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) o<²>k(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 (
  

 

See PSI ..– and Hardy , , with P.Cair.Masp. ...r. Ammonios made payments to the monastery that Hardy took to be donations. The Maspero text (.r.) records a payment e«v t¼ Àrouv Špa %g. ”nio, and on the reverse (.v.) records a payment e«v Anti() di‡ toÓ Ëpod(”ktou). This may be the hypodekt¯es Apollos therein at .r.. The monastery was in the Apollonopolis Mikra nome. On the significance of the “Apa” honorific, see above, n. . P.Cair.Masp. .; see BL . for the restoration. Fournet unpublished gives c. as a potential date for this text by virtue of its connections to PSI .. See below, p.  for the meaning of the phrase huper onomatos in Aphrodito’s cadastre. P.Cair.Masp. ..– (with BL . and .), discussed in the context of inconsistent weights and measures for money and grain at Hardy , . The identification of this count Ammonios with the count Ammonios for whom Apollos worked seems reasonable, as there are no other counts Ammonios in sight. (Fournet ,  does not include P.Cair.Masp. . in his discussion of Ammonios.) The measure had longevity: see P.Vat.Aphrod. . (). P.Cair.Masp. ..–, the formula partially restored: %]pollä. [v]
io[sk»r]ou , ˆxiw[qeªv], ›graya Ë. [p]•r aÉtoÓ. gr†m[m]ata | [mŸ e«d»tov]. P.Cair.Masp. . and ., a pair of fascinating documents to which Maspero himself gave more than his usual spartan level of commentary and analysis.

3 Aphrodito and the strong ties of village society



instance, P.Cair.Masp. . includes a line item for brick transport to the cistern in Peto: (Ëp•r) metafor(Žv) pl©nq(wn) to(Ó) l†kk(ou) P”to. The geographical distribution of the entries in Ammonios’ accounts is striking. By a rough count, Ammonios’ accounts record one entry for Psinabla, two for Antaiopolis, three for Antinoopolis, six for Aphrodito, and twenty-two for Peto. Here I suggest we have quite the opposite situation from that shown for the Apions in Chapter . There, analysis of the Oxyrhynchite topographical network showed that Apionic toponyms had nothing to distinguish them from non-Apionic toponyms. This suggests that Apionic fiscal jurisdiction and landownership was spread throughout the nome in something like a random pattern, not connected to or clustered around any particular ancestral estate or homeland. The same level of topographical network analysis is not possible for the holdings of Ammonios. Nonetheless, the skewed distribution of toponyms in our sample of his records is striking. Such a disproportionate level of references to Peto suggests that Ammonios had stronger ancestral ties to Peto than other locales. But where was Peto? Maspero was surely right when he noted that it “devait eˆtre situ´e dans le nome Antaiopolite, mais en dehors du territoire d’Aphrodit¯o, puisqu’on l’en distingue soigneusement.” The lease from Aphrodito adopting the Ammonian grain standard gives a further hint to its location. The farm in question was called OÉesätov, with the toponym Piap”to written in superscript above it. Ouesotos presumably derives from a variant on the original Egyptian name for Aphrodito, Ouazt. Thus, its association with the field of Peto suggests that the latter was very close to Aphrodito, perhaps even bordering on it. It was presumably Ammonios’ expansion from Peto into Aphrodito that brought him into contact with Apollos and family, this connection in turn bringing these  

  

 

P.Cair.Masp. ...r.. Based on P.Cair.Masp. . and .. Including P.Cair.Masp. ., a much shorter text than either of the first two, adds one or two for Aphrodito, the same number for Antaiopolis, and another for Peto. The papyri do not help. See Calderini (), Diz.geogr.: : three of the four texts attesting to Peto are Ammonios’ accounts. The fourth mentions the place only in passing. See his introduction to P.Cair.Masp. ., p. . Piap”to being Greek for the Coptic, PIA PETO or “Peto field.” The lease is virtually the same as that of P.Lond. ., with the same participants, but there, Piap”to appears alone. Bell took this to mean that it was a correction for ìOuesätov in the Maspero papyrus. For Piapeto, see Calderini, Diz.geogr., (): , treated as distinct from Peto. Calderini, Diz.geogr., ():  and (): . Piapeto and Ouesotos also appear in the eighth-century Aphrodito land registry, P.Lond. ..



Social networks in Byzantine Egypt

members of the Aphrodito elite in closer touch with Peto and the two poleis, Antinoopolis and Antaiopolis. The Aphrodito cadastre and a related Ammonios account show just how recently Ammonios had expanded into Aphrodito itself. The cadastre, a product of Ioannes the scholastikos and k¯ensit¯or c., was based on a land survey made by Mammas early in the previous decade. One unusual feature of the cadastre is eighteen entries in which one person’s holdings appear “in the name of” (½n»matov) someone else. Zuckerman has argued that these entries indicate land that changed hands between the survey of Mammas and the survey of Ioannes. Ammonios himself appears in eight of these eighteen entries, and a document from his own accounts lists all eight in the same sequence in which they appear in the cadastre. One entry on that list stands out: two arourae of Ammonios were entered as having been originally “in the name of Besarion the pr¯otok¯om¯et¯es.” As the cadastre’s editors note, this is probably Dioskoros’ uncle. To summarize: eight pieces of land at Aphrodito, including one originally belonging to the family of Dioskoros, came under Ammonios’ control in the decade following the survey of Mammas in the s. None of his land in the cadastre appears to predate that survey. The family’s appearance as business managers for Ammonios in the decades after he took over a piece of the family’s land is surely no coincidence. Throughout this chapter, we see repeated evidence of Aphrodito’s social ties to other settlements throughout its nome, and beyond it to settlements in other nomes as well. The case of Ammonios provides a detailed look at the structure of these ties. A powerful landholder owned property at and near Aphrodito, which brought him into contact with Aphrodito’s local elite. These elite in turn served him in a management capacity in more long-distance ventures. Put another way, geographically proximate social ties created more long-distance connections. (As an analogy from an earlier period, consider the role of Alypios, himself a prominent local landholder in the Fayum, and in turn the manager of the large estate of Appianus, himself a citizen of Alexandria. ) This is another example of     

 Zuckerman a, . Zuckerman a, –. SB . and P.Cair.Masp. . (= SB .), the latter clearly derived from the former; see Gascou and MacCoull , Appendix I, with Zuckerman a, –. SB .. and SB ... The one ambiguous case appears at line , which reads: %mmÛniov Qeodos©ou Ëp(¼) Prwmaän ìIsak©ou. This is a different format, but may mean much the same thing. An analogy suggested to me by Roger Bagnall; see Rathbone  for Alypios and Appianus.

3 Aphrodito and the strong ties of village society



social connections serving several functions, as well as another example of strong ties shaping Aphrodito’s social network.

the petition to the empress Aphrodito’s petition to the empress Theodora in defense of the village’s autopragia, typically seen as a display of village unity, highlights the importance of social ties in a village setting. Fifty-one signatories appear at the end of the petition, a collection Gagos and van Minnen describe as “a unique snapshot of the adult male e´lite of Aphrodito” c.. Maspero, the text’s original editor, thought that the list of signatories included “toute la kÛmh, clerg´e, fonctionnaires, marchands, ouvriers et propri´etaires ruraux.” The cast of characters includes the religious (Victor, a cousin of Dioskoros; Abraamios the priest; Phe¨us his lector), the wealthy (Flavius Dioskoros, Flavius Theoteknos), the tax-collectors (Apollos son of Iosephios, Hermauos), and a fascinating group of chief craftsmen (Konstantinos the bronze-smith, Phoibammon the carpenter, Ieremias the boatwright). Confronting this list raises any number of questions. Why were these men included? Who was left out and why? In what way were these people connected to each other, and distinct from those not on the list? What explains the order and grouping of the names? According to one recent analysis, the clergy received “pride of place because of their religious and social prestige,” followed by the wealthiest landowners, the notaries, and the craftsmen. So far so good. But why then follow with Senouthes, the village headman (pr¯otok¯om¯et¯es), and then end the list in “more or less random order”? 



  

P.Cair.Masp. .. Theodora is not named, the text reading only ¡män d”spoina and variants: see Index III s.v.
”spoina. For the text’s date, see the following note. See also MacCoull , – and Gagos and van Minnen , , discussing the signatories. Gagos and van Minnen ,  quote Maspero’s original date of / (p.  in his introduction to the text), which is based on Theodora’s death in , and the assumption that the beginning of the affair cannot have been too long before Dioskoros’ trip to Constantinople in . Since Dioskoros’ cousin Victor signs, MacCoull argues (, –) that the text must predate November ; see also Reiter ,  n. , reaching the same conclusion for different reasons. If the Apollos kt¯et¯or therein is Dioskoros’ father, then, as MacCoull (,  n. ) notes, the text must predate March , by which time Apollos is dead. Zuckerman a,  dates the text to “sans doute d´ebut .” P.Cair.Masp. . introduction, p. . For further discussion of the issues raised herein, see Ruffini forthcoming a. Gagos and van Minnen , .



Social networks in Byzantine Egypt

Gagos and van Minnen propose two solutions: () that Dioskoros simply went from one end of town to the other, collecting signatures as he saw fit, or () that the order of signatures represents an unspoken affirmation of the Aphrodito social hierarchy made manifest in a “signing ceremony” immediately after “a general meeting of the adult male inhabitants of Aphrodito.” While the second option has a certain dramatic appeal, parts of it do not ring true. As Bagnall has pointed out, “there is no evidence for such town meetings.” Would Senouthes, the village headman, Dioskoros’ own brother and a man who would himself make the trip to Constantinople in a few years’ time, really have stood off to the side, watching Dorantinoos the wine-merchant and Iosephios the fuller take the stage ahead of him? Would Flavius Theoteknos and Flavius Bessourous have let themselves be mixed with the Aurelii in “more or less random order”? The answer seems to be a combination of the two hypotheses, that Dioskoros went “door to door” to discrete gatherings of men from specific professional groups. With the exception of Palos, the illiterate priest and oikonomos of Apa Ermauos, all of the priests in the text sign in consecutive order, and must have been in the same place. Signatories include Victor from the holy catholic church, presumably Aphrodito’s main church; a different Victor from the new catholic church; Psais from the holy catholic south church; Petros from the holy catholic church of the Apostles, and so on. The ten churches listed at the start of the petition would not have been all in the same part of town, especially if not©nhv refers to Aphrodito’s southern plain, and not simply its location in town. The same is true of the tradesmen, almost all of whose signatures appear in a block as well. We do not know enough about the village geography of Aphrodito, but it stretches the imagination to think that bronze-smiths, wine-merchants, basket-weavers, and boatwrights all clustered in one place. Perhaps this cluster of signatures suggests an informal meeting of the village’s guild heads. Dioskoros would have attended that gathering before he solicited the signatures of the village’s economic elite. Again, I think it unlikely that the village headmen, collective taxpayers, and landowners would have stood aside while the boatmen signed first at the same meeting. Whether the elites    

 Bagnall  n. . Gagos and van Minnen , . For Senouthes, see above, n. . On the Flavii and Aurelii, see Keenan , , . Iezekiel son of Victor, the chief wine merchant, signs on p. , line . Makarios son of Iosephios, a guild chief whose profession is lost, is separated in the text from the other guild chiefs only by the name of the notary signing for him, and Promauos son of Apollos, whom the kaª in the clause may in fact link to Makarios as another guild man.

3 Aphrodito and the strong ties of village society



had a gathering of their own is less clear from the remaining signatures in the petition, but we know from other evidence that they considered themselves to be a single, discrete unit. Another text mentions a koinon of village headmen, collective taxpayers, and landowners (tän prwtokwmhtän kaª suntelestän kaª ktht»rwn). It is not clear “whether the order of the terms of the membership in the group . . . is an ascending (or descending) order of importance.” Whichever it is, that order did not matter to the signatories of the petition itself, which freely intermingles the three ranks. Gagos and van Minnen describe the petition to Theodora as a “show of power,” a “display of civic virtue . . . [designed] to express the will of the community.” For network analysts, the document also shows the importance of Aphrodito’s strong ties. First, seven of the  signatories are relatives of Dioskoros: one cousin also named Dioskoros, another cousin named Victor, Victor’s sons Kallinikos and Kyros, Phoibammon, Dioskoros’ brother Senouthes, and Dioskoros’ father-in-law, Ioannes son of Kornelios. Was this family so important that, counting Dioskoros himself, it accounted for nearly one out of every six of the local elite? Or do they play a disproportionate role here because Dioskoros was already closely connected to them? Second, seven of the signatories also sign for illiterates, in either this petition or other documents. Several other signatories are found elsewhere as witnesses. The point is fairly clear: several of the men in this collection of Aphrodito’s local elite served dual social functions at the same time. We may even speculate that they sign here on behalf of certain individuals because they have done so before. More generally, the petition shows that, thanks to social networks, village and community blend smoothly into family and friends. 

  



P.Cair.Masp. .. For the koinon of village headmen in the general context of Byzantine Egypt, see Gascou b, –, drawing a direct analogy (–) between the professional guilds and kephalai¯otai on one hand and the village and pr¯otok¯om¯etai on the other. See also the remarks on koina more generally in Chapter  above, nn.  and . Keenan a, . For my doubts about the formality of such hierarchies in Aphrodito, see n.  above and a further discussion in Chapter  below, n. . Gagos and van Minnen , . Accepting the identifications proposed by Gagos and van Minnen, , –. Jean-Luc Fournet points out that the signature of Dioskoros son of Megas is in the hand of his cousin, Dioskoros the poet, suggesting that the former was not available when the petition was drafted. Ieremias son of Victor, Pilatos, Apollos son of Iosephios, Triadelphos son of Konstantinos, Flavius Theoteknos (discussed in more detail in Chapter  below, pp. , ), Abraham son of Victor, and Ioannes son of Kornelios. This does not count those in the text who merely sign on behalf of illiterates, but are not signatories in their own right: Phe¨us the lector, Enoch son of Herakleios, Ioannes the monk. For references, see Gagos and van Minnen, , –.



Social networks in Byzantine Egypt the murder mystery

Another well-known event in Aphrodito history, the so-called Aphrodito murder mystery, also highlights the importance of face-to-face social ties in a village setting. I have throughout this chapter traced prosopographical connections with an eye to the multiplexity of the ties involved. The same principle may be applied to the two murders described in the court documents of P.Mich. . and .. We have produced throughout this chapter evidence supporting an intuitive principle, that social actors rely on their personal connections for a number of different reasons. It follows that social ties between any two people documented in Aphrodito’s papyrological record are probably strong or multiplex ties of some kind, even if supporting evidence for the nature of other connections between the two is lost. In a small village society, a victim and his killer would likely have known each other, and indeed may have worked together closely, sharing a variety of social ties. This line of reasoning may hold the key to the murder mystery, allowing us to better identify one of the victims, and get closer to identifying his killers. The texts in question are records of court proceedings, the subject of introductory remarks by P. J. Sijpesteijn and an article each by Leslie MacCoull and James Keenan. In all likelihood we will never understand the full truth behind the events in question. This much we can say more or less with certainty. Heraklios and a priest named Victor had died. Victor’s brother and Heraklios’ wife Maria appeared in court to accuse a soldier named Flavius Menas of a role in these deaths. It is alleged that Menas and another man, Flavius Sarapammon, ordered the death of Heraklios indirectly at the hands of the village kephalai¯otai. But Menas is said to have beaten Victor to death himself, slowly, over the course of nearly a whole day. A wide array of secondary players complicates our understanding of 

 



See P.Mich. .–, SB ., Gascou , MacCoull , Gagos , Keenan , and now Fournet , on a date in the first half of the sixth century. Interpreting this murder mystery, I typically follow Sijpesteijn and Keenan over the argument of MacCoull, on which see below, n. . On the dating clause in , see Bagnall and Worp a, . See also the interpretation of the murder mystery put forth at Sarris , , associating the saga with the village’s struggle for autopragia, and, contra Sarris, Ruffini forthcoming c. See MacCoull , , for a Christian sectarian interpretation of the murder, refuted by Keenan , –. P.Mich. ..: e«ržkasin [sc. o¬ kefalaiwtaª] Þv Sarap†mmwn ¾ –ndox(»tatov) kaª MhnŽv ›grayan ¡m±n aÉt¼n ˆnele±n. Kephalai¯otai: Sijpesteijn leaves the word untranslated in P.Mich. ., and Keenan implicitly takes it as “headmen” at , . But see LSJ s.v.: “capitularius, secretary and treasurer of a group of landowners or artisans, acting as recruiting officer, tax-collector, etc.” For “guild heads” as appropriate here, see below, p. . P.Mich. ..– as printed: MhnŽv ¾ kakosiÛmenov t¼n –m¼n. ˆ. delf. ¼. n B©ktora t¼n kaª presbÅteron –kbib†sav –f»neusen magganik¼n xÅlon b†llon e©v tŸn ˆrister‡n | o[..]ceira

3 Aphrodito and the strong ties of village society



the case. Victor’s brother Theodoros, evidently accused of some role in one or both of the crimes, speaks on his own behalf; Maria and a Flavius Apollos speak for him as well. Three Flavii, Psoios, Kollouthos, and Christophoros, all make appearances too fragmentary to understand. We know hardly any of the characters from elsewhere. The Heraklios kephalai¯ot¯es attested in an account of Ammonios is possibly the Heraklios herein, although the identification is advanced with hesitation. More on him in a moment. The Flavius Apollos who testified on behalf of Theodoros could equally be anybody, were it not for the fact that this fragmentary section shortly thereafter mentions a trip to Constantinople, possibly by the same man. To MacCoull, the identification of this Flavius Apollos with the father of Dioskoros the poet seemed obvious. It is natural to think of him first, as he had made a trip to Constantinople in another context. It may also be worth remembering that Apollos worked for Flavius Ammonios, the very man whose accounts attest to the existence of the Heraklios kephalai¯ot¯es just mentioned. But Keenan, “at the risk of being deemed . . . a ‘nattering nabob of negativism,’ . . . think[s] it best left (for now) a matter of doubt.” What precisely was at stake in these cases? To our minds, a crime had been committed. The money trail might provide a possible motive for

 



   

 

kaª p. lhg‡v poll†v –pagagÜn kat‡ toÓ stom†cou aÉtoÓ ˆ. p¼ p”m. p. thv ãrav m”[cr]i —sp”rav t¦. v. aÉt¦v ¡m”rav, which Sijpesteijn translates as, “the evil-doer Menas forced my brother Victor who also is a presbyter outside and murdered him having thrown a piece of wood of a machine to his left arm and having (placed) many blows on his stomach from the fifth hour till the evening of the day.” But see BL . and Gascou’s proposal (, , confirmed at Keenan , ) that for kakosiÛmenov read kaqosiwm”nov, thus ironically correcting Menas’ epithet from “evil-doer” to “devoted.” But now no longer to be found in the text are a woman named Talis and an unnamed Flavius speaking in defense of Theodoros: see the remarks in Keenan , –. Thus following Keenan’s interpretation , – that P.Mich. .. should be corrected to read “Fl(avius) Theodorus excu(sans) d(ixit)” rather than the scribal “Fl(avius) Theodorum excu(sans) d(ixit).” The correction is reasonable: the line makes no sense without it, and Flavius Theodoros appears in need of defense twice further, at P.Mich. .. and ... The one exception is Sarapammon, on whom see below, n. . See also MacCoull , , where she assumes the identification of the Apollos herein with the father of Dioskoros the poet. Her assertion that the murder mystery “is also the last recorded law case of Flavius Apollos” is baseless. MacCoull’s identification with our Apollos is an opinion she later reaffirmed in MacCoull ,  n. . This opinion leads her to date the texts to  and . P.Cair.Masp. .. P.Mich. ..: ] kaª ˆpelqe±n –pª Kwnstant©nou p»lin kaª proselqe±n t eÉsebest†t ¡män desp[»t]h„ . See MacCoull , , and n.  above. Gascou ,  saw in this reference to a trip to Constantinople indication that this murder case had escalated to the highest level, and did not take it as a damaged reference to one of the trips to the capital in which the Aphrodito elders defended their autopragia.  Keenan ,  n. . See above. But on the possibility that “crime” and “penalty” might be concepts with no weight in Roman Egypt, see Bagnall , –.



Social networks in Byzantine Egypt

the murders. Zacharias, whose role is also unclear, levied upon a group of unspecified villagers (perhaps the kephalai¯otai responsible for the death of Heraklios?) a fine for conspiracy to murder. He in turn gave the gold to an unnamed assistant (bo¯ethos), who passed it on to a second Victor and an unnamed “headman.” From there, the gold ended up in the hands of Sarapammon, a megaloprepestatos and endoxotatos whose specific rank is a mystery. An unspecified “he” promised that same gold to Letoios, who thought the whole affair fraudulent. If two men died, cui bono? Sarapammon, it would seem, who ended up with the gold. Is it too much to imagine that he – implicated with Menas in ordering the death of Heraklios – instigated the crimes with an eye towards the fines he could levy? It might be just this fact that lies behind Letoios agreeing that Sarapammon has acted in bad faith. First, a brief word about these kephalai¯otai of Aphrodito. Of the eleven attested in Girgis, we know just over half of them from one text, the remarkable petition to the empress we have already discussed. These are guildsmen, heads of the wine-makers, the shepherds, the copper-workers and others, which suggests that kephalai¯otai, which Sijpesteijn left untranslated in his edition of P.Mich. ., might best be rendered as “guild heads” here. Maria’s testimony in the murder mystery does not specify what sort of kephalai¯otai we are dealing with. She simply says that they and others “in its [sc. Aphrodito’s] service” came to her husband in jail: o¬ kefalaiw. [t]aª %frod©thv t¦v –m¦v kÛmhv met‡  



 

 

Take P.Mich. .. (crus©on kaª ˆph„ t¦sqai par‡ Zacar©ou) together with P.Mich. .. (kaª Ëp•r toÅtou o¬ ˆ. p¼. t[¦]v kÛmhv [ˆph„t]ž. [qhsan] t. re±v l©tra[v cruv]oÓ [sc. by Zacharias?]). P.Mich. ..–, where the speaker is not identified. Headman: Sijpesteijn’s rendering of meizoteros, although LSJ s.v. m”gav suggests that village “elder” might be more appropriate in this context. Keenan suggests to me that the term could mean “agent” in this context. But who was Sarapammon? Gascou , : “Au moment des faits Sarapammˆon pouvait eˆtre duc de Th´eba¨ıde, mais plus probablement praeses.” The name is extremely rare in the context of sixth-century Aphrodito; see MacCoull b,  and n. . Recent opinion agrees with Gascou: Sarapammon is now considered to be identical to the praeses Thebaidos in P.Oxy. . and PSI .; see BL .. Fournet ,  has proposed that the Sarapammon appearing in P.Aphrod.Lit . as the father of Ioannes, that poem’s central character, is identical to the Sarapammon here. MacCoull ,  finds this “hard . . . to believe.” P.Mich. ..–: a[Ét]¼n Ëposc”sq[ai cr]us©[on] Lhto©; LhtÛiov ¾ s. ofÛtatov p. [†nt]ev sun”q. ento aÉt ã[v ti] par†logon –po©hsa. In my opinion, MacCoull has overinterpreted the money trail. She reads (, ) the first line of P.Mich. . to indicate that the gold originally came from Sarapammon, ultimately returning to him as part of a money-laundering scheme. This is unnecessary, since the text states clearly that the gold had originally been paid to Zacharias: ... P.Mich.  adds only the headmen we have been discussing, without giving their names or occupations. None appear in P.Michael. and P.Vat.Aphrod. at all. For other possibilities, see above, n. .

3 Aphrodito and the strong ties of village society



Table  Kephalai¯otai in Girgis Girgis number

Name

Notes

  

Arsenios Heraklios Ezekiel



Jeremias



Constantine

 

Makarios Menas

  

Patermouthis Plinthos Phoibammon



Horouonchios

P.Cair.Masp. .. P.Cair.Masp. ...v. of the wine-makers: P.Cair.Masp. ... of the ship-builders: P.Cair.Masp. ... of the copper-workers: P.Cair.Masp. ... P.Cair.Masp. ... of the shepherds: P.Cair.Masp. . V (A) P.Cair.Masp. ..,  P.Cair.Masp. .. of the carpenters: P.Cair.Masp. ... of the weavers: P.Cair.Masp. ...

Šllwn ËpourgoÅ. ntwn aÉt
 sfaleis{a}m”noi t¼n –m¼n Šndra | ëH. [r†k]l. eion ˆp”qent[o] –n t
fulak
. It is curious that given the gravity of the situation, she does not make an attempt to identify any of them by name. Maybe their affiliation was evident to anyone familiar with the identity of their victim. It may well be that the murderers are hidden among the kephalai¯otai of Table . But why did Heraklios and Victor die? The exact manner of Heraklios’ death might provide another clue. Keenan found it “peculiar” that these unidentified village headmen of Aphrodito “spend some time drinking wine with the soon-to-be-victim” before beating, killing, and cremating him. Heraklios is not a particularly rare name: we know some two dozen from the Aphrodito area. One interesting possibility does, however, stand out from the list: Heraklios kephalai¯ot¯es, mentioned in passing above. Was this our Heraklios? In other words, were the village guild heads killing one of their own? P.Mich. .. says that the villagers killed Heraklios for being an informer or denouncer (Þv sukofantoÓnta): perhaps his denunciations 



P.Mich. ..–; but who else in Aphrodito is ËpourgoÅ. ntwn or works “in its service” (trans. Sijpesteijn)? Perhaps the pr¯otok¯om¯etai, whom we have seen elsewhere (above, pp. –) and hypodektai, but this is purely speculative.  Girgis –.  P.Cair.Masp. ...v. = Girgis . Keenan , .



Social networks in Byzantine Egypt

came to Sarapammon’s attention, giving him the opportunity to instigate the other kephalai¯otai. This is little more than speculation, but it might be possible that in this murder mystery, one village headman denounced several others for improprieties, and his death was the final retribution. Certainly, this would help explain the “peculiar” pre-murder drinking binge: it may be that the kephalai¯otai went to Heraklios in his jail cell hoping only to talk to him, and the situation escalated out of control. The extent to which multiplex social ties appear repeatedly in both peaceful and contentious social connections throughout this chapter makes this scenario seem quite likely. aphrodito, neighboring villages and beyond A metropolis and its villages are typically tied together in a structure much like the spokes of a wheel. Aphrodito, itself a former metropolis, enjoyed such ties to a number of satellite villages. Aphrodito shared close ties with Thmonachthe, a village on Aphrodito’s eastern borders. Indeed, one route through Aphrodito’s eastern plain was known simply as “the road heading to Thmonachthe.” Apollos farmed just over thirteen arourae in Thmonachthe. Some of the agricultural land on the border between the two villages was shuffled back and forth between jurisdictions: under the land survey of the scholastikos Ioannes, some Aphrodito land had been transferred to Thmonachthe (me. t. [aq]”sewv . . . | . . . e«v tŸn [pedi]†. da [kÛm]hv mon†cqh); the landowners, themselves residents of Aphrodito, later wanted it transferred back. In , when Phoibammon leased a farm from Flavius Samuel, originally from the village of Tanuaithis in the nome of Apollonopolis Mikra, the land in question was either near or had once been within the boundaries of Thmonachthe.    



See Ruffini . Much of the following discussion closely parallels Keenan b. Gagos and van Minnen , , where lines – refer to a lot called Mounlakon –k n»tou t¦v ¾doÓ | t¦v balloÅshv –pª tŸn kÛmhn Tmon”cqh. P.Cair.Masp. .. But, as with Apollos’ involvement in Phthla, Keenan (see below, n. ) casts doubt as to whether that involvement went so far as actual ownership. P.Cair.Masp. ..–. In its ed. princ., the text had been dated to –, but according to Crawford (P.Michael.  n. to lines , ) “this was a slip; the date, a third indiction, was probably .” For the transfer, see Bell , . For its role in dating the Aphrodito cadastre, see Zuckerman a, –. See P.Michael. .–, where Crawford takes –n t
ˆphliwtik
pedi†di t¦v aÉt¦v kÛmhv %frod©thv –n klžr | [ ] . . . o. Åshv kÛmhv Qmon†cqh to be read with “formerly belonged to” or some equivalent in the gap: see commentary at p.  and translation at . This assertion has the land transfer in P.Cair.Masp. . specifically in mind, but is reasonable enough, taken with the redating of : see previous note.

3 Aphrodito and the strong ties of village society



Aphrodito’s village elite also had close economic connections with nearby Phthla, which bordered Aphrodito. While Phthla appears to have been an economic satellite of Aphrodito, it was hardly a socially disconnected backwater. In , Phoibammon leased from Flavius Ioulianos, an absentee illustris, land in Phthla that once belonged to a now-deceased politeuomenos. Ioulianos appears to have been a pagarch from one of the nearby cities. Apollos, the father of the poet, both leased and owned land in Phthla. In his leases there, Apollos dealt with at least four landowners of Flavian rank. Apollos rented from the heirs of Flavius Kyros, who appears to have had landholdings both in the Antaiopolis area and elsewhere. Kyros’ agent assisted in issuing another receipt to Apollos on behalf of Flavius Ioannes, a politeuomenos whose own agent was illiterate. Apollos’ absentee landlords also included a former official from Panopolis, and a scriniarius Kollouthos probably from Antinoopolis. The latter case is another example of multiplex ties playing a role in land management. In /, Flavius Kollouthos issued to Apollos through an assistant another receipt on rent paid for land in Phthla, but did so through Apollos’ son Menas. While this is mere speculation, perhaps Apollos deputized his son Menas to handle business with the agents of Flavius Kollouthos in hopes of extending his son’s social network.



