite villages in this period. Yet, by my count of Pruneti’s topographical register, at least  settlements are explicitly attested as k¯omai in the sixth century; the total is likely to have been higher. Perhaps these other villages were autopract, like Aphrodito far to the south. Until better evidence comes to light, the contours of the fiscal share system I have outlined here should be considered provisional. mapping the apions i: traditional approaches Crucial questions about the formation of the Apionic oikos remain unanswered. Did Apionic jurisdiction spread organically, emerging outward from the original nucleus of a single estate or group of clustered estates? Did the Apions acquire land through pre-existing rural-based social ties? Or were their holdings relatively scattered, indicating patchwork acquisition directed from the central metropolis? The latter option would support the commonplace that the Apions were socially distant from their own 

Hardy , .



Social Networks in Byzantine Egypt

holdings, indeed even removed from the very “Egyptianness” of them. Gascou’s model of fiscal shares provides a complicating factor: we cannot assume that the Apions owned every place appearing in their archive. What then was the relationship between Apionic property and land merely under Apionic fiscal responsibility? Were these two groups clustered or scattered with respect to each other? Modern scholars studying estate formation and land distribution in Roman Egypt have emphasized the importance of inheritance. In his groundbreaking study of the Heroninos archive and the Appianus estate, Dominic Rathbone describes the managerial hierarchy of the third-century Fayum estates as a three-dimensional model, a “complex matrix of hierarchical and personal links.” Clues concerning estate formation are sparse, and the archive is relatively quiet about the buying and selling of land. Instead, family ties seem important: Appianus’ estate probably swallowed land belonging to his father-in-law, and the estate itself ends up in the hands of his daughter and son-in-law before coming into the possession of the state treasury. Jane Rowlandson’s more general work, Landowners and Tenants in Roman Egypt, also notes the importance of family ties. As Keenan observed in his review of Rowlandson, “Most land sales were for small-sized plots, with relatives, friends, and neighbors favored as buyers. Land on the whole remained more likely to change hands by inheritance than by sale.” If this remained true in late antiquity, then the rise of the large estates presents an even greater puzzle than previously imagined. Could oikoi like the Apions have formed on a purchase basis if sales remained the exception? Could the great estates come about purely through the gradual accretions of inheritance and marriage alliance? Estates with widely scattered component parts are a well-known phenomenon. In the Heroninos archive, Appianus had phrontides (administrative units which Rathbone renders as “concerns”) in thirty or more Arsinoite villages. These “scattered landholdings . . . [were] normally centered on a particular village,” and these holdings in turn formed “constellations of phrontides . . . directed from headquarters in Arsinoe.” One such 

 

Hardy: Apionic involvement in their estates was the exception rather than the rule, explicitly noteworthy (Hardy , ). Gascou: Apionic tenures in local Oxyrhynchite offices were, at least early on, filled “sans doute par delegation,” as in the case of Apion I’s defensorship (Gascou , ). See also Gascou on Apion II’s local offices (Gascou , –). Rathbone , . His recommendation at Rathbone ,  of a hand-by-hand approach to the texts of the Appianus archive could usefully apply to the Apionic archive as well.  Keenan , .  Rathbone , –. Rathbone , –.

2 The growth of the Apions



phrontis, that at Theadelphia, had  known constituent parts, comprising a possible total of at least  arourae, some  per cent of Theadelphia’s total land. The phrontis at Euhemeria was even bigger, in the region of , arourae. Despite these enormities of scale, Appianus was not distant from his land: he could “recognize minor estate employees like Sokras.” There is something reminiscent of the Apions here: a large landholder directing scattered holdings from an urban center, but still finding himself involved from time to time at a quite personal level. Jairus Banaji’s recent study of Agrarian Change in Late Antiquity contrasts the Oxyrhynchite model of a “general dichotomy and physical discreteness” between villages and epoikia, and the Fayum model of epoikia named after and in close proximity to nearby villages. He argues that “the large estates of the eastern Mediterranean were neither ‘village estates’ nor mere assemblages of small scattered parcels but agglomerations of compact settlements.” He bases this argument on a survey of toponyms from large estates in the Fayum, including that of pseudo-Strategios. But as Banaji admits, “we simply do not know how many more settlements Strategius is likely to have controlled” in any given area of the Fayum. He thus has no way of knowing how representative his sample may have been. Put another way, Rathbone, Rowlandson and Banaji have useful examples of how estates formed, but without knowing how typical these examples were, we cannot know whether the growth of the Apions followed a comparable pattern. And what of the landowners themselves: were these wealthy elites fundamentally urban or rural creatures? If they were urban, as scholars usually suppose, we may imagine that their land accumulation had no particular geographic focus: absentee landlords are more likely to acquire land “at random” throughout their economic sphere. If rural, it seems more likely that their expansion would have been outwards from their ancestral holdings: local landlords will take a worm’s-eye view, less likely to acquire land throughout the entire Oxyrhynchite nome. The wealth of Oxyrhynchite toponyms attested in Apionic texts can provide the data-set for exploring this question in our case. We can examine       

See Rathbone , Table  and the discussion at .  Rathbone ,  with P.Flor. .. Rathbone , . For personal involvement, see e.g. the case of Diogenes in Chapter  above, pp. –.  Banaji , . Banaji , .  Banaji , . Banaji , –, with Appendix . For a discussion of this phenomenon, see Rowlandson , . Summarized in Mazza , .



Social Networks in Byzantine Egypt Table  Attestations of the Apionic archive over time  Apionic texts per decade

Number of texts

30 25 20 Series1

15 10 5

63 0

61 0

59 0

57 0

55 0

53 0

51 0

49 0

47 0

45 0

43 0

0

Decade 

This graph, prepared with Microsoft Excel, is based on Mazza’s Tabella A, Mazza , –. It counts each indiction year twice: e.g. once in  and once in ; it counts as one attestation more than one text on the same side or different side of a papyrus; and it omits all texts assigned to an uncertain date, an uncertain set of dates, or an uncertain range.

the spread of Apionic land-ownership and fiscal responsibility throughout the fifth and sixth centuries, and see how this expansion compares to our models. Hardy pointed out that it was during the period of Apion II, in the mid-sixth century, “that we begin to have documents from the estate in fairly large numbers,” and thus suggested that “we may probably think of him as actively interested in the management of his property. At least we know that he occasionally visited it.” Table  above highlights the dramatic extent to which Apionic documentation grows in the second half of the sixth century, increasing in the s, hitting a first peak in the s and s, and hitting a second, higher peak in the s. This increase in texts from the Apionic estate correlates to a rise in the number of known Apionic toponyms. Table  following shows the distribution of first attestations of toponyms in an Apionic context; the increase in new Apionic toponyms begins in the s and peaks dramatically in the s. But the majority of new toponyms appears in the second half of the sixth century, when the number of texts from the archive per decade is holding steady. 

Hardy , , citing P.Oxy. ..–.



2 The growth of the Apions Table  First attestations of Apionic toponyms over time  First Attestations of Apionic Toponyms New Toponyms

30 25 20 Series1

15 10 5

60 0

58 0

56 0

54 0

52 0

50 0

48 0

46 0

44 0

0

By Decade 

This graph, prepared with Microsoft Excel, is based on Mazza’s Appendix , Mazza , , substituting for the first text in each entry the date given for that text in Tabella A, Mazza , –. It thus records the first attestations of a toponym within the Apionic archive, excluding all prior non-Apionic attestations. It omits texts with multiple possible dates or texts only assigned to a century or a range within a century. Texts assigned to circa, before, or after a specific date are assigned to the decade of that date, except when a date is given as before the first year of a decade; the previous decade is then used. In addition to the data provided in Mazza, this graph includes the new Apionic toponyms in the recent P.Oxy. , along with the dates of the texts in which they appear.

This suggests that Apionic influence continued to grow throughout this period, after its early burst in the s and s. Where were all of these toponyms in relation to one another? Hickey argued that when mapped against Rowlandson’s reconstructed map of the Oxyrhynchite nome, Apionic viticultural holdings displayed a “northern face” to the Apionic estates. This may be true of Apionic viticulture, but it is otherwise an incomplete picture. In her survey of each known Apionic prono¯esia, Mazza noted that “Procedendo dal limite meridionale del nomo a quello settentrionale, risulta evidence che la famiglia contava possedimenti sparsi per tutto il territorio.” She argued that the papyrological data “dimostrano che l’o²kov era diffuso su tutto il territorio dell’Ossirinchite,   

For discussion of potential reasons for this growth, see below, pp. –. Hickey , . Mazza , . See also her earlier remarks at Mazza , , where she writes that “la propriet`a era costituita da unit`a sparse sul territorio in modo discontinuo.”



Social Networks in Byzantine Egypt

con un andamento discontinuo.” These comments are somewhat qualified by her survey of pronoetic jurisdictions, which shows a concentration in the middle of the nome, and outliers in the north and the south. In the following section, I provide a network analytical technique to derive quantitative support for Mazza’s argument that the estates were spread through the whole nome. But before doing so, I analyze other techniques for studying the relative location of Apionic toponyms. First, and despite Mazza’s more general conclusions, specific Apionic holdings were certainly near each other. Hardy drew attention to Mikra Tholthis and Mikra Teruthis. We know that the Apionic estates had ties to both epoikia or hamlets, and we may presume that both sites were close to their larger counterpart villages, Tholthis and Teruthis respectively. The Apions were certainly present in both of those villages, in different parts of the nome. Tholthis in turn was only a sheep-theft away from still other Apionic toponyms. In the late fifth century, the headman (meiz¯on) of Tholthis wrote to his counterpart in Takona about a squabble between shepherds of the two villages over stolen sheep and other property. Takona is amply attested in Apionic documents from the s on. Estate accounts also make explicit reference to the physical proximity of Apionic toponyms. One describes the jurisdiction of Stephanos as Paggouleeios “with other parts,” and Margaritos, Ambioutos, Maiouma, “and other adjoining places.” Another describes the jurisdiction of Ioannes son of Philoxenos as Teruthis, Theagenes, Eutychias, “and other adjoining places.” But many cases are not as clearly described, and other approaches to locating Apionic holdings are necessary.

 

 

  



Mazza , . Mazza , Appendix . She notes that there appears to be “una maggiore concentrazione dei possedimenti nei dintorni della citt`a di Ossirinco” (Mazza , ) but admits that this impression may be a result of the archaeological provenance of our papyrological finds. For a complicating argument, that proximity to Oxyrhynchos does not correlate to a toponym’s attestation strength, see Ruffini . Hardy , . For Teruthis see e.g. P.Oxy. ., ., .; for Tholthis see e.g. P.Oxy. .. Apionic archival references to all four sites are collected at Mazza , . See also the nome maps at Rowlandson , xiii–xiv. See P.Oxy. .. See e.g. P.Oxy. ., ., with full references at Mazza , . P.Oxy. . and Mazza’s Pronoesia D; see above, p. . For “with other parts”: sÆn to(±v) Šll(oiv) m”r(esi). For “and other adjoining places”: kaª Šll(wn) –xwtik(än) t»pwn, with Hardy’s translation and these references at Hardy , . P.Oxy. . and Mazza’s Pronoesia C; see above, pp. –.

2 The growth of the Apions



One technique is to examine Apionic land we know once belonged to another landholder. In , the hamlet of Piaa was in the hands of a man named Flavius Alexander. Alexander is one of those rare Egyptian landholders also known from non-Oxyrhynchite evidence: he was a former comes of the Egyptian border, former praefectus augustalis, and a former, perhaps honorary, magister utriusque militum. How close he was to his own land in the Oxyrhynchite is hard to tell. A receipt issued in  to Flavius Ioseph, Alexander’s dioik¯et¯es or property manager, acknowledged receipt of a required part for a water-wheel. From this alone, there is no evidence to indicate whether Alexander’s holdings were part of a more substantial and permanent family holding, or if his presence in Egypt was simply transitory. By the s, Piaa appears in the holdings of the Apionic estate; Mazza supposes that it may have been part of her Pronoesia F, which may put it near Petne, modern Itnih. But how it came under Apionic jurisdiction from Alexander’s estate several generations before is unknown. Another approach to mapping Apionic territory is to look for potential Apionic boundaries. A sixth-century list of Apionic coloni who fled from estate property has helpful toponyms. Twenty-three farmers fled from Thaesis to Machauson. Four men fled from Leontos, three of them to Philostratos, the fourth to Machauson. A cattle-thief named Pathon fled from Peleos to Machauson. Without this text, we would have no reason to think Philostratos or Machauson such attractive destinations for late     





For an example of this type of Apionic expansion not discussed here, see the case of Diogenes and the monastery of Abba Hierax, detailed in Chapter  above, pp. –. P.Oxy. ..–: Symphonias is described as “ˆp¼ –poik©ou | Pi†a. toÓ ˆutoÓ –ndoxot†tou ˆndr¼v [%lex†ndrou] toÓ aÉtoÓ nomoÓ.” See PLRE ., as well as entries in the CJ: .., .., and ... P.Oxy. .. Maybe much earlier than the s: see Mazza ,  on P.Wash.Univ. ., which names Piaa, and dates by her analysis to V/VI, contra the editor’s IV/V. Piaa appears in P.Oxy. . in the company of known Apionic settlements in / or /. The text is closely related to P.Oxy. ., which dates to /. On the inclusion of both in the Apionic archive, and for the latest references on their dating, see Mazza , –, with n. . For Piaa and Pronoesia F, see Mazza , . P.Oxy. ., discussed briefly at Hardy , , and described as a Gnäs(iv) gewrg(än) fug(»ntwn) –k toÓ ktžma(tov) Qažsiov. Mazza  does not include this text in her survey of Apionic archival texts, nor does it appear in her papyrological index. But Thaesis was part of what she labels Pronoesia G, under Apionic fiscal administration: see above, p. . The list itself is interesting. Judging solely from consecutive patronyms, we have four pairs of brothers, and one trio of brothers, together accounting for eleven of the . So which is more curious, that a farmer should flee with a relative, or without one? Hardy ,  saw in this list “something parallel to the peasants’ strikes of the Ptolemaic period.”



Social Networks in Byzantine Egypt

antique Oxyrhynchites. The implication seems clear: the two sites were outside the border of the Apionic estates, Thaesis, Leon, and Peles inside that border and relatively nearby. A letter between estate employees Victor and George gives another clue: “the people of Pinuris have got a guard owing to the people of Thmoinepsobthis . . . [who have] made an irruption into the village.” Both sites presumably had some connection to the Apionic estates. In the age-old tradition of Egyptian village warfare, the two places were presumably neighbors. A sixth- or seventh-century letter describes a specific inter-village arbitration, perhaps under Apionic jurisdiction, and paints a somewhat more complicated picture. The people of Kosmou and those of Spania have been reconciled. (The fact of a dispute at all suggests their proximity to one another. ) The people of Ision Kato and Phebicheos mediated between them. (Perhaps they were not quite as nearby? ) The chief messenger of Meskanounis was far enough away that this letter was needed to send for him, so that he can “collect their goods and hand them over to them.” A promissory note addressed to the Apionic endoxos oikos names the villages of Popano and Pakerke, which we know to be Apionic from other contexts. In that note, estate officials were involved in a payment for 

   







Neither Philostratos nor Machauson appears in Mazza . The former site appears in both P.Oxy. . (a fifth-century payment list) and PSI . (); the latter site is otherwise unattested in Pruneti. Thaesis is well attested in Apionic texts; see Mazza , . Leontos is as well; see Mazza , . Peleos appears in P.Oxy. ..–, a brick account of apparent Apionic origins. P.Oxy. ..–: o¬ ˆp¼ PinÅrewv ›cou. s. i fula. k. h. [n | d[i]‡ toÆv ˆp¼ QmoineyÛbqewv . . . ›. b. a. lon g. ‡r. e«v k. [Ûmhn. Mazza  includes P.Oxy. . in her list of Apionic texts, yet excludes Pinuris and Thmoinepsobthis from her list of Apionic toponyms. Village warfare: for earlier examples, see Palladius Lausiac History , Zosimus .., Juvenal Sat. . P.Oxy. .. Mazza  does not assign this text to the Apionic archive, and it does not appear in her papyrological index. But see the text’s topographical ties to Apionic jurisdictions cited in the following note. Mazza ,  and  places Spania and Kosmou in adjacent prono¯esiai. Makarios, me©z(wn) Span©av, informs the unnamed addressee that the mediators from Isiou Kato and Phebicheos are (lines –) e«v t¼ m”son ¡män [sc. Spania] | kaª met’ aÉtän [sc. Kosmou]. Thus when Makarios wrote Âti kat‡ tŸn sžmeron ¡m”ran –d»qhv –n t ¦{n} ƒ e«ržnh | met‡ tän ˆp¼ K»smou (lines –, ed. princ. transl. “you were to-day reconciled with the people of Cosmu”), the addressee is understood to have been on Spania’s side of the dispute. They do not appear in Mazza . Ision Kato is in the Lower Toparchy, as Spania would have been as well, and Phebicheos may have been near the border with the Herakleopolite; see Pruneti s.n. in both cases. P.Oxy. .. Mazza ,  assigns this text to the archive on the basis of prosopography and archaeology. See Mazza ,  and , where Popano is described as “nelle vicinanze di Satyrou e Pakerke.”

2 The growth of the Apions



damages inflicted by the people of Popano upon Pakerke, once again indicating proximity between the two settlements. mapping the apions ii: network approaches These examples might easily lead scholars to conclude that Apionic sites were generally clustered or physically proximate to one another. But this is merely argument by anecdote; quantitative rigor is necessary to support or refute such a conclusion. Surviving Oxyrhynchite texts provide a tremendous amount of detail about that nome’s topography. Pruneti’s topographical index of the nome lists hundreds of toponyms throughout the region. But in only a few cases has it been possible to locate sites in relation to each other and to known sites in the modern world. Network analysis provides a relatively simple way forward. I have demonstrated elsewhere how the data in Pruneti’s Oxyrhynchite topographical register can be used to create a network of Oxyrhynchite settlements. Pruneti’s register functions as a network recording the links between over  sites and over , Oxyrhynchite texts naming them. UCINET analysis of this network provides numerous insights into Oxyrhynchite topography. First, this analysis shows a network built around a series of hubs, in which the center, Oxyrhynchos, links to a small group  





P.Oxy. ..–: prof†si t¦v –t©av toÓ ˆnaireq”ntov ˆp¼ | Pap”rkh toÓ ktžmatov par‡ tän ˆp¼ PwpanÜ. For the translation of anairethentos see note to line  in the ed. princ. For the most accessible examples, see Rowlandson , xiii–xiv. The first map shows the modern equivalents to just under  ancient sites in the Oxyrhynchite, to which can be added Spania at Safaniya (see Chapter  above, n. ). Ruffini . I have not updated Pruneti’s data with topographical references published since. A list of  toponyms creates a network of , potential connections. Additions to that matrix published since Pruneti will have no statistically meaningful impact on the shape of the network as a whole. Nor have I used the network data compiled in Kr¨uger , –. While the connectivity charts in that work clearly explore some of the phenomena relevant here, those charts include no papyrological citations, and thus make impossible the creation of a data-set susceptible to the network analyses used in this chapter. For the original data-files used in this network, see http://www.grr.net/SNBE/. Two versions of my Pruneti data-set are available: the one I used for Ruffini , which retains the spellings of each toponym as presented in Pruneti, and a second, used herein, which strips each toponym of any special characters, some of which cause problems in conversion to UCINET. An astute reader of Ruffini , Bruce Nielsen, noted two problems with that piece after it had been sent to publication: () Pruneti’s references to various texts by P.med.inv. numbers were obsolete, and needed to be changed to improve the data-set; and () the degree strengths I present for various settlements in Tables  and  (Ruffini ,  and ) differ dramatically. To address the first problem, I have modified my data-set to replace P.med.inv. numbers with relevant SB numbers. The changes have modified the results somewhat from my original dissertation. The second problem is explained by the fact that I mistakenly presented the results from two different data-sets, one (for Table ) which included all Pruneti texts, and another (for Table ) which used a version of Pruneti eliminating various nome-wide village lists, as described in the following note.



Social Networks in Byzantine Egypt

of elite nodes (the nome’s major villages) and those nodes in turn link to dozens of less important sites. The chief nodes themselves are well connected to one another. Smaller Oxyrhynchite settlements were less strongly linked to one another, and usually only linked to each other indirectly, through shared ties to the larger centers. If each settlement had been evenly connected to those closest to it, we would have expected each pair of settlements to have roughly the same number of ties of any given strength. Instead, some sites – e.g. Takona, Skutalitidos – had a disproportionately high number of ties in respect to their peers, suggesting that Oxyrhynchite settlement geography was considerably hierarchical. The following network analysis builds on these conclusions, and begins with a crucial assumption, that two places are more likely to appear together in the same text the closer they are to each other. Katja Mueller has made this assumption in her topographical analyses of the data from the Fayum. My recent paper on topographical network analysis has demonstrated this assumption to be true of the Oxyrhynchite data on the level of the toparchy and pagus. This assumption makes sense on an intuitive level; examples abound of papyri naming two or more places clearly near 





One flaw to this approach is that many topographical attestations come from texts listing settlements throughout the Oxyrhynchite nome, when no direct connections between the settlements are implied. Consider the two most connection-rich texts used in the Pruneti register, P.Oxy. . and P.Oxy. .. Both are third-century village lists, organized by toparchy and covering the entire nome, the first perhaps recording adaeratio payments, the second an account of the crowntax. This problem is equivalent to that of distortion caused by payment lists such as the excerpt from Zuckerman’s fiscal register originally published as P.Flor. . in our analysis of the Aphrodito Girgis prosopography in Chapter  below, pp. –, and discussed in the manual above at p. . We can correct this bias by removing texts which are simply nome-wide village lists, leaving only those texts – by far the majority, in any case – which describe interactions between only a handful of connected villages. I perform the following analyses of Oxyrhynchite topography on both the original Pruneti register, and an edited data-set from which I remove the top eleven most connection-rich (degree central) texts, thus correcting for the distortion of large texts. (I remove eleven texts as texts ten and eleven both have the same ranking by degree centrality.) The texts in question are, in order of degree centrality: P.Oxy. ., ., ., ., ., ., ., ., ., P.Mich.inv. , P.Iand. . For more on this technique and its validity, see Chapter  below, pp. –, and Ruffini . In the discussion that follows, I have discussed only the results obtained via correcting for distortions in this fashion. See Mueller a, – for discussion of this assumption, which is also implicit in Rathbone’s “London underground map approach,” which Mueller discusses at ; see also Mueller b, –, and her remark () that “In antiquity and in our papyri economic and administrative relationships are mostly local issues. They correlate significantly to proximity and space.” This assumption is less a factor in Mueller , concerned more with rank-size than proximity and connectivity. Ruffini . My conclusions showed that: () in eight out of ten pagi, settlements had more intrapagus ties than extra-pagus ties, and () in all six toparchies, settlements have more intra-toparchy ties than extra-toparchy ties. Curiously, proximity did not affect the external ties. Over half of the ties to external pagi and toparchies were to non-neighboring pagi and toparchies.

2 The growth of the Apions



one another. (The obviously exceptional cases, usually nome-wide village lists, can be removed from our analysis using a method described at note .) A consequence of this assumption is that two settlements or two groups of settlements with more ties between them than to other settlements or groups of settlements are more likely to be physically closer to one another. Using the known Apionic toponyms in Mazza’s Appendix  and comparing their connectivity levels with each other and with non-Apionic toponyms therefore allows us to test whether Apionic settlements were physically closer to each other than to other settlements. If the Apionic holdings were physically proximate, they would show a pattern of connectivity different from other settlements in the nome. They would have more connections to each other than to non-Apionic settlements. The network density of the Apionic holdings would be higher than the network density of the rest of the network. A network analysis meeting these conditions would strongly suggest that the Apionic estates were on average closer to each other than to other sites, and therefore that at least some of those estates were acquired by rural-based expansion, through the “worm’s-eye view” described above. By contrast, network analysis not meeting these conditions would mean that the Apionic estates show no tendency to be near one another, and therefore, that they were more likely acquired at random, via relatively indiscriminate urban-based acquisition. A brief word about three risks inherent in this method. First, as Hickey put it, “No one would suggest that the [Apionic] family went on a building spree upon its rise to fame, and that there was a stasis thereafter.” In other words, the papyri cover a long period of time in which many of the attested toponyms might have entered or left the Apionic holdings. Treating the entire network as a chronological simultaneity may lead to false impressions. Secondly, this method does not catch those sites we suspect were near each other, but whose names do not appear in the same text. For example, Pamouthios managed farms for the Apions at both Pyleas and Leontos, and is attested at both sites in . Presumably they were near each other. And yet according to Pruneti’s topographical register, they never appear in the same text. Thus, the absence of a link in the network data does not prove the absence of a real tie in the late antique    

See e.g. P.Oxy. ., ., ., and ., all discussed above, pp. –. For network density, see the Introduction above, pp. –. Hickey , . P.Oxy. . (= P.Lond. ., P.Oxy.Descr. .) and P.Oxy. . (= SB ., P.Oxy.Descr. .). The reference to Leontos in the latter is missing from Pruneti.



Social Networks in Byzantine Egypt

Figure  Apionic toponyms and the Oxyrhynchite nome I Unaffiliated sites in circles, sites with known Apionic ties in squares. This image was generated by NetDraw using the affiliations data-set of the Pruneti Oxyrhynchite toponymic register showed through a spring-embedding layout. For this analysis, I retain only those settlement ties of a strength greater than one. Unlike most of this chapter’s analyses, this visualization also retains all settlements, not simply the network’s main components, and shows connections appearing in the top eleven non-Apionic texts, deleted in most of this chapter’s analyses, to show geater detail. Both changes are desirable for a more complete picture of toponyms and ties. For cleanliness of display, isolates and pendants are deleted. 

Oxyrhynchite. Third, and finally, when looking for the Apionic estates within the topographical network assembled from Pruneti’s register, the Apionic texts themselves distort the shape of that network. Assuming that texts are more likely to include place names located near each other will, because of the distortions caused by the Apionic archive, simply produce results more likely to prove that the Apionic estates were all near each other. We can avoid this circularity by generating a second network, one that for corrective purposes removes all texts from the Apionic archive, while still retaining evidence for Apionic sites also attested in non-Apionic texts. The discussion to follow includes the results generated by both sets of networks.

2 The growth of the Apions



Figure  A random large estate and the Oxyrhynchite nome I Unaffiliated sites in circles, sites from hypothetical large estate in squares. As in the previous figure, this image was generated by NetDraw using the Pruneti data-set and retaining all pairs of tie-strength two or more. But the square sites herein were distributed throughout the network using a random number generator. Note the lack of clustering of the squares so pronounced in the previous figure. This random simulation of a large estate uses as many square toponyms as the actual representation of Apionic toponyms in the previous case. 

The results of our first analysis seem striking. Figure  is a visual representation of the ties between Oxyrhynchite toponyms in Pruneti’s register. To focus on the Apions, I provide a key, indicating unaffiliated toponyms as circles, and those with a known Apionic affiliation as squares. This figure shows the Oxyrhynchite’s apparent topographic division into two distinct groups: more of the unaffiliated toponyms are in one segment, more of the Apionic toponyms in the other. Under our assumption that textual ties between toponyms suggest geographic proximity, this chart provides a geographic abstraction of the Oxyrhynchite nome. The apparent concentration of Apionic holdings in one cluster might therefore suggest that Apionic toponyms had general geographic proximity to one another.



Social Networks in Byzantine Egypt

(We might for instance conclude that we are looking at Hickey’s so-called “northern face” (see above, p. ). To be sure of this conclusion, we must answer two questions: () Is there any real correlation between the visual clustering of circles and squares? () Do these two groups still exist in an unbiased data-set, independent of the distorting effect of the Apionic archive itself? One way to test our conclusion is by comparing the actual network against a counter-factual hypothesis. How would our results appear if the Apions had expanded their holdings more or less at random, simply seeking economic opportunities in villages and hamlets throughout the Oxyrhynchite whenever the opportunity presented itself? Pruneti’s topographic register records just over  place names within the Oxyrhynchite nome. According to Mazza’s latest count, the Apions are attested in  different toponyms within the nome itself. Figure  shows what a hypothetical large estate attested in  different, randomly chosen Oxyrhynchite toponyms would look like. The difference between this figure and the figure depicting the actually attested Apionic estates is fairly obvious: the Apionic data places the estate holdings in a cluster, while the random data naturally enough scatters the holdings throughout the nome. The actual numbers appear even more convincing (see Table ). In this preliminary analysis, Apionic settlements do have disproportionately more connections to each other than to non-Apionic settlements, and the network density of the Apionic network is greater than the density of the rest of the Oxyrhynchite network. This would appear to support the hypothesis that Apionic settlements were generally nearer to each other than to non-Apionic sites. Ties between Apionic sites and non-Apionic sites account for , of the network’s total ties; ties between Apionic sites account for ,. Apionic internal ties thus represent a disproportionately high  per cent of total Apionic ties. (Apionic toponyms are only  per cent of the nome’s total.) Much more strikingly, the internal network density of the Apionic settlements is nearly seven times the average network density of the rest of the Oxyrhynchite network. 



Since the original data-set I derived from Pruneti used  of her place names, I used an internet random number generator to pick  random numbers from  to , replacing duplicates with re-generated replacements. Ties between Apionic settlements are also marginally stronger than ties between Apionic and nonApionic settlements: . per cent of the ties between Apionic settlements have a tie-strength greater than one, while only . per cent of the ties between Apionic and non-Apionic sites do. (. per cent of the ties between non-Apionic settlements have a tie-strength greater than one.) Outside of the context of a discrete archive, such increased tie strength would indicate a greater degree of activity connecting the sites in question. What it presumably means in this case is only that Apionic sites



2 The growth of the Apions Table  Distribution of Oxyrhynchite ties I  : Apionic versus non-Apionic settlements

Internal ties Internal density External ties External density

Non-Apionic settlements ()

Apionic settlements ()

, ties of strength  , ties of strength >  . % , ties of strength  , ties of strength >  . %

, ties of strength  , ties of strength >  . % , ties of strength  , ties of strength >  . %

 My thanks to Henning Hillman, who guided me through this section of my work. The analyses summarized here and in the subsequent table below were performed in Pajek, using Pruneti’s Oxyrhynchite data as the network and Mazza’s list of Apionic settlements as a partition. The entire Pruneti network was used, rather than merely its main component. To determine the ties within a group, use Pajek’s Operations→Transform→Remove→Remove Lines→Between Clusters, then Operations→Extract from Network→Partition for each partition. To determine the ties between groups, use Pajek’s Operations→Transform→Remove Lines→Inside Clusters. For Table , the Pruneti network was altered to retain all Apionic settlements, but to remove all ties attested in known Apionic texts.

But the two discrete groups so obvious in the NetDraw visualizations suggest the possibility of distortion. A series of texts such as the Apionic archive, many dealing with the same places, will have the distorting effect of strengthening the connections between those places, and thus making those places appear closer together than they actually are. The geographic proximity of the Apionic estates would be proven, but through a circular and invalid method. Many of the Apionic toponyms are linked through their appearance together in specifically Apionic texts. But removing these links from the data-set and performing a second analysis without the Apionic archive present in the network allows us to see whether ties between Apionic sites exist solely because they are Apionic, or whether they share an independent connectivity as well.

are more likely to appear with other Apionic sites by virtue of their presence in the Apionic archive itself. This is one feature suggesting distortion caused by this archive, discussed herein at pp. –. Removing the top eleven connection-rich non-Apionic texts in the Pruneti register to correct for distortion, as discussed above at n. , only strengthens the conclusions described here. In that corrected analysis, non-Apionic sites have a network density of . per cent, Apionic sites and non-Apionic sites a network density of . per cent between them, and the Apionic sites themselves have a network density of . per cent. Further, Apionic internal density is nearly  times higher than non-Apionic internal density.