  



 



Phthla: potentially to be identified with modern El-Wa’adla, north of Tima,  kilometers from Aphrodito (Sauneron , –, followed by Fournet , ). The modern name may retain the original phonetic component in Phthla, which may be the article p plus thla, stemming from a word dl’, to indicate a cultivated field.  PSI . with Keenan b, –. These remarks follow Banaji , . PSI ..– ed. princ.: F. l. †(uiov) %. l”. x. and. r. o. v. –ndox». t. (atov) [«llo]Åstri[ov] kaª p[†gar-] |c. [o]v. %nt[a©ou ?]. For Ioulianos over Alexandros, see n.  above. P.Cair.Masp. ., a rent receipt issued to Apollos, and P.Cair.Masp. ., where Apollos collected a series of receipts for payments on land he leased in Phthla from absentee landlords. For that text see Keenan b, particularly , with a dating schema placing it in July . Keenan doubted that Apollos owned any Phthla land (see b,  n. , where he cites only P.Lond. .–, without P.Cair.Masp. .), but see now Fournet , . See P.Cair.Masp. ., where the heirs of Kyros are represented (lines –) di –moÓ | B©ktorov Ya©o(u) presb(ut”rou) kaª pronohto(Ó) tän %ntaiopolitikän pragm†twn; mention of “Antaiopolite property” surely implies the existence of non-Antaiopolite property. So the phrase is understood by Keenan b, . See also Keenan b, – for more connections between him and the family of Apollos than I discuss here.  P.Cair.Masp. ... Keenan b, . P.Cair.Masp. .., , and ; Banaji ,  thinks him “based presumably in Antinoopolis” and cites Keenan b, , who takes his role as scriniarius to imply an attachment “either to the ducal or to the praesidial officium in Antinoopolis.” P.Lond. ..–; for Menas, the half-brother of Dioskoros by another mother, see the stemma in van Minnen , . The date, from Fournet’s unpublished list of texts, relies on the presence of Apollos the hypodekt¯es, found in P.Cair.Masp. .– (–).



Social networks in Byzantine Egypt

Ties between Aphrodito and neighbors such as Phthla overlapped with Antaiopolite and Antinoopolite connections to Shenoute’s White Monastery in the Panopolite. Phoibammon, a relative of Apollos and Menas by marriage, leased land in Phthla belonging to Shenoute’s White Monastery, proving that Phthla was tied not only to Aphrodito, but also to urban centers a day or two to the south. The will of Flavius Theodoros, who served in the office of the dux of the Thebaid, specifies Antinoopolis as its provenance, three times mentions property in the Hermopolite nome, and includes a bequest to “the foundation of the same undefiled monastery of Saint Shenoute” (t¼ d©kaion to(Ó) aÉto(Ó) eÉ. agoÓv monas. thr©o(u) Špa SenoÅqou) in the Panopolite. The Aphrodito land cadastre also records an entry for the monastery of Apa Shenoute, presumably the White Monastery. Perhaps, like Flavius Theodoros, some other absentee landlord in the area willed land to the monastery on their death. Finally, one of Dioskoros’ poems includes a complaint about the pagarch’s decision to deposit money in Shenoute’s monastery. Residents of Aphrodito had further ties to places in other nomes. A group of Aphrodito landholders began their unfortunate tour of Upper Egypt’s prisons described earlier in this chapter during a trip to the market at Thinis, in the Hermopolite nome. The trip seems to have been a regular one for the group; their narrative describes events “about the time of our absence in Thinis for the local animal market, it being usual to go there each year” (perª t¼n kair¼n toÓ ¡mŽv ˆpi”nai –n žnei t
| –mfÅt ˆgorŽ„ tän zÛwn, € e«wq»tav kaq’ –niaut¼n –ke±se gen”sqai). Aphrodito villagers also had ties to Tanuaithis in the Lesser Apollonopolite nome. Phoibammon’s economic connections to Flavius Samuel from Tanuaithis have  

   



MacCoull , , citing P.Cair.Masp. .. P.Ross.-Georg. ., where line  reads: T¼ d©kaion toÓ ‰g©ou monasth]r©ou *pa SenoÅq[ou] ˆrcim[andr(©tou) À]ro[uv Tri]fi.[o]Ó [? toÓ Panopol(©tou) nomoÓ. The restoration is somewhat more secure than it looks, by virtue of comparison with P.Cair.Masp. ..–, where Flavius Theodoros leaves a piece of his inheritance to the same monastery. P.Cair.Masp. .., –, and . SB .. (Gascou and MacCoull ) with other references both potential and certain collected in the notes to that line. P.Aphrod.Lit. .c.R, originally published as P.Cair.Masp. .: Âtti crus¼n ˆpe©rona koinob©oisin ›q[hken | qespes©ou te SenoÅqou Sabor©ou melh. [ ].v. For Thinis (Thennis), see Drew-Bear , . The market: we know little about it, although Keenan has noted (a,  n. ) its place in a hypothetical future calendar of the rhythm of life at Aphrodito. For the trip, see above, p. , and P.Cair.Masp. .. This Thinis was also the site of an annual feast, believed to be the celebration of Christ’s resurrection; see Mandilaras . I thank James Keenan for redirecting my attention to the Hermopolite Thinis and providing me with the last reference. P.Cair.Masp. ...–.

3 Aphrodito and the strong ties of village society



already been well documented. Apollos leased land in Aphrodito to another Apollos, who describes himself as living in Tanuaithis: o«kän –n t¦„ not©nh„ pedi†di t¦v | kÛmhv Tanua©qewv. Late in the s, Dioskoros’ wife Sophia leased land at Aphrodito to two shepherds from Tanuaithis. Psinabla is another such village outside of the nome to which Aphrodito’s elite had economic ties. In the Panopolite nome, Psinabla was probably the site of a camel-troop recorded in the Notitia Dignitatum. A fragmentary receipt addressed to Charisios, Bottos, and Apollos, pr¯otok¯om¯etai of Aphrodito, was issued by someone whose name is lost (an absentee landlord?) from the village of Psinabla. In two consecutive years in the s, Dioskoros leased land in Psinabla to a shepherd named George. And Psinabla was in a different nome, not especially far away, but far enough to stand apart from Thmonachthe and Phthla. If Phoibammon managed land for absentee landholders from nearby villages, does Dioskoros’ connection to Psinabla here represent the opposite process in action, Aphrodito’s landed elites becoming absentee landholders slightly further afield? aphrodito and its pagarchs The region’s political elite, the Flavii we see throughout these nearby nomes, were in regular contact with Aphrodito society. If Aphrodito anywhere came in contact with an Oxyrhynchite superstructure of administrative ties, this contact happened through the regional pagarchs. While this   

 

 

 P.Cair.Masp. ..– with BL .. See above, pp. –. P.Strasb.inv.gr.  (/). See Bell’s summary of references to Psinabla in P.Lond. , in the introduction to , and Calderini Diz.geogr., : , for attestations from  through . See Gauthier , –, s.n. Psoumbeledj = PCNBLLE, which “signifie quelque chose comme ‘le passage des tessons’” (). Nestorius died in the area. P.Cair.Masp. .: Maspero does not give a complete edition of the text. P.Lond. .. Pastoralism was a documented part of the economy of Psinabla as far back as the fourth century. Three or four shepherds appear by name in P.Lond. ., comprising nearly a quarter of the total entries in the first column of this fourth-century account, the nature of which is unclear. The most useful survey of the elite as seen through Dioskoros’ poems is Fournet , –. For the pagarchy, see Gascou b; Liebeschuetz  and ; Mazza ; and Banaji , passim in Chapters –, with his prosopography of Aphrodito pagarchs, –. One pagarch per pagarchy is specified in Justinian’s thirteenth edict, but for simultaneous pagarchs, see Gascou b. Comparing P.Oxy. ., P.Cair.Masp. ., and P.Flor. ., he writes of “un cas de fractionnement de l’autorit´e pagarchique” (b, , emphasis in original), and adds, “Cela limitait les risques encourus” (). Attestations of two pagarchs holding the office collegially: see e.g. P.Lond. . and other examples in Fournet . SB . (= Hassanein , P.Cairo inv. S.R.  ()), one of those examples, has an apparent attestation of three pagarchs holding



Social networks in Byzantine Egypt

contact may have come on the heels of a change in Aphrodito’s fiscal status, it did little to disrupt the sort of multiplex face-to-face ties we have seen throughout this chapter. In some cases, Aphrodito’s ties to the pagarchs followed these very same social patterns. Evidence implicating the pagarchs in the fabric of Aphrodito’s landed society suggests that many of the local elite were not outsiders imposed on the village from above, but were in some ways little different from Dioskoros himself. For these pagarchs, I borrow the description of Liebeschuetz: A pagarch was often a local man, a big landowner, with honorary or retired rank in the imperial service. The office might become hereditary . . . The pagarch’s basic function was . . . collection of imperial taxes from villages and estates of the city territory that were not specifically exempted from his authority.

While cities remained in charge of their own financial affairs, the pagarchs collected from the surrounding countryside. The origins of this institution are murky. Bell suggested long ago that the creation of the pagarchy dates to the same era as Leo’s grant of autopragia to Aphrodito. More recent scholarship has dated the creation of the pagarchy to the reign of Anastasius (–). Pagarchs were at the lower levels of an administrative pyramid culminating with the region’s counts and dukes. These office-holders were often connected through family ties. According to Jean-Luc Fournet: “le duc Athanasios avait pour fr`ere le pagarque Phoibammˆon et . . . le duc Kallinikos e´tait le fils et le petit-fils de praesides ou de ducs de Th´eba¨ıde ainsi que le fr`ere du pagarque Kollouthos et de Dˆoroth´eos, magister dans son officium.” If these proposed family ties are correct, one can easily

 







the office together, Ioulianos, Kometes, and an interesting third, corrected to E. [É]qÅm. [iov] by Klaas Worp: see BL .. For a correction to the original date proposed in Hanafi , , see Fournet , . Liebeschuetz , . As Bell put it (, , cited by Gascou b, ), “The whole rural area now forms a single district, financially administered by an official called a pagarch.” See also Mazza , – for the evidence of the pagarchy’s administrative purview, with her comments on its relationship to the city at . Bell , , citing P.Cair.Masp. ..– for the origins of Aphrodito’s autopragia in the reign of Leo: aÉt»praktov oÔsa. kaª aÉtotelŸv tän eÉseb. ä. n. kaª dhmos©wn (Ëmän) | e[«]s. forän, mhd”pote k. lh. [rw]q. e. ±.[s]a. Ëp[¼ p]a. g. a. rcikŸn –xous©an ˆp¼ gon”wn | a. É. tä. n kaª prog»nwn, –c»ntw. [n] t¼ pron»mio. n. aÉt. [ä]n ˆp¼ qe©o. [u] tÅpou | toÓ t¦v qe±av lžxewv L”. [on]t. o. v. This last word is perhaps not as secure a reading as one would like. Liebeschuetz  argues that the creation of the pagarchy relates to Anastasius’ reforms of the imperial taxation system, and indicates a further stage in the weakening of civil government and its replacement with appointments from above. Mazza , – largely follows his lead. Fournet , , first explored in Fournet , –, and reiterated at Fournet , –, in his discussion of the “monopole familial.” Fournet publishes his editions of the encomia to the

3 Aphrodito and the strong ties of village society



imagine why Dioskoros and the other landowners of Aphrodito had difficulties with their pagarchs. The village’s most obvious recourse when they felt the pagarch had overstepped his bounds was to the duke, a man sharing strong ties, even biological relationships, with the subjects of complaint. We cannot tell from the surviving evidence how long a pagarch usually remained in office. The pagarchy was not part of a strict administrative hierarchy, but rather fell to people largely without regard to their rank. The strongest evidence for this proposition is the fact that pagarchs are attested at various status levels: lampr»tatov, megaloprep”statov, –ndox»tatov. Presumably, if their rank as pagarch were the determining factor in their status, each would have the same honorific. What power and social influence any pagarch might hold given this variance probably depended in some degree on their own social proximity to the local dukes; some pagarchs were closer than others to these centers of authority. Starting in the s, when Zuckerman has argued that Aphrodito lost its autopract status, we have extensive documentation of several consecutive pagarchs. A local notable and apo archont¯on named Flavius Ioulianos, whose career highlights the considerable interconnectivity among local elites, was one of the first. His attestations in office in the early s document various social ties to Aphrodito. One example is particularly revealing of the interconnectivity between the pagarchs and the family of Dioskoros.



 

pagarchs Phoibammon and Kollouthos sequentially, as P.Aphrod.Lit.  and . For Kallinikos and Dorotheos, see Fournet , and Fournet , –. In P.Cair.Masp. ..–, Menas, pagarch in , is said to have had eight predecessors: kaª ¢dh diagenam”nwn ½ktÜ pag†rcwn m”cri nÓn | t¦v %nta©o(u). Zuckerman a,  takes this to refer to the number of pagarchs since Aphrodito lost its autopragia. He must therefore argue for a relatively short tenure for each pagarch (not “strictement annuel”), since Aphrodito’s autopragia was clearly intact in the late s. Thus the upsurge of documentation on regional pagarchs would be due directly to Aphrodito’s change in fiscal status. For earlier approaches to the puzzle of pagarch tenure, see Liebeschuetz ,  n. ; Liebeschuetz ,  n. ; and Mazza , , who asks: “come interpretare quegli otto pagarchi, visto il fenomeno della detenzione collegiale della carica?” (For concurrent pagarchs, see above, n. .) Mazza , , inclines towards “una lunga durata” for a pagarch’s tenure. Her question regarding collegial pagarchies must still be answered if Zuckerman’s argument is to hold.  Gascou b,  and Mazza , –. Gascou b, . Ioulianos’ career has produced considerable confusion. See Mazza , – for a complete list of certain and possible attestations; Fournet , , following Gascou b,  n. ; Fournet , , following Mazza, but only treating the references from the s without discussing any other references. I follow his implicit assumption that there was more than one Ioulianos. See entries in the PLRE: Iulianus , the man we discuss here, at vol. , p. ; Iulianus  at ; and Iulianus  at . The editors of the PLRE rightly thought the latter two might be identical, but thought them not the same as the first of the three, “unless the same man held the office [of pagarch] again, later in his career” (). Mazza , – and Banaji ,  take all three men as one.



Social networks in Byzantine Egypt

Ioulianos leased land at Phthla to Phoibammon himself in . A papyrus from  was addressed to Ioulianos and Menas, both pagarchs, through two village headmen from Aphrodito on behalf of Pesalos and Menas, two Aphrodito liturgists, who acknowledged a -solidi debt to the two pagarchs. Pesalos and Menas were described as ˆpaithtaª tän leitourgän, which may indicate that they collected money from people liable to specific liturgical obligations. Probably the pagarchs had shifted responsibility for certain collections onto Pesalos and Menas, who acted as their subordinates. The intermediary headmen in this case are Apollos son of Ioannes, who had gone to Constantinople with Dioskoros two years before, and Dioskoros himself, whose role in the debt acknowledgement presumably explains why this text survives. If this text is an indication of the activities of the pagarchy in a post-autopract world, it is also evidence that pagarchs had to work through intermediaries, in this case Dioskoros, whose familiarity with Aphrodito’s social space could help them utilize the resources of Aurelius Pesalos and Aurelius Menas. Flavius Ioannes was one of the next pagarchs in the years following Ioulianos. The family of the pagarch Ioannes may have had property interests in Aphrodito. Flavius Ioannes, described as an endoxotatos pagarchos and kom¯es, appears in a collection of tax-receipts from the mid-s. This kom¯es Ioannes “could . . . have been the megaloprepestatos komes John who was asked to investigate the affairs of the monastery of Stratonikis” in P.Fouad. . He may also be the figure we see in a lease document from  involving his daughter, whose name is lost. She leased land to Aurelius Ioannes, son of Hermeias and Rachel, land located just north of the monastery of Abba Victor. It is not going too far to suggest that where some land is found, more may be hidden. The pagarch Ioannes may therefore have had rather more extensive ties to Aphrodito than his role in the tax-receipt series suggests.    

 

PSI ., based on an unpublished reading; see n.  above.  Thus the interpretation of Bell in his introduction to the text, –. P.Lond. .. For prosopographical notes on this Apollos, distinct from Apollos son of Isakios, see Palme , –. For the argument placing P.Flor . and P.Cair.Masp. . in the late s, see Fournet , with further treatment of both texts starting at p. , including tables detailing the payers, intermediaries, and pagarchs in each of the receipts therein. For Ioannes as pagarch, see Mazza , , Fl. ìIw†nnhv . See also Banaji , . Banaji ,  n. , citing Gascou . P.Cair.Masp. .. Banaji ,  n.  suggests restoring Patricia, on whom see below, pp. –. But see also n.  below.

3 Aphrodito and the strong ties of village society



Flavius Ioulianos, Ioannes’ predecessor, shared the pagarchy with a woman named Patricia. This Patricia was the recipient of an epithalamium from Dioskoros in /. It is typically assumed that she held the pagarchy by virtue of succession from her father, or more generally, by virtue of her tremendous wealth. And yet, the epithalamium names her father only in passing (a Kallinos unknown to us from any other source ), and gives only slightly more attention to her new husband, the equally obscure Paulos. Patricia appears in her capacity as pagarch c. in a text Bell described as “a contract concerning the collection of taxes.” The unnamed addressee therein, one of the village tax-collectors, received a guarantee from another unnamed party to secure the addressee from claims by the pagarchs Patricia and Ioulianos on the d¯emosia. Noteworthy for our purposes is the clause explaining that Patricia acted “through the most brilliant lord Menas her deputy” (dioik¯et¯es). In essence, Patricia deputed her pagarchic responsibilities to Flavius Menas. We do not know the exact relationship between Patricia and Menas, or why she chose him to handle her responsibilities. His position as dioik¯et¯es is suggestive: perhaps he already served as her administrator for lands from which she was absent, much as Apollos did for Flavius Ammonios, or even as Phoibammon did for such lesser figures as Flavius Samuel. If this argument is correct, Menas may have risen to a position of greater responsibility in the region precisely because of his multiplex ties. Further, the social lines between Dioskoros and his arch-enemy Menas start to blur rather dramatically. Both served the local aristocracy – Patricia in one case, Ammonios in the other – in an analogous fashion. One ends up as the  



  



P.Lond. ..–; and see Banaji , . For collegial pagarchies, see above, n. . Fournet , , the commentary to his text number , P.Lond.Lit.  C. Fournet’s identification is surely correct: the recipient of the epithalamium must have been important in the region, and no other Patricia is known to us who fits the description. Her wealth: Liebeschuetz , . Her father: Bell’s introduction to P.Lond. .. Banaji ,  suggested that Patricia is the unknown daughter of the kom¯es Ioannes, known from P.Cair.Masp. .. But this is already obsolete: see Fournet , –. See Bell’s introduction also for the argument that since Patricia deputed Menas to serve for her as pagarch in , Menas did not formally hold the office in his own right until much later. This chronology alleviates a problem that had troubled Maspero, namely that P.Cair.Masp. . implies that Menas was not appointed pagarch until . Fournet , , where he suggests in his commentary to line  that the name may be an example of the archaizing tendency common in the onomastics of the Byzantine upper class. P.Lond. .. P.Lond. ., introduction. If the reading in line  of [F]q. l. a. is correct, then the addressee was therefore a Phthla tax-collector, but presumably not the only one: see line , where t¼ epib†llon soi m”rov is mentioned. P.Lond. ..: d(i‡) toÓ lampro(t†tou) kur©ou MhnŽ aÉt¦v dioikhtoÓ.



Social networks in Byzantine Egypt

alleged oppressor of Aphrodito, the other its defender, perhaps by virtue of nothing more than the political and blood ties of the landholders they served. For this reason, Banaji was right when he described Menas as a breed apart from the Patricias of the world; he “was of distinctly lower rank, lamprotatos, conceivably the third and lowest level of the aristocracy.” We can follow his career from his association with Patricia in the early s to the late s. According to the standard chronology, he took office as pagarch in his own right just prior to the well-known crisis of /. Dioskoros and the villagers of Aphrodito addressed a petition against Menas that very year to a patrician named Athanasios, dux and augustalis of the Thebaid in the s. Menas, often cast as one of the arch-villains of Aphrodito at least partially because of the contents of this petition, appears there in coordination with a count named Serenos. If we believe Dioskoros’ version of the story, a group of small landholders from Aphrodito were “ambushed by the stewards of the endoxotatos illoustrios Serenos and we were thrown from there to prison.” Contrary to Banaji’s assertion that “the villagers of Aphrodito were careful not to criticize [Serenos] despite a certain association with the pagarch  









Banaji , . See PLRE .– for references. Our Menas  is not PLRE .’s Fl. Menas , a kom¯es at Aphrodito in : see P.Cair.Masp. ., dated to that year. If that text did relate to our Menas, it would push his career twenty years back in time, an unlikely proposition implicitly adopted at Banaji , . His assumption that the Menas who acted as pagarch for Patricia in the s is the same Menas acting for the unnamed daughter of kom¯es Ioannes in the s is exploded by Fournet’s discovery of the name of Patricia’s father: see above, n. . For a summary of the attestations, see Mazza , –, where his attestations come in two groups: () as pagarch with Ioulianos twice in the s, in one instance acting for Patricia, and () from c. on, without any references to him acting as an intermediary on behalf of another. P.Cair.Masp. .. See PLRE .; Maspero  (Flavius Marianos); MacCoull , ; Fournet , –; Laniado , ; and Laniado ,  for his polyonymy. Gascou b,  n. ; Mazza ,  with n.  citing Gascou; and Banaji ,  all adduce P.Antin. . for Athanasios and the Serenos appearing with him in . But this is mistaken. The text dates to the seventh century. The Athanasios and Serenos in P.Antin. . are simply two among several in an account, with no connection being drawn between them or the others. The coincidence of names in P.Antin. . is no more than that. Not the Serenos who was one of Ioannes’ colleagues as pagarch in the mid-s, for whom see P.Flor. . passim; Fournet , –; Mazza ,  under Fl. Ser¦nov; Banaji , –; and PLRE ., which gives two entries: Serenus , the kom¯es and scholastikos of P.Cair.Masp. ., and Fl. Serenus , the pagarch of P.Flor. , P.Cair.Masp. . III V, et al. Banaji ,  treats these two as identical, without providing any reference for the identification. According to the PLRE entry for Serenus , Serenos served as pagarch after  CE. This is Maspero’s original dating of P.Cair.Masp. ., which was corrected in Fournet . The PLRE entry should thus be changed: Serenos was pagarch nearly thirty years before the date it gives. P.Cair.Masp. ...: kaª kairothrhq”ntev t»te par‡ tän dioikhtän to(Ó) –ndox(ot†tou) «ll(oustr©ou) Seržno(u) kaª –blžqhmen e«v tŸn –ke±qen oÎsan e¬rktžn.

3 Aphrodito and the strong ties of village society



Menas,” Serenos received open and full indictment for his role as a co-conspirator with Menas against the people of Aphrodito. Serenos was the recipient of grammata (letters? orders?) from Menas which Dioskoros implied were directly connected to the ambush Serenos’ dioik¯etai laid for the villagers at the market-place of Thinis. Serenos took a number of animals from Aphrodito landholders, keeping the best for himself. His meizoteros Victor took clothes and equipment from the same landholders. The list goes on. Serenos was certainly well implicated, maybe more so than we realize. Dioskoros notes that the duke before Athanasios had given explicit orders that the Aphrodito landholders be released and freed from their payments (Ka© tinev –x ¡män tote pros¦lqan t e«rhm(”n) –ndox(ot†t) douk©. | ìEk”leusen ¡mŽv ˆzhm©wv ˆpoluq¦nai.), but that those orders remained unobeyed (oÉk’ ˆpelÅqhmen). We may imagine the reason why: the dukes, counts, and pagarchs we have seen so far all had close prosopographical ties. Menas and his ilk would have had various connections to Athanasios and the previous dukes, and were both geographically and socially well placed to influence their actions when it came to the conflict with the Aphrodito landholders. This would seem to be a significant challenge to Banaji’s model of an imperial aristocracy composed of disproportionately powerful social outsiders. If these aristocrats did not act to halt the aggressive acts of Menas and others against Aphrodito, we may suppose it was because they did not want to. Known principally through this petition of Dioskoros, Menas does not look too pleasant at first glance. He stands accused of arresting villagers, fining them, torturing them, and recruiting the shepherds of Phthla to encroach on village land. But we should see through this rhetoric and remember that Dioskoros, his accuser, had a normal working relationship with him in less trying times. We have already seen a document from  addressed to Menas and Ioulianos in their capacity as pagarchs from

 



 

Banaji , . ¾ e«rhm”(nov) lampr(»tatov) MhnŽv [g]r†mmata –car†xen t peribl”pt k»me(ti) kaª «llostr© meg[al]opre(pest†t) Seržn; P.Cair.Masp. ..., with the ambush shortly thereafter at line . P.Cair.Masp. ... refers to tän zän Àntwn \¡män/ –n ka[to]c¦„ Ëp¼ t¼n e«rhm”(non) «llo(Å)strion [sc. Serenos] ˆtr»fwn æn kaª ple±ston m”rov o«keiws†menon \–klekt¼n/ —aut Šneu tim¦[v]. kaª B©ktwr ¾ meiz»terov aÉto(Ó) t‡v –sq¦tav ¡män ˆfe©lato kaª p†nta t‡ skeÅh ¡män; P.Cair.Masp. ....  All from P.Cair.Masp. .: see discussion at n. . P.Cair.Masp. ...–.



Social networks in Byzantine Egypt

two liturgy-collectors (ˆpaithtaª tän leitourgän). Dioskoros and his colleague Apollos son of Ioannes acted as intermediaries to Menas and Ioulianos on behalf of the two apait¯etai, and there is no indication in the text that this was an extraordinary or otherwise strained arrangement; indeed, the liturgy in question, for the guards, was presumably a regular obligation. Here, the relationship between Menas and Dioskoros was merely that of a landholder acting as surety against an official debt owed to one of his peers. The pagarchy of Menas provides other examples of the tangled web of connections between Aphrodito, the pagarchs, and the local elite. A letter which Menas, presumably the pagarch, wrote to Dioskoros and Apollos in – mentioned the estate of Ioulianos the apo eparch¯on at Aphrodito. The estate, ¡ oÉs©a toÓ end[ox(ot†tou) ˆp¼] –p†rcwn ì Iouli†nou, was sunteloÓsa e«v t¼ kwmhtik»n, paying into the village account. The use of the term sunteloÓsa suggests that he played a role in Aphrodito analogous to the village’s suntelestai, if he was not one himself. As Menas pointed out, Ioulianos the apo eparch¯on was not alone in paying into the village accounts. “The city’s great landowners” (o¬ meg†loi ktžtorev t¦v p»lewv; sc. Antaiopolis) were doing so too. Menas was displeased, and as Maspero politely put it, “il invite Dioscore a` venir le trouver (`a Antaiopolis?) pour parler des affaires.” In this letter, Menas named Ioulianos’ estate as the prime example of a trend diverting tax payments from the pagarch’s purse into Aphrodito’s k¯om¯etikon or village account. Modern scholars have always agreed that in late antique Egypt, the line between public and private was a difficult one to locate with certainty. Menas may have been faced with considerable confusion caused by this ambiguity. It is an intriguing feature of Aphrodito’s hostile relationship with the regional pagarchs that several of the pagarchs appear to have been closely 

   

P.Lond. ., which serves circumstantially as the basis for the dating of . In  Menas was presumably acting for Patricia, who deputed her pagarchy to him: see above, pp. –. For apait¯etai see Palme ; Pesalous and Menas are entries  and . Following Bell’s reading of P.Lond. ..: tän aÉtän [l]eitourgän \¢toi ful. [ak]ä. n/ t¦v aÉt. ¦. v kÛ[m]hv %frod. ©.t. h. v. P.Cair.Masp. .. See Fournet unpublished for the date, and n.  above on Apollos. P.Cair.Masp. ..–; and see Banaji , . P.Cair.Masp. ., with W.Chr.  and Banaji , . But which Menas? Which Apollos? Maspero took this Menas to be the pagarch, and Wilcken followed him. Given his interest in the diversion of tax payments into Aphrodito’s village account, this identification makes sense. A date in the s means that Apollos could not be Dioskoros’ father. Presumably we are dealing with the Dioskoros and his brother-in-law Apollos, both known to be pr¯otok¯om¯et¯es, and both shown dealing with Menas the pagarch in P.Lond. .. Alternatively, the Apollos in question may be Apollos son of Ioannes, the pr¯otok¯om¯et¯es who appears with Dioskoros as an intermediary to Menas in P.Lond. ., for which, see above, p. .