Social Networks in Byzantine Egypt Table  Distribution of Oxyrhynchite ties II: with all Apionic texts removed

Internal ties Internal density External ties External density

Non-Apionic settlements ()

Apionic settlements ()

, ties of strength  , ties of strength >  . % , ties of strength  , ties of strength >  . %

 ties of strength   ties of strength >  . % , ties of strength  , ties of strength >  . %

After making these corrections, let us revisit the initial conditions: did Apionic settlements have more connections to each other than to nonApionic settlements, and was the network density of the Apionic network greater than the density of the rest of the Oxyrhynchite network? From the numbers in Table , it is clear that neither of these conditions is satisfied by this final analysis. Ties between Apionic sites and non-Apionic sites account for , of the network’s total ties; ties between Apionic sites account for only , a much more realistic, even under-representative, . per cent of the total number of Apionic ties. Finally, the internal network density of the Apionic settlements is nearly the same as the average network density of the rest of the Oxyrhynchite network. Once the distorting effects of the Apionic archive have been removed, it is necessary to reject the hypothesis that Apionic settlements were generally nearer to each other than to nonApionic sites. Visual representations support this conclusion. The corrected Oxyrhynchite network, in which we remove the Apionic archive from the analysis, produces the visualizations seen in Figures  and  below. With Apionic texts removed from consideration we have a purified look at the textual connections between various sites in the Oxyrhynchite nome. Figure , like Figure  above, indicates the known Apionic holdings with a different shape. Right away, the viewer is struck by a prominent difference between this figure and the first two, namely the lack of a bipartite structure. By 

Nor does removal of the top eleven most connection-rich non-Apionic texts to correct for distortion (see above, n. ) make any difference here either. In that corrected analysis, non-Apionic sites have a network density of . per cent, Apionic sites and non-Apionic sites a network density of . per cent between them, and the Apionic sites themselves have a network density of . per cent. Apionic internal ties are a proportionate . per cent of Apionic total ties.

2 The growth of the Apions



removing all Apionic texts from Pruneti’s data-set, we remove the pronounced bulge in visualizations of that data-set. In the process, we remove the artificial clustering of the Apionic toponyms. The result, in Figure , is one large cluster with the Apionic holdings scattered more or less randomly throughout, their arrangement not manifesting any apparent pattern. On a purely visual level, the presence of the Apionic texts in the first figures distorted our attempts to discern patterns in Apionic estate formation. The apparent clustering of Apionic toponyms in Figure  is an illusion based on that distortion. Further verification of that conclusion comes in Figure . That figure takes the corrected data-set of Figure , but rather than displaying Apionic holdings, it uses the same method employed in Figure  above, and provides a visualization for a hypothetical, randomly generated large estate. The results are consistent with the visualization in Figure : we see one large cluster of Oxyrhynchite toponyms, rather than two, and the hypothetical holdings of our artificial large estate are scattered throughout, without apparent pattern. By this method, the hypothetical large estate is not distinguishable from the actual Apionic estate at the level of these visualizations. To confirm this visual impression, let us turn again to hard numbers. Table  shows the distribution of ties between the complete Oxyrhynchite network and our hypothetical, randomly generated large estate, equal in size to the Apionic oikos. In this random network, ties between estate sites and non-estate sites account for , of the network’s total ties; ties between estate sites account for only ,, . per cent of the estate’s total. The internal network density of the estate settlements is nearly the same as the average network density of the rest of the Oxyrhynchite network. These results are virtually indistinguishable from those derived from the actual Apionic oikos with our corrected data-set above. The conclusion is clear: in terms of network connectivity between various toponyms, there is no difference between the Apionic estates and a large estate generated at random. This suggests that the Apionic estates did not grow outward from a topographically central holding, but were widely scattered, and came under Apionic responsibility more or less at random. To demonstrate that Apionic influence grew evenly throughout the Oxyrhynchite nome, without reference to a single geographic focal point, the previous analyses are incomplete. I have () demonstrated that the Apion estates show the same characteristics as the rest of the Oxyrhynchite network, and () demonstrated that a randomly generated large estate produces the same result. One possible counter-argument remains, that the



Social Networks in Byzantine Egypt

Figure  Apionic toponyms and the Oxyrhynchite nome II: without Apionic texts  This image was generated in NetDraw by removing all Apionic texts from the original, two-mode Pruneti Oxyrhynchite data-set, creating a new affiliations network from this altered data-set, and only then replacing known Apionic toponyms with squares, as in Figure  above. This method corrects for distorting effects caused by the presence of the Apionic archive within Pruneti’s topographic register. Note the lack of the bipartite structure caused by Apionic documents in the previous figures. (As with Figure  above, only ties of strength greater than  are retained.)

provenance and urban focus of the evidence dictates this sort of result, and may thus prevent any group of toponyms from producing different results. Network analysis of the nome’s administrative sub-districts is the best place to look for an example refuting this counter-argument, a collection of places that do appear linked to one another organically, through their rural ties, and not via the “random” links I describe for the Apionic estates. We know that the nome’s sub-districts were organized by geographic order, so these sub-districts would be the best place to find topographical networks displaying the characteristics we have not found in our Apionic and random estates.

2 The growth of the Apions



Figure 

A random large estate and the Oxyrhynchite nome II: without Apionic texts  This image was created using the same method as Figure  above, removing all the known Apionic documents from the data-set. But it also employs the method used in Figure , using a random number generator to create a hypothetical large estate in an attempt to compare the results against the actual Apionic estates. Note the lack of discernible difference between this randomly generated distribution, and the actual Apionic distribution in the previous figure.

My recent study of the ties between Oxyrhynchite pagi and toparchies showed that these administrative sub-districts had a disproportionate tendency towards internal ties to the detriment of ties to settlements in other administrative sub-districts. A closer look at one specific example highlights this tendency. The eighth pagus is the best attested of the nome’s ten pagi. Its  toponyms share  ties of a strength of ,  ties of 

See Ruffini  with Table  on p. .



Social Networks in Byzantine Egypt Table  Distribution of Oxyrhynchite ties III: on a randomly generated large estate

Internal ties Internal density External ties External density

Non-estate settlements ()

Estate settlements ()

, ties of strength  , ties of strength >  . % , ties of strength  , ties of strength >  . %

 ties of strength   ties of strength >  . % , ties of strength  , ties of strength >  . %



This analysis was performed on the complete Pruneti affiliations network, without removing the top texts or the Apionic texts. Removing both groups of texts from consideration produces results similar to our randomly generated estate: ties between estate sites and non-estate sites account for , of the network’s total ties; ties between estate sites account for only , . per cent of the estate’s total.

a strength greater than , and an internal network density of . per cent. This is in contrast to the external ties of the eighth pagus:  ties of a strength of , only  ties of a strength greater than , and an external network density of less than . per cent. The implications here are clear: the papyrological record does produce topographical networks in which high internal ties correspond to geographical proximity. This is true despite the uniformly urban provenance of nearly all our sources. The absence of such ties in the Apionic topographical networks does therefore affirm the absence of geographical proximity among that network’s holdings. tracking apionic chronological expansion Tracking the evolution of the Apionic topographical network over time confirms our impressions that the family’s influence was evenly distributed throughout the nome. Figure  shows the network of Apionic toponyms when those toponyms are disconnected from their ties to the rest of the Oxyrhynchite network. This network can be sliced into finer components by showing only the toponyms that appear in the Apionic archives 

Analysis performed on a data-set in which Apionic texts and the top eleven non-Apionic texts have been removed.

2 The growth of the Apions





Figure  Apionic toponyms extracted from the Oxyrhynchite nome This visualization is produced via NetDraw extracting all Apionic toponyms from the complete Pruneti affiliations data-set. The image has been altered to delete isolated toponyms and retain only ties stronger than .

for the first time in any given decade. Given the chronological data presented in Tables  and  above, it is not surprising that the Apionic toponyms do not begin to coalesce into a unified network until the s and s, the decades when documentation of the family’s reach increases dramatically. Before that period, attested Apionic toponyms appear isolated from each other, lacking any ties to one another in the Pruneti data-set. Figure  shows the topographical ties between places first attested in the Apionic archive in the s. This is, in a single decade, the sudden emergence of  locations linked together by their ties to the family of the Apions. This mini-network emerging in the s has no single geographic focus to it. Its toponyms come from the fourth, eighth and tenth pagi. Still other pagi would no doubt be included if they were known for



Social Networks in Byzantine Egypt

Figure  Toponyms new to the Apionic archive in the s  As with other figures in this chapter, Figure  is generated in NetDraw retaining only ties of strength greater than .

all sites. (Another Apionic toponym, appearing in the s, was located in the seventh pagus. ) In other words, before the s, our documentation attests to no particular ties between places in the Apionic archive. In the s, when those ties begin to appear in force, they are between places located across the entire length of the nome. A similar, if somewhat more traditional, approach can be taken to each individual Apionic prono¯esia. Table  shows the known Apionic prono¯esiai or jurisdictions, proceeding in rough order from south to north. Here, no clear chronological sequence is apparent in the geographic growth of Apionic influence. The southernmost prono¯esia is first attested midway between the two possible dates for the first attestation of the northernmost prono¯esia. The four prono¯esiai located in the nome’s more central regions 

Adaiou (), Skutalitidos (), Sesphtha ().



Sepho: see Pruneti s.n.



2 The growth of the Apions Table  First appearances of each Apionic prono¯esia Prono¯esia

First attestation of prono¯esia

First toponym attested in archive

First Apionic attestation of toponym

A B C D E F G H I L

   or later   before  late VI–VII VI ? ? ?

Kissonos Adaiou all toponyms Paggouleeios Notines Paroriou Papsau Leonidou Skutalitidos Euangeliou Spania

 or earlier  or earlier  or later  ?   or earlier  or earlier  or earlier 



Table proceeds from south to north.

are first attested over a span of nearly fifty years. This evidence is slender, not enough to support a definitive claim that one prono¯esia was organized before or after another. It does not, however, support any suggestion that Apionic influence spread chronologically across the nome from or to any particular direction. The evidence compiled in Table  also seems to suggest that Apionic presence in a region predates the creation of formal bureaucratic structures in that region. In one case, Pronoesia C, its toponyms appear in an Apionic context for the first time only when the prono¯esia itself is first attested, in  or some indiction year  or  years after. In two cases, Pronoesiai G and H, the prono¯esiai are first attested in texts with vague dates, so certain comparisons are not possible. In the remaining seven cases, toponyms from the prono¯esiai are attested within the Apionic archive before the prono¯esia itself. In six of these seven cases, the gap is pronounced, ranging from a generation in one case to over a century in another. Here again, this evidence, while slender, suggests no particular geographic sequence to the spread of Apionic influence throughout the nome. These dates highlight what we have already seen, the emergence of many toponyms in an Apionic context in the mid-sixth century. Perhaps the gap between first attestation of a toponym within the prono¯esia and the first attestation of the prono¯esia itself is a glimpse of the Apionic bureaucracy in motion; Apionic ownership or fiscal share spreads to a new region, and a



Social Networks in Byzantine Egypt

generation later, a formal jurisdiction exists to manage expenses and receipts in that region. Places could fall out of Apionic fiscal jurisdiction as well. Apionic growth was not solely a one-way street. One recently published text appears to show that the epoikion called Aspida, part of the Apionic estate c., had passed into the hands of Flavius Ioannes, son of Euphemia, by . Another appears to show that the kt¯ema called Monimou, then a part of the Apionic estate (), had passed into the hands of the domus divina by the s. comparing the oxyrhynchite and apionic topographical networks A comparison of the Oxyrhynchite topographical network’s structural characteristics with those of the Apionic topographical network reveals a number of interesting differences (see Tables  and  below). To begin with, the Apionic estates produce centrality results quite divergent from those of the Oxyrhynchite nome as a whole. Compared to the centrality results from nome-wide analyses omitting Apionic archival material, that archival material clearly distorts the nome’s network characteristics. On the one hand, the estate’s toponyms have a much lower betweenness centrality than the rest of the nome, which suggests that Apionic estates are connected in a rather nebulous fashion: few Apionic toponyms serve as connecting nodes between otherwise unconnected parts of the estate. On the other hand, the estate’s toponyms have a much higher degree and closeness centrality than the rest of the nome. What this suggests is not immediately clear. It may be that any discrete institution will show a greater tendency towards centrality than the nome as a whole. It may, however, mean that the Apionic estates represent a centralizing tendency for the nome’s social geography when compared to the rest of its history through the Greco-Roman period. This divergence between the degree and closeness centrality of the Oxyrhynchite nome generally and the Apionic estates in particular draws attention to the centrality scores of individual toponyms. Table  shows the top five most connected toponyms in the Oxyrhynchite network both excluding and including Apionic texts, and in the Apionic network itself. There is considerable overlap between the second two lists, which shows  

See P.Oxy. . and P.Oxy. . respectively. See P.Oxy. .. and P.Oxy. .. respectively.



2 The growth of the Apions Table  Measuring topographical centrality I 

Degree centrality Betweenness centrality Closeness centrality

The Oxyrhynchite without Apionic texts

The Oxyrhynchite with Apionic texts

The Apionic estates

. % . % . %

. % . % . %

. % . % . %



For this analysis, I use the largest component of the Pruneti affiliations network, both with and without Apionic texts, which allows us to measure the network characteristics only of those toponyms connected to other toponyms in the Oxyrhynchite data-set. Using the largest component of the Apionic affiliations network allows us to disregard Apionic toponyms we are unable to connect to the rest of the Oxyrhynchite topographical network. In all cases, I use a corrected network in which the top eleven most connection-rich texts are removed.

Table  Measuring topographical centrality II  Degree centrality

The Oxyrhynchite without Apionic texts

The Oxyrhynchite with Apionic texts

The Apionic estates

    

Seruphis Sinaru Pela Dositheou This

Takona Teruthis Skutalitidos Tampeti Sinaru

Takona Teruthis Skutalitidos Paggouleeios Phakra



These analyses use the same data-sets as described in the previous note.

just how greatly the Apionic archive can affect the rank of individual toponyms. The top three most degree central sites are the same in each case, and none of these three toponyms ranks comparably in an Oxyrhynchite network where the Apionic texts have been removed. This is revealing. It suggests that when the influence of the Apionic oikos spread, it did so throughout sites that were not the most connected in our record, and therefore perhaps not the most important in the nome. But during the spread of Apionic influence, the sites in which it grew the most became the most connected sites in the nome. If the rate of connectivity is related to a site’s importance, the rise of the Apions would appear to have altered the shape of the nome’s social geography, at least within the surviving evidence, by changing which settlements in the nome were the most important.