3 Aphrodito and the strong ties of village society



tied to Aphrodito society. The suggestions of landowning and estate management we have seen here likely hint at deeper ties. This raises a number of interesting questions. Aphrodito’s leading men clearly felt that autopragia was in their own self-interests, a privilege worth defending. Yet the pagarchs, despite their close ties to Aphrodito society, apparently did not share this opinion, and therefore presumably lacked a sense of shared corporate identity with Aphrodito’s headmen. If pagarchs such as Ioulianos, Ioannes and Patricia served a dual role as local landowner and Antaiopolite office-holder, then perhaps here, as with so many of the crucial events this chapter documents, the explanation for specific actions lies in the multiplex roles of the actors. The tensions Aphrodito residents experienced in their contacts with the region’s larger social network stemmed directly from the dual functions served by those connecting them to that outer world. conclusion Even after these pagarchs arrived on the scene, Aphrodito still looked much the same, with agricultural and financial activities conducted through multiplex ties between social peers. An account book compiling texts from the s to  provides hints of some of these relationships within Dioskoros’ extended family. This is the account book that has revealed the identity of Dioskoros’ wife, Sophia. The compilation includes payments to the pagarchs listed “under the name” ((Ëp•r) ½n»(matov)) of both Dioskoros the poet and Kornelios son of Philantinoos, Sophia’s grandfather, through someone named Pteros. It likewise records payments to the pagarchs by both Kornelios son of Philantinoos and the heirs of Bekios son of Theodoros. These payments were made alternately by Sophia or a man named Victor. In the context of Dioskoros’ family, this Victor may well have been another relative, but we cannot be sure. A network in miniature emerges from the tangle. Pteros made payments for both Dioskoros and Kornelios; perhaps he had purchased from both men the land still registered in their names. Both men in turn were relatives by the poet’s marriage to Sophia, who in turn makes payments to the pagarchs for both of them. The appearance in the family archive of a fragmentary lease agreement () naming Sophia’s father Ioannes, the son of Kornelios, suggests    

P.Cair.Masp. ., with Fournet  and forthcoming for the dates involved. P.Cair.Masp. . IV.v., VIII.r., VIII.r.. P.Cair.Masp. . III.v., VII.v., IV.v., VII.v., VII.v., VIII.v., et al. Compare entries with ½n»matov in the Aphrodito cadastre, discussed above, p. .



Social networks in Byzantine Egypt

even closer intermingling of social and economic resources in this extended family than we yet know. This is not the only example of interconnections between extended family members in this period. A lease agreement drafted in  between Dioskoros and the other heirs of Apollos and an Aurelios Psempnouthios for land belonging to those heirs describes it as bordering a plot belonging to Apa Kyros, son of Besarion, an “ancestor” (toÓ | Ëmän progÛnou, sic) of Dioskoros. Besarion was presumably Dioskoros’ great-grandfather on his father’s maternal side. The two pieces of land must have been part of Besarion’s ancestral holdings, which at this point may date back to the beginning of the century, if not even further into the past. Their continuity within different lines of the same family is intriguing. Aphrodito’s surrounding fields must have been filled with plots rich in such interlocking ties going back generations into the past. In this chapter, I have argued for the applicability of the concept of multiplexity to our understanding of Aphrodito society in the sixth century. Multiplexity is merely the extent to which one’s social ties to any given individual comprise several different social roles. It is the phenomenon in which one’s father-in-law might also be one’s landlord, next-door neighbor, and security on a recent loan. For several aspects of this discussion, particularly our analyses of land acquisition in Aphrodito, this concept serves mainly to give theoretical support to what may have been intuitively clear. For other aspects of this discussion, such as the dispute settlement presented by Gagos and van Minnen, multiplexity can provide a ready explanation for the presence of certain social actors. Finally, in such cases as the murder mystery, an assumption of multiplexity can be a useful aid towards identifying otherwise unknown actors and understanding the significance of their actions. At the end of Chapter , I outlined reasons why the evidence presented there in support of a highly centralized Oxyrhynchite social structure would appear to exclude any hypothetical contradictory evidence. A similar argument holds in the case of Aphrodito. A skeptic might object that our conclusions about Aphrodito are skewed by our reliance on archival data found in the village itself. But how different would the view look from Antaiopolis? As I have suggested, Aphrodito’s pagarchs may be the most visible element of a nome-wide superstructure connecting the  

 P.Cair.Masp. ..–. P.Cair.Masp. .. An observation I owe to Peter van Minnen’s revised stemma presented at the Colloque Dioscore in Strasbourg, .

3 Aphrodito and the strong ties of village society



village to the nome center. We have seen evidence of large landholders on Aphrodito’s horizons: Flavius Ammonios, Flavius Ioulianos, Flavius Kyros, Flavius Ioannes. Although the evidence is slender, Zuckerman has proposed that the apo eparch¯on Flavius Ioulianos made payments accounting for the majority of Aphrodito’s cultivated land. Maybe these payments were indeed for land he owned, but it seems more likely that this merely represented a fiscal share of some sort. Either way, these landholders might be all we would ever see of Oxyrhynchite-style great estates when viewed from the vantage point of a village-level archive. But nothing we see these landholders doing in Aphrodito impacts the social ties the villagers form amongst themselves. They would still give loans, lease land, and make dispute settlements along the lines of previously existing face-to-face social connections. It is hard to see how the presence of a pagarch in their midst would change any of that. Indeed, it seems entirely likely on the face of existing evidence that some of the pagarchs shared similar ties to the village elite or needed to use those elite’s social connections to steer their way through Aphrodito’s social space. 

See n.  above.



See further discussion in the conclusion below, p. .

chapter 4

Quantifying Aphrodito’s social network

introduction Evidence for social life in Byzantine Aphrodito comes almost entirely from the archive or archives of Dioskoros and his family. Dioskoros and his relatives were certainly at the center of these archives. However, network analysis of this evidence indicates that he and his family were by several standards of measurement less socially important than we would have expected. These same analyses draw our attention to previously understudied members of Aphrodito’s lower classes, and invite us to ask why such individuals – hidden from view because of our attention to Dioskoros and his family – might appear so structurally important for Aphrodito’s social connectivity. The analyses that measure features of the network as a whole – particularly distance, centrality, and network cutpoints – indicate that Aphrodito’s social network had a very low degree of hierarchy and was relatively decentralized. These results lend support to the picture of Aphrodito’s social network drawn in the previous chapter. Coming from a century we picture as having rather rigid social hierarchies, this evidence opens the door to reconsidering the realities of village life throughout late antiquity. I have already argued for the importance of strong, multiplex ties in Aphrodito village society. But the evidence supporting those results was incomplete at best, based upon a handful of well-known people. Faced with the riches of Aphrodito’s papyrological corpus, modern scholars have naturally gravitated towards the most obvious characters and storylines: Dioskoros the lawyer-poet; Phoibammon, his entrepreneurial in-law; Aphrodito’s interactions with Upper Egypt’s political elite. Research into these subjects, important though they are, draws attention away from other avenues of investigation. Who beyond the circle of Dioskoros played a central role in the affairs of the village? Is it possible to get a better sense of the 

4 Quantifying Aphrodito’s social network



structural importance of Aphrodito’s non-elite inhabitants and learn what roles they played in that society’s social connectivity? testing the methodology Sixth-century Aphrodito is more susceptible to quantified social network analysis than any other place in late antique Egypt, and perhaps more susceptible to it than anywhere else in late antiquity. From the point of view of network analysis, the provenance of our Aphrodito documentation – all apparently from the archive or archives of a single family – provides us with a unique methodological opportunity. Although Byzantine Oxyrhynchos provides a wide range of material, we cannot employ it to form a single social network without considerable speculation. Even the texts with obvious interconnections, such as those related to the Apionic estates, relate to a dispersed geographic area. The people named in those texts have few demonstrable connections to each other. Aphrodito, by contrast, is known to us from a tight package of interlocking texts: hundreds of Aphrodito villagers can be connected to one another without any speculation at all. This rich prosopographical cluster permits effective use of UCINET, a leading computer program for network analysis. The text of the Girgis prosopography can be imported directly into UCINET. This data-set permits exploration of quantitative questions about Aphrodito society, including the identity of its most central members and the degree of hierarchy in its social structure. Iterative use of these techniques will also permit controlled experiments on the nature of our evidence, in which the removal of certain key texts and players will determine the extent to which our results are biased or distorted by features inherent in the Dioskoros archive itself. While Girgis’ prosopography suffers from typographical and interpretative errors, these do not ultimately affect our analyses. One feature  

 

Archive or archives: see Chapter  above, p.  with n. . This is based on analysis of the Girgis prosopography of Aphrodito using the UCINET computer program. For a more detailed description of this process, see below, pp. –. The largest component within the Girgis prosopography is  people. This excludes such tax registers as Zuckerman a (and its previously published component piece, P.Flor. .); on their elimination from the data-set due to their distorting effects, see below, pp. –. Borgatti et al., , discussed in the Introduction above, p. . See Ruffini , Appendix  for a discussion of the errors in Girgis’ prosopography, and their (minimal) impact on the results of my network analyses in this chapter. Girgis entries  and  are two exceptions that do affect analyses; inclusion of P.Cair.Masp. . in the former entry and P.Cair.Masp. . in the latter entry cannot be defended. Both errors result in the inclusion of the respective Victors towards the top of various centrality measures. They are omitted from discussion throughout this chapter.



Social networks in Byzantine Egypt

separating network analysis from the micro-history typical of much recent Aphrodito scholarship is a network’s ability to capture all of the evidence at once. This means that the errors in the Girgis prosopography have little statistical significance. A network data-set mapping the potential connections between nearly , people is essentially a grid with , rows and , columns, totalling ,, squares. Analyses to determine distance, centrality, density, and other measurements of a network this size will not easily be altered by the presence in the original prosopography of a handful of undiscovered errors. Girgis’ selective inclusion of attestations of Dioskoros and the most visible of his relatives does affect centrality measures for those individuals. Since this chapter is more concerned with people those measures highlight outside of the poet’s family, this flaw likewise is not a hindrance. The fact that much Aphrodito material has been published since the release of the Girgis prosopography is equally unimportant for this technique. A revised Aphrodito prosopography is no doubt needed. Many of the same people in the Girgis prosopography appear in the more recently published texts, and those later texts would make a useful body for analysis in their own right. One could cull material from the indices of each volume of Aphrodito material published since the Girgis prosopography and use it as a supplement to the Girgis data-set. But without thorough prosopographical analysis, such supplements would not provide the social connections necessary to add them to our existing network. Further, a serious objection stands in the way of adding this later material to my data, an objection centering around the nature of the papyrological finds themselves. In Chapter , I noted a growing scholarly consensus that more recently published Aphrodito material comes from a separate archive or archives, perhaps those of Dioskoros’ in-law, Phoibammon son of Triadelphos, and other members of the extended family. In keeping with that consensus, we can call these two sources the archive of Dioskoros and the archive of Phoibammon, if only for the sake of convenience. On the present evidence, this division goes back to antiquity, representing two groups of apparently distinct provenance, the first found in  and the second not brought to public eye until . The “Dioskoros archive” is generally the material excavated and published prior to the publication of the Girgis prosopography, and the “Phoibammon archive” is generally the material published since. Both groups appear to have documents unrelated to their  

See Chapter  above, n.  and Ruffini forthcoming b. See Chapter  above, p.  and n. .

4 Quantifying Aphrodito’s social network



namesakes, and include material both predating and postdating them. A bilingual Coptic and Greek group of texts from the seventh century has only a few documented prosopographical links to either of the two main archives, providing still another subgroup to material once thought to be unified. For these reasons, excluding the prosopographical material published since Girgis is highly appropriate. If some or all of the texts published since  do indeed constitute a later find of a separate archive, then adding all of their information to the material published in the Girgis prosopography would be the network analytical equivalent of combining apples and oranges. Studies of modern social networks have been based on e.g. high-school friendship surveys and anthropological field notes from African villages. For our purposes, one ought no more to combine the preGirgis material with the post-Girgis material, if they do in fact come from two distinct archives, than modern network analysts ought to combine two separate friendship surveys, or the field notes of two different anthropologists from the same village, to derive a single data-set. The results would be almost inherently distorted; by definition, the most centralized people in a combined network would be those people linking the two distinct archives. The combined network’s overall centrality would likely be much higher as well, focused on a small set of artificial centralizing connections. Under such conditions, the network analyses used in this chapter would produce almost entirely circular results. For these reasons, I proceed on the assumption that the Girgis data, inadequate as it is by modern prosopographical standards, provides the right material for the network analysis of a single archive. However, other objections are still at hand. To be confident that the figures for average distance and density, presented below, are meaningful, we must take into account our network’s potential for change over time. In the field’s infancy, network analysts were aware of a major flaw in their approach, the tendency to create a composite picture lacking in the ability to capture diachronic evolution. More recently, longitudinal studies have permitted progress in this area. Unfortunately, the chronological  

 

See Bagnall and Worp . Nor should the material of the Aphrodito cadastre and fiscal register be considered either; while the two texts are historical goldmines, they do not represent social connections between everyone appearing in them, and their sheer size would create a distortion in our analyses. I demonstrate this point in detail with tests on a previously published component of that fiscal register, P.Flor. ., which itself accounts for a considerable portion of the names in the Girgis prosopography; see below, pp. –. See Boissevain’s remarks in Boissevain and Mitchell , xi. See Wasserman and Faust , –.



Social networks in Byzantine Egypt

data from Aphrodito poses serious hurdles. Many texts remain completely undated. Others can only be assigned a possible set of dates based on their use of the fifteen-year indiction cycle. The number of securely dated texts is imbalanced, documenting considerably more of Aphrodito’s villagers in the first half of the sixth century than the second half. This imbalance complicates efforts to test change over time. However, rough comparisons of the village’s network characteristics in the four decades before and after  suggest that any change in the village’s fiscal status in that period had no impact on the shape of its social structures. The removal of individual villagers in our network to simulate their deaths also demonstrates the stability of the network’s characteristics over time. We will discuss in greater detail the concept of cutpoints, those members of the network whose absence would cause it to fragment into smaller pieces. As we will see (pp. – below), performing a “worstcase scenario” analysis in which we remove all of the cutpoints from the Aphrodito social network does produce higher figures for density and network centrality. The resulting figures serve as a boundary marker of sorts, indicating the highest levels of density and centrality Aphrodito’s network could achieve under a counterfactual extreme fragmentation. But more interesting still is the fact that even in that “worst-case scenario,” the largest intact component of Aphrodito’s surviving social network produces essentially the same central figures that the network produced prior to its fragmentation. In other words, using the removal or inclusion of cutpoints in our analysis to model the death of certain figures in our network, and thus to model chronological change, suggests that for any given period the general network characteristics we have analyzed may be too low, but the important figures who emerge from these analyses remain unchanged. The fact that the Girgis prosopography does not produce a completely unified network is another methodological concern. A survey of that prosopography’s sources shows that a noticeable percentage of the entries come from a small cluster of texts. My Girgis data-set includes  separate papyri,  of which are isolates, meaning they do not mention any people appearing in other papyri. The remaining  papyri form a series of components  

On Zuckerman’s argument for a change in Aphrodito’s fiscal status in this period, see Chapter  above, p. . A figure I derive by creating an affiliations network of texts, rather than of people, and using UCINET’s Network→Cohesion→Reachability function on the resulting data-set. Isolates will be those texts reachable only by themselves. Importing the results from that function into Excel and calculating the reachability sums for each row will reveal isolates, whose sums will only total . These isolates reminds us of a perpetual problem in papyrology, the identification of homonyms. Many

4 Quantifying Aphrodito’s social network



of varying size. The largest such component identified by UCINET is  papyri connected to each other by the prosopographical links between the people attested therein. Other components are much smaller, the next largest consisting of only seven prosopographically interconnected papyri. The largest component, of  texts, provides documentary evidence of social links between  Aphrodito villagers. Network analyses will be most useful when they focus on this component. Analyses taken across the full, disconnected network will produce averages of various network characteristics across isolated parts of that network, producing less meaningful results. Nonetheless, for comparative purposes, the following section, “Testing the data-set,” will show the results of all analyses for both the largest component and the network as a whole. In both cases, particularly large texts cause distortions in our analyses. A small number of the papyri in question are tax registers, essentially long lists of Aphrodito villagers and the amounts collected from them. These texts are anomalous, naming several hundred people, when the average Aphrodito text names only eleven. To use a modern example of network analysis, imagine what any analysis of the network of American corporate boards would look like if one or two corporate boards had ten times the typical number of directors. Attempts to measure that network’s most central players by degree and closeness would be distorted, giving disproportionate centrality to any director who happened to sit on one of those larger boards. The same happens with the Aphrodito prosopography. But these are meaningless results: if two people appear together in a tax register, it may indicate only the most indirect social connection, that of a shared tax-collector or fiscal official of some kind. The sequence of names in a register may also hint at connectivity – if for instance two acquaintances or neighbors pay and are listed in sequence – but those connections are hidden in the text, and will not have been uniform from one case to the next. One such tax list cited in Girgis, and since republished as part of a larger fiscal register, lists  villagers; a second such text lists  villagers and

   

of the dozens of men named e.g. Apollos may one day be identified with one another, thus linking some of these isolated texts. Using UCINET’s Networks→Regions→Components→Simple Graphs function on the affiliations network of texts described in the previous note. Using UCINET’s Data→Extract Main Component function on the affiliations network of my original Aphrodito network. See particularly P.Flor. ., cited as such throughout the Girgis prosopography, but now a part of the larger fiscal register published in Zuckerman a. An exception, in which individual entries record one person making a payment on behalf of another, is frequently attested in Zuckerman a.



Social networks in Byzantine Egypt

their payment amounts for an unspecified collection. At the time Girgis compiled his prosopography, none of the hundreds of individuals attested in those two texts had any known prosopographical links to the rest of the Aphrodito network. Prosopographical links between the list of  names in the fiscal register and the Aphrodito network’s central cluster of  names have been discovered since Girgis published his prosopography. But adding artificial links to our data-set, and keeping texts like the fiscal register in our analyses, can cause serious problems. These long texts, in essence no more than name-lists, act as distorting zones of  per cent density and  per cent centrality large enough to skew the averages provided for those measurements by standard texts of smaller size Therefore, the network analyses I outline below show considerable distortion when these larger texts are part of the network. To counter these effects, I will perform multiple iterations of each analysis, both with and without these texts, as described in the following section. testing the data-set One of this chapter’s chief contentions is that social network analysis can improve our understanding of Aphrodito’s social landscape by allowing us to see beyond Dioskoros and his family. But quantitative attempts to support this contention face two methodological objections, the potential distorting effects caused by Aphrodito’s largest texts on the one hand and the potential distorting effects caused by Dioskoros and his family on the other. To be sure that the results of UCINET analyses on the Girgis prosopography are meaningful and reasonably free from distortion, I will run each analysis on multiple versions of the data-set. This procedure will include () an analysis on the complete Girgis data-set, () an analysis removing the entry for Dioskoros, () an analysis removing the entries for both Dioskoros and his father Apollos, and () an analysis removing the entries for Dioskoros, Apollos, and Phoibammon. The same analysis will 



P.Flor. . and P.Cair.Masp. . respectively. The latter text has not received much scholarly attention, collecting only three BL entries, two on individual names at BL . and ., and a third, BL ., citing Fikhman’s Russian-language Egipet. I am reserving Zuckerman’s fiscal register (including P.Flor. .) and the Freer cadastre (SB .) for separate treatment at a later date: I have come to the conclusion that network analysis on these texts does not yield meaningful results, but that other statistical analyses not relevant to the discussion here will be useful. E.g. Apollonides in P.Flor. . and that in P.Michael , linking the former to Phoibammon, and thus to the entire cluster of thirty-seven texts. See PLRE ., with p.  below for results of an analysis including this proposed link.



4 Quantifying Aphrodito’s social network

Table  The Aphrodito Girgis prosopography I: correcting for Dioskoros and his family

Girgis prosopography Average distance Network density Network centrality: by degree by betweenness Individual degree centrality, top five Individual closeness centrality, top five Individual betweenness centrality, top five



Without Dioskoros

Without Dioskoros and Apollos

Without Dioskoros, Apollos, and Phoibammon

. . %

. . %

. . %

. . %

. % . % Five names appearing in the Aphrodito fiscal register P.Flor. 3.297 As above

. % . % Five names appearing in the Aphrodito fiscal register

. % . % Five names appearing in the Aphrodito fiscal register

. % . % Five names appearing in the Aphrodito fiscal register

As above

As above

As above

The complete network

As previous Enoch (), Apollos, Fl. Apollos, Abraam (), Dioskoros, Fl. Victor (), Ieremias (), Enoch (), Victor (), Pilatos (), Abraam (), Enoch (), Ieremias () Ieremias () Kallinikos ()

Parenthetical numbers are Girgis citations for each individual.

then be tried on () a data-set removing the two large village payment lists, which I will refer to as the “corrected” data-set throughout. Finally, I will compare against all of these () analyses run on the largest component of the prosopography, and () analyses run on that largest component with Dioskoros, Apollos, and Phoibammon removed. The results as presented in the following tables confirm that individuals such as Dioskoros have only limited impact on the characteristics of the Aphrodito network, but that large texts such as tax registers can cause considerable distortion. Tables  and  show the results of basic network analyses performed on several different versions of the Girgis prosopography. Table  shows the results of the same network analyses performed solely on the largest component of that prosopography. The results of



Social networks in Byzantine Egypt Table  The Aphrodito Girgis prosopography II: correcting for large distorting texts

Girgis prosopography Average distance Network density Network centrality: by degree by betweenness Individual degree centrality, top five

Individual closeness centrality, top five Individual betweenness centrality, top five



The complete network . . %

Without P. Flor. .

Without P. Flor. . and P. Cair. Masp. .

Without P.Flor. ., P.Cair. Masp. ., Apollos, Dioskoros, and Phoibammon

. . %

. . %

. . %

. % . % . % . % Five names Five names appearing in appearing in another the Aphrodito Aphrodito fiscal register payment list, P.Flor. 3.297 P.Cair.Masp. 3.67288 As above Enoch (), Fl. Victor (), Abraam (), Dioskoros, Pilatos () As previous Apollos, Dioskoros, Fl. Victor (), Enoch (), Ieremias ()

. % . % . % . % Enoch (), Enoch (), Abraam (), Abraam (), Apollos (), Apollos (), Ioannes (), Hermaouos Hermaouos (), Ioannes () () As previous

As previous

Enoch (), Abraam (), Pilatos (), Kallinikos (), Theoteknos () Enoch (), Abraam (), Ieremias (), Pilatos (), Kallinikos ()

Parenthetical numbers are Girgis citations for each individual.

these iterations indicate that inclusion of the largest texts and inclusion of disconnected smaller components of the network both cause considerable distortion, which will lead us to favor the results we obtain from working solely on the largest component. However, analyses on both that component and the network as a whole show that Dioskoros and family do not greatly distort the evidence, their removal only altering one of five calculated network characteristics. 

Closeness centrality, one of these five, cannot be computed over an incompletely connected network.

4 Quantifying Aphrodito’s social network



Table  The Aphrodito prosopography’s largest component: 364 names linked by 37 texts Girgis prosopography

The largest component with Dioskoros and family

The largest component, removing Dioskoros and Apollos

Average distance Network density Network centrality: by degree by closeness by betweenness Individual degree centrality, top five

. . %

. . %

. % . % . % Enoch (), Abraam (), Ioannes (), Hermaouos (), Kallinikos () Enoch (), Fl. Victor (), Abraam (), Dioskoros, Pilatos () Apollos, Dioskoros, Fl. Victor (), Enoch (), Ieremias ()

. % [∗ ] . % Enoch (), Abraam (), Hermaouos (), Ioannes (), Kallinikos () Enoch (), Abraam (), Pilatos (), Kallinikos (), Theoteknos () Enoch (), Abraam (), Ieremias (), Pilatos (), Kallinikos ()

Individual closeness centrality, top five Individual betweenness centrality, top five



Parenthetical numbers are Girgis citations for each individual. Removing Dioskoros and Apollos fragments this component, thus making closeness centrality incalculable on this network.



The average distance, network density and degree centrality analyses for both the full Girgis data-set and the largest component suggest that Dioskoros and his family do not greatly affect these features of the Aphrodito social network. (The same is not true of individual centrality measures, discussed below.) A number of explanations for this somewhat counter-intuitive result are close at hand. In Chapter , we documented an apparent tendency towards strong ties, in which Dioskoros and his relatives are likely to appear with the same people again and again. This means that those people, their peers, will be nearly as connected in our network as Dioskoros and his family. Further, removing Dioskoros, Apollos, or Phoibammon from our analyses does not eliminate from our data-set the social events of which they were a part. A lease agreement between a landowner and a tenant farmer, with Dioskoros witnessing, and a scribe writing the agreement, would still appear in our analyses as a tie between the landowner and the tenant, and between both men and the scribe. Thus, removing Dioskoros need not make the scribe less connected, if the landowner himself matches Dioskoros in connectivity. Nor, imagining the process in reverse, is adding



Social networks in Byzantine Egypt

Dioskoros back into the data-set likely to dramatically improve the connectivity of others. Adding Dioskoros as a social link does not necessarily connect an Aphrodito villager to new people, but merely gives him another route to villagers to whom he is likely already connected. Although Dioskoros and his family do not introduce significant distortion into our analyses, network density and network centrality measurements (see Introduction, pp. –, –) produce meaningless results when we retain the largest distorting texts. Thus the Aphrodito fiscal register mentioned above, naming  people, includes over a quarter of the entire Girgis prosopography. The degree and closeness tests on the Girgis data-set name as the most central members of that network people who appear in that fiscal register, even if they are otherwise unimportant to us. But running further analyses on the Girgis prosopography without excerpts from the fiscal register dramatically changes the results. The five most central people by degree and closeness in this analysis are entirely different. The top five by degree all instead appear in another Aphrodito payment list, and the top five by closeness all appear in that village’s petition to Theodora. These two texts are quite different in content. The petition to the Empress Theodora, which we discussed in Chapter , has prosopographical links with other texts in the largest component, and in any case documents genuine social connections, while mere payment lists do not. The Aphrodito village account, however, does not document any more social connections than did the excerpts of the fiscal register which we have just removed. Removing this village account as well brings our attention to a new list of Aphrodito’s top five most connected people, including two shepherds named Apollos (Girgis ) and Hermaouos (Girgis ). Note that analyses of closeness and betweenness centrality after removing the largest payment list, and then after removing both it and the second largest such list, provide the same top-ranked figures. This suggests that if we remove enough distorting material, our analyses converge on accurate results. In all of these analyses, Dioskoros and his father Apollos are missing from the top five as measured by degree centrality; Dioskoros is in the top five as measured by closeness centrality, but not at the very top. These measures do not particularly privilege Dioskoros or his family members,  



P.Flor. ., now part of Zuckerman a. The table does not give names here, because every person in P.Flor. . ties for first place, and every person in P.Cair.Masp. . ties for second. A “Top Five” list of actors for these centrality tests cannot truly be generated. P.Cair.Masp. . and . respectively.

4 Quantifying Aphrodito’s social network



no doubt owing to the Girgis prosopography’s selective inclusion of their attestations. But betweenness centrality, a measure commonly held to be a good gauge of an actor’s importance in the area of information flow, privileges that family to a remarkable degree. Running a betweenness centrality test on the complete Girgis prosopography places Apollos, his son Dioskoros, and his cousin Flavius Victor in order in the top three slots. If we run the same analysis with the two largest village payment lists removed (as explained above) we meet the same top players. The remaining seven in this measurement’s top ten include a number of names appearing on other centrality tests as well: Enoch the kt¯et¯or (Girgis ), Ieremias () the priest, Aurelius Abraam () the scribe and kt¯et¯or, Mousaios () the deacon, Victor () the shepherd and fieldguard, Kallinikos () the kt¯et¯or, and another Kallinikos (), also a kt¯et¯or. Counting the first Victor as a priest and Dioskoros as a kt¯et¯or, religious men make up a third of this group, and kt¯etores fully half. Unlike the degree and closeness measures, these betweenness centrality measures remain undistorted by both important people and large texts. While Dioskoros and his family rank highly by betweenness centrality, their presence at the top does not distort the relative rankings of other actors in the network. This much we can see by removing each member of the family one by one: each actor below them in the list merely rises to take their place in the rankings. Enoch, Ieremias and Abraam, previously ranking fourth through sixth, move up to take the top three slots by betweenness centrality.  

See the Introduction above, p. : betweenness centrality measures the extent to which an actor sits atop the highest number of routes between other actors in the network. For the title of kt¯et¯or, see recent remarks in Laniado , : “Dans les sources des Ve–VIIe si`ecles, les termes possessor et ktžtwr renvoient non seulement aux propri´etaires stricto sensu, mais aussi aux notables municipaux,” and , where he remarks, in regard to appearances of the term and geoÓcov/geoucän from Antaiopolis, with a few exceptions “ces termes sont utilis´es au sens priv´e.” Keenan a,  saw kt¯etores at the top of a local cursus honorum which progressed from pr¯otok¯om¯et¯es to suntelest¯es to kt¯et¯or. This refined his position from the previous year, in which he was not sure whether that order went from least to greatest or greatest to least. See also his comments at b, , esp. n. , in which he argues that “most traditional societies have definite rules of thumb” delineating such designations. I doubt the existence of a formal hierarchy involving these terms. Given the well-documented tendency for Aphrodito’s landholders to represent themselves to the outer world (see e.g. Keenan a, ) with terms of false modesty, can we be sure their internal terminology was more precise? See Bagnall , – for examples of remarkably misleading terminology: a petitioner describing his own business partner as tän dunatot”rwn ˆnqrÛpwn, a former magistrate describing himself as “modest” and his fishermen opponents “powerful,” and so on. Gagos and Van Minnen’s  Settling a Dispute text complicates Keenan’s picture. Apollos appears as a pr¯otok¯om¯et¯es, but his client and nephew by marriage, the well-known Phoibammon son of Triadelphos, appears as a suntelest¯es. Surely he had not risen above Apollos at such an early date in his career? Nor could the hierarchy have been in the reverse order, for Dioskoros is attested as a kt¯et¯or after he is attested as a pr¯otok¯om¯et¯es and suntelest¯es: see Chapter  above, n. .



Social networks in Byzantine Egypt

Nor do large texts play much of a distorting role in this betweenness measurement. When Dioskoros, Apollos, and Phoibammon are removed, the top five actors are the same – Enoch, Abraam, Ieremias, Pilatos and Kallinikos – whether or not the large fiscal registers and accounts are included. As we have already noted, much Aphrodito material has been published since Girgis compiled his prosopography. To give an example critical to our previous analyses, the excerpts from the Aphrodito fiscal register have links to the rest of the network in material published since Girgis, and should not simply be cast aside in hopes of finding less distorted results. Does inclusion of links revealed since the original prosopography significantly alter our results? It is possible to find out by performing a crude corrective surgery on our data-set, adding an extra entry linking the excerpts from the Aphrodito fiscal register to the texts making the network’s largest component. For example, Apollonides appears in both the Aphrodito fiscal register and a text involving Phoibammon son of Triadelphos, the relative of Dioskoros. Since Phoibammon appears in our Girgis data-set, we can simply edit our data-set to link the two names, and see if any changes result. Analyses on this artificially altered data-set provide an average distance of . and a network density of ., hardly much different from our unaltered complete set. Apollonides, the subject of our artificial linkage, naturally shoots to the top of the centrality measurements, but the others with him at the top of the list either appeared at the top of these lists in the unaltered version, or are other actors in Aphrodito’s large accounts. Thus the distortion caused by the sheer bulk of names in these texts does not go away simply by linking that text to others; many dozens or hundreds of such additional links would be necessary to offset the disproportional effects of the two large payment lists we have removed. But if the size of a text can prove so important, as the material from the fiscal register shows, does the nature of the text prove equally so? Degree centrality is merely a measure of who has the most first-degree or direct connections. Hence the distortion caused by texts like the fiscal register: anyone appearing in the text listing the most names will appear at the top of a degree centrality list. But this raises institutional or corporate issues as well: note the shepherds, Apollos and Hermaouos, who ranked so highly by degree centrality after we removed the two large payment lists. A 

P.Flor. . and P.Michael . Zuckerman a, “Index du texte” s.n., makes this identification.

4 Quantifying Aphrodito’s social network



document involving a given organization – in Aphrodito, the shepherd and fieldguard guilds come quickly to mind – will, if it names many members of that organization, privilege those members in a degree centrality test. Thus, if we use degree closeness as a measure of connectivity, we must remember that texts documenting corporate membership might skew the results. For this reason, the closer look we give below to our most central players will focus instead on the measurements of betweenness and closeness centrality. To summarize: () adding social connections published since the Girgis prosopography, even if that material links two large components, will not significantly alter the results, but simply fills in minor gaps; () appearing in a text which names a disproportionately high number of names can dramatically increase one’s apparent importance by measures of degree and closeness centrality; () deleting material from such overpopulated texts can correct this distortion (and ultimately removes evidence for what are little more than pseudo-connections); () measuring degree centrality can privilege even men of socially modest background, such as the shepherds Apollos and Hermaouos, perhaps because of their corporate identity; () the family to which most of the Aphrodito material belonged – Dioskoros, Apollos, Phoibammon, etc. – does not cause distortion to the evidence comparable to that caused by large texts. While members of that family may be highly ranked by and affect measures of betweenness centrality, removing them from the analysis does not otherwise alter the rankings of the other actors. With these conclusions in mind, further analyses in this chapter are solely on the Girgis data-set “corrected” by removing the two large payment lists, unless otherwise stated. aphrodito’s most central social actors Social network analysis is most interesting when it highlights people we had otherwise deemed unimportant, or when it demonstrates certain positions or roles to be structurally more important than we would have thought otherwise. The centrality tests we outlined above produced several unfamiliar 

In terms of social network theory, a large text with many names is the equivalent of inserting into a network a clique, or subset of  per cent density, membership in which will automatically privilege any actor in that subset. Nor are all such cliques artificially distortive: a text naming (for instance) all the shepherds in the shepherd guild agreeing to serve as fieldguards probably does represent a clique in real-world terms, even if it appears to distort the evidence.



Social networks in Byzantine Egypt Table  Closeness and betweenness centrality I: based on the corrected Girgis prosopography

Excerpted from Table  above

Most-connected members of the network

By closeness centrality

Enoch (), Fl. Victor (), Abraam (), Dioskoros, Pilatos () Apollos, Dioskoros, Fl. Victor (), Enoch (), Ieremias ()

By betweenness centrality



Parenthetical numbers are Girgis citations for each individual.

Table  Closeness and betweenness centrality II: based on its largest component Excerpted from Table  above, removing Dioskoros and Apollos By closeness centrality By betweenness centrality



Most-connected members of the network Enoch (), Abraam (), Pilatos (), Kallinikos (), Theoteknos () Enoch (), Abraam (), Ieremias (), Pilatos (), Kallinikos ()

Parenthetical numbers are Girgis citations for each individual.

names. It remains to look at each individual actor to see if patterns emerge. Tables  and  list the top five actors by betweenness and closeness centrality in both the corrected Girgis data-set and its largest component with Dioskoros and Apollos removed. The seven top-ranking individuals drawn from the network’s largest component (Table ) highlight the importance of landowners, scribes, and shepherds. (1) Enoch, son of Hermaouos and Thekla, a kt¯et¯or (landowner) in Aphrodito, is attested in two texts from  and c., and ranked as the most central man in our data-set by the measures of closeness 

I do not include the top actors by degree centrality for the reasons discussed above, p. . That measure does not provide meaningful results in the network I have constructed. In a single-mode network, one linking people directly to other people, created by running a UCINET affiliations transformation on a two-mode network originally linking people indirectly via the texts attesting them, a person’s degree will depend on the number of people appearing in those texts with him. For this reason, degree centrality artificially privileges people appearing in larger textual “events” such as the petition to the empress Theodora.

4 Quantifying Aphrodito’s social network



and betweenness. Nothing is known about either parent. In the first text, Enoch guaranteed to Apollos the riparius that Isakos, a shepherd and fieldguard, would “keep watch over matters of peace” (prosedreÅein kaª Ëpourge±n . . . tŸn parafulakŸn tän e«rhnikän | pragm†twn) for the year to come. The second text is the Aphrodito petition to the Empress Theodora complaining about the aggression of the Antaiopolite pagarch. As we saw in Chapter , the signatories on this document are a veritable “who’s who” of Aphrodito: by Maspero’s count, eleven priests, twenty-two kt¯etores, the “chefs de plusieurs corporations industrielles,” and so on. Enoch’s appearance in this text tells us little about him as an individual: each kt¯et¯or has the roughly same formula following his name. But the text as a whole does impress us with the corporate identity of Aphrodito’s landholding elite. Presumably the corporate nature of life in Aphrodito accounts at least partially for Enoch’s centrality here. His role as a kt¯et¯or put him in close social proximity to Aphrodito’s elite; yet, at the same time, he served as a guarantor for a shepherd, who may be presumed to have indirect contact with all of Aphrodito’s other shepherds as well. (2) Abraam son of Victor we know from a career spanning some three and a half decades; he ranked as the second most central man in our dataset by the measures of closeness and betweenness. He appears in  in a fascinating agreement between the community (koinon) of Aphrodito’s village pr¯otok¯om¯etai, suntelestai, and kt¯etores on one hand, and Aphrodito’s

  

   

P.Cair.Masp. . and . respectively. Neither appeared in the pages of Girgis or since in P.Michael., P.Mich. , or P.Vat.Aphrod. P.Cair.Masp. ...–; for prosudreÅein read prosedreÅein. Enoch’s is only one of a dozen guarantees made to Apollos in his capacity as village police officer (riparius) which Maspero published together in this text. For Apollos the riparius, see Palme  with full citations at –. For more on these guarantees, see Keenan a, . The translation is his. See also below, pp. , –. Because P.Cair.Masp. . is in fact a collection of texts with a disproportionately high number of names, it might appear to introduce the same sort of distortion as P.Flor. . and P.Cair.Masp. ., the two texts I have already removed from consideration. But since all of the texts in  are addressed to the same riparius, and since the figures therein share direct social connections both with each other and with others from the Aphrodito documents, I have retained them throughout these analyses. See above, Chapter , pp. –. In Enoch’s case, at P.Cair.Masp. ...–: ktžtw. r k. Û. m. [(hv)] %f. rod©[th]v. [para]kalä. n –p[id”dwka] |Ëm±.n. ta. Å. thn tŸn didaskal©an pr[¼]v e­dh[sin] t¦v ¡män d[espo©nhv. Shared membership in the koinon of shepherds and fieldguards would make this all but certain. For that koinon, see below, pp. –. To Girgis  add P.Mich. ., P.Ross.Georg. ., and P.Cair.Masp. ..



Social networks in Byzantine Egypt

shepherds on the other. There, Abraam signed the agreement for a group of shepherds who were unable to do so themselves. He then appears thirtythree years later in a land agreement between Psais, a deacon, and Dioskoros himself. He served as a witness to the agreement, which itself was drawn up by Pilatos, whom we will meet below. Abraam appears also in the petition to the empress which we have already discussed, there in his capacity as a kt¯et¯or. The formula following his name is nearly the same as that given for Enoch. At some point, Abraam also served as a bo¯ethos. He is attested in that capacity in a series of rent receipts issued to Apollos, father of the poet Dioskoros. In one of those receipts, issued by Flavius Kollouthos through another Apollos, Abraam is described as an assistant (bohq(¼v)) who wrote on behalf of the intermediary, Apollos (not Kollouthos himself  ). He then appears as an ex-bo¯ethos, for a third time writing for someone who could not, in this case the woman named Talos who was party to the dispute we discussed in Chapter . All told, this is an interesting biography: a kt¯et¯or, a bo¯ethos, and an intermediate in a transaction for a Flaviate absentee landlord. Yet, Abraam’s high centrality is not simply due to factors of social status, but also to his ability to write: for the wife of a small landowner, for the receipts of a large landowner, and even for shepherds. (3) Pilatos the grammateus places third on our list of Aphrodito’s most central by closeness and fourth by betweenness. As with others we have seen, he appears in the petition from Aphrodito to the empress, described there as a tabelli¯on (for the Latin tabellio or notary). In  he drafted a guarantee by four Aphrodito villagers through Dioskoros the poet and Apollos son of Ioannes to the pagarch Ioulianos that they would provide security against the departure from Aphrodito of an Enoch son of Ioannes. Pilatos also appears as the drafting nomikos of a receipt for rent issued in  or

       

P.Cair.Masp. .. See the useful discussion at Keenan a, –. For koina, see Chapter  above, nn.  and . P.Cair.Masp. .. P.Cair.Masp. ..., with the addition of the t¦. v. e«[rž]m. enhv epithet to Aphrodito: see n.  above for Enoch’s signature. P.Cair.Masp. .,  CE; BL .. This text is missing from his Girgis entry. See Maspero’s note to line , and his description of Kollouthos’ signature as “une e´criture ronde tr`es e´l´egante.” P.Mich. ..–, on the date of which, see Chapter  above, note . P.Cair.Masp. ..., . For Pilatos’ subscriptions, see Diethart and Worp , –. One would like a more secure reading: see P.Cair.Masp. .., di’ –moÓ [P]il. †. t. [ou nomik(oÓ)] –gr†f(h). For this Apollos versus Apollos son of Isakios, see Palme , –.

4 Quantifying Aphrodito’s social network



 from a landowner named Paulos to the much better-known Phoibammon son of Triadelphos (on whom see Chapter  above). He twice appears as the signatory in receipts issued to Kornelios son of Philantinoos, and is recorded as a nomikos eight other times in a modern catalogue of scribes. Pilatos’ high rank by the measures of closeness and betweenness presents a more explicit example of the pattern we saw with Abraam son of Victor, whose writing skills, not simply his high social status, made him so central. In the case of Pilatos, we see no particular evidence of any high social status. Indeed, he seems to have been well connected solely because of his notarial role. He signed the petition to the empress in his own right, but also on the behalf of three other men, one a kephalai¯ot¯es and another a kt¯et¯or, who are all illiterate. He had ties to the family of Dioskoros and other members of the local elite also owing to his literacy. That literacy improved one’s social connectivity is a pattern we will see repeated throughout this chapter’s discussion of Aphrodito society. (4) Ieremias was son of Iakubios and father of Psates, and ranked as the third most central man in our data-set by betweenness. Maspero saw his family story as evidence that Dioskoros the poet was annexing the land of the family of Ieremias one piece at a time. In , Ieremias, described as a priest, ceded land (paracwrž[s]ewv) to Dioskoros, who formally requested that it be registered “in my name” (e. «.v –m¼. n Ànoma). In another text from perhaps the same period, his son Psates, described as a lector, was doing the same thing. Ieremias also appears decades earlier, in , as a witness to a pledge of surety issued to Apollos the riparius by Apollos son of Hermeias, grandson of Pchichos, on behalf of the shepherd and fieldguard named Ieremias son of Iosephios. An interesting picture of connectivity emerges in this case. On the one hand, Ieremias and his family appear to have been losing economic ground to Dioskoros and his family. On the other hand, Ieremias’ social ties, and perhaps his credibility as a priest,



   

P.Mich. .. For more on Pilatos, see Diethart and Worp , –, which includes references missing from Girgis, and several published since; including these in the analysis would make Pilatos even more centrally connected than he already is. See also Fournet , –, where Pilatos figures in Fournet’s attempts to date various indiction entries in P.Flor. .. For the date of P.Mich. ., see Gascou , –. Kornelios: P.Flor. ..– and –; Diethart and Worp , –.  P.Cair.Masp. vol. , p. .  P.Cair.Masp. .. See above, n. . Psates son of Ieremias appears in P.Cair.Masp. ., but Ieremias himself does not figure in the text. For the date of , see below, n. . P.Cair.Masp. ... Apollos the riparius: n.  above.



Social networks in Byzantine Egypt

made him a useful witness in the ongoing mediation between Aphrodito’s elites and their shepherd community. (5, 6) Two of our top-ranked players from Aphrodito were Kallinikos son of Victor (fourth by closeness) and Kallinikos son of Hermauos (fifth by betweenness), both kt¯etores attested from the  petition to the empress. Modern commentators have suggested that the first Kallinikos was the son of Dioskoros’ cousin Victor. If true, his high closeness centrality here would make him the best-connected member of his extended family after the removal from our data-set of Dioskoros, Apollos, and Phoibammon. He also appears in P.Cair.Masp. ., an agreement from  which he, Dioskoros, Senouthes and a fourth Aphrodito villager reached with Palladios, count of the sacred consistory, and Epigonos comes, concerning Aphrodito’s autopragia. His last appearance is the least often noticed: in April , he signed as a witness to a lease of a drawn wagon. The leasing party was (perhaps) his father’s cousin, Dioskoros himself, the lessors two local villagers. Kallinikos was presumably a witness by virtue of his relation to Dioskoros. Kallinikos son of Hermauos, the second kt¯et¯or named Kallinikos in the petition to Theodora, also served as witnesses, in this case to a lease agreement from July , P.Cair.Masp. .. That lease was between Psais the potter on the one hand, and the heirs of Helen and Maria, two daughters of Romanos, on the other. The other witness was a Ioannes son of Kyriakos, and the scribe was named Kyros: both men appear in the same roles in another lease agreement from the same month involving Dioskoros the poet and the other heirs of Apollos. This Kallinikos seems to show the same sort of connectivity as his homonym, whom we have just discussed: a certain amount of prestige or economic clout by virtue of landowning  

   

Kallinikos : P.Cair.Masp. .... Kallinikos : P.Cair.Masp. .... Thus Gagos and van Minnen , , but without citations. Nor does there appear to be any previous discussion of the issue to provide a foundation for their assertion. MacCoull ,  only describes Kallinikos son of Victor as one of Dioskoros’ “colleagues” on the trip to Constantinople. Kallinikos is not in later Aphrodito volumes, e.g. P.Michael., P.Mich. , and P.Vat.Aphrod. Neither Martin  nor Salomon  comment on Kallinikos. Maspero attempted no identification in his ed. princ. of texts mentioning Kallinikos, nor does Bell mention Kallinikos in any of his relevant articles (e.g. Bell , , ). For a brief discussion of this text, see Martin , . P.Cair.Masp. ..–; for —lkist»v read —lkust»v. Menas son of Psaios is from Tanuaithis in the lesser Apollonopolite nome, but is described as t‡ nÓn di†gwn –ntaÓqa –pª kÛmhv %frod©thv toÓ | %ntaiopol©tou nomoÓ: P.Cair.Masp. ..–. P.Cair.Masp. ., although Maspero admits that the reading at line  of ì I.w†nnhv Ku. ri.a. ko(Ó) is “douteuse.” For Kyros, see Diethart and Worp , , s.n. ..–. Ioannes appears as a witness again in P.Mich. ..

4 Quantifying Aphrodito’s social network



status put both men in the center of Aphrodito’s autopragia defense, at the same time presumably accounting for their desirability as witnesses in other economic transactions. (7) Flavius Theoteknos, a former praepositus, we know from documents spanning thirty years; he ranks fifth by closeness centrality. As a Flavius, he is somewhat out of place in Aphrodito; Keenan once remarked on the “generally ‘Aureliate’ composition of Aphrodito society,” noting that “Flavii in the Aphrodito papyri tend by and large to be outsiders.” The activities of Theoteknos in a number of texts fit a larger pattern. In P.Cair.Masp. ., from , he helped a shepherd named Apollos write an acknowledgement of debt to Dioskoros because Apollos could not do it himself. He appears with Pilatos and Abraam in P.Cair.Masp. ., already discussed, and appears as a landowner in the petition to the empress. One modern scholar observed that Theoteknos “evidently made a practice of writing hypographai for illiterates.” James Keenan has drawn attention to “numerous errors” throughout the dozen examples of his writing: presumably these illiterates would not have known that Theoteknos suffered from poor spelling and grammar. At any rate, we may reasonably suppose that Theoteknos was in demand for social reasons: illiterates sought his services not only for his literacy but for the power of his social connections as well. From these brief biographies, one might conclude that in order to achieve social connectivity in late antique Aphrodito, one should: () own land; () learn to write; and () make friends with shepherds. It is perhaps unsurprising that Aphrodito’s landowners should appear well connected in the papyrological record. They had widespread economic clout and were engaged in activities that relied heavily on the keeping of written records. It is this latter detail that leads us to a somewhat less obvious structural foundation of social connectivity, the ability to write. When we note the relatively small average distance between any given pair of people in our Aphrodito network (less than three degrees of separation: see Tables – above), it seems that the indirect connections latent in the activities of the scribe helped make this possible. (Other indirect connections no doubt existed as well, but we are throughout this discussion constrained by our 

 

See PLRE . and Girgis . Theoteknos’ career is considerably under-represented in the Girgis proposopography: see corrections to the prosopography in the appendix to Ruffini . Analyzing a data-set including these missing citations increases his importance by these rankings.  P.Michael.  intro., cited by Keenan a, . Keenan b, . Keenan a, –.



Social networks in Byzantine Egypt

reliance on the written record. ) Finally, the role of the shepherd, a figure undeniably from Aphrodito’s lower social registers, is a surprising one, but presumably attributable to their curious dual role as pastoralists with ties to many different regions, and as fieldguards implicated in the economic well-being of the landholding elite. This role in fostering connectivity will be explored in the following section. shepherds and fieldguards Quantitative analyses alert us to the importance of shepherds in Aphrodito’s social networks and suggest the need for a closer look at the prosopographical evidence. For earlier scholars, attention to pastoralism has heightened appreciation of Aphrodito’s economic adaptability. As James Keenan put it: more villagers are identified as shepherds in the Aphrodito documents than in documents for any other town or village of Roman or Byzantine Egypt, [which] suggests a degree of specialization greater than that attested for anywhere else by the papyri; and this very specialization points toward production for exchange rather than for subsistence alone.

The importance of Aphrodito’s shepherd community is affirmed by a late seventh-century village account in which the shepherds as a group pay  solidi  keratia, more than any of the ten other associations (suntecn©tai) also recorded in the account; only the t”ktonev come close to matching the size of the shepherds’ payments, paying  solidi. From this alone, we might expect Aphrodito’s shepherds to play a large role in its social network. This is affirmed by our earlier discovery, that providing financial security or scribal assistance to shepherds partially accounted for the high connectivity of several of the most central players in Aphrodito’s social network. Chapter  drew attention to the importance of dual roles in contributing to high social connectivity. With this in mind, we should remember that some Aphrodito shepherds also appear in our records as agricultural landowners. Perhaps more important – as our degree centrality tests have suggested – was the shepherd’s dual role as fieldguard, a role greatly complicated by the field damage Aphrodito’s pastoralists might frequently cause. Documentation of this dual role appears in texts from  and    

People knowing each other by sight or by reputation alone were probably quite common.  Keenan , . For much of what follows, see Keenan a and .  Keenan , –. P.Hamb. .; see also R´emondon .

4 Quantifying Aphrodito’s social network



respectively. The first document outlines what James Keenan has called a “wickedly clever” agreement the Aphrodito landowners made with the shepherd guild outlining that guild’s permanent responsibilities for guarding the landowners’ fields. The second text is a “series of twelve guarantees submitted to the village police officer (riparius) in the year , all in the month of Epeiph ( June– July),” about which we will hear more below. The guarantees are individual and more detailed elaborations of the collective agreement the shepherds had made seven years before. A collection of accounts dating after /, which Maspero suggested may relate to the holdings of the monastery of Michael the Archangel, provides information vital to our understanding of pastoralism in Aphrodito. A number of hands can be identified in the text, one of the most central being another Dioskoros, on palaeographical grounds clearly not the poet. Because he writes in his own hand a declaration of the size of a sheep flock, beginning with the phrase ìEmo(Ó) toÓ Diosk»ro(u), sÆn Q(e)w‚ô¯ , pr. (»)b(ata) k. g ˆr(renik‡) (kaª)[qhlik(‡)], Maspero took him to be a shepherd. Keenan has doubted this assertion, arguing that a shepherd’s activities were not likely to generate such a lengthy document. Rather, he sees Dioskoros as controlling land in which the flocks in question pastured a portion of the time. Certain characteristics of the text – particularly the labeling of certain sheep as black-wooled (m”lai(na)) – indicate that the sheep were raised for their wool, not for their milk or meat. The implication, presumably, is that shepherds were a part of the market economy, socially connected to purchasers in Aphrodito and other villages.  

 

  

P.Cair.Masp. . and .. And yet how clever? Keenan may be right, but it is hard to escape the feeling that this was an arrangement of necessity, successful as often as it was not, owing to the shepherds’ natural tendency to see the land around them as thoroughfare rather than obstacle. See for instance the attitude of Mousaios’ son in P.Cair.Masp. . (discussed at length in Keenan a and below, p. ): how different could he have been from the shepherds Dioskoros expected to protect his land? Keenan a, , and below, p. . For an interesting deconstruction of the different types of riparii often assumed in the historiography of this period, see R´emondon , –. See P.Cair.Masp. . and Keenan , . This paragraph follows Keenan ,  closely. For the date of the text, see Fournet’s unpublished “Liste des papyrus e´dit´es de l’Aphrodite Byzantine.” Maspero dated the accounts to the end of the sixth century by virtue of the appearance therein (P.Cair.Masp. ...v.) of Ioannes son of Kornelios, who appears in P.Cair.Masp. . () in the midst of a highly fragmentary reference to the diakonia “to(Ó) ‰g©o(u) Ârouv Mica]hl©o(u) ˆrcag’g”lo(u)” (line ), but Ioannes belongs to a slightly earlier period.  P.Cair.Masp. ...r.. Girgis . Keenan , , although he admits the attraction of Maspero’s suggestion on the same page. As Keenan , – puts it, “Given the prevalence of sheep and wool at Aphrodito and the evident importance of sheep husbandry, and given the existence of the livestock fair, we should be



Social networks in Byzantine Egypt

Shepherds roamed the fields on every side of Aphrodito. Residents divided the area around the village into north, south, east, and west plains. Shepherds are attested in at least three of the four quadrants, as are shepherds doubling as fieldguards. But their presence in separate divisions of village territory did not mean that these shepherds were socially isolated from each other. The shepherd Aurelius Victor, son of Psaios, worked as a fieldguard “of the division of Phennis,” north of Aphrodito. Aurelius Psenthaesis, son of Mousaios, was a fieldguard of the fields of Pherko, south of Aphrodito. But they were both members of the same guild and appear as co-signatories together in that guild’s contract with the village headmen in . Shepherds not only came to town, as we imagine they must have done to sell their wool, they owned houses there and probably lived there as well. A house sale dating from the reign of Heraklios (–) describes the property’s southern boundary as “the house once of Abraam, son of Horuogchius, the vetch-seller, now of the heirs of Andreas, the shepherd.” Nor would distance play much of a factor in keeping these shepherds socially isolated. Aphrodito had in its jurisdiction some , arourae of land registered for tax purposes, as both astika and kom¯etika, nearly  per cent of it arable, and spread more or less evenly on all sides of the village. But this amounts to hardly . square miles: unburdened by his flock, any shepherd could have crossed the entire village’s holdings in under an hour.



  



inclined to say that if there were no wool market for Aphroditan products, it would have to be invented.” North: In P.Cair.Masp. .., from , Enoch’s agreement with the shepherd named Isak concerns Phennis. In P.Cair.Masp. . (following Keenan a, where the text is re-edited, with BL .), from , Dioskoros has problems with a shepherd trampling the lot of Phenis (klžrou Føhøénø eø.[w]v). South: Pherko, where we find Aurelius Psenthaesis son of Mousaios in P.Cair.Masp. ... East: P.Michael. , where we find Kollouthos and Victor selling land to Phoibammon (see Chapter  above, pp. –). West: The one doubtful reference, P.Cair.Masp. ...–: Ieremias son of Iosephios, shepherd and fieldguard t¦v dekane©av to(Ó) libik. o. (Ó) | Porqm. ©o(u). Porthmiou is a place in the Aphrodito region: see Calderini, Diz. geogr., (): , citing P.Lond. .. (a doubtful reading: Po]r. q. m©o(u)) and P.Lond. ., but not this text. Does libikoÓ place Porthmiou in Aphrodito’s western plain? The usage seems unlikely, but if the author had intended to specify the western plain of Porthmiou, it would have been easy enough to do so. North, south, west: see references to P.Cair.Masp. . in the previous note.  P.Cair.Masp. ...  P.Cair.Masp. .. P.Cair.Masp. ... SB ., first partially published with the given translation in P.Mich. ..– (n»tou pote ½rbiop»lou | %braam©ou ërouw. g. c©ou nunª d• o«k©av tän klhron»mwn %ndr”av poim”(nov)), from which I conclude that Andreas himself must have owned the property after Abraam, in order for his heirs to now be in possession. See the Aphrodito cadastre in Gascou and MacCoull  (= SB .) for the village’s arable territory, and Calderini Diz.geogr., ():  for Ptol. ..’s description of Aphrodito as mes»geion.

4 Quantifying Aphrodito’s social network



The sheer mobility of the pastoral lifestyle was presumably the source of Aphrodito’s problems with its shepherd community. In his first of two articles dealing with pastoralism in Roman Egypt, James Keenan produced a re-edition and translation of an –ksfr†gisma or deposition dating from . The deposition is a clever rhetorical trick, through which Dioskoros the poet gives credence to a complaint he has lodged against a local shepherd by getting Kollouthos, bo¯ethos at the office of the defensor in Antaiopolis, to affirm that Dioskoros came to him with his complaint, which Kollouthos in turn repeats verbatim. In that complaint, we hear that a son of Mousaios, whose name is lost in a lacuna, had “destroyed” (–lumžn[at]o) Dioskoros’ crops by crossing his fields. Despite later having recourse to the claim that Dioskoros’ land “was formerly [or: sometimes] in a category of thoroughfares for lambs ‘in our time and in that of our ancestors,’” the shepherd nonetheless appears to have asked Dioskoros for permission. Dioskoros’ refusal to grant passage does not suggest that the shepherd was a stranger or socially unknown to him. On the contrary, the shepherd wished to pass because fields belonging to his brother Makarios were “just below” (parak†tw) those of Dioskoros himself. From this case, it appears that the tension between Aphrodito’s landed elite and its shepherds could manifest itself on a highly personal level. But Aphrodito’s pastoral problems were not merely social and economic. They may have been epidemiological as well. MacCoull has published a murky Coptic text missing crucial details, but including the following tempting fragments: “I inform you that the pustular disease has not settled upon me . . . They took them, they went out of the high fields . . . Our shepherds said . . . the place of lice (?) . . . ” MacCoull takes this text to be a reference to the plague, and thus to date from the s. Maybe the appearance of the plague and the shepherds within a few lines of each other is sheer coincidence. On the other hand, sheep and other animal populations suffered greatly from the plague in the fourteenth century. If    





P.Cair.Masp. .; Keenan a, and revisited by him briefly in Keenan . Trans. Keenan a, , of P.Cair.Masp. ..–: Ëfo. rwm”n t¼n gen†menon a. «.kism¼n | [›fhsen e²n]a© (?) pote t¼n t»pon –n ¾dhpor©av t†xei qremm†twn –pª ¡män kaª –pª prog»nwn. This would seem to be implied in pr¼ ¡merän g‡r §lqen pr¼v –m•. at line . Trans. MacCoull , –, from the Coptic TTAMW DE MMOK E MPI E ECOUW N MMAI . . . AU ITOU AUMAA E NCETIO E N. BOL N AACT. . . TN OOC O MMOOU. . . MN PMA NKOKTE. Although she had earlier written (MacCoull , ) that the world of Dioskoros “seems to have suffered hardly any effects from the sixth-century plague,” and suggested (ibid. n. ) as a result that the remarks in P.Cair.Masp. .. are “figurative, not literal.” For epizootic outbreaks of the plague described in al-Maqrizi, see Dols , –.