Social Networks in Byzantine Egypt conclusion

At the end of my census of the Apionic population, I suggested that their fiscal share had a population of at least , Oxyrhynchites, which may have been a considerable underestimate. Rowlandson estimated the population of the Oxyrhynchite nome in her period to be about ,, excluding Oxyrhynchos itself. The devastations of the plague could have lowered the figure by as much as a third. I have cautioned against using payment ratios as substitutes for population ratios, but it is worth looking at the numbers to see where they lead. If the Apionic population was roughly one-third of the total population under the jurisdiction of the Oxyrhynchite large estates, and if my Apionic census was at all correct, then the leading oikoi together must have had combined responsibility for at least , Oxyrhynchites. By these estimates, it looks as if the plurality of the Oxyrhynchite’s rural population was financially accountable to a mere half-dozen of the region’s largest landholders. If the payment ratios we have seen in various nome-wide accounts are any indication, the holdings of the church and the imperial house could have added comparable numbers to this total. These numbers affirm the impression developed in Chapter  of considerable nome-wide centralization. If some , Oxyrhynchites fell under the fiscal responsibility of half a dozen families, the Oxyrhynchite social pyramid would have been very steep indeed. Certainly, when a hamlet in the nome’s far north and one in the nome’s far south fell under the same fiscal share, Farmer X in the north and Farmer Y in the south would have been only a few degrees of separation apart. Even if these hamlets fell under different fiscal shares, that small circle of Oxyrhynchite elite all knew each other, hardly adding much more social separation between our two random farmers. And between a random farmer and the elite landowner? In the s, when Diogenes managed to use his property as security in bad loans to two different people, he caused a momentary conflict between the Apionic estates and a local monastery. The solution was easy enough: write to Apion in Constantinople and have him sort it out. Using Apion may be arguing by exception, but nonetheless, in this case, the fate of the residents of Mao, near Ophis, was decided by men in the capital who enjoyed imperial appointments and access to the highest levels of imperial



Rowlandson , .



P.Oxy. ..

2 The growth of the Apions



confidence. The missing variable in this equation is whatever land may have fallen outside the fiscal jurisdiction of the great houses. I have discussed above (p. ) the possibility that a dozen villages, if not more, may have remained outside of these fiscal structures. This part of the nome may have been filled with autopract villages, the Aphroditos of the Oxyrhynchite. But the dramatic increase in attested toponyms for the Apionic estates in this period leaves doubt as to whether any fiscally independent villages could have long avoided the growing influence of the great houses. To return to a striking point: a substantial percentage of the rural residents of the nome dealt with the Apionic bureaucracy. This Apionic sphere of influence – whether it was ownership or mere fiscal responsibility in any given location is not relevant for our purposes – did not spread outward from hamlet to hamlet, from village to village. On the contrary, network analysis demonstrates that Apionic toponyms shared no special connectivity with each other when compared to the rest of the Oxyrhynchite nome. This fact is still further indication of the nome’s social centralization: the one feature of connectivity that Apionic settlements shared with their nonApionic peers was a tie to the urban center at Oxyrhynchos, from which they were administered. Anouthios, a priest from Tarousebt, and Petros, a cripple from Loukiou, probably never met, but they both made payments to the same man, working for the same estate, run from just outside the gates of the nome’s capital city itself. Hardy once wrote that the Apionic scribes must have enjoyed a “great feeling of pride” to write out orders from “our master the consul.” If the , people or more under Apionic responsibility shared these sentiments, their pride would have bound them less to each other than to a central figure most of them never saw. The influence of the Apionic house grew dramatically in the s and the decades that followed, a process I suspect had much to do with the plague, which spread so violently across the eastern Mediterranean from Egypt in . We know the effects of the plague from a number of dramatic eye-witness accounts in the literary sources, and yet the papyri themselves are strangely silent. John of Ephesus describes vast tracts of farmland 

  

Appointments: see Procopius and the wars against the Persians under Anastasius, with references above at n. . Confidence: Strategios II for example is said to have been close friends with Justinian. For the evidence on Strategios, see above, p. . P.Oxy. ., lines  and . Outside the gates: see above, p. . Hardy , . For a general discussion of the Justinianic plague, see Allen  and the studies collected in Little . For the perspective of the history of medicine, see Atkinson . For the fiscal impact of the plague, see Sarris , – and Sarris .



Social Networks in Byzantine Egypt

abandoned for lack of harvesters, and entire villages lost to the disease. It is hard to see why the same should not have held in Egypt. Malalas and Theophanes mention grain shortages in Constantinople in  and again a decade later; agricultural disruption in Egypt was surely behind at least part of this. Justinian’s Novel . () shows that the government’s reaction when confronted with the loss of tax revenue caused by abandoned land was to force local elites to pick up the burden. If the owner of the land does not appear, or is unable to pay the tax . . . We order that documents fully describing the nature and condition of the said land and its appurtenances, shall be drawn up before the judge of the province, in order that the decurions, collectors, or other officials may receive it.

We know that the house of Flavius Apion exercised the authority of a pagarch in . If the plague had any effect in the Oxyrhynchite whatsoever, and Justinian’s Novel was obeyed in any way, Apion and his peers must have been required to shoulder additional fiscal responsibility in this period. The considerable increase in toponyms newly associated with the house of Apion throughout the nome in the period following this legislation is presumably not a coincidence.







See also Procopius Anecdota .–. Novel  is not one of the legal texts typically adduced in discussion of the great estates. See both Hardy  and Gascou , where portions of CJ . are central. Zuckerman a treats Novel .–, but not .. Nor do I find discussion of the legislation in the relevant chapters of Banaji . Novel . appears only in Atkinson , . Trans. Scott : E­ pote d• sumba©h desp»thn o¬asdžpote ktžsewv £ mŸ fa©nesqai £ pr¼v tŸn tän dhmos©wn katabolŸn mŸ ˆrke±n . . . keleÅomen Ëpomnžmata gen”sqai par‡ t t¦v –parc©av Šrconti dhloÓnta tŸn poi»thta kaª kat†stasin t¦v aÉtŸv ktžsewv kaª p†nta t‡ –n aÉt ¦ƒ eËrisk»mena, ¯na oÌtwv aÉtŸn o¬ politeu»menoi ¢goun –x†ktwrev £ b©ndikev £ taxeätai paralamb†nousin. P.Oxy. ..

chapter 3

Aphrodito and the strong ties of village society

introduction Social networks in Aphrodito look much different from those of Oxyrhynchos. Aphrodito villagers acquired new property from, established lease agreements with and married people already familiar to them through previous social and economic connections. These were arrangements between social equals, men who held the same status or belonged to the same guild. The Aphrodito evidence suggests a relatively even distribution of horizontal social ties throughout the village’s social network. This is in sharp contrast to the hierarchical, vertical ties found in the Oxyrhynchite evidence. In Chapter , I argued that missing evidence from hypothetical Oxyrhynchite village archives would not overturn the impressions formed about the nome’s social networks based on the view from the city itself. Here, I make a parallel argument: Aphrodito’s village-level social structures were independent of larger nome-wide structures in the Antaiopolite. Discovery of new papyri from Antaiopolis would still leave us with an Aphrodito rich in multiplex ties between social equals. What specific patterns in Aphrodito’s social interactions are documented in the papyrological record? How did prominent figures in Aphrodito conduct their business? To whom did they turn for assistance and advice? What social connections were behind their land acquisitions and other economic activities? Questions like these are the stock in trade of the sociologists and anthropologists who developed social network analysis. Time and again, our study of Aphrodito answers these questions by highlighting strong, multiplex social ties.  

See the survey in the Introduction above at pp. –. For multiplexity and “strong ties,” see the Introduction above, p. . See also White, ,  and, in the same issue of Semeia, Clark , , on the variety of connection-types (“marriage/kinship; religious mentorship; hospitality; traveling companionship; financial patronage, money, and gifts,” etc.) attested between players in the Origenist controversy.





Social networks in Byzantine Egypt

We find repeated evidence for what we might intuitively suspect, that Aphrodito’s key players relied heavily on people they already knew. This is a characteristic Jeremy Boissevain noted in his early work on social network analysis, when he wrote of a “tendency for single-stranded relations to become many-stranded if they persist over time, and for many-stranded relations to be stronger than single-stranded ones, in the sense that one strand – role – reinforces others.” Specifically, in the case of Aphrodito, its inhabitants turned time and again to those they already knew in one social capacity to fulfill functions required in other capacities. The networks of several well-known figures from Aphrodito have this principle at their heart. The “entrepreneur” Aurelius Phoibammon’s connections to the family of Apollos and Dioskoros could serve as textbook examples of multiplex ties. So too could Phoibammon’s pattern of land acquisition. In the dispute settlement analyzed by Traianos Gagos and Peter van Minnen, multiplex or multi-stranded ties explain the presence and behavior of several key actors in the settlement. Even Aphrodito’s ongoing struggle with the local pagarchs, long the subject of modern scholarship, becomes clearer when reconsidered in the light of the multiplex ties some pagarchs themselves had to Aphrodito. Much of this is intuitive: we might well expect social actors to conduct business through people already known to them. But, as Boissevain has pointed out, this is more a feature of village societies than of urban or specialized ones. Indeed, those of us living in societies of the latter sort may have hundreds of social ties that never become strong or multiplex. This chapter’s conclusion, that Aphrodito’s social network has a high degree of multiplexity, suggests future directions for a comparative approach that would place Aphrodito in the larger context of Mediterranean village society. A major consideration that must be addressed in making this argument is the question of change over time. Our understanding of Aphrodito’s fiscal status and the importance of large landholders there is still in flux. We know that Aphrodito enjoyed the right of autopragia, which entitled it to collect taxes on its own behalf. The emperor Leo had granted Aphrodito autopragia in the fifth century; and in the sixth, to protect that right, the villagers placed themselves under the direct protection of the Empress Theodora. Perhaps the grant of autopragia to Aphrodito was an attempt   

 See below, pp. –. Boissevain , .  Boissevain , –. See below, pp. –. Leo: P.Cair.Masp. ., on which see below, n. . Theodora: P.Cair.Masp. ., on which see below, pp. –. See Geraci  and Salomon  for the village’s relationship with the

3 Aphrodito and the strong ties of village society



to curb the growth of large landowners in the region. Indeed, it was long thought that large landholders like the Apions were missing from the village. But this impression may be incorrect. Theodora’s protection may indicate a larger imperial landholding presence in the region than previously supposed, an intriguing possibility when we recall that the Apions themselves began as estate managers for a previous empress. Constantin Zuckerman has recently argued that the former prefect Ioulianos owned up to three-fifths of Aphrodito’s land. And in a parallel argument, Zuckerman has reinterpreted a well-known Aphrodito petition to suggest that the village lost its autopragia at some point in the s. These arguments force a reappraisal of Aphrodito’s relationship with the pagarchs responsible for regional tax collection and also lead us to ask whether the village’s social networks were impacted in any way. The pagarchs who appear periodically in the Aphrodito papyri resorted to bizarre and seemingly desperate tactics to get Aphrodito fully under their control. If Zuckerman is right – and more concrete proof is still needed – the strained relationship between Aphrodito and the nearby pagarchs is evidence not of persistent village independence, but instead of friction after that independence was lost. This chapter’s primary assertion is that Aphrodito displayed strong, multiplex, evenly distributed social ties. Its secondary assertion – explored in more detail in Chapter Four as well – is that this situation remained unchanged throughout the sixth century. If Aphrodito lost its independence to the regional equivalent of a Flavius Apion, as Zuckerman argues, that fact did not alter the village’s social networks. Multiplex ties among relative social equals drove Aphrodito’s internal social landscape both before and after the s, Zuckerman’s posited transitional period, even though a nome-wide superstructure like that of the Oxyrhynchite may have ultimately controlled the village’s fiscal status.

 

 

empress, and for the restructuring which must have taken place after her death. For autopragia in general, see Hardy , ; Gascou , ch. ; and R´emondon , for a distinction between what he calls “l’autopragie simple” of landlords who deliver their collection to the local provincial officials, and “la super-autopragie” of those who are responsible for an infrastructure which oversees delivery to Alexandria. For a recent historiographical survey of Byzantine “feudalism” generally and autopragia more specifically, see Banaji , .  A comparison I owe to Roger Bagnall. Bell , . Zuckerman a, , based on a figure in P.Ross.Georg. .; but the figure has no secure context, and may refer to fiscal share rather than ownership. See the remarks in Bagnall forthcoming and in my conclusion below, p. . Sarris , like Zuckerman, argues for a major role for large landholders in Aphrodito. Zuckerman a, . The argument, based on assumptions about the length of pagarch office tenure, is not without difficulties: see below, p. . For more on the office of the pagarch, see below, pp. –, and nn.  and .



Social networks in Byzantine Egypt aphrodito: the village and its archives

Aphrodito, modern-day Kom Ishqaw, would still be considered an obscure and unimportant village in Upper Egypt but for a series of spectacular papyrus finds from the late antique and early Arab periods. Instead, we now know it as a fascinating rural community with a considerable degree of Hellenistic literacy and culture. Much of the evidence for sixth-century Aphrodito relates somehow to the life, family, and career of Dioskoros, a poet and landholder who has intrigued modern commentators. Almost all of the surviving texts from that period come from the archives of his extended family. The papyri once in their possession detail the activities of several members of this extended family group and of various colleagues and neighbors. Thus a single group of texts illuminates a wide array of social connections. Local affairs otherwise lost come into view because a local 







See Bell  for the discovery of the papyri, and for a description of Kom Ishqaw early in the twentieth century; Keenan a for a survey of the papyri and their locations today; Bagnall and Worp  for the chronology of Aphrodito’s Coptic texts, pushing some texts in P.Mich.  well into the seventh century; and Fournet ,  with n.  for the role of museum archaeology in Aphrodito studies. As Fournet points out (, ): “Un guide dans l’esprit de celui des archives de Z´enon ne serait pas inutile.” The re-editions, editions of unpublished texts, collected images and database of the Aphrodito material he is currently undertaking, along with a prosopography in preparation by the current author (Ruffini forthcoming b), will go a long way to meeting this need. For Dioskoros, see MacCoull  and Fournet  and their reviews, Keenan c and MacCoull  respectively. See also MacCoull  on Dioskoros’ Coptic archive; MacCoull  on Dioskoros’ relations with the Egyptian dukes; MacCoull  for the monastery under Dioskoros’ management; Fournet  for the convergence of poetic, epistolary, and petitioning styles in the Dioskoran corpus; Fournet  for Dioskoros’ Iliad and the Scholia minora, and the grammatical papyri from the Dioskoros archive, both revisited in Fournet ; Fournet  on new material from Dioskoros’ library and archive; Fournet , discussing Dioskoros’ place in the literary culture of late antiquity; van Minnen  on Dioskoros and the legal culture of his time; Bell’s introduction to P.Lond.  and Fournet , – on the hands of Dioskoros; and the works of Baldwin, Kuehn, and Maspero, cited in the bibliography and, where relevant, below. The exceptions are P.Michael. –, from a different findspot from the other published texts. In his introduction to P.Lond. ., Bell wrote, “In all probability . . . these sixth-century documents from K¯om Ishgau all come from the archive of Dioscorus.” The consensus is now more nuanced. Fournet ,  n.  doubts that the texts of P.Mich.  and P.Vat.Aphrod. come “des archives de Dioscore stricto sensu.” The former go with P.Michael. –, found c., a point recognized in Gascou , . The latter probably belong to that group, potentially the archive of Phoibammon. Keenan has adopted this perspective, writing in ,  of the “seemingly independent survival” of the papers of Phoibammon. Neither P.Michael., P.Mich. nor P.Vat.Aphrod. contain any texts in the hand of Dioskoros, all of which were  finds. P.Mich. . ( or ? see CSBE ) and  (/) add a further wrinkle. No certain prosopographical ties exist between these later texts and the main archive of Dioskoros and his family; see Bagnall and Worp . For network analysis over the full chronological spread of the archives, see Chapter , pp. –, –. The characterization is that of Gagos and van Minnen , .