Social networks in Byzantine Egypt

the same problem surfaced in Aphrodito in the sixth century, the villagers may have seen their local shepherds as part of the problem. Why do we have so much documentation about these local shepherds? We have already met the shepherd named Victor, son of Psaios, whose activities suggest an answer. He is an incidental character in an article by James Keenan about another man of exactly the same name. He also appears in a text dated to , in which he leased an estate in the western fields of Phthla called “the Sarapammon” from Flavius Panolbios, the son and heir of a Flavius Ioannes, politeuomenos, who appears in a number of Aphrodito texts. We find him a few years later (?), still in the western fields of Phthla, due east of Aphrodito. But why do these texts survive at all? Other authors have noticed that they have no immediate connection to the rest of the Dioskoros archive. One possibility presents itself. We know that several years later, Apollos, father of the poet Dioskoros, leased land in the Phthla area from the heirs of a politeuomenos named Kyros and another politeuomenos named Flavius Ioannes, not the father of Panolbios. If Apollos leased land from one city councilman and the heirs of another, he may have been involved in the business of a third, Panolbios, in his lease to the shepherd Victor at an even earlier date. Economic involvement in the estate might explain the presence of the paperwork in his family archives. (At any rate, Dioskoros had reasons to keep a paper trail about other shepherds: he received a midsixth-century letter from a companion of the pagarch Menas, writing that a certain shepherd had been ordered to desist from harassing Dioskoros’ tenants. )  

   



See above, p. . Keenan a, where the shepherd appears in n. . The Victor son of Psaios Keenan concerns himself with throughout is a scribe of poor skill who drafted PSI . and P.Cair.Masp. .. The discussion that follows on PSI ., P.Flor. ., and P.Cair.Masp. ., ., and . follows this Keenan article closely. For P.Cair.Masp. ., see the latest re-edition at Keenan b. P.Flor. .. For Panolbios, see Salomons , . P.Cair.Masp. .. The date of the text is lost, but see Keenan a,  n. , suggesting  as probable. Keenan a, . See PSI . (/?), SB . (= P.Cair.Masp. ., c./) and P.Cair.Masp. . (the following year). Keenan a,  initially supposed that this Ioannes was the father of Panolbios, but this soon fell on chronological grounds owing to the redating of PSI . from the s in Keenan b,  n. . See also Salomons , , particularly for his argument that Kyros is in fact Panolbios’ son, based on a “virtually certain” restoration of Kyros in his edition of P.Lond. .. From the plate he provides () such certainty seems to me unwarranted. P.Lond. .. And P.Cair.Masp. ., an account detailing the theft of numerous sheep, and the apparent suspects, may also be in the hand of Dioskoros.

4 Quantifying Aphrodito’s social network



But this is not all we see of Victor, son of Psaios. We have already met him in P.Cair.Masp. ., where in  he appears as both shepherd and fieldguard of Phennis, in Aphrodito’s northern plain. In other words, Victor’s activities took him from the western plains of Phthla, through Aphrodito’s eastern plains, in , to the region’s northern border in . (And, if Keenan is right about the date of P.Cair.Masp. ., back again to Phthla a few years later, c.. ) But what sort of activity is this, really? Is it evidence of the seasonal or annual cycle of transmigration we might have imagined to be a part of Victor’s life even without the evidence? Or is it something more permanent, an indication that a shepherd like Victor might lease land in one area while keeping his flocks somewhere else entirely? Certainly, shepherd families enjoyed ongoing relations with various Aphrodito institutions, suggesting a certain degree of permanency and settled familiarity. The Aphrodito cadastre provides an example in the hospital of Apa Dios of Sunoria. The hospital had four landholdings, two of them very small, registered in the cadastre. Two of them were in the hands of (Ëp¼) a shepherd named Victor, son of Hatres. The third was in the hands of a shepherd named Tsouroose, daughter of Hatres, and presumably Victor’s sister. (Female shepherds are rare. Does the reference mean that Tsouroose tended her own flock, or only that she had inherited animals of her own left in the care of others?) Unfortunately, the references are scattered in three different sections of the cadastre, providing us with no clue whether the holdings were near each other, or, again, represented some sort of migratory pattern. Meeting two siblings, both of whom were shepherds with economic ties to the same institution, leads us to the question of hereditary tendencies in Aphrodito’s pastoral community. Keenan thinks that the Oxyrhynchite herdsman in P.Oxy. . might have been legally bound to follow his father’s occupation. In another context, MacCoull has also asserted the inheritability of professions in the Aphrodito area. However, there is no definite evidence that Aphrodito’s shepherds were obligated to follow the family practice. Of all the shepherds listed in Girgis, only Victor son of Psenthaesis and Ieremiah son of Pathelpe were definitely both shepherds and sons of shepherds. Pathelpe himself may have been the son of a    

 See above, n. . Northern plain: see above, n. . SB .; Gascou and MacCoull , , s.v.  See her arguments in MacCoull , . Keenan b,  n. . Victor: Girgis . Ieremiah: Girgis . The following appear in Girgis as mothers or fathers of shepherds: Girgis , , , , , , , , , , , , , , , , , ,



Social networks in Byzantine Egypt

shepherd in turn: in the same text in which Pathelpe is reported for the theft of two sheep, his father Mesmois was reported for or at least suspected of the theft of twenty-five sheep and thirty goats. Considering that we have over forty attestations for the parents of shepherds, this is not an impressive total. But none of the remaining attestations include any occupations or titles whatsoever. This leaves no reason to doubt that they were shepherds, and that, as in any village society, a strong tendency existed for the sons of shepherds to become shepherds in turn, as did Victor and Ieremiah. Continuity of occupation within a family has implications for social connectivity. In , when Psenthaesis son of Mousaios appears as a shepherd and fieldguard, Ioannes the deacon, son of Beskos, appears in the same text, signing for Psenthaesis and several other illiterate shepherds appearing therein. In , when Psenthaesis’ son Victor appears as a shepherd and fieldguard of an unspecified location, Ioannes the deacon appears again, providing surety for both Victor and his father. We have seen that landholders in Aphrodito formed institutional relationships with neighboring families. Here, we see shepherds doing something similar, returning to the same service personnel years later, despite their relatively unsettled lifestyles. We will discuss Ioannes’ strong ties to the shepherd community in further detail below at pp. –; it is interesting to note that he signs for one group of shepherds in P.Cair.Masp. ., while Abraam son of Victor, whom we have also met before, signs for the other. Perhaps these two shepherd groups represent extended families, or some other as yet unrecognized clusters or factions within the larger shepherd network. These further analyses of the shepherd community make it clear that the liminal status of shepherds in no way prevented those in Aphrodito from being socially well connected. I propose that the prominence of the



 

, , , , , , , , , , , , , , , , , , , , , , , , , , . See the remarkable P.Cair.Masp. ., where at line  begins an “account of those who stole my lambs,” Gnäs(iv) tän kl”yantwn t‡ –m‡ qr”mmata; at line  is an entry p(ar‡) toÓ poim(”nov) Paqelp. e. pr»(bata) b to(Ó) Mesmo(u)©tov; and at line  is an entry p(ar‡) Mesmou©t. o. v pr»(bata) ke k(aª) a«g(©dia) l. This depressing account is in what Masp´ero describes as a “Cursive ronde soign´ee, tr`es analogue a` l’´ecriture du po`ete Dioscore.” He notes that this could be the same hand found in P.Cair.Masp. ., where reference is made to a ëIerhm©o(u) uªo(Ó) Paqlpe poim”(nov), giving us two generations of the same family. For this text, see only BL . and .; it has received no close analysis, but strongly suggests that Dioskoros, if the hand is his, had flocks of his own in the care of shepherds, revealing yet another tie between him and the shepherd community.  P.Cair.Masp. ... P.Cair.Masp. .. and . For discussion of potential factionalism in Aphrodito, see Ruffini forthcoming b.

4 Quantifying Aphrodito’s social network



shepherds in our analysis of Aphrodito’s social network is an example of the “strength of weak ties” originally noticed by Granovetter in . The essence of his proposal (on which, see the Introduction above, p. ) was that bridges between various sections of a large network are most likely to be peripheral figures: not family members and close friends, but those liminal figures one encounters only rarely. Shepherds fit this billing well: they wandered widely through Aphrodito’s rural surroundings, dealt readily with its village elites, and performed multiple different social functions. We have seen that shepherds played a role in connecting Aphrodito to its satellite villages. This role was not always a positive one. When the “wretched small-holders and inhabitants of the thoroughly wretched village of Aphrodito” protested their rights of autopragia to Flavius Athanasios in , in the face of the depredations of Menas the pagarch, shepherds lurked on the edges of the story. Part of the pagarch’s technique for obtaining or ensuring tax-collecting rights over Aphrodito involved using the shepherds of Phthla to harass owners of Aphrodito borderland: one of the “wretched small-holders” to feel the effects of this was Dioskoros himself. It is hard to imagine that the pagarch encouraged this encroachment in a socially disconnected vacuum. Dioskoros received what amounted to a restraining order against one shepherd from Mounkrekin, a village in the Antaiopolite nome, and may have had ties to other shepherds in Phthla associated with land once in his father’s hands. On a purely speculative level, we might imagine that the shepherds in league with the pagarch were precisely those such as the son of Mousaios (discussed above, p. ) who could plausibly claim an ancestral right to cross the lands in question. In other words, it is likely that when Menas sent the shepherds to encroach on Aphrodito land, their actions involved moving through social and physical space with which they were already well familiar. A final word about shepherds, in a comparative Mediterranean context. Aphrodito’s decision to hold the shepherd community collectively responsible for security is not by any means unique in pastoral history. In the 

  

Shepherds of Phthla: see P.Cair.Masp. ..., where Menas “turns over harvested arouras to the bo¯ethos of the village of Phthla and to its shepherds to settle” (–pitr”yai t t. e boh q t¦v kÛmhv FøqøløŽ kaª to±v taÅthv poim”si o«keiÛsasqai —auto±v t‡v ˆroÅrav aÉtoÓ karpoum”nav). Wretched small-holders: see P.Cair.Masp. .., “ˆql©wn leptoktht»rwn te kaª o«kht»rwn t¦v pant†qliav k[Û]mhv %frod©thv.” And cf. P.Lond. ..–: ˆp¼ leptoktht»rwn g‡r | sÅgkeitai ¡ kÛmh, cited in Chapter  above, n. . P.Cair.Masp. ...–. Restraining order: P.Lond. .; cf. P.Cair.Masp. . for Mounkrekin. Associations with Apollos in Phthla: see above, pp. –. P.Cair.Masp. ..



Social networks in Byzantine Egypt

highlands of Sardinia, for instance, the barracelli live a double existence, working as shepherds, yet also serving as mediators in shepherd conflicts and as an armed front-line of defense against animal theft. Much like the shepherds of Aphrodito, legally obliged to curb their own peers through their contracts with the area’s chief landholders, the barracelli, “who are both state officials and shepherds, have more strength than shepherds alone, for behind them is the constantly menacing presence of the state.” In Sardinia too, shepherds are not social isolates, but well integrated into, and indeed, dependent on local social networks: [social] networks, often linked by casual acquaintances, extend through entire zones, and form, in the minds of the shepherds, social maps with which they can chart their journeys. These networks are consciously established, and the need to maintain and reinforce them . . . is consciously recognized.

If the analogy holds, the prominence of shepherds in the social circles of the most prominent men in Aphrodito’s social network may not be a mere fluke or a symptom of the shepherds’ corporate identity, but a calculated social survival strategy with implications for comparative village history throughout the Mediterranean world.

network features i: centrality and average distance The network measurements of centrality and distance suggest that Aphrodito’s social network was relatively decentralized. As I discussed in the Introduction, the concept of centrality does not apply solely to individuals, as a means to measure whether they are more or less central in a network. It applies to networks as a whole, as a way to measure the degree to which one actor or a small group of actors has considerably higher centrality than others in the network. It serves in other words as a measure of how hierarchical or centralized a network is. Aphrodito’s social network was not highly centralized. That network’s largest component has a degree centrality of . per cent, a closeness centrality of . per   

 Ruffini, J., , .  Ruffini, J., , . Ruffini, J., . See the Introduction above, pp. –. Compare the figures given here to one standard comparandum for network analysis, Padgett’s Florentine families. Their degree centrality measure of . per cent is considered “a rather small value, indicating that the difference between the largest and smallest actor-level indices were not very great” (Wasserman and Faust , ); their betweenness centrality measure of . per cent, which indicates the disproportionate importance of the Medici and Guadagni families, is much greater than the betweenness centrality measure given here.

4 Quantifying Aphrodito’s social network



cent, and a betweenness centrality of . per cent, with the last figure dropping considerably when Dioskoros and Apollos are removed from the analysis. Without these figures, we might have imagined that Aphrodito’s elite – its landholders and office-holders – were disproportionately better connected than the other villagers. Instead, these figures tell us that those who were better connected – scribes and shepherds as well as landholders – were not far enough ahead of their peers to create a disproportionate social hierarchy, or aristocracy of connectivity. In the Introduction, I also mentioned Stanley Milgram’s famous social network experiment which ultimately gave rise to the “six degrees of separation” clich´e now part of modern parlance. This clich´e refers to the concept of distance, the number of social connections separating any two people, or the average of such figures across a complete social network. The figures for “average distance” which our analyses produce are surprisingly low. The average distance between villagers in the largest component of the Aphrodito network is . degrees with Dioskoros and Apollos removed. This indicates that the average Aphrodito resident could connect to the majority of the others in only two or three social steps. If Phoibammon son of Triadelphos needed something from villager X, but did not know him personally, he certainly knew villager Y, who could then put him in touch with X. According to network theory, smaller average distances imply a larger degree of randomness and a smaller degree of formalized order in the shape of the social network. A network in which Aphrodito residents knew only their closest neighbors or those closest to their own social status – in mathematical terms, a highly ordered network – would have a much higher average distance. A less ordered network, in which a peasant farmer in the northeast fields of Aphrodito knew a powerful landowner from the southwest fields, is the most likely way to generate such a low average distance. As with Aphrodito’s network centrality, this relatively low average 

 

See above, pp. –. Nor is this merely a symptom of the incomplete state of the Girgis prosopography. While we find many cases in which he does not list all references for any given individual (see Ruffini , Appendix ), Dioskoros and Apollos both have considerably more citations than average listed therein. Discussed and defined in more detail in the Introduction above, pp. –. For a brief summary in layman’s terms, see Buchanan , –. It is important to note that when I describe a social network as more or less ordered, I am not speaking of order in the social sense of class, group, family and hierarchy. Rather, I am speaking of order in terms of graph theory. Visualizing a network of  nodes as a grid of  nodes by , a highly ordered network is one in which each node connects only to those nearest to it. A highly disordered network features nodes with connections all over the graph. Proximity and distance in this context of graph theory do not necessarily correlate to any obvious feature of the Egyptian social landscape.



Social networks in Byzantine Egypt Table  Ties of strength greater than 1

Names and Girgis Numbers

Notes

Kallinikos () to Senouthes ()

Senouthes was a ge¯orgos and pr¯otok¯om¯et¯es, Kallinikos a kt¯et¯or Enoch was a monk, Dioskoros and Apollos both pr¯otok¯om¯etai The first two are shepherds, the latter a deacon All four were kt¯etores

Enoch () to Dioskoros (), Dioskoros () to Apollos () Victor son of Psaios () to Hermauos () and Ioannes (), both of the latter to each other Abraam () to Flavius Theoteknos (), Theoteknos to Amais () and Mathias (), both of the latter to each other 

Parenthetical numbers are Girgis citations for each individual. Strength of ties in a dataset can be determined by exporting the data-set from UCINET into a text file in which the data is stored in DL edgelist format, with row and column labels included and the diagonal excluded.

distance suggests that Aphrodito society enjoyed a certain fluidity, that its social connections were not particularly stratified. network features ii: tie strength The strong ties we find in the Aphrodito network are quantitative manifestations of the important multiplex ties we identified in Chapter . In a recent analysis of the topography of the Oxyrhynchite nome, I emphasized the importance of the strength of a tie, or the number of links between any two members of a network. In Oxyrhynchos, this method reveals the relative importance of the connections between any two villages. In Aphrodito, it can reveal which men and women were most frequently connected to one another. But the Girgis data-set cites relatively few texts per person, and thus does not give a particularly nuanced look at tie strength. Of the nearly , names, most of the connections have only one tie between them. Remove all of these ties from the analysis, and only four groups of people appear connected to one another through multiple texts. Why are these people better connected to one another in the surviving evidence   

Ruffini . For examples of citations missing from Girgis, see Ruffini , Appendix . This small number does not contradict my argument in Chapter  that Aphrodito’s villagers manifest a tendency towards multiplex ties. The ties I discuss in Chapter  are very often ties – kinship, business partnership, proximity of owned land, etc. – that appear all together in a single text, and thus do not appear among the “strong ties” herein.

4 Quantifying Aphrodito’s social network



than their peers? Do these four groups have anything in common that might explain this increased connectivity? Let us look at each case. (1) Kallinikos and Senouthes: Senouthes was the son of an Apollos, and thus possibly the brother of Dioskoros the poet. We know Kallinikos son of Victor from three texts, c., , and  respectively, and have discussed him before: he is one of our network’s most-connected members by betweenness centrality. Two of the three Kallinikos texts mention Senouthes. Both appear as signatories to the petition to Theodora in defense of Aphrodito’s autopragia. They appear together again a few years later in an even more remarkable document in defense of the same right, written in Constantinople in the summer of . In it, Senouthes represented Kyros son of Victor, an Aphrodito kt¯et¯or who had signed the petition to Theodora a few years earlier. He appears with Kallinikos, Apollos son of Ioannes, and the poet Dioskoros, securing an agreement from a person of no less importance than Flavius Palladios, count of the sacred consistory. If Kallinikos’ father Victor was indeed Dioskoros’ cousin of the same name, as has been proposed, the higher than average tie strength between Kallinikos and Senouthes no doubt came as a result of that family tie. (2) The second group with a high internal tie strength was Enoch, the poet Flavius Dioskoros and Apollos. This Enoch is not the kt¯et¯or we have already met with the same name: this man appears in two different texts two decades apart, described as a reader (eÉ. l. [ab]”stat. on ˆ. n. agnÛsthn) in , and as a monk in /. Dioskoros appears in both texts as well. In the first text, a group of Aphrodito villagers wrote to the pagarch Ioulianos through Dioskoros the poet and Apollos son of Ioannes, pr¯otok¯om¯etai, to offer a six-solidus guarantee on Enoch’s continued presence in Aphrodito: m»nhv kaª –mfan[e©a]v ìE[nÜc ì Iw†n]n. o(u) eÉ. l. [ab]”stat. on ˆ. n. agnÛsthn | –pª t aÉt¼n –mme±nai kaª eËrhq[¦n]ai –n t
aÉth„ kÛm %frod©thv.     



See Chapter  above, p.  with n. . P.Cair.Masp. ., ., .. See above, pp. –. P.Cair.Masp. .. See Chapter  above, p. . For the proposal of Kallinikos’ family tie, see p.  above. P.Cair.Masp. . and ., where Maspero made the identification in his note to line . He is also presumably the same man listed in line  of P.Flor. . as *pa ìEnÛc EÉlab( ), a presbÅterov. As is clear from that text’s re-edition in Zuckerman a, the original editor mistook EÉlab( ) for a patronymic when it probably ought to be read as an abbreviation for eÉlab”statov. Maspero read the same epithet next to Enoch’s name in P.Cair.Masp. . and .. Girgis  (“Eulab/”) thus disappears. P.Cair.Masp. ..–; the nominative name is established by comparison to lines  and , the patronymic by comparison to P.Cair.Masp. ... Masp´ero notes at . that “on distingue encore le tr´ema qui surmonte souvent les i dans ces papyrus. La comparaison avec le n◦ , l. , rend cette lecture tout a` fait certaine.”



Social networks in Byzantine Egypt

In the second text, Enoch and Dioskoros appear together as joint intermediaries handling a donation to the Apa Apollos monastery. Dioskoros, of course, was that monastery’s curator (frontist[o]Ó kour[†to]r. ov), while Enoch, by then a monk, had become its oikonomos ([mon†]zontov kaª o«ko[n»]mou). It is surely the same man: the donation is made di‡ ìEnÜc ìIw†nnou eÉlabest†tou. These brief glimpses, twenty years apart, suggest the expanding scope of an ongoing relationship between Dioskoros and Enoch. Enoch, previously only a reader, must have joined the Apa Apollos monastery and demonstrated sufficient competence to become the oikonomos. Apollos son of Ioannes, the final member of this triad, is linked to Enoch only once, in the guarantee of surety filed on Enoch’s behalf, but appears with Dioskoros twice. Both Apollos and Dioskoros were intermediaries in that guarantee of surety, and both made the trip to Constantinople in , attested together with Dioskoros’ proposed relatives Kallinikos and Senouthes, in the autopragia agreement with Flavius Palladios. (3) The third group with a high tie strength was Victor, son of Psaios, Hermauos, son of Iosephios, and Ioannes, son of Beskos. Victor is the shepherd we have already discussed. It is no surprise that one of his strongest ties should be to Hermauos, another shepherd: they both appear in two of the collective agreements the shepherds make with the landowners and various officials of Aphrodito, in  and . Ioannes, a deacon, is the odd man in this triad. In , he signed for one group of shepherds, including Hermauos, who were themselves unable to do so. In , the connections were more varied: both shepherds were referred to Apollos the Aphrodito riparius in their capacities as shepherds and fieldguards and Flavius Theoteknos provided his services in both cases, as a literate signatory and witness. Here, in P.Cair.Masp. ., the shepherds agreed to certain specific fieldguard duties, Victor in the dekania Phennis (lines .–: t¦v dekane©av kalo(u)m”n(ou) | F.ø eø nø nø eø wv), Hermauos in the dekania Terkeos (line .: t¦v dekane©av kaloum”n(ou) Thrk. e. [wv?]). As he did in , Ioannes again signed for shepherds, this time offering security for a father and son in Pherko (line .: pedi†d(i) Føeø røkø oø). These links, although indirect, are not coincidental. There is no indication in either text that Ioannes worked    

P.Cair.Masp. ..–. P.Cair.Masp. .. For this Apollos, see n.  above.  P.Cair.Masp. ..–. P.Cair.Masp. . and .. Signatory: P.Cair.Masp. ...–. Witness: P.Cair.Masp. ...–. Apollos: n.  above.

4 Quantifying Aphrodito’s social network



for the village headmen, kt¯etores, or the riparius; his role as a writer in these texts is due to his connections to the shepherd community. The father he signed for in , Psenthaesis, was in the group he signed for in  as well. Why this deacon would have particular links to the region’s shepherds is not obvious, but one possibility suggests itself. Victor son of Promauos has been proposed as a shepherd and fieldguard in his youth who became a priest in his old age. On a sheer guess, this deacon Ioannes may be another example of a shepherd entering the religious life, but maintaining his ties to the shepherd community. (4) The fourth group with a high tie strength comprised Abraam, Flavius Theoteknos, Amais and Mathias. Abraam and Theoteknos we have met before. They were two of our network’s most central figures, by multiple different measures. Abraam and Theoteknos both served as witnesses to an acknowledgement of indebtedness issued to Flavius Dioskoros by a Psais, son of Besios and Tasais, in . They also appeared together in the petition to the empress Theodora. (In , Abraam also joined Ioannes in signing for shepherds who were unable to do so: see (3) above.) Amais son of Abraamios and Mathias son of Iosephios both appear as kt¯etores in the same petition to Theodora. They also appear with Theoteknos in  in a sale on credit to Flavius Dioskoros: Mathias served as a witness; Amais wrote the text; and Theoteknos signed for Apollos, a shepherd now in debt to Dioskoros for three solidi. These four examples of strong ties are suggestive. The first pair, Kallinikos and Senouthes, may have enjoyed a family tie and certainly had a direct documentary connection, the document from the trip the four landholders took to Constantinople in defense of their village’s autopragia. They and two men in the fourth group, Abraam and Theoteknos, also enjoyed a direct connection, their involvement in the petition addressed to Theodora a few years before, which raised the very same issue. The third group and the fourth in turn enjoyed parallel connections, as Aphrodito landholders signing on behalf of illiterate shepherds. In short, the most strongly connected pairs and triads in the Girgis prosopography formed 

  

So Girgis , citing P.Cair.Masp. .., , and P.Cair.Masp. ., thirty-nine years later. The main objection to this identification would be Victor’s illiteracy in the former text and his signature in the latter. Given the difficulties apparent in reading his given name in a digital image of , this may be an erroneous identification by Girgis.  P.Cair.Masp. .. See above, pp. –, . For the genre and this text as an example of it, see J¨ordens . P.Cair.Masp. ..: –gr†fh di’ –moÓ ìA. m†itov %braam©ou, his only appearance in Diethart and Worp , .



Social networks in Byzantine Egypt Table  First attestations of Aphrodito villagers by decade

First attestations

First attestations of people by decade 700 600 500 400 300 200 100 0 500s

510s

520s

530s

540s

550s

560s

570s

580s

Decade

bonds between them through their corporate identities, such as landholders acting on behalf of their collective rights or shepherds fulfilling their collective responsibilities. This need not have been the case. Strong ties could and no doubt did emerge from patron–client or lessor–lessee relationships spanning decades, if not generations. Indeed, I have discussed evidence for such ties in the previous chapter. But they are not the ties that emerge most strongly from quantitative analysis of the evidence. Does this mean that Aphrodito society was structured less around power-relations and more around corporate ties? It may be objected that Dioskoros’ role as a village headman would privilege corporate entities in his archive. But it is hard to imagine how the hypothetical archives of some other Aphrodito villager might look different. A shepherd’s life no doubt centered at least in part around the activities of his guild, as did the life of the typical Aphrodito wine-maker or weaver. Indeed, the transactions of villagers at the lower social levels might have looked even more corporate than those of Dioskoros: consider the boat-builders, who presumably dealt not only with their own guild, but those of the carpenters and the metalworkers as well. change over time i: isolating sub-networks by decade Analyzing the Aphrodito network data on a decade-by-decade basis reveals considerable stability in that village’s network characteristics. Table  shows 

For references to all of these guilds, see Chapter  above, pp. –.



4 Quantifying Aphrodito’s social network Table  Network characteristics of Aphrodito by decade I (first attestations ) Period

Number of people first attested

Average distance

Degree centrality

Betweenness centrality

s s s s s s s s  to 

        

. . . . . . . . .

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

. % . %  . % . %    . %



Both this and the subsequent analysis are performed on the complete and unmodified affiliations network of the Girgis prosopography.

the first appearance of Aphrodito villagers in the Girgis prosopography by decade. (People appearing in texts with unknown or imprecise dates are left out.) More people appear for the first time in the s than any other decade, and the majority appear for the first time in the period from  to . This substantially limits our ability to test change over time in Aphrodito’s social networks; the generation after the date at which Zuckerman argues that Aphrodito lost its autopragia is much less well attested than the generation before. Merely isolating periods of equal chronological length will produce sub-networks of radically different size. One alternative is to compare periods of unequal length – the decade of the s, for instance, versus the period from  to  – which are roughly comparable in the number of villagers first attested therein. Network analysis of Aphrodito focusing only on those villagers first attested in each decade (Table ) reveals considerable consistency in the network’s average distance and degree centrality. In a social network constructed only out of the people first attested in the s, each one of those people is on average just over a degree of separation away from all of the other people first attested in that decade. The same holds more or less true for every decade to follow, with a slight increase approaching two degrees of separation in the s. In the social network constructed only from the people first attested in the s, degree centrality is just over  per cent. The figure is identical six decades later, in a social network constructed only from the people first attested in the s. The average figure for the three



Social networks in Byzantine Egypt

Table  Network characteristics of Aphrodito by decade II (all attestations) Period

Number of people attested

Average distance

Degree centrality

Betweenness centrality

s s s s s s s s  to 

        

. . . . . . . . .

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

. % . % . % . % . % . %   . %

decades following  is just over  per cent. The same level of consistency is apparent with betweenness centrality, where the average figure for the three decades following  is the same as the figure for the s. Network analysis of Aphrodito including all villagers attested in each decade (Table ) does not produce a substantially different picture. The network of villagers attested from the s to the s shows an average distance somewhat higher than the network of villagers at the start of the century, but one somewhat lower than the network of villagers midcentury. Similarly, the period from the s to the s produces degree and betweenness centrality measures between either extreme in the first half of the century. In short, analyzing Aphrodito’s social network on a decadeby-decade basis does not show any particular trendlines, either towards centralization or away from it. change over time ii: testing network cutpoints Simulating the deaths of key players in the Aphrodito network also confirms that the village’s network characteristics remain largely unchanged over time. One way to do this is by examining the Aphrodito network’s cutpoints and levels of cohesion. As discussed in the Introduction, cutpoints are the crucial bridges in a network, the ties that bind each component part into one unified whole. If we remove a network’s cutpoints, that network fragments into isolated pieces. The more cutpoints a network has, the greater its cohesion, the more evenly distributed its information flow can 

See the Introduction above, pp. –.