3 Aphrodito and the strong ties of village society



notable asked Dioskoros to serve as a witness or keep a copy of someone else’s loan agreement. During a critical stretch from  to , the archive also includes documents for which Dioskoros served as notary in Antinoopolis. Further, Dioskoros, his father Apollos, his uncle Besarion, and his cousin-in-law Phoibammon were all active in the purchase, lease, or management of land throughout the Aphrodito area. Their business activities made them what James Keenan has called “keys in the dialectic between ecclesiastical institutions and the secular economy, between the Byzantine Egyptian poleis and their satellite agrarian villages.” But they were politically active as well. Zuckerman has recently described the family as one of the “clans” that filled the village offices. While this family’s importance within the village is obvious, the size of the village is not. Estimates of Aphrodito’s population range from several thousand to ,, but these are only educated guesses based on the number of people in the published papyri. In light of Aphrodito’s former status as a nome capital, one might expect a higher figure rather than a lower one, but in the absence of better evidence, the jury remains out. No one has estimated the potential population of the village’s surrounding land. Aphrodito had farm- and pasture-land on all sides, the fields labeled by the points of the compass. The Aphrodito cadastre indicates that at one point in the sixth century, the village registered , acres of cultivated land. This figure included both land owned by villagers (kom¯etika) and land appearing in Antaiopolite civic registers (astika). Rowlandson estimated the rural population of the Oxyrhynchite nome at roughly ,, a population density of – persons per square kilometer. For Aphrodito,

  

 



 Keenan b, . Generally, see Keenan b, particularly . Zuckerman b, . For the family, see an abridged stemma at Figure . The lower figure: Gagos and van Minnen , , roughly based on the number of individuals attested in Girgis. The higher figure: MacCoull , , where she deduces “some three thousand tax-paying male heads of household” by consulting “the indices of personal names recorded in published papyri.” But earlier (MacCoull , ) she thought Aphrodito’s “population may have been about –.” Zuckerman argues (a, ) on the basis of the carrying capacity of Aphrodito’s land for a low number of cultivators, around ,, and accordingly thinks that the village was not larger than c., people, closer to MacCoull’s earlier estimate. For brief notes on the topography of the village, see Keenan a,  n. . , arourae (, acres) of arable land,  arourae (. acres) in vineyards,  arourae ( acres) of thruis, and . arourae (. acres) of garden land, some ,. acres of land all told. See the summary table at Gascou and MacCoull , , with line  of the cadastre (SB .): ¾(moÓ) kÛm(hv) %frod©t(hv) spo(r.) (Šr.) / Es qr(u©d.) (Šr.) ld L d h’ ˆm(p.) [(Šr.)] ra par(ad.) (Šr.) fo L. An aroura as . acres: Rowlandson , . Rowlandson , , by comparison to what she calls “a relatively low maximum population for Roman Egypt as a whole of no more than five million.”



Social networks in Byzantine Egypt

comparable density would give a rural population of no more than a few thousand people. The people of Aphrodito called their home a village of small landowners, and modern scholars have to some degree accepted this description. While Oxyrhynchos demonstrates the importance of the large estate, Aphrodito proves the importance of small and mid-sized landholders in this period, especially when taken with evidence from Hermopolis. Quantitative analyses of the distribution of landholdings among outsiders owning land at Aphrodito support a comparison with Hermopolis. The two sites together have been taken as evidence of a “significantly different and more diverse picture of the society and economy of late antique Egypt . . . [in which the] landed wealth of the medium and small landowners combined exceeded that of the really big landowners.” apollos, dioskoros, and the village elite Apollos, Dioskoros, and the other headmen or pr¯otok¯om¯etai of Aphrodito are case studies in the importance of strong, face-to-face ties in the village’s social structure. These prominent landowners and pr¯otok¯om¯etai were Aphrodito’s inner circle. While their holdings were small relative to the large estates of Oxyrhynchos, their relative financial comfort nonetheless gave them pride of place among their village peers. They were the local  



 



P.Lond. ., intro. and lines –: ˆp¼ leptoktht»rwn g‡r | sÅgkeitai ¡ kÛmh, cited by Keenan a, . For Oxyrhynchos, see above, Chapters  and . For Hermopolis, see most recently Banaji , –, and his characterization of the region as typified by “small aristocracy,” “owners who clearly resided in the areas near their estates,” () something quite unlikely in Oxyrhynchos. Sarris  represents the backlash, skeptical of peasant autonomy in Byzantine Egypt generally and Aphrodito specifically. See Bagnall , : the Gini index for individual landholders in the category of astika or urbanowned land in the Aphrodito cadastre (.) compares favorably to the analogous example of the Gini index generated from a register of individual Antinoite landholders in Hermopolis in the fourth century (.). But this is an incomplete picture; Bagnall’s forthcoming analysis of Zuckerman’s Aphrodito register produces a Gini index for Aphrodito kom¯etika, land owned by villagers, of ., and a combined figure for both villager- and urban-owned land of .. Gagos and van Minnen , . For Dioskoros, see n.  above with PLRE .–. For Phoibammon, a relative by marriage (see Gagos and van Minnen , with their commentary to lines – on p. ), see Keenan . For the career of Dioskoros, see MacCoull , – and Bell . For stemma, see Gagos and van Minnen , Figure , with revisions at van Minnen , , and Figure  herein. For pr¯otok¯om¯etai, see above, Chapter  n. , with Gascou b and more recently, the thorough survey in Harrauer , including a list of all papyri relevant to pr¯otok¯om¯etai throughout Egypt and a prosopography of the known holders of the office.

3 Aphrodito and the strong ties of village society



administrators who ensured that Aphrodito collected its own taxes and paid its administrative bills out of the sums collected. Apollos, the father of Dioskoros the poet, served in  and for much of the s as pr¯otok¯om¯et¯es in Aphrodito, alternating with his brother Besarion. The family seems to have been less prominent in the s, when the headmen Charisios and Bottos are more visible, but in  a crisis thrust it back into the spotlight. Apollos and his son Senouthes – Dioskoros’ elder brother – received a letter from the governor of the Thebaid at the behest of Count Flavius Ammonios, their apparent patron, informing them that Bottos and his son had resigned their positions in Aphrodito, and that Apollos and Senouthes were now the village’s sole headmen. The explanation for these resignations may have been fiscal in nature; the same year, Ammonios wrote Aphrodito’s headmen asking them to regulate the gold in their tax payments more carefully. Apollos’ public career culminated in a trip to Constantinople in late . With him went a priest named Victor, the son of Besarion, his now dead brother. This trip is usually seen as part of the ongoing saga of Aphrodito’s relations with the pagarchs, in which Aphrodito repeatedly asserted its right to autopragia and placed itself under the protection of the empress Theodora. Indeed, Apollos’ son Dioskoros would travel to the capital twice on just such a mission. Aphrodito’s rights were apparently a matter of family interest. But Apollos’ actions on his arrival in Constantinople are mysterious. Keenan proposed that a twentysolidi loan Apollos received from a Flavius Anastasios in Constantinople 

  

 



 and earlier: P.Flor. . (= Mitthof , .) and P.Cair.Masp. .. s: P.Cair.Masp. ., ., P.Lond. .. Besarion: P.Lond. .. See Zuckerman a,  with Keenan’s original reconstruction in b, . P.Cair.Masp. .. For date and discussion, see Zuckerman a, , and b, . Fournet , – with P.Cair. SR  (). P.Cair.Masp. ., from Constantinople in January , with Bell , Keenan b, Zuckerman b. This is the only attestation of Apollos as a Flavius: Keenan b, . Maspero restored a Flau©ov %]pollä. [v]
io[sk»r]ou to P.Cair.Masp. ., which dates to  or  (see CSBE  with n. ), but this has been contested: see BL . and Malz , . MacCoull ,  cites P.Cair.Masp. . for Apollos as a Flavius, but neither Apollos nor the Dioskoros mentioned as a scholastikos is a Flavius in that text. P.Cair.Masp. ..–: u¬¼v | [Bhsar©]wnov toÓ t¦v makar©av mnžmhv. Most recently, Zuckerman b, . As Keenan points out (b, ) this is “more assumed than proved.” His alternative suggestion (b, ) – that the trip was a religious pilgrimage by “the priestly nephew in train of his older uncle, recently . . . turned monk” – does not seem likely, especially in light of recent doubts that Apollos became a monk: see below, n. . For the texts relating to all three trips, see Zuckerman b, with Keenan a, –. For the latter two, see below pp. –.



Social networks in Byzantine Egypt

in January of  suggests that he ran into unanticipated delays while confronting the clamor of the big city. One or two details can still be filled in. The social connections implied by the loan suggest that, whether or not Apollos and Victor were the “country bumpkins” of Keenan’s portrayal, they nonetheless hit the ground running. The lender, Anastasios, was not merely a rich man, but an imperial banker: ¾ kaqosiwm”nov kastrhsian¼v t¦v qe©av trap”zhv kaª ˆrguropr†thv. In addition, he had interesting friends: the agens in rebus Flavius Olybrios, Flavius Maximos of the Dalmatian archer corps and Flavius Sanos all witnessed the loan. Sanos was Egyptian, a boatman and assistant (bo¯ethos) of “the most God-loving lord Menas.” A plausible scenario based on pre-existing social networks begins to emerge. Far from risking an uncertain journey into the imperial unknown, Apollos and Victor may have left for Constantinople with Egyptian connections already established in the capital city to smooth their arrival. The loan of twenty solidi – not a trivial sum – may have been among the fruits of these Egyptian connections. Indeed, the sum was to be payable four months later, to Anastasios’ agent Thomas, in Alexandria. The Menas whom Sanos served held a monastic title, which may connect his boat to the ecclesiastical fleet in Alexandria. Sanos himself may have been in charge of the boat bringing Apollos to Constantinople. This loan may thus be less the work of a capital-city financier expanding his network to unsuspecting Egyptians than evidence of pre-existing Egyptian networks functioning in the imperial capital. Back in Aphrodito, Apollos had an estate substantial enough to support the foundation of a monastery bearing his own name and surviving for   





Keenan b,  and with more detail in .  PLRE . from P.Cair.Masp. .. Bumpkins: Keenan b, . Fl. Olybrius : PLRE .. The other two do not appear in the prosopography. Keenan b,  poses a number of questions about these witnesses: “Were they acquaintances of Anastasius’? Were they on hand by chance or had they been specifically invited to witness the loan? Were they paid a fee for their witnessing?” Whatever the answers to the second two questions, the answer to the first is almost certain to be affirmative. For Sanos in preference to Sonos, see Fournet and Gascou ,  n. . The name is known in ´ Upper Egypt: see P.Bala’iza ., . Maspero called Sanos “un Egyptien” “`a en juger par le nom de son patron,” no doubt thinking of the many Egyptian office-holders named Menas attested in the sixth-century (s.n. in volume  of the PLRE). The popularity of the name in Egypt was due to the popularity of the cult of the Egyptian saint of the same name; see Papaconstantinou , . Maspero read “Lord Menas” instead of seeing reference to a Kyros son of Menas; see Fournet and Gascou, ibid., for affirmation of this reading. For diakon¯et¯es as a “moine au service d’un moine plus aˆg´e,” see Fournet and Gascou , . The suggestion of a connection to the ecclesiastical fleet is Keenan’s: ,  and n. .

3 Aphrodito and the strong ties of village society



years after his death. In contrast to Phoibammon, his niece’s husband, about whom we will read more below, Apollos’ holdings seem to have been concentrated in Aphrodito’s southern plain. For instance, in  AD, Aurelius Ioannes leased from Apollos a plot of land located in that region. Some years later, the sale of a plot of land in the southern plain of Aphrodito gave as a border marker a plot named Pharatopos, belonging to the heirs of Apollos (t¼ kt¦ma tän klhron»mwn %pollätov
iosk»rou Faratopov). The land being sold was near the lane leading to the hamlet of Psiniou. A lease dating to  mentions another plot of land belonging to the heirs of Apollos near Psiniou. In the s, a land-lease from Dioskoros to a shepherd of Psinabla mentions land in the southern plain, presumably part of Dioskoros’ paternal inheritance. Apollos died by /, and his nephew Victor probably died by November . The many references to the heirs of Apollos in the family papers suggest an initial period of uncertainty before his property was finally divided. His son Dioskoros, usually believed to have been born c., was by no means the family’s sole survivor. Dioskoros’ uncle Besarion had a number of descendants, as did an unnamed aunt who had died and left her children in the care of Apollos as they grew up. One of these children 

  

    

For Apollos’ holdings, see Keenan b, . For the Apa Apollos monastery, see P.Cair.Masp. ., , , P.Mich. .–, PSI ., SB ., and MacCoull . Dating the monastery to  (e.g. Gagos and van Minnen , , , and  and MacCoull , ) relies on Apollos’ appearance as Apa in PSI . (). In fact, any foundation could predate that text. The orthodoxy that Apollos became a monk is weak. Apollos appears first as an Apa (PSI ..) and then as a Flavius (P.Cair.Masp. .). Apollos appears as an “Apa Apollos the pr¯otok¯om¯et¯es” in a Coptic letter published in MacCoull b, – (= P.Ismailia inv. ). In , Derda and Wipszycka list (p. ) apparent cases of the title “Apa” applied to non-monastic office-holders. Most recently, van Minnen ,  n.  rejects altogether the notion that Apollos became a monk, claiming, without citing Derda and Wipszycka, that PSI .’s “Apa” “does not mean he was a monk,” and adding that the adjective sebasmios “refers to his status as a pr¯otok¯om¯et¯es.” The latter point is not certain: Apollos is not specified as a pr¯otok¯om¯et¯es in the PSI text, any more than he is in the pasturage agreement P.Cair.Masp. ., where he is also addressed as sebasmiotatos. Nonetheless, Apollos’ under-attested career as a monk appears to have vanished. Keenan b, , citing P.Lond. ., , , P.Cair.Masp. ., ., P.Michael. , and P.Flor. . (= Mitthof , .). Four of these seven examples are in the southern plain. P.Flor. .. P.Michael. ., which probably dates to . The date: an eighth indiction, which Crawford (p. ) took as “probably either  or .” Since Apollos died in / (see below, p. ) and is dead herein, the former date is excluded. P.Cair.Masp. .. But there is no particular reason to identify these plots as identical: see the section on ‘Property management and social connectivity’ below, p. . P.Lond. .. For Victor’s death, see MacCoull , –. For the legal problems attendant to Apollos’ death and the division of his property among his heirs, see van Minnen , Zuckerman b, and nn.  and  below. P.Cair.Masp. .–, with translations at van Minnen , –.



Social networks in Byzantine Egypt

is the other Dioskoros attested in the Aphrodito papyri, a cousin of the poet and a classic example of the confusion that naming after a grandfather can cause. Dioskoros the poet also had an unnamed half-sister and a half-brother named Menas, with whom he was at times on bad terms. Dioskoros’ other brother Senouthes is well attested in the papyrological record, but is only now receiving notice from modern scholars. This Senouthes was an active participant in the affairs of his village: he served with his father as pr¯otok¯om¯et¯es and went with Dioskoros to Constantinople in . Senouthes was presumably older than Dioskoros, for he appears first in the papyrological record and disappears first from it. According to Zuckerman’s reconstruction, he even acted alone in “le rˆole d’homme fort du village” for a brief period after his father’s death. If Senouthes was older, this should give us pause: despite our fascination with Dioskoros, appropriate because of his papyrological output, he may not have been the senior member of the family after the death of his father. In the years before his father’s death, Dioskoros worked for Count Flavius Ammonios, one of the most powerful men in the region and one with various ties to Aphrodito. This relationship was presumably passed on from his father, also an agent for Ammonios. The oldest datable text attesting to Dioskoros appears in this period, a few years after his father’s trip to the imperial capital. With the coincidental deaths of both Ammonios and Apollos, Aphrodito’s situation became more precarious. While still a young man, Dioskoros followed in his father’s footsteps and journeyed to Constantinople in /, securing his inheritance from seizure  



 

 

See van Minnen , – and Zuckerman b on P.Cair.Masp. .–. See van Minnen , , with Dioskoros’ list of damaged pasturages including an entry –x Ëpobol¦v MhnŽv ˆdelfo(Ó) mo(Ó) appearing at P.Cair.Masp. .., and van Minnen  passim for an analysis of Dioskoros’ inheritance dispute with these half-siblings. For his career, see Zuckerman a, –. For a list of texts, see the introduction to P.Mich. .; for the assertion that Senouthes was older than Dioskoros, see Gagos and van Minnen , , and lines –, with the note on p. . Identification of Senouthes as such began with P.Hamb. .; the editor’s proposal has usually been followed since. Van Minnen , – rejects the identification, although he admits that some other family tie between Dioskoros and Senouthes is possible. His objection is unconvincing. He supposes that if Senouthes had indeed been the brother of Dioskoros, the latter would in P.Hamb. . have identified him more explicitly than “Senouthes the son of Apollos.” But why? The phrase is certainly accurate enough as it stands. Pr¯otok¯om¯et¯es: e.g. P.Cair.Masp. . (). Constantinople: P.Cair.Masp. .. Zuckerman a, –. Senouthes appears inter alia in P.Thomas  (–, including P.Michael. ), P.Cair.Masp. .. ( or ; BL .) and as pr¯otok¯om¯et¯es in P.Cair.Masp. . (also ). He cannot be identified with certainty in texts after the s (e.g. P.Cair.Masp. ., ), although Sijpesteijn proposed a possible appearance in P.Mich. . ( or ). Zuckerman a, , citing P.Cair.Masp. ., an account of money payments “p. ar’ –mo(Ó) Seno(Å)qo(u)” (line ). The identification is tempting, but not necessary.  P.Cair.Masp. ., . See pp. –.

3 Aphrodito and the strong ties of village society



by the pagarch Ioulianos. Perhaps a year or two prior to this trip, he composed one of Aphrodito’s most fascinating documents, a petition from Aphrodito’s civic elite to the empress Theodora in defense of the village’s cherished autopragia. Count Ammonios’ son Theodosios inherited his father’s position of power over the village, but apparently not his good relations with the villagers. In , Dioskoros traveled again to Constantinople, where he registered complaints against Theodosios and the same pagarch Ioulianos, who were attempting to put collection of Aphrodito’s taxes under their own authority. Dioskoros was apparently successful, receiving in response an imperial rescript which orders the Duke of the Thebaid to investigate the situation, and if Dioskoros’ claims are warranted, to put a stop to the activities of Theodosios and Ioulianos. It was in this period that Dioskoros’ literary output began, some fifty poems written over the course of the next twenty years. For the next fifteen years, before the next major crisis in Aphrodito’s history, Dioskoros’ activities are unremarkable. We have an account in his hand detailing payments from a large estate. We have a series of receipts for payments issued by the pagarchs to Dioskoros and Kornelios son of Philantinoos, as well as further accounts in the names of both men. The receipts to Kornelios were all issued through a woman named Sophia, now 

  

  

P.Cair.Masp. ..–: to. u. v. ìIoulian t –ndox(ot†t) | pros. ž. k. [on]ta. v. –pelqe±n kaª ˆfel”sq. a. i. pŽsan. For this first trip to Constantinople, which scholars have typically overlooked because of their focus on his second trip in , see Fournet , , with n. , and P.Cair.Masp. ., SB ., and P.Cair.Masp. ..., with a translation at Fournet , . Nor was Dioskoros alone; SB .. refers to multiple sons of Apollos. P.Cair.Masp. .; see below, pp. –. Identifying Theodosios as the son of Ammonios, himself a son of Theodosios, follows Zuckerman b,  and his “sch´ema onomastique banal.” P.Cair.Masp. .. For this text and Dioskoros’ trip to Constantinople, see van Minnen ; MacCoull , ; and Bell ,  (with translations of the key portions of P.Cair.Masp. . in Bell and van Minnen). The two transgressors are Fl. Iulianus  (PLRE .) and Theodosius  (PLRE .). The former we know to have been a pagarch, but no office is specified for the latter. Fournet ,  thinks Theodosios “peut-ˆetre dioec`ete de la domus divina.” Fournet sees this “Theodosius affair” as the first of two “life crises” stimulating Dioskoros’ poetic output; see Fournet , – for Theodosios, and  for Dioskoros as a “crisis poet,” summarized by MacCoull , –. As with his earlier trip, he took advantage of his presence in Constantinople to secure assistance in personal matters as well: Fournet , . See Martin ; Salomon ; Geraci . Note how the author of SB . describes (line ) the men of Aphrodito as coming to Constantinople and having “made themselves bothersome,” di’ Àclou geg»nasin ¡m±n. For more on , see Geraci , . Fournet  publishes them in chronological sequence, with a table of dates given at –. P.Cair.Masp. . I–II, from an eighth indiction perhaps –; see Fournet forthcoming. This paragraph largely follows his reconstruction of  therein. P.Cair.Masp. . II–VIII, again with Fournet forthcoming.



Social networks in Byzantine Egypt

believed to be both Kornelios’ granddaughter and Dioskoros’ wife. This is the first hint we have that in the s Dioskoros established a family of his own. We will explore the social connections suggested by these texts at the end of the chapter. From  to perhaps as late as  Dioskoros was away from Aphrodito, serving as a notary (nomikos) for the dukes in Antinoopolis. MacCoull has suggested that the move to “an up-and-coming capital” was motivated by Dioskoros’ need for “a better job and better money.” This is no doubt true, but only part of the story. Presumably the ongoing conflict back home between Aphrodito and the pagarchs left the situation there increasingly unsettled. In the most extreme example of this crisis, thirteen of his landholding peers were moved from one regional prison to another over the course of half a year or more, presumably in the period just prior to . In , Dioskoros wrote a petition in which he complained about the alleged misdeeds of Menas the pagarch and the imprisonment of his peers, and called life in Antinoopolis a period of exile in a foreign land. We do not know whether his ancient audience was fooled by this daring bit of rhetoric, but we should not be: certainly Dioskoros and his father before him had already been much further away from home than this, asserting their rights on much the same issues. This petition is one of the most important pieces of evidence to survive from late antique Aphrodito.  

 

 



 See below, p. . Fournet forthcoming. For Dioskoros and the dukes, see MacCoull . This move to Antinoopolis after the “Menas affair” is the second of the two life crises contributing to Dioskoros’ literary output: see Fournet  and MacCoull . For the chronology of Dioskoros in Antinoopolis, see Fournet , : his departure from Aphrodito was “sans doute a` l’extrˆeme fin de ,” as he is attested in Aphrodito in November by P.Lond. ., and his return was at some point between his last attestation in Antinoopolis in November  (P.Cair.Masp. .–) and his first attestation back in Aphrodito at the end of  (P.Cair.Masp. .). MacCoull , . See P.Cair.Masp. ...–; MacCoull , – with the imprisonment only in passing (); Bell ,  for a conservative treatment of Dioskoros’ chronologically ambiguous prose; and the description of the events outlined at Fournet , : this period was the second “crise” inspiring Dioskoros’ literary output. P.Cair.Masp. ...: ¾ Šqlio[v] ™wv nÓn –pª x”nhv sÆn t”knoiv. For trips by Apollos and Dioskoros to Constantinople, see above, pp. – and . For Apollos and associates “finding themselves” in Antinoopolis on more than one occasion, see below, p. . For speculation that Dioskoros lived in Alexandria as a young man, see MacCoull ,  and a. Dijkstra ,  also finds legal training in Alexandria to be “likely” for Dioskoros. P.Cair.Masp. .: see Masp´ero  for discussion of titulature; Masp´ero  for a redating of relevant texts; Hardy , –; Bell  and ,  for partial translations; R´emondon ,  for a defense of Menas; Geraci ,  for a translation of col. II lines – at ; Gascou ,  for a curious comparison between the arrests of Aphrodito’s major landholders in  and the arrest of coloni fleeing from the land to which they are tied; MacCoull , – for partial translations and narrative, with emphasis on under-appreciated literary aspects of the piece, in which “his prose takes wing” (); Fournet , , in which he argues for a Homeric influence

3 Aphrodito and the strong ties of village society



Despite a certain (deliberate?) vagueness in chronology, Dioskoros painted a very vivid picture. Thirteen of his fellow Aphrodito landowners were visiting Thinis for the market, as they did each year, when the dioik¯etai of Serenos the scholastikos ambushed them and threw them in prison. The implication is that Serenos had done so at the instigation of Menas the pagarch, who had just recently handed some of Dioskoros’ land to Kyros the bo¯ethos of Phthla and the shepherds of Phthla. Dioskoros leaves the connections between some of these players murky, either because they needed no stating, or because it would have been impolitic to be more specific. We do not know why Menas and Kyros took Dioskoros’ land, we do not know what Kollouthos the augustalis had to do with the seizure, and we do not know why Serenos was involved at all. In P.Lond. ., Dioskoros says that Menas acted “apparently owing to arrears in, or (alleged) nonpayment of, the taxes,” but naturally contests that Menas had a right to claim owed taxes of any kind. Nonetheless, one thing seems clear. Dioskoros mentioned one piece of land only, and we may assume from his silence that his other properties remained intact. He was therefore not in Antinoopolis as an uprooted exile, but as a man very much intent on taking his case to the highest powers at hand. A large portion of Dioskoros’ documentary work survives from this period in Antinoopolis, perhaps brought back on one or more trips home before his death. His legal work included involvement in an inheritance case, the drafting of various wills, handling loans and sales of land, and so forth. In this period he also composed many of the poems for which he has acquired a dubious reputation in modern times, including the epithalamium for Kallinikos, duke of the Thebaid, and the encomia to Athanasios and Romanos. To this period also belongs what MacCoull

      

in ...–; Fournet , , relating this influence to Dioskoros’ Homeric library; and Banaji , , against which, see below at pp. –. P.Cair.Masp. . should be taken in conjunction with P.Lond. ., which Bell argued, no doubt correctly, refers to the same events. Following his suggestion, we may guess that  was Dioskoros’ first attempt to seek official recourse upon reaching Antinoopolis, and that  followed shortly thereafter. For another text with possible connections to this episode, see MacCoull . For more on Serenos and his role in this affair, see below, pp. –. Cf. P.Lond. ... For a summary, see Fournet , –, with his comment () that the “pr´etendue expropriation n’est sans doute qu’une forme d’intimidation.” Quoting Bell in his introduction to the text. A point Bell understands, but takes as “only conjectural” in his introduction to P.Lond. .. For references, see MacCoull , –. For Dioskoros’ literary output, see the works cited above at n. , particularly Fournet . For Kallinikos, Athanasios, and Romanos, see P.Aphrod.Lit. , , , , , and , with a reconstruction of the career and family of the former in Fournet , – and , –.



Social networks in Byzantine Egypt

called “his most ambitious poetic effort,” two encomia to Ioannes, dux of the Thebaid in the mid to late s. A petition he addressed to Kallinikos’ brother Dorotheos on behalf of one of the latter’s tenant farmers is an example of the blurring of genres in Dioskoros’ writing. Dioskoros returned to Aphrodito by , when we lose sight of him presumably as he approached sixty. Although we do not know when he died, we can be reasonably certain he left his wife Sophia and their children behind when he did. In P.Cair.Masp. ., Dioskoros says that he went abroad – that is, to Antinoopolis – with his children. An unpublished Berlin papyrus contains an acknowledgement of debt issued by Dioskoros in  to a Ioannes of Lykopolis in payment for a course the latter gave to his son Petros. An unpublished Coptic letter from Dioskoros’ sister sends greetings to Sophia and “all your children.” Copies of the many receipts issued to Dioskoros and to Kornelios through his granddaughter Sophia appear in the same papyrus as a land-lease addressed to Sophia c.. What began earlier in the century as the archive first of Apollos and then of his son Dioskoros may have ended its life as the archive of Sophia and Petros. phoibammon son of triadelphos Phoibammon son of Triadelphos, whose wife was Dioskoros’ cousin, exemplifies the extent to which Aphrodito’s social networks involved strong, multiplex ties. His case also suggests that these social networks are unlikely to have been altered by any loss of Aphrodito’s autopragia in the mid-sixth 

 

 

 

MacCoull , . The poems are P.Cair.Masp. ..v and P.Berol.  + P.Cair.Masp. . = BKT ..–, MacCoull’s H and H at –, now P.Aphrod.Lit.  and . This Ioannes is Ioannes  at PLRE .; see Fournet , –, correcting the earlier opinion (e.g. MacCoull , ) which placed Ioannes in the mid-s. Fournet  on SB ., a petition originally published as a letter and using some of Dioskoros’ own poetic vocabulary. The statement at MacCoull ,  that Dioskoros was alive in  by virtue of his appearance in P.Cair.Masp. . no longer holds. He appears only in the portions of that collection of texts dating to the s: see Fournet , –. Dioskoros may have still been alive, but we cannot be certain. The last text in Dioskoros’ hand is P.Cair.Masp. . (), the last text showing him alive P.Cair.Masp. . ( see BL . and .). See Fournet forthcoming for a detailed reconstruction of this period. P.Cair.Masp. .... I am indebted to Jean-Luc Fournet for this information. The son in question, Petros, also appears in a fragmentary context in P.Cair.Masp. . (/). Ioannes of Lykopolis also appears in P.Cair.Masp. . (), itself written by Dioskoros, and presumably in the lacuna of P.Lond. ... -P.Mus´ee Copte inv. .– (ined.): INEETCOF[IA MN NEC-] | HRE TH. ROU. P.Cair.Masp. . IV; Fournet forthcoming corrects Gascou’s proposed reading of line  at BL ..

3 Aphrodito and the strong ties of village society



century. Phoibammon, whose career spans a large portion of our period, has acquired a reputation as one of late antique Egypt’s entrepreneurs. James Keenan has told the story of Phoibammon’s “economic advance . . . at the expense of an Egyptian soldier who was plunging ever more deeply into debt.” He has also argued that Phoibammon’s economic activities represent social connectivity: “Phoibammon fulfills, at least in part, some of the specifications of the entrepreneur, particularly in his ‘forging a link’ between ‘unconnected circuits’, in Phoibammon’s case, the link between absentee landlords, monasteries, churches and the village work force.” The best-known instance of Phoibammon’s entrepreneurial spirit is an early one, involving the soldier, Flavius Samuel. Samuel, a so-called “slow writer,” son of Kollouthos and father of several daughters, owned  arourae east of Aphrodito which he leased to Phoibammon in , simultaneously borrowing from him substantial sums in gold and grain. For Samuel, the following year only made his situation worse: he borrowed still more from Phoibammon, ultimately giving up any right to reclaim his own land until the debts were repaid. Victor son of Paul, actuarius from Thmonachthe, the scriniarius Apollonides, and Ioannes son of Phrer provided grain measures for these transactions. Ioannes son of Phrer is described as Samuel’s former tenant: ìIw†nnou Frhr©o[u] t. oÓ –moÓ progewrgoÓ. Perhaps his continued presence on the land explains his involvement in Phoibammon’s business affairs. Keenan has rightly disputed descriptions of Samuel as a small farmer and Phoibammon as a rich landholder. Samuel’s  arourae are referred to as t¼ geÛrgion (“field” or “farm”), and were therefore presumably in one plot. For a single holding, this is not inconsiderable: it is larger than any individual plot recorded in the Aphrodito cadastre from the same period. 