4 Quantifying Aphrodito’s social network



become and the more robust it is in the face of the death of various members. Computer analyses identify twenty-eight cutpoints in Aphrodito, sixteen of which are cutpoints in the network’s largest component as well. The Aphrodito data-set, omitting the two large payment lists, contains a total of , villagers. If only twenty-eight cutpoints emerge from this network, then only . per cent of the names in that corrected data-set distinguish themselves as having characteristic connectivity much different from that of hundreds of their peers, connectivity that makes them bridgebuilders between different social sets. But a network of several disconnected components does not provide a good basis for counting cutpoints, which are themselves nodes whose absence would fragment the network. It is better to look at the largest component of interconnected names in the network and see which cutpoints emerge. The Aphrodito network’s largest component is  people, sixteen of whom, . per cent, act as cutpoints. Is this a large number of cutpoints for a network of this size, or a small one? Does it suggest a particularly robust social network, or a weak one? Very little work has been done by network analysts on this particular question, but one possible approach is to generate a random network with the same general characteristics, and compare the results. I use the network analysis computer program called Pajek to generate a random undirected graph with  nodes, the size of Aphrodito’s largest component, and an average of  connections per node, roughly the figure that Aphrodito’s largest component produced as well. The results are striking: after half a dozen iterations, none of these random graphs produced more than three cutpoints. In other words, the largest component in Aphrodito’s social network produced five times as many cutpoints as the most robust network I generated at random. While real-world comparanda are lacking owing to the relative lack of attention this network measure has received, the results seem clear: Aphrodito’s social network is stronger than statistically likely, and its fragmentation results only from the removal of a higher   



For a list of the cutpoints of the corrected Aphrodito prosopography, see Table  below. Cutpoints can be found using UCINET’s Network→Regions→Bi-Component function. See Moody and White  for recent work on cohesion; the suggestion of comparing this measurement against that of a random graph I owe to James Moody via personal communication. Pajek comes packaged with UCINET by Borgatti et al. I used the function Net→Random Network→Erdos-Renyi→Undirected→General, specifying  vertices and an average degree of . The first six iterations produced random graphs with , , , , , and  cutpoints respectively.



Social networks in Byzantine Egypt

number of cutpoints than would be present in a comparable random network. These cutpoints in the corrected Aphrodito data-set and its largest component include some predictable figures, particularly Apollos, Besarion, Dioskoros, and Flavius Victor, whom I will not discuss here. It also includes others who ranked high in our earlier connectivity tests: the priest Ieremias, who ranked third by betweenness centrality, and the shepherd Apollos, fourth by degree centrality. Interestingly, however, twenty-one of the cutpoints from the corrected data-set are people we have not otherwise encountered in our discussions. The occupations and roles attested more than once in this group are suggestive: four pr¯otok¯om¯etai, three ge¯orgoi, two kt¯etores, two strati¯otai, two suntelestai, two witnesses, and two men not explicitly named kt¯etores but accurately described by Girgis as landowners. (Several of these have multiple attested roles, including all of the pr¯otok¯om¯etai, one of whom is a ge¯orgos, another a suntelest¯es.) The prominence of those with ties to the land is striking: farmers, landowners, and joint tax-collectors all play a role as cutpoints in turning the Aphrodito social network into a unified whole. Those whose occupations or roles are unique in this list are also an interesting category. As one example, take Flavius Artemidoros, an understudied figure attested as a despoinik¼v dioikhtŸv t¦v kÛmhv. The LSJ defines a despoinikos as “belonging to the Imperial household” but has only this example to cite. More accurately, Artemidoros would have been an estate manager (dioik¯et¯es) for the local estates of the empress, the d”spoina. But his appearance in that text is not specifically to fulfill that role; rather, he appears as a witness to a deposition in an apparent dowry dispute between Stephanous daughter of Abraamios and Psates son of Ieremias. Others appear with him as witnesses: a priest named Daueid, two kt¯etores whose names are lost, and the two men who sign for them, Isaak son of Kollouthos and Senouthes son of Apollos. Artemidoros appears as a   



Neither of the two texts in which he is attested appears in the BL. P.Cair.Masp. .. But when? Fournet’s unpublished “Listes des papyrus e´dit´es de l’Aphrodit´e Byzantine” places this text prior to June , when Theodora died, by virtue of the reference to a despoinikos therein. Psates son of Ieremias appears at line  therein; both Psates and Ieremias appear in P.Cair.Masp. ., which dates to October . Maspero had suggested a date c. for  on strength of the appearance therein of Senouthes son of Apollos, who is attested in P.Cair.Masp. . (), one of the documents pertaining to Dioskoros’ trip to Constantinople. But Senouthes is now known from a much wider range of dates: see Chapter  above, n. . In his notes to , Maspero takes Ieremias to be dead by c., because he erroneously assumes him to be dead by , and dates that text to c. only by association with Senouthes. On Senouthes, potential brother of Dioskoros the poet, see Chapter  above, p.  and n. .

4 Quantifying Aphrodito’s social network



Table  Cutpoints in Aphrodito’s corrected prosopography Panel 1. Dioskoros and family Apollos Besarion∗ Dioskoros Fl. Victor  Panel 2. The well-connected Ieremias  – third place by betweenness centrality – priest Victor  – high strength of ties – shepherd Apollos ∗ – third place by degree centrality – shepherd Panel 3. Others Hadrian ∗ – strati¯ot¯es Kallinikos  – kt¯et¯or Isakios  – kt¯et¯or Ioannes  – kt¯et¯or Ioannes  – witness Ioannes  – suntelest¯es Fl. Victor ∗ – ge¯orgos and strati¯ot¯es of the Hermopolite Moors Victor  – shepherd and fieldguard Bottos ∗ – witness; landowner Mousaios  – ge¯orgos and deacon Abraamios ∗ – sumbolaiographos Victor  Sansneus ∗ Senouthes  – ge¯orgos and pr¯otok¯om¯et¯es Apollos ∗ – bo¯ethos and pr¯otok¯om¯et¯es Apollos ∗ – scribe Fl. Artemidoros  – despoinikos dioik¯et¯es Phoibammon  – pr¯otok¯om¯et¯es and oikonomos of the Apa Apollos Phoibammon ∗ – singularios Charisios ∗ – suntelest¯es and pr¯otok¯om¯et¯es Psim(p)anobet ∗ – landowner ∗

Indicates a cutpoint outside the network’s largest component (with Girgis numbers)

dioik¯et¯es again in a land rent receipt he issued to the sons of a Phoibammon on behalf of an unspecified monastery. It is easy to see why someone like Artemidoros might be a cutpoint in Aphrodito’s social network: he served as a dioik¯et¯es for more than one estate with local holdings, he was a witness with two local landowners in a dispute involving Psates, who was a church reader (ˆnag. [n]Ûstou [‰]g©av kaqolik¦v not©nhv –kklhs©av) and son of 

P.Lond. .. The identification is that of Bell, who points out in his note to line  that “the name was hardly so common at this period that we need assume him to be different.”



Social networks in Byzantine Egypt

Ieremias, a priest who himself was one of our top actors by betweenness centrality and had his land re-registered under the name of Dioskoros the poet. Apollos the scribe is the other cutpoint with an occupation otherwise missing from our list. He appears in an extensive account from an unspecified large estate in an entry that reads L. ». g. (ov) c. edr©av d»s(ewv) t¦v kaª ˆgorasq(e©shv) p(ar‡) %pollät(ov) | ìI.w. s. h. f©ou toÓ pi.ssÅg. g. (ou). Girgis thought that this entry, an “account of fruit contributions and sales for Apollos, son of Iosephios, the shoemaker,” indicated that Apollos “Ha comprato del foraggio per il bestiame.” Chedrias is known only from here and two other texts, in the second of which Bell thought it might “= pulse?” But the entry is difficult to read; examination of a digital image of the papyrus confirms the patronymic but cannot confirm pissugg(ou) for shoemaker. If he really was the son of a shoemaker, Apollos was on his way up in life. He could write in Greek, and thus worked drafting rent receipts for landholders such as Flavius Kollouthos, whose agent (logof»rov) is described as being ignorant of letters. Kollouthos was one of the absentee landlords we discussed in Chapter , whose business relations with Apollos, father of the poet Dioskoros, brought that family into closer economic contact with neighboring Phthla. Appearing in the large estate account with Apollos the shoemaker’s son is Ioannes, a kt¯et¯or who also proves to be a cutpoint in the Aphrodito social network. If Apollos were as socially humble as Maspero’s reading takes him, he presents quite a different model from Flavius Artemidoros, who presumably appeared as a cutpoint at least in part by virtue of his higher social status. Here, Apollos highlights a recurring trend, the relatively high social importance someone of even modest social origins can enjoy by virtue of knowing how to write. Admittedly, some of our analytical categories are not mutually exclusive: a witness, for instance, might be a farmer, a soldier, or nearly anything else. Nonetheless, we can draw some general conclusions. Some of Aphrodito’s     

Anagn¯ost¯es: P.Cair.Masp. .. and P.Cair.Masp. ... Ieremias himself: see above, p. . His land: P.Cair.Masp. ..  P.Cair.Masp. . and P.Lond. .. P.Cair.Masp. ...v.–. I do not identify this Apollos son of Iosephios with the Apollos son of Iosephios appearing in P.Michael.  and , a landowner and former bo¯ethos.  See Chapter  above, p. . P.Lond. ... P.Cair.Masp. .. Girgis omitted this text in his entry for Ioannes, but inclusion of it does not change the analysis: both men still appear as cutpoints.

4 Quantifying Aphrodito’s social network



cutpoints are men who filled patently dual roles, e.g. Artemidoros, the manager of two different estates, and Senouthes, the farmer who also served as a village headman. By virtue of their multiplex roles, people like these could be cutpoints in any network at any time, and will always help to lower the average social distance between different portions of a network. The landowners perform a similar role, connecting the average farmer to the circles of the village elite. Take Isakios and Ioannes, the two kt¯etores from Panel  on our list of Aphrodito cutpoints. Both of them appear in the petition to Theodora, a text full of men at the heart of the village elite. Yet both of them also appear in land leases which implicitly tie them to rural peers of more modest social status. Briefly, a word about the presence of Dioskoros and his family members on our list of cutpoints. Dioskoros’ father Apollos may be a cutpoint at least in part for chronological reasons. His dated career spanned over thirty years, and the undated or only generally dated nature of so many of the Aphrodito texts may mask an even wider range of attested dates among those with social connections to Apollos. In other words, cutpoints in a network spanning several generations, as Aphrodito’s does, may be the links joining the village forefathers to its younger generations, the social glue binding the past to the present. If so, using cutpoints to fragment our network may serve, as I suggested earlier in the chapter, as a substitute for rigorous longitudinal modeling. When Aphrodito’s largest component of  names is fragmented along its  cutpoints, the result is the creation of  separate social groups. This fragmentation is less striking than it may appear at first glance, because the majority of these groups () have five members or less, but one of them has . In other words, even after severing the network’s largest component at all of its cutpoints, it is still  per cent intact. If removal of crucial bridges in a network can serve as an adequate substitute for mortality in that network, then Aphrodito’s evidence appears to be fairly robust. In this surviving group of , the most central actors by degree, closeness, and betweenness centrality are the names we are already familiar 



P.Cair.Masp. . and . respectively. P.Cair.Masp. . is a three-year lease from Isakios to the Aurelii Phoibammon son of Victor and Ermauos son of Pamonios. P.Cair.Masp. . is too fragmentary to permit a description of Ioannes’ role in that text. I import Aphrodito’s largest component into Pajek, use it to identify bi-components (Net→Components→Bi-Components), extract the largest bi-component from the network (Hierarchy→Extract Cluster), and re-import the results back into UCINET for easier analysis (Operations→Extract from Network→Cluster).



Social networks in Byzantine Egypt

with: Dioskoros, Enoch, Abraam, Hermaouos, Ioannes, Victor, Kallinikos, Pilatos, and Flavius Theoteknos. To be sure, some network characteristics do change under this fragmentation. Density, degree and closeness centrality measures all increase considerably. But these results do not reflect conditions in a real network: they represent an artificial worst-case scenario, with Aphrodito’s social network cut to its component pieces all at once. What this provides is the high end of a range. If complete chronological data were available, density and network centrality measures in a longitudinal study would probably be in between these results and those initially presented above at pp. –. The high degree of multiplex ties we surveyed in Chapter  suggests the actual results would be closer to those we first presented than to the worst-case model. Aphrodito’s social network shows a considerable degree of redundancy: its members tended to take social action in groups whose members were all familiar to each other. As a result, the effects of the death of any given member, and therefore of chronological change in this network over time, were likely to have been minimal.

conclusion The network analyses presented in this chapter show that Dioskoros and his family were not the only central figures in Aphrodito’s social network, even viewed through the evidence of their own archives. Instead, various measures of network centrality produced figures – Enoch, Abraam, Pilatos, etc. – who were relatively unknown, but in turn connected a wide variety of people performing an array of social functions. The importance of landowning and literacy among the prominent members of Aphrodito’s social networks might raise the suspicion that such prominence is purely artifactual in nature, based on the importance of writing and participating in rent and lease agreements. But the appearance of shepherds in these circles challenges such a conclusion, and instead suggests that network analysis might even be a viable technique for highlighting the importance of the non-elite in a village’s social structure. 



Note that some of these figures are cutpoints themselves. The fragmentation process I describe here does not remove the cutpoints, but simply cuts the network into the smaller components which the cutpoints themselves serve to connect. In the largest surviving block of Aphrodito’s fragmented largest component, the network density is . per cent, and degree and closeness centrality respectively rise to . per cent and . per cent. Betweenness centrality drops to . per cent.

4 Quantifying Aphrodito’s social network



Analyses of potential change over time show a considerable degree of consistency in the Aphrodito social network. Decade by decade crosssections removed from the network as a whole show that relevant network characteristics – distance and degree centrality, for instance – remain largely unchanged over time. Simulation of the deaths of certain key players by the removal of Aphrodito’s network cutpoints also shows the network’s high degree of durability. In sum, network analyses give no reason to think that any remarkable change in the village social structures took place over the course of the sixth century. This study has generally remained agnostic on Zuckerman’s argument that Aphrodito lost its autopragia in the s. However, network analysis suggests that if Zuckerman is right, the change had little impact on the shape of the relations the villagers had with one another. In essence, this argument complements that made regarding Oxyrhynchos in Chapters  and  above. There, I argued that the hierarchical, centralized ties unifying the Oxyrhynchite nome could not be invalidated by the publication and analysis of future evidence, as the ties I described were exclusive in their content. But the nature of the Oxyrhynchite evidence prevents us from seeing the view at the village level, in the same way that the nature of the Aphrodito evidence prevents us from seeing the view at the nome level. The multiplex ties we see throughout the Aphrodito evidence – relative social equals connected by multiple types of social interaction – are themselves unlikely to change no matter what the nature of Aphrodito’s ties to the rest of the nome may have been. If Aphrodito had to bow to the authority of the pagarchs in the s, in essence finding itself in a situation similar to many Oxyrhynchite villages under the Apions, this change would have altered none of Aphrodito’s multiplexity. Aphrodito villagers would still have entered business agreements with their family and friends, would still have acquired land near to their own, and would still have continued with all of the daily business we cannot see in Oxyrhynchos because our urban bias alters the view.

Conclusion

From the start, I have stressed that both traditional prosopography and network analysis create divergent social pictures of Oxyrhynchos and Aphrodito. Aphrodito village society had a powerful face-to-face element, in which social ties developed along pre-existing lines. Analysis of the strongest ties in the Girgis prosopography highlighted the importance of corporate links – particularly among the community of landholders and the community of shepherds – and did not uncover particularly strong links of a subordinating nature, such as patron–client. Analysis of Aphrodito’s petition to the empress Theodora strengthened this impression. The Aphrodito villagers presented themselves by group, naturally taking social action with those to whom they already had corporate ties. Contrary to what Gagos and van Minnen have argued, an unprejudiced reading of the sequence of names in that petition suggests a relatively relaxed approach to whatever social hierarchies existed. The structural characteristics of the Aphrodito network as a whole support this impression. An average distance through the entire network of under three degrees of separation suggests that Aphrodito’s social world was relatively small. We must also remember that distance only measures ties attested in the documentary record. It is for that reason impressive that Aphrodito should appear so small on paper. Certainly, in any village society, most people would be familiar to one another by sight or by reputation. By this measure, Aphrodito’s world would have been even smaller than the documentary evidence suggests. This is one of the reasons why Aphrodito’s network centrality is not particularly high: there can be little hierarchy or centralization in a village in which each person has close connections to nearly everyone else. One virtue of network analysis of Aphrodito is that it provides a quantitative technique with which to see past Dioskoros and his family. Although scholarly interest in his life and work are not misplaced, Aphrodito’s other major characters, those who were its network cutpoints and centrality 

Conclusion



leaders, are in some ways more interesting for the depth they add to our view of Aphrodito society. Our centrality analyses repeatedly brought the literate to the foreground. Some of Aphrodito’s most central social actors were not simply scribes or notaries, but those like Flavius Theoteknos, who, in his own fumbling way, was able to sign for others who could not. To some extent this is an artifact of the documentary evidence. But it is also true that the ability to write increased one’s social connectivity in this society; this point has not been stressed in the past. It is thus curious to find centrally placed villagers in our analyses whose chief shared characteristics are their ties to the shepherd community, who we do not imagine were among the village’s most literate. But a closer look at the shepherds themselves reveals a group well integrated into village life, and occupying a particularly important place in it by virtue of their dual role as pastoralists and fieldguards. Chapter ’s discussion of Aphrodito society emphasized the importance of these multiplex roles, and argued that much of Aphrodito’s social interaction focused on pre-existing social ties. A scribe used once was likely to be used several times more. The economic ties between lessor and lessee would often become ties of lender and borrower as well. The processes of land acquisition would not take place sight-unseen, but through social ties to neighbors, relatives, and previous business acquaintances. This emphasis on face-to-face multiplexity may seem like common sense, but these ties are absent from our prosopographical study of the Oxyrhynchite, and our quantitative network analyses of land acquisition there. Network analysis shows that the Apions acquired property and fiscal responsibility far and wide, without any quantitatively demonstrable regard for the physical proximity of these properties. Indeed, it is difficult to imagine the elite of the Oxyrhynchite great oikoi being familiar enough with their distant estates to be able to form multiplex ties with them at all. The prosopographical connections traced in the first chapter on Oxyrhynchos very explicitly lacked a face-to-face element. Take the example of farmers in a rural settlement who needed an estate middleman to secure supplies from a third party in the city. This sort of social tie, connecting a small village to the city of Oxyrhynchos, is manifestly different from the social ties that connected one villager of Aphrodito to another. Where strong, multiplex ties seem absent in Oxyrhynchos, we find vertical, centralizing ties instead. The large landholders have left documentation of indirect social ties linking Oxyrhynchos to outlying villages only through their estate administrators and other bureaucratic subordinates.



Social networks in Byzantine Egypt

This documentation creates the “spokes in a wheel” effect I have described above, in which the social activities of a wide variety of people across the nome are directed towards the nome capital. The population estimates I have derived in Chapter , rough as they may be, confirm this sense of centralization. If the Apions themselves typically made up some one-third of the payments made by the nome’s great houses, and in turn had something on the order of , Oxyrhynchites under their fiscal jurisdiction, then a considerable plurality of the nome’s population would have been part of a fiscal administrative pyramid culminating in only some half a dozen landlords. I have, to the extent possible, argued that the characteristics of the Aphrodito and Oxyrhynchos social networks followed consistent patterns over time. The centralizing, hierarchical networks created by large estates in Oxyrhynchos are present in evidence from the fifth and early sixth centuries, and remain in full force through the sixth century into the early seventh. The one case where we come closest to tracking the decade-bydecade evolution of such centralizing networks – the house of Apion – shows, if anything, an increasing degree of centralization over time. As the long career of Phoibammon and other examples show, multiplex ties remain at the heart of village-level transactions in Aphrodito throughout the sixth century, both before and after any putative loss of village autopragia. Aphrodito’s network characteristics show the same consistency, appearing in analyses by decade and in tests for the deaths of key players. In short, prosopographical and network analyses show real and persistent differences in the social structures of the two sites. In the introduction, I suggested that this variety left us with two possible ways forward. One approach would accept these results as evidence of regional differences in social structures, and begin exploring other datasets for further evidence of regional differentiation in Byzantine Egypt. A second, homogenizing approach would question what is missing from our evidence in Oxyrhynchos and Aphrodito, and ask whether that missing evidence might have made the two places seem more similar. (Or, as Zuckerman might have it, whether Aphrodito became more like Oxyrhynchos over time.) This approach would focus on the nature of the evidence, and suggest that evidence from an Oxyrhynchite village archive or from Aphrodito’s equivalent of the house of Apion would easily erase any apparent regional differentiation. At first glance, the regionalist thesis seems the more attractive of the two options. We know that different places in Byzantine Egypt had different cultural habits: various villages show an intriguing range of naming

Conclusion



practices, to give but one example. Social and economic structures may vary from place to place as well. Certainly, villages in late antique Egypt show disparities in Gini indices measuring the equality of their land distribution. Temseu Skordon, a Hermopolite village, had a Gini index of ., considerably higher than Aphrodito’s at .. By this measure, at least, Aphrodito is demonstrably more egalitarian in its wealth distribution than other villages of the era. This analysis is intriguing, complementing as it does our conclusion that Aphrodito’s social networks were themselves not inherently vertical or hierarchical. It becomes easy to imagine a full range of villages up and down the Nile valley, some with relatively diffuse social networks and even wealth distributions, and others with strict social and economic hierarchies that previous generations would have called feudal. But as I have suggested, regionalist arguments are always vulnerable to attacks focusing on the haphazard nature of the surviving evidence. A homogenizing approach can easily retort that Aphrodito and Oxyrhynchos appear different to us only because we have evidence from a family basement in one case and a city dump in another, or evidence detailing a village in one case and covering a nome-wide scale in another. This objection to the regionalist thesis can never be overcome, except by discovering bodies of evidence sharing unlikely similarities. Other than that of Dioskoros, we have no other large archive of a single landowning family operative at the level of a large village in sixth-century Egypt. The archive of Theodoros may one day provide a useful comparandum from nearby Petra, but not on the same scale or with the same level of detail. Indeed, proponents of the homogenizing approach do not seem likely to embrace evidence of Aphrodito-like villages elsewhere; they are, I suspect, more actively looking for evidence of Apion-like mega-families hidden in every nome. This, at any rate, is the impression we get from Zuckerman’s work on Aphrodito, in which the author argues that the village’s distinctive independence is ultimately destroyed, its landscape dominated by a shadowy super-landowner who remains almost completely unknown. But would a homogenizing model emerge to erase the differences between Oxyrhynchos and Aphrodito if more evidence were found at both the nome and the village levels in each place? At the end of both  

 

Ruffini . See Bagnall forthcoming for these figures. That from Aphrodito combines data from the village register and its cadastre, the former detailing land registered as k¯om¯etika, village property, and the latter detailing Aphrodito land registered as astika, urban property. See p.  below. See also Sarris , particularly his Chapter  on Aphrodito, with Ruffini forthcoming c.



Social networks in Byzantine Egypt

prosopographical chapters on Oxyrhynchos and Aphrodito, I argued that the evidence we have for both places prohibits by its very nature contradictory evidence which may now be lost. Time and again, the evidence for Oxyrhynchos shows indirect connections between that nome’s elite, connections formed through estate managers, scribes, priests, tenant farmers and a whole host of characters scattered through dozens of settlements. We have evidence of these ties, centering as they did on the metropolis, precisely because they did not exist at the village level: they were ties that needed the mediation and institutional capacity of the city and its landowners in order to be fully realized. The converse is true of the evidence from Aphrodito. Rent agreements, loans, property divisions and the village elite’s entire array of business transactions could sustain themselves on face-to-face strong ties precisely because activities at this low level of scale would have been irrelevant to any large estates active in the region. The result is a logical dilemma: the regionalizing thesis is vulnerable to a critique about variations in the nature of the evidence, while a homogenizing thesis awaits lost evidence that cannot erase the distinctions I have demonstrated. The only solution to this dilemma, I believe, is a model in which Oxyrhynchos and Aphrodito do not contradict each other or await some sort of curative that will erase the differences between them, but merely coexist on different levels of scale. This is, in essence, a universalizing thesis, in which the differences between the two sites are both real and complementary. As Keenan has recently noted, “Implicit in all this is a sense of a massive loss of documentation.” This missing evidence is the dark matter of Byzantine Egypt, the evidence we cannot consult directly and the contours of which we are only now beginning to measure. Egypt’s dark matter is different from site to site. In Aphrodito, the dark matter is evidence for land in the village owned by Ioulianos and other regional elites, land found in neither Aphrodito’s register nor its cadastre. This is the evidence for a vertical centralizing hierarchy connecting Aphrodito to Antaiopolis. In Oxyrhynchos, the dark matter is villages, in their entirety, which do not show up in our census of the Apionic estates, or, for that matter, in documentation of any great estate at all. This is the evidence at the village level, of strong, multiplex ties among neighbors, family and friends, ties which must have existed, but have no place in evidence from Oxyrhynchos itself.  

Keenan b, . Dark matter: I am indebted to Roger Bagnall for discussion of what follows, and owe the use of the phrase in this context to him.

Conclusion



Consider how much evidence is missing even from Aphrodito, where our documents give such an illusion of completion. Either our impression of the size of Aphrodito is badly wrong, or thousands of people have gone missing. In the Aphrodito register, only a few hundred contributors paid into the village’s k¯om¯etika, arable land registered in the village account. These landowners would account for families amounting to barely , people. Girgis lists under , villagers, and a revised prosopography is not likely to produce a number more than  per cent higher, spanning a century or more. Even accounting for what must have been a large tenant population (on the land registered as astika, on the city account, and perhaps all the people on the estates of Ioulianos) and the various guilds from shepherds to boat-builders, it is hard to reconcile these figures with the five-digit population estimates for Aphrodito that some scholars have put forth. Even Zuckerman’s modest population estimate of , – based in part on the carrying capacity of Aphrodito’s agricultural land – leaves a tremendous number of people undocumented, and a tremendous amount of documentation lost. We face the problem of missing evidence in Oxyrhynchos as well. We have no Oxyrhynchite equivalents for the Aphrodito register and cadastre. We thus have no way of knowing how the fiscal share of the Apionic house might look in the village register of (say) Spania. The large payment attributed to Ioulianos, and taken by Zuckerman to show that he owned vast amounts of Aphrodito land, may be no more than his fiscal share in the village’s post-autopract era. We must assume that even in the regions for which the Oxyrhynchite great estates held fiscal shares, land registers and cadastres still recorded the names of individual small landowners. Thus, a massive presence such as Ioulianos might go unnoticed in our Aphrodito documentation, even if we had a cadastre or register contemporary to Ioulianos. The fiscal share of a great estate and a relatively egalitarian Gini index might coexist in the same place, the former invisible in the same body of evidence that produces the latter. Apionic Prostasia B, discussed in Chapter  above, includes specific reference to “your property in villages” Episemou and Adaiou as part of the jurisdictions of the prono¯et¯es Serenos. This is precisely how Aphrodito would look from the vantage point of the great estate of Ioulianos. We would have no way of knowing whether those villages had any unusual characteristics. In fact, there would be no way to tell whether Episemou  

See n.  above. A point I owe to Roger Bagnall; see his discussion of Ioulianos in Bagnall forthcoming.



Social networks in Byzantine Egypt

and Adaiou were otherwise autopract or not. With so many hundreds of villages from Byzantine Egypt completely undocumented, how can we assume autopract status to be exceptional? Certainly, we do see villages that are specifically not autopract: Takona is described as being under the Apionic pagarchy. But such explicit references are themselves quite rare. Lacking certain evidence for Aphrodito exceptionalism, is it so implausible to imagine the same autopract phenomenon in the Oxyrhynchite? On surviving evidence, the Apions probably did not have a major fiscal role in more than a quarter of the forty-eight or more villages attested in the Byzantine Oxyrhynchite. The great estates as a whole may not have been able to shoulder the fiscal burden for all of these places either. Even those villages attested in documentation from great estates are sometimes hard to categorize. Spania, one of the best-documented villages of the Byzantine period, appears to emerge fully formed in the fifth century. The Apions were present there – it was part of their Prono¯esia L – but it does not appear to have been completely subject to the Apionic pagarchy in the way that Takona was. And even here, there are subtle hints of an Aphrodito hidden below the surface. When Makarios, meiz¯on of Spania, wrote a letter to his “master,” perhaps a member of the Apion family, his tone suggests that Spania fell under the jurisdiction of the master (despot¯es) in some unspecified way. The letter itself explains that the despot¯es has been reconciled with the people of Kosmou. But reconciled over what? The vague reference to “their goods” (perhaps farm implements or livestock: ta skeu¯e aut¯on) suggests that some theft was at stake. It is easy to imagine this sort of thing in the context of a border dispute between two villages, a less amicable version perhaps of the ambiguous border between Aphrodito and Thmonachthe. The  landowners from Spania in a different text, obliged to provide restitution for a theft from Kuriakos, another meiz¯on, are also reminiscent of Aphrodito. To begin with, it is a large group of landowners, unexpected 

   

P.Oxy. . nn. – on the term (personal name) ˆp¼ kÛm(hv) (village name) pagarcoum(”nhv) | par‡ t¦v Ëm]et”ra Ëperfue©av: Gonis points out that the usage appears only for k¯omai. They “were normally under the control of the pagarch, but it appears that their fiscal administration could be exercised by great landowners, whose authority was similar to that of a pagarch.” A search of the Duke Databank reveals only six instances of pagarchoumen¯es, three from the sixth century and three from the seventh. Payments from Spania: P.Oxy. ... Prono¯esia L: P.Oxy. . with Mazza’s pronoetic list above, p. . Takona under the Apions: P.Oxy. ., with p.  above. I am not the first to compare Spania to Aphrodito: see Rowlandson forthcoming. P.Oxy. .. Arguments for an Apionic context: see Chapter  above, n. . Elsewhere in the sixth century (P.Oxy. .), Spania is described as tou lamprotatou kuriou I¯oannou kathol(ikou?).