 

  

Keenan . For Phoibammon, PSI ., P.Michael. –, , P.Mich. ., and many others, listed in the introduction to the latter text. The documentation concerning Phoibammon has a history of its own: see Gascou ,  and n.  above for P.Michael. and P.Mich.  as products of the same period, belonging to Phoibammon’s personal archive. Keenan a,  n. , citing I.M. Lewis, Social anthropology in perspective (Harmondsworth ), –. See P.Michael. – and P.Mich. ., the former with Keenan b, –, and the latter with Gascou , –. Samuel’s progressive indebtedness to Phoibammon is a striking counterexample to Banaji’s general argument that the advanced monetization of the late antique economy contributed to soldiers and bureaucrats asserting their economic power in new ways. Victor: P.Mich. .. Apollonides and Ioannes: P.Michael. .  Keenan , .  P.Michael. ., .. P.Michael. .. For the cadastre, see Gascou and MacCoull  (= SB .). Although the cadastre records land registered only on the astika or town account (line  summarizes the entire cadastre, ¾(moÓ) ˆstik(än) ½nom(†twn), with numbers following that derive solely from the individual entries therein), and some individual plots on the village account may well have been larger, it is still



Social networks in Byzantine Egypt

Indeed, Samuel seems to have been an absentee landlord, perhaps with land he no longer found sufficiently profitable. This fits Phoibammon’s pattern: Keenan counts nine different lessors to Phoibammon, all of them unable or unwilling to farm the land they own. These lessors ranged from shepherds to locals with moderate holdings to monasteries to prominent office-holders like Flavius Ioulianos. This last case is a curious one: Ioulianos was a vir illustris and pagarch of Antaiopolis in . This is a telling detail: Dioskoros’ own relative by marriage gives proof that the family enjoyed close relations with at least some of the local pagarchs. Indeed, this case shows that Phoibammon was leasing from a pagarch at the start of the very decade in which Zuckerman has argued that Aphrodito lost its independence from those pagarchs. This example is in keeping with a larger pattern we discuss below, in which the pagarchs display a series of multiplex ties to local society, ties more complex than their tax-collecting responsibilities alone would entail. It is hard to see why deeply embedded and mutually convenient lease arrangements like those between Phoibammon and the pagarch Ioulianos would have been altered by the village’s loss of autopract status. One of the most remarkable aspects of Phoibammon’s career is his considerable longevity. The first firmly dated text in which he acts on his own, making loans and renting land, dates to ; the last, in which he leases pasturage to shepherds, dates to . The nature of his activities – land-based economic entrepreneurship – is consistent over nearly half a century. If any radical changes in Aphrodito’s fiscal status or patronage took place, it did little to alter Phoibammon’s economic activities. One scholar has suggested that Phoibammon was the prot´eg´e of Apollos, the poet’s father, adding that Apollos helped Phoibammon “stand on his feet.” The publication of part of a dispute settlement shows that Phoibammon married Apollos’ niece Anastasia, also called Tekrompia, “the dove.” In the s a series of receipts acknowledged tax payments in the

  

  

striking that Samuel’s individual plot stands out so prominently against the largest available set of comparanda. Keenan , . Keenan , , where “one of Phoibammon’s practices” should be read for “one of Samuel’s practices.” PSI ., where Florence Lemaire reports that Ioulianos is to be read in preference to the original Alexandros, whose PLRE entry (.) now disappears. For the PSI text discussed in more detail as an example of an antimisth¯osis, see Keenan b, –, who thinks Antinoopolis more likely than Antaiopolis. For more on Ioulianos, see below, pp. –. For Aphrodito, Dioskoros, and the pagarchs, see below, pp. –.  P.Michael. .  Gagos , . P.Michael. . P.Vat.Aphrod. , now part of P.Mich.Aphrod. (= SB ., ?). I follow Zuckerman b,  over van Minnen , ; the Peristera in P.Cair.Masp. . (ed. princ. Maspero ) can now be read as Anastasia as well.

3 Aphrodito and the strong ties of village society



name of the now-deceased Apollos through Phoibammon. This activity came years after Phoibammon had married Anastasia, c. or earlier. As Gagos points out, in the years leading up to this marriage, Apollos would not yet have the help of his children in business matters. Perhaps Apollos drew Phoibammon more and more into his confidence after he married Anastasia, ultimately delegating some fiscal matters to him. The reverse sequence of events is also possible, but would not change the essential interpretation: the coincidence of Phoibammon’s business and family ties displays precisely the pattern one expects to find in a small village, that of single-stranded relations growing into multi-stranded strong ties. property management and social connectivity Phoibammon’s career also suggests that land acquisition in the Aphrodito area proceeded along social lines. Phoibammon showed a natural tendency to expand in areas in which he already had a presence, and was therefore socially connected. When he leased and agreed to pay the taxes on land belonging to the monastery of Psentouses, he noted that the land was located “to the west of my own holding called ‘the dyers.’” The proximity was presumably causal. The pasturage he bought from two shepherds in  bordered his own land to the southeast. He had already bought land from the same shepherds immediately to the east of his new pasturage. The new purchase in  thus gave him three contiguous properties in Aphrodito’s eastern plain. Indeed, Phoibammon shows a certain preference for the eastern plain: the  arourae he leased from Samuel were on that plain’s far border, the plot of land he and a Victor, son of Kollouthos, leased out to Besis and a shepherd named Tabes in  was also in the eastern plain, and the plot they leased out to shepherds named Anouphis and Pakouis in  was there as well. The topographical descriptions of these plots have details that do not mean anything to the modern reader, but nonetheless make clear that we are dealing with three separate plots.   

   

SB . (= Pintaudi and Sijpesteijn, ) and SB . (= Gagos ), dating to / and / respectively, with P.Princ. . (/): see Fournet . Gagos and van Minnen , , with references at n.  above. P.Mich. ..–. The Psentouses monastery seems to have been under the jurisdiction of the better-known Apa Sourous monastery. As Gascou puts it (, ): “Il semble que . . . le monast`ere d’Apa Sourous se pr´esente en h´eritier de sa fortune” (i.e. the “monast`ere dit de Psentous`es”). Psentouses, and thus Phoibammon’s holdings in this case, were located in the western plain “of the same village” (t¦v (aÉt¦v) kÛm(hv)), presumably meaning Aphrodito. See P.Michael. . P.Michael. : the allotment formerly belonged to the village of Thmonachthe. Our records show other land transfers between Thmonachthe and Aphrodito: e.g. P.Cair.Masp. .. P.Michael. . Victor is likely Phoibammon’s nephew; so the editors of P.Bingen . P.Michael. .



Social networks in Byzantine Egypt

Heirs of Charisios son of Hermauos

the oil-maker

Kollouthos and Victor new sale

Phoibammon

pasturage that Kollouthos and Victor sold to Phoibammon

Ama Maria hamlet

Figure  P.Michael. : Kollouthos, Victor, and neighbors The hamlet of Ama Maria: P.Michael. . begins: . . . .] . . . . –p. ». kou [sic] *ma Mar©av. If this reading stands (this line of the text has received no BL entries), it and Pseketos (P.Cair.Masp. ., ) present us with epoikia on village territory, another wrinkle to Banaji’s typology of epoikia: see Banaji , , where he distinguishes between the “general dichotomy and physical discreteness of village and epoikion” in the Oxyrhynchite on the one hand and the Fayum’s “substantial villages” and the epoikia named after them “which must clearly have been in their immediate vicinity” on the other.



Phoibammon’s own social connections may be at the heart of this pattern. We get one clue from the description of the pasturage he purchased in  (see Figure ). The land he bought from the shepherds is just south of estates in the hands of the heirs of Charisios, son of Hermauos. Charisios himself would have been well known to Phoibammon during his lifetime. Phoibammon’s relatives by marriage, Besarion and Apollos – uncle and father of Dioskoros, respectively – worked with Charisios for many years. Charisios served as pr¯otok¯om¯et¯es with first one, then the other in a period spanning over two decades. It is natural that when Phoibammon 

 

Charisios is Girgis , where Girgis records attestations in  and . He also appears in P.Cair.Masp. ., which was not published in full, but which recent arguments place in either / or /. If the later date holds, we can place the death of Charisios in /, since P.Michael. ., dating to the summer of , mentions klhron»mw. n Ca. ris©ou ë Ermauä. tov prw. tokwmžtou. See Zuckerman a,  and the same argument at Ruffini ,  developed without reference to Zuckerman. Following Keenan b, . Besarion and Charisios together: P.Cair.Masp. ., which Keenan b,  takes to be prior to . Apollos and Charisios together: P.Cair.Masp. ., . desc., . desc., P.Flor. ., all of which Keenan (ibid.) thought belong to the s, but Zuckerman now dates to the late s: Zuckerman a, –. Keenan’s reconstruction is the more compelling; based as it is on the twin inclinations to place Apollos’ career as a headman prior to his attestations as a suntelest¯es in the s, and to make the most chronologically compact sequence possible given his attestation as a headman in .

3 Aphrodito and the strong ties of village society



started to acquire land in Aphrodito’s eastern plain, he bought plots bordering the estates of a family long known to his own. We also know Charisios from the Aphrodito cadastre, where two parcels of his land, of five and nine arourae respectively, were later registered under the name of a city councilman named Elias son of Nemesianos. Elias we know from a lease agreement in two parts from  and  between a shepherd named George and Dioskoros himself involving land that used to belong to Elias. Elias’ father Flavius Nemesianos appears in  receiving one in a series of receipts from a prono¯et¯es working for Dioskoros’ father Apollos. Once again, we have a picture of land ties developing between people all connected in tight social clusters, and spanning the course of several decades without any apparent disruption. Thanks to a fascinating text recently published by Bagnall and Keenan, we now know a little about Phoibammon’s activities in the village itself as well. A property division dating c. to  gives us the first autograph of Phoibammon’s signature, which “shows him to be a slow writer, but not hopelessly so.” The division is between two sisters, Aurelia Elisabeth and Aurelia Eudoxia, on one hand, and Phoibammon and his brother Kollouthos on the other. Damage to the text makes it difficult to tell exactly what sort of property is at stake, but mention is made of a number of rooms, including a bread storage-room (ˆrtoqžkhv), and “the rights and appurtenances of the entire house” (tän dika©wn kaª crhsthr©wn Âlwn | t¦v p†shv o. «k©av). This text is tremendously useful for the prosopographical connections it provides. Two of the five witnesses to the division are known from elsewhere. Phoibammon son of Hermauos may be identical to the one appearing in a number of Aphrodito texts, most importantly as the foremost representative of the local shepherd guild. In that capacity, he and the other shepherds made a land-guarding agreement with the Aphrodito    



 

Gascou and MacCoull , ll. ,  (= SB .). Elias himself now presumably being dead: see P.Lond. .a , reading Bell’s ëHliopol. [it]e. us†meno[u] as ëHliou pol. [it]e. us†meno[u] with BL .. Dates: BL .. P.Cair.Masp. ..; BL .. P.Lampros = P.Bingen  dates to c.–, a range dictated by the first appearance in the papyri of Phoibammon himself in , and the last appearance of the text’s notary, Abraham, in . The quote is from Bagnall and Keenan , , commentary to lines –. This text is the first to tie Phoibammon to his brother, Kollouthos son of Triadelphos. Paulos, apparently a third son of this same Triadelphos, appears in P.Michael. . P.Bingen .–. Mention of a fifth part (p”mptou m”rouv) in both this text and P.Michael. , where Kollouthos appears as well, leads Bagnall and Keenan to conclude that the two texts belong together, but they are unable to untangle the sequence of pieces. For the connections in this paragraph, I follow Bagnall and Keenan’s commentary on p. . For other ties from Dioskoros and family to local shepherds, see Chapter  below, pp. –.



Social networks in Byzantine Egypt

landholders, a group in which Dioskoros and Phoibammon were both central. Mouses son of Hermauos, a doctor, appears in the Aphrodito cadastre of / as a property manager for land belonging to the Apa Sourous monastery. (The fact that Dioskoros served as a middleman lessee for the same monastery suggests the presence of further multiplex ties here. ) Phoibammon son of Phib, who wrote on behalf of the two sisters, may be the kt¯et¯or who signed the petition to the empress Theodora prior to . Phoibammon’s brother Kollouthos himself suggests new connections: the Victor son of Kollouthos appearing in two other texts as Phoibammon’s partner seems likely to have been his nephew as well. The point here is minor, but worth making: when residents in Aphrodito settled their village domestic affairs, they used the same social connections which mapped their activities in Aphrodito’s rural space as well, namely, family members, fellow landholders, and, in the case of the shepherds, associates whose pastoral lifestyles had brought them into close business contact. Phoibammon’s connections to his fellow landowners and the shepherds who owned portions of and moved through the same rural terrain were also employed in the village itself. For sake of comparison, it is difficult to imagine the Apions working the same way: we would not expect a business agent in one of their rural properties to appear in a deal related to their home in Oxyrhynchos proper. Moving beyond Phoibammon, P.Michael.  gives more hints about the process of land acquisition in the Aphrodito area. The diagram below, Figure , based on the description of the landholdings in that text, suggests a similar process at work. Isaac son of Beskouis held three plots of land. Two of the plots were next to each other. The third was on the far side of a plot his son Ioannes sold to Apollos the bo¯ethos in this very text. We do not know the order in which Isaac came to own each plot, or specifically how, but we can imagine a process analogous to that of Phoibammon’s expansion. His son’s sale provides some of the details. Ioannes held his plot through his mother, who had married twice, once to Isaac, and once to a      

P.Cair.Masp. .. Gascou and MacCoull , col. , l. , with note (= SB .). P.Cair.Masp. ., with a discussion of the Apa Sourous holdings at Keenan b, . Girgis , , citing P.Cair.Masp. .; see also P.Vat.Aphrod. , where Bagnall and Keenan think Phoibammon and Kollouthos might also appear paired together. See P.Michael.  and . Interpretation of this text is admittedly very difficult. I follow the translation in its rendering of the breakdown of the various plots of land therein.

3 Aphrodito and the strong ties of village society



the heirs of Apollos, son of Dioskoros Isaac

loannes

Isaac

Isaac

(to Psiniou)

Figure  P.Michael. : Ioannes and neighbors

man named Mathias. Ioannes’ title was to one-half share of his mother’s land, which was in turn only a third-share of a jointly held farm (¡m©souv mou | m”rouv toÓ mhtrou. m. o. u. tr©tou m”rouv ktžmatov). His halfbrother has already sold his half-share of the third to Apollos, who now stands to gain the whole, through his purchase from Ioannes. This alone is complex enough. But Ioannes added that he will also sell to Apollos his “sixth share of the lodge and of the house and of all the rights appertaining thereto.” The scenario probably runs as follows. The mother, who remained unnamed, inherited a title to a third of what we may imagine to have been an estate originally belonging to her parents. Two unknown siblings presumably received the other third-shares. Her two children through two different husbands then in turn received one sixth share each, presumably upon her death. This Apollos son of Iosephios, about whom we know little else, represents through his purchases the partial defragmentation of the title to this estate. But what of the other two-thirds? No one other than Isaac appears in this text. It may be that the other shares simply go unmentioned because they are irrelevant to the transaction at hand. It is equally possible that the third-shares were not abstractions, but real plots of land. If they were contiguous, the boundaries of Ioannes’ estate are clear: the only possible owner of the other shares is Isaac himself. This reconstruction suggests that land acquisition in late antique Aphrodito followed social and familial lines. Isaac may have purchased    

P.Michael. .–, with the m. o. Ó. certain from a parallel phrase in lines –. P.Michael. .–. Although this was clearly not an isolated act by Apollos: P.Michael.  shows him acquiring land elsewhere, in the village of Thmounameris. This point is not a new one, having been made for Roman Egypt generally in Rowlandson . Rowlandson begins () with Montevecchi’s observation that “many sales of all kinds of real property involved more than one member of the same family,” and points () to the likelihood that “information about land for sale passed by personal contact.” This chapter’s inclusion of social



Social networks in Byzantine Egypt

two of his plots from his wife, her family, or her heirs. Indeed, the marriage itself may have been logical because of his land’s proximity to hers. This scenario may imply that Mathias was the first husband and Isaac the second, benefiting financially from the fragmentation of his wife’s estates upon her death. On this limited evidence, land acquisition on the village level seems to have followed a different pattern than it did on the nome level. In the Oxyrhynchite nome, the Apionic evidence suggests that large estates acquired new holdings at random, not spreading outward from initial centers. Here, on a much more localized scale, estate expansion is anything but random, proceeding one contiguous plot at a time through pre-existing social connections. settling disputes The  volume Settling a Dispute publishes the longest text related to Phoibammon, and highlights the importance of strong, multiplex social networks for Aphrodito’s villagers. As the most detailed analysis of a late antique dispute settlement to date, it is a guide to the important role that social connections played in the mediation process. Phoibammon and his wife were one party to the dispute, Aurelius Nikantinoos the other. The events, summarized in a dialysis or settlement text drafted early in the reign of Justinian, are hard to follow. The parents of Nikantinoos had borrowed money using a piece of their property as security. They died, leaving the debt to Nikantinoos, but the property itself to his nephew and nieces, Kollouthos, Eudoxia, and Antonia. (Or so it seems at first glance:

  



multiplexity in this discussion merely adds theoretical rigor and additional examples. Rowlandson’s point is made in the context of an extended assertion that the “land market” as such in Roman Egypt was rather small; Keenan’s remarks on the “land market” at Aphrodito in , –, do little to change this impression. See Chapter  above, pp. –. P.Mich.inv.  + P.Vat.Aphrod.  = P.Mich.Aphrod. (= SB .), cited herein as Gagos and van Minnen ; reviewed in Bagnall  and Delia . Nikantinoos is otherwise unknown: see Gagos and van Minnen , , where they argue that his name “suggests that the connection of his family with Antinoopolis is of long standing – maybe an ancestor won a victory in the games that were held there.” For the date as c., see Gagos and van Minnen , . They propose a general range from  to : Justinian is mentioned in the text, which dates it to  or later, but Apollos does not yet appear as the “most worshipful Apa Apollos,” an appellation he has in . This is not a strong argument. Apollos does not for instance appear as “Apa” in Constantinople in : see above, n. . Their argument is helped by their adduction (in commentary to lines –) of SB ., which dates to  and gives a parallel to the oath formula in their text. But Bagnall  n.  points out that the oaths are not exact matches, and finds the chronological speculation of the authors unfounded.