Conclusion



in a nome we have been conditioned to imagine as populated only by great estates and enapographoi ge¯orgoi. The list includes figures as important as two politeuomenoi; figures as predictable as another current and a former meiz¯on; and figures as lowly as oil-makers and house-builders. The group as a whole must have had considerable financial resources of their own. The  solidi for which they are responsible are presumably distributed between them in amounts of one solidus or less because such sums would have made little impact. We can imagine Spania’s Dioskoros at work on this account, listing every item stolen from Kuriakos, much the way Aphrodito’s Dioskoros had to write an account of livestock stolen from him, presumably in hopes of receiving similar restitution. In the passage quoted at the start of this work, James Keenan remarked that the older “feudal model is also by necessity for Egypt an ‘Oxyrhynchus model,’” and in the face of the Aphrodito evidence does not suffice to represent Egypt as a whole. It is too simple to argue whether Byzantine Egypt looked more like Oxyrhynchos or Aphrodito, as the homogenizing thesis might have it. I rather suspect it looked quite a bit like Aphrodito in Oxyrhynchos, in which village networks formed from strong horizontal ties connected to nome-wide networks through centralizing vertical ties. Put another way, I would argue that Aphrodito probably looks like what we might see in an Oxyrhynchite village below the level of the evidence of the urban archives, and the Oxyrhynchite nome probably looks like what we might see in Antaiopolite urban archives. I have stressed throughout this work that our pictures of Oxyrhynchos and Aphrodito are not mutually exclusive. If tomorrow we discovered an archive that made a large and previously under-documented Oxyrhynchite village – Spania again, for example – suddenly look much like Aphrodito, the shape of the already published Oxyrhynchite evidence would be unchanged. The vertical hierarchies we have mapped would still be an important part of the nome superstructure. And if tomorrow we discovered the Antaiopolite garbage dump and were then able to connect Aphrodito to a nome-wide set of hierarchical vertical ties very much like those of the Oxyrhynchite, none of the village-level strong multiplex ties would be changed by this discovery.   

P.Oxy. .. See p.  above, with the remarks on onomastics in Ruffini . Keenan , . Although the Antaiopolite’s nome-wide network, smaller than that of the Oxyrhynchite, was likely to have been somewhat less hierarchical in nature, an observation I owe to Roger Bagnall. Differences of scale would have applied in comparisons between individual nomes as well.



Social networks in Byzantine Egypt

Let us paint a picture from pure fantasy. For sake of argument, let us choose an Oxyrhynchite village about which we know nearly nothing in this period, Iem¯e, for example. Our ignorance about affairs there will let us remain purely in the realm of the hypothetical. We might imagine a family of landholders of moderate means whose most prominent members – Ioulios and Ioannes, perhaps – provide Iem¯e’s village headmen and other officials in consecutive generations. We might imagine that the brightest among them have received a smattering of rhetorical and philosophical education. In this hypothetical scenario, we speculate that some of the land they lease in Iem¯e is just on the border of land registered to their neighboring village, land which they in turn come to lease through extended collaboration with their neighbors just across village lines. After too many parcel subdivisions and recombinations, after too many complicated mortgage agreements and bad loans made on the security of land on that village border, after too many land surveys redrawing the boundaries of uncultivated plots, a prono¯et¯es managing land in the neighboring village on behalf of Flavia Anastasia may soon realize that land he considers part of his jurisdiction has in fact been registered into Iem¯e’s cadastre since the last major survey. Neither Ioulios nor Ioannes is very interested in his efforts to collect on that land, whether his employer holds the pagarchy for the village next door or not. They might be perfectly happy to go all the way to Constantinople to sort out the legal mess over land they had come to consider their own, or over land on which collections help to pay their salaries. Whether Flavia Anastasia is happy to buy them off to make the affair disappear, or instead sees fit to have one of her subordinates run livestock over their land in return will no doubt depend on the strength of their social networks, and on the social capital they bring to the table. If this hypothetical has any plausibility, it rests on elements we have seen elsewhere, in both Oxyrhynchos and Aphrodito. Here, we have a family much like that of Dioskoros, a land dispute similar to the case of Diogenes 



Iem¯e or Gem¯e is well attested in the Roman period (Pruneti s.n.) and appears three times in the Byzantine period, in P.Ryl. . (fifth century), P.Oxy. . (s) and . (late sixth century). The first text is a short grain account, the second a list of contributions to the public bath discussed passim in Chapters  and . For the third, see n.  below. The absence of Iem¯e from any large estate accounts and its presence in the form of payments “through those from Iem¯e” (d(i‡) tän ˆp¼ E«”mh) side by side with payments by large estates suggests it as an appropriate choice here. And the appearance of an enapographos ge¯orgos accountable to Flavia Anastasia and living in Iem¯e in P.Oxy. . (late sixth century) would in this case suggest to us how our hypothetical scenario came to an end, with the growing influence of a large landholder over a once (hypothetically) independent village.

Conclusion



and re-registration of land similar to that on the border between Aphrodito and Thmonachthe. We have at the same time the strong, face-to-face social ties through which small village landowners grow their business influence, and the remote centralized hierarchies of larger landowners looking out from the city onto their rural estates. This is the model of Aphrodito in Oxyrhynchos. Nothing in this hypothetical case has any distinctive regional characteristic that makes its portability to other regions inherently implausible. It merely combines evidence from two different levels of scale, and begins to see them as part of one unified picture. final thoughts Generally, this study presents a methodology capable of sparking a wideranging dialogue between historians of the ancient world and those of any time or place. Despite the growth of social network analysis in recent decades, there remain relatively few historical communities analyzed using these techniques and available as comparanda. Given the attention we have paid to city and village elites in this study, the most relevant example is fifteenth-century Florence. Network analysts accept John Padgett’s data-set of Florentine families from this period as a standard comparandum. This data-set has been used to study the wealth of Florentine families, the extent of business and marriage ties between those families, and how those factors affected the struggle for control over Florence itself. Oxyrhynchos and Aphrodito can certainly join the short list of sites that can be analyzed in analogous terms. A close inspection of the epigraphic and papyrological evidence from other sites may open comparable lines of inquiry elsewhere, and permit real comparisons between networks of the ancient, medieval and modern worlds. Similarly, this work shows the applicability of network analysis to other fields of microhistory. Surely the most detailed study of a single community anywhere remains Emmanuel Le Roy Ladurie’s work on Montaillou, a village in southern France in the early s. Students of Byzantine Egypt see in Aphrodito something analogous to Montaillou, where the immediacy of the documents gives considerable fuel for microhistory. But 

 

See particularly Wasserman and Faust , – and also Erickson , a review of Wasserman and Faust, discussing both the Florentine example and Peter Bearman’s  work on the Norfolk elite in Relations into Rhetorics: Local Elite Social Structure in Norfolk, England, 1540–1640, New Brunswick, NJ. Le Roy Ladurie . See Keenan a,  and n. , relaying remarks by P. J. Parsons perhaps coming ultimately from R. Duncan-Jones, and since cited approvingly by Gagos and van Minnen , .



Social networks in Byzantine Egypt

the similarity between the two sites goes even deeper: rarely are any two sources of data so susceptible to network analysis. When Le Roy Ladurie writes that “all roads . . . sooner or later led to Clergue the priest,” that the Belot clan was central in the spread of Cathar heresy in Montaillou, or that Pierre Maury – the work’s shepherd protagonist – “stood at the intersection of a whole news network stretching from one mountain pass to another,” he is making network analytic claims without knowing. A return to Jacques Fournier’s Inquisition Register with the methods outlined in this work could put Le Roy Ladurie’s claims to quantitative test and again add to the list of comparanda available for historical network analysis. More specifically, this study has considerable implications for the social history of the late antique and early Byzantine worlds. Evidence of regionally varying social structures within Egypt neutralizes the utility of the Egyptian evidence. If Aphrodito’s small landholders were atypical even within Egypt, they can more easily be ignored on a wider stage. The same is true for Oxyrhynchite large estates. The model I have proposed is one with potential to save the Egyptian evidence for use on that wider stage. If the evidence from both sites can exist at different levels of scale in a single model, then there is much less reason to imagine Egyptian particularism as an explanation for what we see at either site. Since we have seen that the Aphrodito model and the Oxyrhynchite model are not mutually exclusive, we can more readily expect to find both in abundance in the ancient world. The network analytical techniques we have developed herein can help us measure the differences between these types of sites, and provide quantitative comparanda for the results we should expect at each level of scale. The picture that emerges from the Egyptian papyri of the Byzantine period probably had thousands of analogues around the Mediterranean world. To take but one possible topic of analysis, a comparison of texts from Aphrodito to texts from Petra, has led one scholar to suggest that private, “out of court” dispute settlements had an increasing “pan-Mediterranean appeal” in our period. A closer look at the Petra texts is likely to uncover further cultural similarities between Egypt and its neighbors in this period. Extending the “Aphrodito in Oxyrhynchos” model beyond the borders of Egypt, we may look to Petra as a potential Oxyrhynchos.  

 Gagos , . Le Roy Ladurie , , – and . This comparative approach to Aphrodito, Petra, Nessana et al. is logical, but has so far been the domain chiefly of scholars working on the Petra papyri: see Gagos  and Koenen  in particular.

Conclusion



The world of the Petra papyri is a world of wealthy landowners, Flavii and politeuomenoi. Flavius Theodoros, one of the central figures in the Petra papyri, owned land on his maternal side in the village of Serila, and property on his paternal side in other cities nearby. The dim outline of a regional network can emerge with the recovery of pieces such as these, a network linking metropoleis both to each other and to their peripheral villages. The excavations at Nessana, a small late antique frontier village in the Negev, uncovered sixth- and seventh-century papyri in many ways strikingly reminiscent of those from Aphrodito. A division of property between two soldiers ( July ), the Flavii Ioannes and Zonainos, mentions property they divided that borders other property already belonging to each of them. The presence of Ioannes’ sister in the text raises the natural suspicion that the two soldiers were brothers-in-law. They were finally both members of the same military unit, based at Nessana. Here we have a classic case of multiplex social ties, two men whose bonds included military service, joint landholding, and possible marriage ties. This pattern of land acquisition – attention to adjoining plots acquired through pre-existing social ties among apparent equals – is the same we see in Aphrodito. Evidence of close social ties between propertied neighbors is elsewhere in the Nessana papyri as well. A notice of land transfer ( November ) involves two brothers on one hand, the Flavii Abraam and Abou-Zonainos, soldiers in the camp at Nessana, and Thomas, another soldier in the same camp, on the other. The land the brothers transferred bordered on the west property belonging to their paternal cousin. Another sixth-century text details the division of property between three brothers of the extensive holdings belonging to their father Eulais. The text lists a series of holdings all divided in three, with their respective borders delineated in each case. In one case, the share assigned to one brother, Biktor, borders the east side of unrelated property already in his possession through purchase from an unrelated third party. We can assume that Biktor had made that earlier purchase precisely with an eye to that property’s proximity to his family’s holdings. In much the way that Antaiopolis remains relatively peripheral to the daily business affairs of villagers at Aphrodito, Nessana’s nearest city,    

See the remarks of Lehtinen . Lehtinen ,  for the logical conclusion that “his family probably did not originate in Petra.” For Serila, see Daniel , .  See P.Ness. .. with introduction p. .  P.Ness. .. P.Ness. ..  P.Ness. ..–. P.Ness. ..



Social networks in Byzantine Egypt

Elusa, impinges on the villagers only rarely, as in the summons to Elusa issued to a soldier in . Additional reflection about similarities between Egypt and the rest of the eastern Mediterranean cannot be far behind. We can apply network analysis at any level of scale, from village to empire, and the analyses can quantify the differences in social structure we see at each level of scale. We can apply network analysis to any evidence that creates connections, between people and places, people and things, or even people and ideas. Here, we have used the connections papyrus records, but the technique would apply equally well to Greek civic inscriptions, Roman brick-making stamps, Egyptian amphora labels or Byzantine seals. The result is a considerable step forward in ancient social history. We are no longer forced to rely on impressionistic amalgamations of evidence from multiple sites at conflicting levels of scale, but can instead develop quantitative comparanda for each type of site at each stage of development. Some scholars object to these methods on the grounds that they measure only the surviving evidence, or some smaller section thereof. Ultimately, this objection does not get us very far. To begin with, it applies to traditional prosopography and qualitative analyses no less than quantitative ones. Further, it does no good to wait for the discovery of some hypothetical body of missing evidence. The surviving evidence has shape. Its chronology has length, its geography has breadth, its characters map trajectories through that time and space. The connections between these characters also have shapes, hierarchical or diffuse, sparse or dense. Measuring these shapes improves our understanding of the evidence, whether it is incomplete or not. Without these attempts, many of the features of our evidence will be relegated to silence.  

 For Roman brick-making stamps, see p.  above. P.Ness. .. This is for example the objection Edward Watts raises (, p.  n. ) to my attempts at network analysis of late antique philosophers in Ruffini .

Stemmata

Strategios I

Isis = Apion I

Heraklidas

Strategios II = Leontia Apion II

Strategios III

Theognosia

Praiecta = Strategios Paneuphemos

Apion III = Eusebia

Strategios IV

Figure  Stemma of the Apion family



Georgios



Social networks in Byzantine egypt Psimanobet

Dioskoros

Besarion

Apollos Dioskoros of Aphrodito = Sophia

Senouthes

Victor

Anonymous Daughter = Megas

Petros

Figure  Abridged stemma of the Dioskoros family

Anastasia = Phoibammon

Bibliography

Alexander, M., and J. A. Danowski. () “Analysis of an ancient network.” Social Networks : –. Allen, Pauline. () “The ‘Justinianic’ plague.” Byzantion : –. Alston, Richard. () The City in Roman and Byzantine Egypt. London. Alston, Richard, and Robert D. Alston. () “Urbanism and the urban community in Roman Egypt.” JEA : –. Antonini, L. () “Le chiese cristiane nell’Egitto dal IV al IX secolo secondo i documenti dei papyri greci.” Aegyptus : –. Atkinson, John E. () “The plague of : not the birth of the clinic.” Acta Classica : –. Azzarello, G. () “P.Oxy. XVI  e la nascita della domus gloriosa degli Apioni.” ZPE : –. Bachrach, B. () “Was there feudalism in Byzantine Egypt?” Journal of the American Research Center in Egypt : –. Bagnall, Roger. () “Official and private violence in Roman Egypt.” BASP : –. () “Landholding in late Roman Egypt: the distribution of wealth.” JRS : –. () Egypt in Late Antiquity. Princeton. () Review: Traianos Gagos and Peter van Minnen, Settling a Dispute: Towards a Legal Anthropology of Late Antique Egypt. BMCR (http://ccat.sas.upenn.edu/bmcr). () “Monks and property.” GRBS (): –. () “Village and city: geographies of power in Byzantine Egypt,” in Les Villages dans l’empire byzantin, ed. Cecile Morrisson and Jean-Pierre Sodini. Paris: –. (Forthcoming) “Village landholding at Aphrodito in comparative perspective,” ´ in Les Actes du colloque Dioscore (Etudes d’arch´eologie et d’histoire ancienne). Bagnall, Roger, and Bruce W. Frier. () The Demography of Roman Egypt. Cambridge. Bagnall, Roger, and K. Worp. () “Chronological notes on Byzantine documents, II.” BASP (): –. (a) “Chronological notes on Byzantine documents, IV.” BASP (): –. 



Bibliography

(b) “Papyrus documentation in Egypt from Constantine to Justinian,” in Miscellanea papyrologica, ed. R. Pintaudi. Florence: –. () “Dating the Coptic legal documents from Aphrodite.” ZPE : –. Baldwin, Barry. () “Dioscorus of Aphrodito and the circus factions.” ZPE : –. () “Dioscorus of Aphrodito: the worst poet of antiquity?” in Atti del XVII congresso internazionale di papirologia. Naples. Banaji, Jairus. () “Rural Communities in the Late Empire.” DPhil thesis. Oxford. () “Agrarian history and the labour organization of Byzantine large estates,” in Agriculture in Egypt from the Pharaonic to Modern Times, ed. A. K. Bowman and E. Rogan. Oxford. () Agrarian Change in Late Antiquity: Gold, Labour, and Aristocratic Dominance. Oxford. Barab´asi, Albert-Laszlo. () Linked: The New Science of Networks. Cambridge, Mass. Barison, P. () “Ricerche sui monasteri dell’Egitto bizantino ed arabo secondo i documenti dei papiri greci.” Aegyptus : –. Barnes, J. A. () “Networks and political process,” in Social Networks in Urban Situations, ed. J. C. Mitchell. Manchester: –. Beaucamp, J. () “Apion et Praejecta: hypoth`eses anciennes et nouvelles donn´ees.” REByz : –. Bell, H. I. () “The Aphrodito papyri.” JHS : –. () Review: Jean Maspero’s Catalogue g´en´eral Tome iii. JEA : –. () “The Byzantine servile state in Egypt.” JEA : –. () “The episcopalis audientia in Byzantine Egypt.” Byzantion : –. () “Proposals for a social history of Graeco-Roman Egypt,” in Atti del IV congresso internazionale di papirologia. Milan. () “An Egyptian village in the age of Justinian.” JHS : –. () Egypt from Alexander the Great to the Arab Conquest: A Study in the Diffusion and Decay of Hellenism. Gregynog Lectures, . Oxford. Bell, H. I., and W. E. Crum. () “A Greek–Coptic glossary.” Aegyptus : –. ´ Bogaert, Raymond. () “La Banque en Egypte byzantine.” ZPE : –. Boissevain, Jeremy. () Friends of Friends: Networks, Manipulators and Coalitions. New York, London. Boissevain, Jeremy, and J. Clyde Mitchell (eds.). () Network Analysis: Studies in Human Interaction. The Hague. Bonneau, D. () “L’Administration de l’irrigation dans les grands domaines en ´ Egypte au VIe si`ecle de notre e`re,” in Proceedings of the XIIth International Congress of Papyrology. Toronto: –. Borgatti, S. P., M. G. Everett, and L. C. Freeman. () Ucinet for Windows: Software for Social Network Analysis. Harvard: Analytic Technologies. Bowman, Alan K. () Town Councils of Roman Egypt. Toronto. () “Landholding in the Hermopolite nome in the fourth century ad,” JRS : –.

Bibliography



Bowman, Alan K., Revel Coles, Dirk Obbink, and Peter Parsons (eds.). (Forthcoming) Oxyrhynchus: A City and its Texts. London. Buchanan, Mark. () Nexus: Small Worlds and the Groundbreaking Science of Networks. New York. Calderini, Aristide. () “Richerche topografiche sopra il nome Ossirinchite.” Aegyptus : –. () “Censimento topografico delle banche dell’Egitto greco-romano.” Aegyptus : –. () “Bouleutika.” Aegyptus : –. Cameron, Alan. () “Polyonomy in the late Roman aristocracy.” JRS : – . Carri´e, J.-M. () “Les Finances militaires et le fait mon´etaire dans l’empire romain tardif,” in Les “D´evaluations” a` Rome: ´epoque r´epublicaine et imp´eriale, ed. G. Vallet. Paris: –. () “Le Colonat du bas empire: un mythe historiographique?” Opus : –. () “Un roman des origines: les g´en´ealogies du colonat du bas empire.” Opus : –. ´ () “Figures du ‘colonat’ dans les papyrus d’Egypte: lexique, contextes.” Atti del XVII congresso internazionale di papirologia. Naples: –. ´ ´ () “Economie et soci´et´e de l’Egypte romano-byzantine.” AntTard : –. () “Antiquit´e tardive et ‘d´emocratisation de la culture.’” AntTard : –. () “The contribution of papyri,” in Fifty Years of Prosopography: The Later Roman Empire, Byzantium and Beyond, ed. Averil Cameron. Oxford: –. Cherf, W. J. () “What’s in a name? The Gerontii of the later Roman Empire.” ZPE : –. Chow, John K. () Patronage and Power: A Study of Social Networks in Corinth. Sheffield. Clark, Elizabeth A. (a) “Elite networks and heresy accusations: towards a social description of the Origenist controversy.” Semeia : –. (b) The Origenist Controversy: The Cultural Construction of an Early Christian Debate. Princeton. Claus, A. ()


Bibliography

Diethart, Johannes, D. Feissel, and J. Gascou. () “Les Prˆotokolla des papyrus byzantins du Ve au VIIe si`ecle.” Tyche : –. Dijkstra, J. H. F. () “A world full of the word: the biblical learning of Dioscorus,” in MacDonald et al., eds., –. () “A cult of Isis at Philae after Justinian? Reconsidering P.Cair.Masp. I .” ZPE : –. Dols, Michael W. () The Black Death in the Middle East. Princeton. Drew-Bear, Marie. () Le Nome Hermopolite: toponymes et sites. Missoula, Mont. Eger, O. () “Papyri der Giessener Univers¨atsbibliotek.” AFP : . Elm, Susanna. () “Virgins of God”: The Making of Asceticism in Late Antiquity. Oxford. Emmett, A. M. () “Female ascetics in the Greek papyri.” JOB : –. Erickson, Bonnie H. () “Social networks and history: a review essay.” Historical Methods : –. Falivene, Maria Rosaria. () The Herakleopolite Nome. Atlanta, Ga. Fararo, T. J. and M. Sunshine. () A Study of a Biased Friendship Net. Syracuse, N.Y. Fikhman, I. F. () “On the structure of the Egyptian large estate in the sixth century,” in Proceedings of the XIIth International Congress of Papyrology. Toronto: –. () “Kurienland in Oxyrhynchos?” in Festschrift zum 150j¨ahrigen Bestehen des Berliner Aegyptischen Museums. Berlin: –. (a) “Large landowners and municipal organisation in late Roman Egypt (based on the data in the Oxyrhynchus papyri).” VDI : – (in Russian, with English synopsis). (b) “Quelques donn´ees sur la gen`ese de la grande propri´et´e fonci`ere a` Oxyrhynchus,” in Le Monde grec: hommages a` Claire Pr´eaux, ed. J. Bingen et al. Brussels: –. () Oxyrhynchus, City of the Papyri. (In Russian, introduction in French) Moscow. ´ () “Esclaves et colons en Egypte byzantine.” AnPap : –. () “Sur quelques aspects socio-´economiques de l’activit´e des corporations professionelles.” ZPE : –. Fonseca, James. () The Urban Rank-Size Hierarchy: A Mathematical Interpretation. Published online at http://www.zanesville.ohiou.edu/geography/ urbanrank/. Foss, C. () “The sellarioi and other officers of Persian Egypt.” ZPE : –. Fournet, Jean-Luc. () “Appendice sur le nom d’%frod©thv kÛmh.” REG : –. () “A` propos de SB XIV  ou quand la po´esie rencontre le document.” BIFAO : –. () “Du nouveau dans la biblioth`eque de Dioscore d’Aphrodit´e,” in Akten des 21. Internationalen Papyrologenkongresses. Berlin. (a) “Un nouvel e´pithalame de Dioscore d’Aphrodit´e adress´e a` un gouverneur civil de Th´eba¨ıde.” AntTard : –.

Bibliography



´ () Hell´enisme dans l’Egypte du VIe si`ecle: la biblioth`eque et l’oeuvre de Dioscore d’Aphrodite. Cairo. () “Le Syst`eme des intermediaries dans les rec¸us fiscaux byzantins et ses implications chronologiques sur le dossier de Dioscore d’Aphrodit´e.” AFP (): –. () “Du nouveau dans les archives de Dioscore d’Aphrodit´e,” in Atti del XXII congresso internazionale di papirologia. Florence: –. () “Between literary tradition and cultural change: the poetic and documentary production of Dioscorus of Aphrodite,” in MacDonald et al., eds., –. () “Entre document et litt´erature: la p´etition dans l’antiquit´e tardive,” in La P´etition a` Byzance, ed. D. Feissel and J. Gascou. Paris: –. (Forthcoming) “Archive ou archives de Dioscore?” in Les Actes du colloque ´ Dioscore (Etudes d’arch´eologie et d’histoire ancienne). (Unpublished) “Liste des papyrus e´dit´es de l’Aphrodit´e byzantine.” Distributed at the colloquium “Les Archives de Dioscore d’Aphrodit´e: cent ans apr`es leur d´ecouverte,” Strasbourg, December . Fournet, Jean-Luc, and Jean Gascou. () “Moines pˆachomiens et batelleri.” Alexandrie m´edi´evale : –. Frankfurter, David. () Review: Arietta Papaconstantinou, Le Culte des saints ´ en Egypte des Byzantines aux Abbasides. BASP : –. Franzosi, Roberto. () “A sociologist meets history. Critical reflections upon practice.” Journal of Historical Sociology : –. Franzosi, Roberto, and John W. Mohr. () “New directions in formalization and historical analysis.” Theory and Society : –. Freeman, L. C., () “Visualizing social networks.” Journal of Social Structure , online at http://www.cmu.edu/joss/index.html. Freeman, L. C., and C. R. Thompson. () “Estimating acquaintanceship volume,” in The Small World, ed. M. Kochen. Norwood, N.J.: –. Gagliano, Calogera Liliana. () “A proposito di –n m†ndraƒ in P.Oxy. .” Aegyptus : –. Gagos, Traianos. () “The Aphrodito murder case (P.Mich. XIII  and ) and the ghost-epithet kakosiwm”nov.” ZPE : . () “A new papyrus from the family archive of Dioskoros of Aphrodito and the mysterious Phoibammon (P.Mich. inv. ).” APF (): –. () “Negotiating money and space in sixth century Petra,” in Atti del XXII congresso internazionale di papirologia. Florence: –. Gagos, Traianos, and Peter van Minnen. () “Documenting the rural economy of Byzantine Egypt: three papyri from Alabastrine.” JRA : –. () Settling a Dispute: Towards a Legal Anthropology of Late Antique Egypt. Ann Arbor. Gascou, Jean. (a) “Notes critiques sur quelques papyrus des Ve et VIe si`ecles.” CE : –. ´ (b) “La D´etention coll´egiale de l’autorit´e pagarchique dans l’Egypte byzantine.” Byzantion : –.



Bibliography

(a) “P.Fouad 87: les monast`eres pachˆomiens et l’´etat byzantin” BIFAO : –. ´ (b) “Les Institutions de l’hippodrome en Egypte byzantine.” BIFAO : –. () Review: P. J. Sijpesteijn, The Aphrodite Papyri in the University of Michigan Papyrus Collection (P. Mich. XIII). CE : –. (a) Review: R. Pintaudi, I papyri vaticani greci di Aphrodito. Aegyptus : –. (b) “Documents grecs relatifs au monast`ere d’Abba Apollˆos.” Anagennesis (): –. ´ () “Les Grands Domaines, la cit´e et l’´etat en Egypte byzantine (Recherches d’histoire agraire, fiscale et administrative).” Travaux et M´emoires : –. () “La Table budg´etaire d’Antaeopolis (P.Freer 08.45 c–d),” in Hommes et richesses dans l’empire byzantin, vol. : IVe–VIIe si`ecle. Paris. () Un codex fiscal Hermopolite (P.Sorb. II 69). Atlanta, Ga. () “Ducs, ‘praesides’, po`etes et rh´eteurs au Bas-Empire.” AntTard : –. () “Les Papyrus lycopolites de l’Acad´emie des Inscriptions,” in Atti del XXII congresso internazionale di papirologia. Florence. Gascou, Jean, and L.S.B. MacCoull. () “Le Cadastre d’Aphroditˆo.” Travaux et M´emoires : –. Gascou, Jean, and Klaas A. Worp. () “CPR VII : r´ee´dition.” Tyche : –. Gauthier, H. () “Notes g´eographiques sur le nome Panopolite.” BIFAO : –. Gelzer, Matthias. () “Altes und Neues a.d. byzant.-¨agyptischen verwaltungsmisere.” APF : –. Geraci, G. () “Per una storia dell’amministrazione fiscale nell Egitto del VI secolo d.C.: Dioskoros e l’autopragia di Aphrodito,” Actes du XVe congr`es international de papyrologie, part 4. Brussels: –. Geremek, H. () “Les politeuomenoi e´gyptiens, sont-ils identiques aux bouleutai?” Anagennesis (): –. Girgis, V. A. () Prosopografia e Aphroditopolis. Berlin. Goma`a, Farouk, Renate M¨uller-Wollermann, and Wolfgang Schenkel. () Mittel¨agypten zwischen Samalut und dem Gabal Abu Sir. Wiesbaden. Gonis, Nikolaos. (a) “Six notes on documentary Papyri.” ZPE : –. (b) “Notes on Oxyrhynchus papyri II.” ZPE : –. (c) “Three notes on governors of Byzantine Egypt.” ZPE : –. (d) “A symmachos on mission and his paymaster: P.Herm.  enlarged.” ZPE : –. (e) “P.Oxy. XVI : A lease of a room from Byzantine Oxyrhynchus.” ZPE : –. (f) “A late Byzantine land lease from Oxyrhynchus: P.Oxy. XVI  revised.” Tyche : –. (g) “Some Oxyrhynchite villages in the eighth century.” CE : –. () “Studies on the aristocracy of late antique Oxyrhynchus.” Tyche : –.