3 Aphrodito and the strong ties of village society



Roger Bagnall has suggested that Nikantinoos may have been a party in this dispute, if not in fact its initiator, specifically because he too was a partial heir. ) Meanwhile, the original lenders had also died, leaving the loan contract to their son Iosephios. Kollouthos, Eudoxia, and Antonia went on to sell the property to none other than Phoibammon himself and his wife Anastasia, cousin of the young Dioskoros. As Gagos and Van Minnen point out, “selling property under lien was illegal according to Greco-Roman law.” In practical terms it left Nikantinoos without the property he owed debt on, and it left Iosephios with a deed of surety against land now in the hands of an irrelevant third party. The settlement process implied in this text shows the importance of social networks in Aphrodito. Nikantinoos agreed to pay the debt owed to Iosephios – the amount is nowhere specified – in exchange for a payment of seven solidi two carats, and twenty artabae of wheat from Phoibammon. This is not precisely a small sum, but in the previous decade, Phoibammon had been able to loan Samuel over twice that amount without difficulty. It must have been but a fraction of the cost of the land for it to have been a reasonable price for the settlement. More importantly, in Nikantinoos’ own words, “the judgment of the mediating friends” influences his decision to accept the payment offered him: di‡ t¼ –m• e«lhf”nai –k plžrouv t‡ d»xanta d©dosqa© moi pr¼v | tŸn kr©sin tän m”swn f©lwn. Apollos played a central role in the settlement, appearing in the Antinoopolite courthouse on behalf of his niece Anastasia and her husband Phoibammon: di‡ soÓ toÓ prolecq”ntov %pollätov | prwtokwmžtou –n t kat‡ cÛran politik dikasthr©. Apollos was a village headman (pr¯otok¯om¯et¯es), Phoibammon and Nikantinoos both Aphrodito collective taxpayers (suntelestai). Headmen and taxpayers were hardly far apart    

 

 Gagos and van Minnen , . Bagnall . For the settlement in a legal-anthropological context, see Gagos and van Minnen , –. See above, n. . “Because I have received in full what has been agreed upon to be given to me in accordance with the judgment of the mediating friends.” Gagos and van Minnen , , lines –, with trans. on , and earlier at , line . “By you, the aforesaid Apollos, village chief, in the local civil courthouse.” Gagos and van Minnen , , lines –, with trans. on . Gagos and van Minnen , , lines –. For the suntelestai in general, see Liebeschuetz , , against the traditional definition as “a technical term for a member of the consortium of landowners responsible for the taxes of a city or village,” and asserting a much more simple meaning of “taxpayer.” But from the following year, and without reference to him, Mirkovic , affirmed () the standard opinion that “Die Verbindung der syntelestai mit der Steuereintreibung ist in verschiedenen Texten klar bezeugt.” See also Laniado  and Sarris , . For a summary of the traditional definition, see Gagos and van Minnen , , drawing on Gascou , ; see



Social networks in Byzantine Egypt

socially: some individuals are attested as both. Given the fact that the settlement took place away from Aphrodito, in Antinoopolis, and given Nikantinoos’ already precarious situation, he may have wondered whether Apollos could really make a settlement on Phoibammon’s behalf. The witnesses and the mediators themselves presumably provided part of that guarantee. The identity of the mediators, whom Nikantinoos describes as “good friends” (lines –: f©loi | ˆgaqoª), is not specified, but Gagos and Van Minnen make a reasonable guess that: these friends are presumably adult male inhabitants of Aphrodito who happen to be in Antinoopolis at the time of the settlement. It is likely that they are mutual friends of both Nicantinous and Apollos, and it is not unlikely that some of them are among the witnesses to the present settlement, who appended their signatures below.

Who were the witnesses? Senouthes was a son of Apollos, “presumably [the] older brother of Dioscorus” himself. Mouses son of Psaios and Hermaos son of Bottos both appear elsewhere as witnesses to the agreements made between Aphrodito’s landholders, police officials, and its shepherds and shepherd guild, which we will discuss in detail below. Apollos son of Besios we also know reasonably well. He appears in  as a witness to a debt owed to Apollos, the poet’s father, by a David son of Victor. He also appears with an Aurelius Abraam, the two men issuing an



  

 

also the corresponding LSJ entry s.v.: “member of a land-owners’ union which is responsible for the collection and payment of its taxes.” Keenan a,  is content with an agnostic “contributary.” See e.g. Charisios: Girgis , P.Cair.Masp. . and ., and herein, pp. , . For Keenan’s “suspicion” or “hunch” that the two titles represent stages in a cursus honorum from pr¯otok¯om¯et¯es to suntelest¯es to kt¯et¯or, see Keenan a, , and below, Chapter , n. . The chronology for such a cursus would be tight. Dioskoros is a pr¯otok¯om¯et¯es in P.Cair.Masp. . in , the same year in which he appears as a suntelest¯es in P.Cair.Masp. ., a text Keenan does not cite. The settlement dispute is a significant challenge to his interpretation. Apollos appears as a pr¯otok¯om¯et¯es while Phoibammon is a suntelest¯es. One might therefore suppose that the sequence ran in the other direction, from possessor to contributary to headman, but this falls when we encounter Dioskoros as a kt¯et¯or later, in P.Cair.Masp. ., in . Perhaps more than one position could be held at once; Mirkovic ,  affirms the latter point: “Man konnte gleichzeitig syntelestes und pr¯otok¯om¯et¯es sein.” In this case any order or progression would have held much less significance. Gagos and van Minnen , line  and the note to it on p. . Gagos and van Minnen , cited at n.  above and discussed at p. . Hermaos: P.Cair.Masp. ... Mouses: P.Cair.Masp. .... He also appears in P.Hamb. ., but with no obvious prosopographical connections. For the shepherd agreements, see Chapter  below, pp. , . Gagos and van Minnen , lines – and the note to them on page , which cites Gascou and MacCoull ,  n. , where Apollos appears in the Aphrodito cadastre. P.Flor. . (= Mitthof , .).

3 Aphrodito and the strong ties of village society



acknowledgement of debt to Apollos, again the poet’s father. He may appear in the settlement because of these connections to Apollos, an example of a prominent figure in Aphrodito turning to his strong ties in the process of resolving social tension. In short, the witnesses all had social ties to Apollos, the representative of one of the disputing sides. If the witnesses were the mediators as well, Nikantinoos may have found himself simply surrounded, whether they were in fact “good friends” or not. If, as Bagnall has suggested, Nikantinoos initiated this dispute proceeding to disentangle himself from the entire affair, he was not socially well armed. To my mind, it stretches the imagination to believe that when Apollos came to Antinoopolis, perhaps specifically intending to settle this dispute, four of his relatives and business associates “found themselves” in town (all of the witnesses describe themselves as eËreqeªv –n %ntin»ou), while Nikantinoos seems to have produced no men of his own, despite himself living in Antinoopolis: di†gwn –ntaÓqa –pª t¦v %ntino”wn kallip»lewv. No doubt the witnesses had some connection to him in order to be acceptable; Mouses’ role as suntelest¯es may have made the two well known to each other. Nonetheless, this seems a clear-cut case of Phoibammon benefiting from the strength and the geographical reach of his in-law’s social networks. Another Aphrodito dispute settlement from the late s or s manifests the same sorts of multiplex social connections between the parties involved. The dispute at hand is as complex as that just discussed, and because the settlement is fragmentary, some of the details remain unclear. The prosecuting parties, Psaios son of Mousaios and Talos, his wife, daughter of Heraklios, had a series of complaints against Apollos, Paulos, and Maria, the children and heirs of an otherwise unidentified Ioannes, perhaps himself a priest, now dead. Victor son of Besarion and Senouthes son of Apollos represent the defending parties. Senouthes is almost certainly Dioskoros’ brother, making Victor his cousin.   

  

P.Lond. ., not after /: see SB ., n. to line . Whether the debt in  is a public one or private is not clear from the fragmentary text. Gagos and van Minnen ,  line . P.Mich. .. For the date of this text, see Gascou , : the terminus ante quem, the death of Apollos, is /, not . If Maspero was right about the date of P.Cair.Masp. . and P.Mich. . comes after it because Abraam is no longer a bo¯ethos, then  dates to between  and /, not Sijpesteijn’s range of  to . Zuckerman a, – places the settlement in the late s and the original sale in the late s. For the short summary that follows, see Sijpesteijn’s introduction to P.Mich. .. t”knwn kaª klhron»mwn ìIw†nnou | toÓ t¦v eÉlaboÓv mnžmhv: P.Mich. ..–. Sijpesteijn missed this identification. For Senouthes, see above, p.  and n. .



Social networks in Byzantine Egypt

Why Senouthes and his cousin Victor were reasonable choices to represent the heirs of Ioannes is not clear, but once again, multiplexity of ties may be the answer. Psaios himself appears in a highly fragmentary land lease in / involving Besarion, Victor’s father. In that text, prior to the current dispute settlement, Ioannes had leased to Besarion land registered in the name of Psaios, perhaps the very land at the heart of the dispute. Note also that Apollos, father and uncle to the two representatives, was himself one of the witnesses: as with his own witnesses in the previous dispute, he “found himself ” (eËreqeªv) in Antinoopolis at the right time to lend a hand. The connections of the other witnesses are more obscure. Two were Aphrodito residents only in Antinoopolis on a visit. Flavius Victor, another witness, was in town more permanently as a scribe of the ducal office. Although nothing is known about the one unnamed pr¯otok¯om¯et¯es appearing as a witness, he presumably knew Senouthes and Apollos, both themselves pr¯otok¯om¯etai. The complaints Psaios and Talos brought against the heirs of Ioannes make it clear that Psaios himself had strong multiplex ties to Ioannes and his family. Psaios had sold to Ioannes a holding called Tausiris, and then complained that he had not received full payment for the sale (t¦v tim¦v toÓ ktžmatov kaloum”nou Taus©rewv). Psaios had also sold two other plots to the children at a low price: ™tera dÅo ktžmata ˆrourän | Âswn –st©n p”praken aÉto±v ½l©gou timžmatov. Furthermore, he had “executed a security for them on a house that belongs to him for the security and freedom of encumbrances of the fourth part of the holding of Theodosius from Pakerke located in the plain of the village of Aphrodite.” Presumably, the multi-stranded complexity of this case is part of what drove it to mediation. But the social ties between Psaios and the defending parties did not stop there. With amusing honesty, Psaios and Talos admitted that “they had often used loud complaints [poll†kiv –kbožsesi kecr¦sqai] in the Holy Church against John of discreet memory.” Was the Holy    

   

P.Cair.Masp. .; see Gascou and MacCoull , note to line  for the date. So Zuckerman a, –, restoring P.Cair.Masp. .. as [Taus]iriov, the name of the plot in P.Mich. .. But Maspero was confident that Irios was complete on its own. P.Mich. ... P.Mich. .., with Sijpesteijn’s note on p. . But he, as a Flavius while Apollos was still alive, is presumably not, contra Sijpesteijn’s introduction, the Victor son of Ioannes in P.Cair.Masp. . and ., both of which refer to an apparent Aurelius in the s. P.Mich. ..–. Given by name first at P.Mich. .., then again at ...  Sijpesteijn’s trans., P.Mich. ..–. P.Mich. ..–. Sijpesteijn’s trans., P.Mich. ..–, which begins o¬ toÓ diÛkontov m”rouv –d©daxan, “the people of the prosecuting party proved.”

3 Aphrodito and the strong ties of village society



Church where Ioannes himself was a priest? It is easy to imagine this sort of thing in a village context: when multiplex ties between two actors start to deteriorate, the conflict might move to a more public stage where the complaining party can find the defendant and plenty of others close to him. These two examples of dispute settlement display the importance of multiplex ties in Aphrodito’s social networks. But we have no reason to imagine that these characteristics were not also found in, say, villages throughout the Oxyrhynchite. In fact, these two disputes may hint at circumstances in which villagers otherwise used to dealing with their own affairs found it necessary to go up to the city, in this case the regional capital, Antinoopolis. Perhaps here we see the occasions when a relatively decentralized, non-hierarchical village integrated into the larger region’s centralizing hierarchy. These disputes may have been settled because the relevant villagers’ strong networks provided adequate social capital to bring the affairs back to the village from the regional capital.

flavius ammonios comes Count Ammonios, one of the largest landowners in Aphrodito’s social landscape in the early sixth century, appears to have been a buffer between Aphrodito and the outside world. Modern authors have seen him, a count of the sacred consistory, as the patron of autopract Aphrodito, perhaps the head of one of its major factions. Though one of Aphrodito’s largest landholders, he did not live there, but in Antaiopolis, and was possibly a praeses of the Thebaid. He is thus another example of Aphrodito’s economic ties to the outer world, and perhaps the closest thing to a Flavius Apion among Aphrodito’s protagonists in the first half of the century. But in sharp contrast to what I have argued about the Apions in previous chapters, the estates of Ammonios appear to have emerged from a specific rural nucleus, Peto, near the fields of Aphrodito.  

 

Zuckerman a,  takes this “sans doute.” For the management of Ammonios’ estates, see Hardy, passim. See also Maspero’s introduction to P.Cair.Masp. .; P.Ross.-Georg. .. note for a list of Ammonios texts; PLRE .; Fournet , , where Ammonios is on his list of possible (“Dubii”) praesides of the Thebaid; Fournet , , which discusses the headings to a number of unpublished Aphrodito letters, some of which include Ammonios. See also MacCoull forthcoming. Patron: Zuckerman b, . Factions: Ruffini forthcoming a. Hardy , , citing P.Cair.Masp. ., affirmed in detail at Fournet , .



Social networks in Byzantine Egypt

As I have already mentioned, Ammonios had business ties to Dioskoros and his family. A fragmentary document from  addressed to the foundation or corporation (dikaion) of the Apa Agenios monastery in the Apollonopolis Mikra nome describes the monastery as under the stewardship of Ammonios through the hands of Apollos himself. Apollos also received a letter from representatives of the same monastery (