Bibliography



(a) “P.Bingen  and Flavius Apion I.” ZPE : –. (b) “Notes on Oxyrhynchus papyri III.” ZPE : –. Goudriaan, Koen. () “Ethnical strategies in Graeco-Roman Egypt,” in Ethnicity in Hellenistic Egypt, ed. P. Bilde. et al. Aarhus: –. Goudriaan, Koen, et al. (eds.) () Prosopography and Computer. Louvain. Graham, Shawn. () “Ex Figlinis: The Complex Dynamics of the Roman Brick Industry in the Tiber Valley during the st to rd Centuries ad.” PhD thesis, University of Reading. (Forthcoming) “Who’s in charge? Studying social networks in the Roman brick industry in central Italy.” Graham, Shawn, and Giovanni Ruffini. () “Network analysis and GrecoRoman prosopography,” in Prosopography Approaches and Applications: A Handbook, ed. K. S. B. Keats-Rohan. Oxford. Granovetter, Mark S. () “The strength of weak ties.” American Journal of Sociology : –. () “The strength of weak ties: a network theory revisited,” in Sociological Theory 1983, ed. R. Collins. San Francisco: –. Grenfell, Bernard P. (/) “Oxyrhynchus and its papyri.” Archaeological Report 1896–1897: –. Habermann, W. () “Zur chronologischen Verteilung der papyrologischen Zeugnisse.” ZPE : –. Hanafi, A. () “A document from the archives of Dioscorus,” in Atti del XVII congresso internazionale di papirologia. Naples: –. () “Two documents from the archive of Dioscorus.” BACPS : – . () “Note from the archive of Dioscorus: Apollos son of Dioscorus in P. Cairo Mus. inv. () again.” BACPS : –. () “Four unpublished documents of the papyrus collections in Cairo and Copenhagen,” in Akten des 21. Internationalen Papyrologenkongresses. Berlin: –. () “Two contracts of marriage of papyrus collections in Cairo and Copenhagen,” in Atti del XXII congresso internazionale di papirologia. Florence: –. Hardy, E.R. () The Large Estates of Byzantine Egypt. New York. () Christian Egypt: Church and People. New York. Harland, Philip A. () “Connections with elites in the world of the early Christians,” in Handbook of Early Christianity: Social Sciences Approaches, ed. Anthony J. Blasi, Paul-Andr´e Turcotte, and Jean Duhaime. Walnut Creek, Ca. Harrauer, H. () “Neue protokometen-papyri. Mit einer dokumentation der protokometen.” Aegyptus : –. Harrauer, Hermann, and P. J. Sijpesteijn. () “Verkauf von Wein gegen Vorauszahlung.” CE : –. Hassanein, Alia H. () “Two documents from Aphrodite.” BIFAO  Supplement: –.



Bibliography

Hatzilambrou, Rosalia G. () “A lease of house-property from Byzantine Oxyrhynchus.” BASP : –. Herrmann, Johannes. () “Exegese von PSI  (Gestellungsb¨urgschaft f¨ur einen Hirten).” APF : –. Hezser, Catherine. () The Social Structure of the Rabbinic Movement in Roman Palestine. T¨ubingen. Hickey, Todd M. (a) “A fragment of a letter from a bishop to a scholastikos.” ZPE : –. (b) “P. Oxy. X  descr.” ZPE : –. () “Notes on some Cairo papyri from Byzantine Oxyryhnchus.” ZPE : –. () “A Public ‘House’ but Closed: Fiscal Participation and Economic Decision Making on the Oxyrhynchite Estate of the Flavii Apiones.” Diss. University of Chicago. () “Reuniting Anastasia: P.bibl. univ. Giss. Inv.  + P.Erl.  (with some notes on other Erlangen papyri.” APF (): –. Hickey, Todd M., and James G. Keenan. (–) “More from the archive of the descendants of Eulogius.” Analecta Papyrologica –: –. () “P.Lond. V : classification and date,” in Papyri in honorem Johannis Bingen octogenarii. Louvain: –. () “P.Lond. V  descr.: which landowner?” CE : –. Hoffman, Piotr, and Bartek Klin. () “Careful with that computer: on creating maps by multidimensional scaling of papyri in Katja Mueller’s recent studies on the topography of the Fayum.” JJP : –. Husson, Genevi`eve. () “Recherches sur les sens su mot pro†stion dans le grec d’Egypte.” Rech.Pap. : –. Johnson, A. C., and L. C. West. () Byzantine Egypt: Economic Studies. Princeton. Jones, A. H. M. () The Later Roman Empire, 284–602: A Social, Economic, and Administrative Survey. Oxford. J¨ordens, Andrea. () “Die a¨gyptischen symmachoi.” ZPE : –. () “Kaufpreisstundungen (sales on credit).” ZPE : –. Judge, E. A. () “Fourth-century monasticism in the papyri,” in Proceedings of the 16th International Congress of Papyrology, New York, 1980. Chico, Ca.: –. Keenan, James. () “The Nomina Flavius and Aurelius: A Question of Status in Byzantine Egypt.” Diss. Yale. () “Flavius John, scholasticus: a note on PSI .” BASP (–): – . () “The names Flavius and Aurelius as status designations in later Roman Egypt.” ZPE : –. (a) “The date of P.Warren .” ZPE : . (b) “On law and society in late Roman Egypt.” ZPE : –. () “The case of Flavia Christodote: observations on PSI I .” ZPE : –.

Bibliography



(a) “From the archive of Flavius Eulogius and his descendants.” ZPE : –. (b) “Two Byzantine papyri from the Michigan collection.” ZPE : – . () “Aurelius Phoibammon, son of Triadelphus: a Byzantine Egyptian land entrepreneur.” BASP : –. (a) “Victor, son of Psaios, and three Aphrodito rent receipts.” BASP : –. (b) “An afterthought on the names Flavius and Aurelius.” ZPE : – . (a) “The Aphrodito papyri and village life in Byzantine Egypt.” BSAC : –. (b) “Aurelius Apollos and the Aphrodite village elite,” in Atti del XVII congresso internazionale di papirologia. Naples: –. (a) “Village shepherds and social tension in Byzantine Egypt.” YCS : –. (b) “Notes on absentee landlordism at Aphrodito.” BASP : –. (a) “On languages and literacy in Byzantine Aphrodito,” in Proceedings of the 18th International Congress of Papyrology, vol. . Athens: –. (b) “P.Cair.Masp. II .” BASP : –. (c) “L. MacCoull: Dioscorus of Aphrodito.” BASP : –. (d) “A Christian letter from the Michigan collection.” ZPE : –. () “Pastoralism in Roman Egypt.” BASP : –. (a) “On the Cairo Maspero papyri,” in Proceedings of the 19th International Congress of Papyrology, vol. . Cairo: –. (b) “A Constantinople loan, ad .” BASP : –. () “Papyrology and Byzantine historiography.” BASP : –. () “The Aphrodito murder mystery: a return to the scene of the crimes.” BASP : –. () Review: Jane Rowlandson, Landowners and Tenants in Roman Egypt. The Classical Bulletin (): –. () “‘Die Binnenwanderung’ in Byzantine Egypt.” GRBS (): –. (a) Review: Denis Feissel and Jean Gascou, eds., La P´etition a` Byzance. BASP : –. (b) Review: Constantin Zuckerman, Du village a` l’empire: autour du registre fiscal d’Aphroditˆo (525/526). BASP : –. () “Byzantine Egyptian villages,” in Egypt in the Byzantine World, 300–700, ed. Roger S. Bagnall. Cambridge: –. Kehoe, Dennis P. () “Aristocratic dominance in the late Roman agrarian economy and the question of economic growth.” JRA (): –. Kochen, M. (ed.) () The Small World. Norwood, N.J. Koenen, L. () “Neue Lesungen zu kairener Urkunden – papyri.” ZPE : –. () “Preliminary observations on legal matters in P. Petra ,” in Atti del XXII congresso internazionale di papirologia. Florence: –.



Bibliography

Kr¨uger, J. () Oxyrhynchos in der Kaiserzeit: Studien zur Topographie und Literaturrezeption. Frankfurt. Kuehn, Clement A. () “Dioskoros of Aphrodito and Romanos the Melodist.” BASP : –. () “A new papyrus of a Dioscorian poem and marriage contract: P.Berol. inv. no. .” ZPE : –. () Channels of Imperishable Fire: The Beginnings of Christian Mystical Poetry and Dioscorus of Aphrodito. New York. Laniado, Avshalom. () “Suntelesthv. Notes sur un terme fiscal surinterpr´et´e.” JJP : –. () Recherches sur les notables municipaux dans l’empire protobyzantin. Travaux et m´emoires du Centre de recherche d’histoire et civilisation de Byzance. . () “L’Onomastique romaine dans le monde protobyzantin: quelques t´emoignages n´eglig´es.” AntTard : –. Lapin, Hayim. () Economy, Geography, and Provincial History in Later Roman Palestine. T¨ubingen. Le Roy Ladurie, Emmanuel. (). Montaillou: The Promised Land of Error, trans. Barbara Bray. New York. Lehtinen, M. () “Preliminary remarks on the prosopography of the Petra papyri.” Atti del XXII congresso internazionale di papirologia. Florence: – . Liebeschuetz, W. () “The origin of the office of the pagarch.” BZ : – . () “The pagarch: city and imperial administration in Byzantine Egypt.” JJP : –. () “Civic finance in the Byzantine period: the laws and Egypt.” BZ : –. Little, Lester. (ed.) () Plague and the End of Antiquity: The Pandemic of 541–750. Cambridge. MacCoull, Leslie. () “The Coptic archive of Dioscorus of Aphrodito.” CE : –. () “Notes on the social structure of late antique Aphrodito.” BSAC : –. () “Dioscorus and the dukes: an aspect of Coptic Hellenism in the sixth century.” Byzantine Studies : –. () “P.Cair.Masp. II  verso –. The Gnostica of Dioscorus of Aphrodito.” Tyche : –. () Dioscorus of Aphrodito: His Work and his World. Berkeley. () “The Aphrodito murder mystery.” JJP : –. (a) “A Cinderella story from Byzantine Egypt: P.Cair.Masp. I  and III .” Byzantion : –. (b) “More missing pieces of the Dioscorus archive,” in Actes du IVe congr`es copte, ed. M. Rassart-Debergh and J. Ries. Louvain. () “The Apa Apollos monastery of Pharoou (Aphrodito) and its papyrus archive.” Mus´eon : –.

Bibliography



´ () Review: Jean-Luc Fournet, Hell´enisme dans l’Egypte du VIe si`ecle: la biblioth`eque et l’oeuvre de Dioscore d’Aphrodite. BASP : –. () “P.Mich.inv.  revisited: a sixth-century Coptic contract from Aphrodito.” ZPE : –. MacDonald, A. A., M. W. Twomey, and G. J. Reinink. (eds.) () Learned Antiquity: Scholarship and Society in the Near-East, the Greco-Roman World, and the Early Medieval West. Louvain. MacLennan, Hugh. () Oxyrhynchus: An Economic and Social Study. Princeton. Malz, G. () “The papyri of Dioscorus: publications and emendations,” in Studi in onore di A. Calderini e R. Paribeni, vol. . Milan: –. Mandilaras, Basil G. () “The feast of Thynis, ìEn —ort¦ƒ QÅnewv.” Tyche : –. Maresch, Klaus. () Nomisma und Nomismatia: Beitr¨age zur Geldgeschichte ¨ Agyptens im 6. Jahrhundert n. Chr. Opladen. Martin, V. () “A letter from Constantinople.” JEA : –. ´ Maspero, J. () “Etudes sur les papyrus d’Aphrodit´e II–IV.” BIFAO : –. () “Les papyrus Beaug´e.” BIFAO : –. Mayerson, Philip. () “The economic status of potters in P.Oxy. L – and XVI , .” BASP : –. Mazza, Roberta. () “Ricerche sul pagarca nell’Egitto tardoantico e bizantino.” Aegyptus : –. () “Fl. %p©wn gen»menov prwtopatr©kiov.” Simblos : –. () “P.Oxy. XVI  e i conti annuali dei pronoetai.” ZPE : –. () L’archivio degli Apioni: terra, lavoro e proprieta senatoria nell’Egitto tardoantico. Bari. Milgram, S. () “The small world problem.” Psychology Today : –. Mirecki, Paul. () Review of Timothy Teeter, Columbia Papyri XI. BASP : –. Mirkovic, Mira. () “Dioskoros als syntelestes,” in Akten des 21. Internationalen Papyrologenkongresses, Berlin 1995, Archiv f¨ur Papyrusforschung, Beiheft . Stuttgart: –. Mitchell, J. Clyde. () “The concept and use of social networks,” in Social Networks in Urban Situations: Analyses of Personal Relationships in Central African Towns, ed. J. Clyde Mitchell. Manchester: –. () “Networks, norms and institutions,” in Boissevain and Mitchell, eds., –. ¨ Mithof, Fritz. () Annona militaris: die Heeresversorgung im sp¨atantiken Agypten. Florence. Montevecchi, O. () La papirologia. Turin. Moody, James, and Douglas R. White. () “Structural cohesion and embeddedness: a hierarchical concept of social groups.” American Sociological Review (): –. Morris, Royce L. B. () “Reflections of citizen attitudes in petitions from late Roman and Byzantine Oxyrhynchus,” in Akten des 21. Internationalen Papyrologenkongresses. Berlin: –.



Bibliography

Mueller, Katja. () “Ptolemaic settlements in space. Settlement size and hierarchy in the Fayum.” APF (): –. (a) “Places and spaces in the Themistou Meris (Fayum/Graeco-Roman Egypt). Locating settlements by multidimensional scaling of papyri.” Ancient Society : –. (b) “Mastering matrices and clusters, locating Graeco-Roman settlements in the Meris of Herakleides (Fayum/Egypt) by monte-carlo-simulation.” APF (): –. () “What’s your position? Using multi-dimensional scaling (MDS) and geographical information systems (GIS) for locating ancient settlements in the Meris of Polemon/Graeco-Roman Fayum.” APF : –. () “Geographical information systems (GIS) in papyrology. Mapping fragmentation and migration flow to Hellenistic Egypt.” BASP : –. () Settlements of the Ptolemies: City Foundations and New Settlement in the Hellenistic World. Studia Hellenistica . Louvain, Paris, Dudley, MA. Mullett, Margaret. () Theophylact of Ochrid: Reading the Letters of a Byzantine Archbishop. Aldershot. Newman, M. E. J., Albert-Laszlo Barabasi, and Duncan J. Watts. () The Structure and Dynamics of Networks. Princeton. Newman, M. E. J., S. H. Strogatz, and D. J. Watts. () “Random graphs with arbitrary degree distributions and their applications.” Physical Review E , . Niemeijer, Rudo. () “Some applications of the notion of density to network analysis,” in Boissevain and Mitchell, eds., –. Noble, Mary. “Social network: its use as a conceptual framework in family analysis,” in Boissevain and Mitchell, eds., –. Nock, Arthur Darby. () “Later Egyptian piety,” in Essays on Religion and the Ancient World, ed. Z. Stewart. Cambridge, Mass. Nooy, W. de, A. Mrvar, and V. Batagelj. () Exploratory Social Network Analysis with Pajek. Cambridge. ¨ Palme, Bernhard. () Das Amt des ˆpaithtžv in Agypten. Vienna. (a) “Die domus gloriosa des Flavius Strategius Paneuphemos.” Chiron : –. (b) “Neuedition von SB I  + SB I  + SB XVI  (= I ).” ZPE : –. (a) “Flavius Strategius Paneuphemos und die Apionem.” ZSavRom : –. (b) “Flavius Epiphanius, comes domesticorum.” Eirene : –. () “Die officia der Statthalter in der Sp¨atantike. Forschungsstand und Perspektiven.” AntTard : –. () “Die riparii des Dorfes Aphrodite: Claudius bzw. Flavius Apollos, Aurelius Apollos und Flavius Victor,” in Festschrift f¨ur Gerhard Dobesch zum f¨unfundsechzigsten Geburtstag am 15. September 2004, Vienna. Papaconstantinou, Arietta. (a) “Sur les e´vˆeques byzantins d’Oxyrhynchos.” ZPE : –.

Bibliography



(b) “La Liturgie stationale a` Oxyrhynchos dans le premi`ere moiti´e du e si`ecle. R´ee´dition et commentaire du POxy XI .” REByz : –. ´ () Le Culte des saints en Egypte des byzantins aux abbassides. L’Apport des sources papyrologiques et ´epigraphiques grecques et coptes. Paris. Par´assoglou, G. M. () Imperial Estates in Roman Egypt. American Studies in Papyrology . Las Palmas. Pintaudi, R., and P. J. Sijpesteijn. () “Papiri tardo-bizantini conservati alla British Library.” ZPE : –. Preger, Theodorus. () Scriptores Originum Constantinopolitanarum. New York. Pruneti, Paola. () I centri abitati dell’Ossirinchite: repertorio toponomastico. Florence. Rathbone, Dominic W. () “Villages, land and population in Graeco-Roman Egypt.” Proceedings of the Cambridge Philological Society n.s. : –. () Economic Rationalism and Rural Society in Third-Century ad Egypt: The Heroninos Archive and the Appianus Estate. Cambridge. Reiter, F. () “Einige Bemerkungen zu dokumentarischen Papyri.” ZPE : –. R´emondon, R. () “Soldats de Byzance d’apr`es un papyrus trouv´e a` Edfou.” Recherches de papyrologie : –. () “P.Hamb.  et P.Lond.  (notes sur les finances d’Aphroditˆo du VIe si`ecle au VIIIe).” CE : –. (a) “Papyrologica.” CE : –. ´ (b) “L’Egypte au Ve si`ecle de notre e`re; les sources papyrologiques et leurs probl`emes,” in Atti dell’XI congresso internazionale di papirologia. Milan. () “Le Monast`ere alexandrin de la M´etanoia. Etait-il b´en´eficiaire du fisc ou a` son service?” in Studi in onore di Edoardo Volterra,  vols. Milan: vol. , –. () “L’Acte de cautionnement byzantin P.Varsov. .” CE : –. () “Les Contradictions de la soci´et´e e´gyptienne a` l’´epoque byzantine.” JJP : –. Remus, Harold E. () “Voluntary association and networks: Aelius Aristides at the Asklepieion in Pergamum,” in Voluntary Associations in the Graeco-Roman World, ed. John S. Kloppenborg and Stephen G. Wilson. London: –. ´ Rouillard, Germaine. () L’Administration civile de l’Egypte byzantin. nd edn., Paris. () Vie rurale dans l’Empire byzantin. Paris. Rowlandson, Jane. () “Sales of land in social context,” in Proceedings of the Sixteenth International Congress of Papyrology. Missoula, Mont.: –. () Landowners and Tenants in Roman Egypt: The Social Relations of Agriculture in the Oxyrhynchite Nome. Oxford. (Forthcoming) “Oxyrhynchus and its hinterland,” in Oxyrhynchus: A City and its Texts, ed. Alan K. Bowman et al. Ruffini, Giovanni. () “Late antique pagan philosophers from Athens to the Thebaid,” in Ancient Alexandria between Egypt and Greece, ed. W. V. Harris and Giovanni Ruffini. Leiden.



Bibliography

() “Social Networks in Byzantine Egypt.” Diss. Columbia University. () “The commonality of rare names in Byzantine Egypt.” ZPE : –. () “New approaches to Oxyrhynchite topography,” in Proceedings of the 24th Congress of Papyrology, Helsinki, 1–7 August, 2004. Helsinki: –. ( forthcoming) Review: Katja Mueller, Settlements of the Ptolemies: City Foundations and New Settlement in the Hellenistic World. Topoi . (Forthcoming a) “Factions and social distance in sixth-century Aphrodito,” in ´ Les Actes du colloque Dioscore (Etudes d’arch´eologie et d’histoire ancienne). (Forthcoming b) An Aphrodito Prosopography. (Forthcoming c) Review of Peter Sarris, Economy and Society in the Age of Justinian. Topoi. Ruffini, Julio L. () “Disputing over livestock in Sardinia,” in The Disputing Process: Law in Ten Societies, ed. Laura Nader and Harry F. Todd Jr. New York: –. Salomon, R. G. () “A papyrus from Constantinople (Hamb. inv. no. ).” JEA : –. Salomons, R. P. () “Testamentaria.” ZPE : –. Salway, Benet. () “What’s in a name? A survey of Roman onomastic practice from c. bc to ad .” JRS : –. Sarris, Peter. () Economy and Society in the Age of Justinian. Cambridge. () “Bubonic plague in Byzantium: the evidence of non-literary sources,” in Plague and the End of Antiquity, ed. Lester Little. Cambridge: –. ´ Sauneron, S. () Villes et l´egendes d’Egypte. Cairo. Schmitt, Oliver. () “Die buccellarii.” Tyche : –. Scott, J. () Social Network Analysis: A Handbook. London. Scott, S. P. () The Civil Law. Cincinnati. Sijpesteijn, P. J. () “Five Byzantine papyri from the Michigan collection.” ZPE : –. () “The monastery of Abbas Andreas.” ZPE : –. () “P.Mich.inv. : a reconsideration.” ZPE : –. () “Varia papyrologica III.” ZPE : –. Stark, Rodney. () The Rise of Christianity: A Sociologist Reconsiders History. Princeton. Stark, Rodney, and William Sims Bainbridge. () “Networks of faith: interpersonal bonds and recruitment to cults and sects.” American Journal of Sociology : –. Syrcou, Angeliki. () “Receipt of wine: P.Oxy. X  descr. (V AD).” APF (): –. () “Six Byzantine documents.” APF (): –. Tacoma, L. E. () “Replacement parts for irrigation machine of the divine house at Oxyrhynchus. P.Columbia inv. , October , ad  (?).” ZPE : –. Talbert, Richard J. A. (ed.) () Barrington Atlas of the Greek and Roman World. Princeton. Thomas, J. D. () “The office of exactor in Egypt,” CE : –.

Bibliography



Thomas, J. P. () Private Religious Foundations in the Byzantine Empire. Washington. ¨ Timm, Stefan. () Das christlich-koptische Agypten in arabischer Zeit: eine Samm¨ lung christlicher St¨atten in Agypten in arabischer Zeit. Wiesbaden. Torallas, Tovar. () “The police in Byzantine Egypt: the hierarchy in the papyri from the fourth to the seventh centuries,” in Current Research in Egyptology 2000, ed. A. McDonald and C. Riggs. Oxford: –. Turner, E. G. () “Roman Oxyrhynchos.” JEA : –. Van Haelst, J. () “De nouvelles archives: Anastasia, propri´etaire a Oxyrhynchus.” CE : –. () “De nouvelles archives: Anastasia, propri´etaire a Oxyrhynchus.” CE : –. () “De nouvelles archives: Anastasia, propri´etaire a Oxyrhynchus.” Atti dell’XI congresso internazionale di papirologia. Milan: –. Van Minnen, Peter. () “Dioscorus and the law,” in MacDonald et al. eds., –. Vinson, Steve. () The Nile Boatman at Work. Mainz. Wasserman, Stanley and Katherine Faust. () Social Network Analysis: Methods and Applications. Cambridge. Watts, Duncan J. (a) Small Worlds: The Dynamics of Networks between Order and Randomness. Princeton. (b) “Networks, dynamics, and the small-world phenomenon.” AJS (): –. () Six Degrees: The Science of a Connected Age. New York. Watts, Duncan J., and S. H. Strogatz. () “Collective dynamics of ‘small-world’ networks.” Nature : –. Watts, E. () City and School in Late Antique Athens and Alexandria. Berkeley. White, L. Michael. (ed.) (a) “Editor’s preface” to Social Networks in the Early Christian Environment: Issues and Methods for Social History. Semeia : vii– viii. (b) “Finding the ties that bind: issues from social description.” Semeia : –. (c) “Social networks: theoretical orientation and historical applications.” Semeia : –. ´ Wipszycka, E. () Les Resources et les activit´es ´economiques des ´eglises en Egypte du IVe au VIIIe si`ecle. Brussels. () “Les Terres de la congr´egation pachˆomienne dans une liste de payements pour les apora,” in Le Monde grec: pens´ee, litt´erature, histoire, documents. Hommages a` Claire Pr´eaux, ed. J. Bingen et al. Brussels: –. (“) %nacwrhtžv, –rhm©thv, ›gkleistov, ˆpotaktik»v. Sur la terminolo´ gie monastique en Egypte.” JJP : –. Worp, K. A. () “Flavius Strategius: some notes.” ZPE : –. () “A checklist of bishops in Byzantine Egypt (ad –c.).” ZPE : –. () “On the aureliate of clergy and monks.” ZPE : –.



Bibliography

Zuckerman, Constantine. (a) Du village a` l’Empire: autour du registre fiscal d’Aphroditˆo (525/526). Monographies du Centre de Recherche d’Histoire et Civilisation de Byzance . Paris. (b) “Les Deux Dioscore d’Aphrodit`e ou les limites de la p´etition,” in La P´etition a` Byzance, ed. D. Feissel and J. Gascou. Monographies du Centre de Recherche d’Histoire et Civilisation de Byzance . Paris: –.

Subject index

Abraam son of Victor  Adaiou , , , , , ,  affiliations networks , , , , ,  Alexander, Flavius ,  Ammonios, Flavius , , , , , –, , ,  Anastasia, Flavia , , –, , –, , , , , ,  Antaiopolis , , , –, , , , , , ,  Antaiopolite , , , , , , , ,  Antinoopolis , –, , , –, , , ,  Aphrodisias  Aphrodito cadastre , , , , , , , , , , , , , , , –,  cutpoints of  estate formation , –, ,  fiscal register , –, , ,  geography of , , , , ,  guilds of , , ,  murder mystery  and neighboring villages  network characteristics of , , , , , , , , , , , –, –, , , ,  network’s largest component , , , , , ,  and the pagarchs , , , , –, ,  population of , ,  and its shepherds  strong ties in , , , ,  Apion I , , , ,  Apion II , –, ,  Apion III , ,  Apion estates archive of , , , , , , –, 

growth of , , , , , , , , , , , –, ,  income of ,  location of –, –, , , ,  network characteristics of , , –, , , ,  population , , , , ,  prono¯esiai of –, –, ,  size of  Apollos and Ammonios  career of  and dispute settlements ,  monastery of ,  property of , , ,  Appianus , , ,  Arcadia ,  Artemidoros, Flavius – attributes network  autopragia , , , , , , , , , , , , , , –, , , ,  Bagnall, Roger , , , ,  Banaji, Jairus , , –, , ,  barracelli  Boissevain, Jeremy , , , ,  Bottos , , ,  centrality , , , , , , , –, –, , , , ,  betweenness centrality , , , ,  closeness centrality , , ,  degree centrality , , , , , ,  prestige centrality  Christodote, Flavia , –, , ,  church, Oxyrhynchite –, , , , ,  clique  cluster ,  clustering coefficient 





Subject index

cohesion  Constantinople , , , , , , , , –, , , , , , , –, , , ,  cutpoints , , , –, ,  cutset 

Ioulianos the pagarch , , , –, –, ,  isolates , , 

dark matter  degree  density , , , , , , , –, , , , –, –, ,  Diogenes , , , –, , , ,  Dioskoros and Aphrodito’s network , , –, , , ,  archive of , , , , , , –, , , , ,  career of – family of , ,  poems of , , ,  and shepherds , , , ,  directed graph  distance , , , , , , , ,  DL nodelist 

Keenan, James , , , , , , , , , , , , , , , , , –, –, ,  koina , , , , ,  Kometes , –, –,  Kosmou ,  kt¯et¯or , , , , , –, , , ,  Kyria, Flavia –, 

effective network  Elusa  Enoch son of Hermaouos  Episemou , ,  Erd¨os, Paul  Eulogios, Flavius , , , –, , ,  Euphemia, Flavia –, ,  extended network  Florentine families , ,  Gabrielia, Flavia , , , , , ,  Gascou, Jean , , , , , –, – Gini , ,  Girgis prosopography , , , , –, , , , , , ,  Herakleopolite , , , , ,  Hermopolis ,  Hermopolite nome , , , , , , ,  Iem¯e  Ieremias son of Iakubios  Ioannes son of Timagenes – Ioannes the prono¯et¯es ,  Ioulianos the eparch , , , , , , , , , , , , 

Justinian , , , , , , , , , 

Le Roy Ladurie  literacy, importance of , , ,  Mazza, Roberta , , , , –, ,  Menas son of Apollos , ,  Menas the oiket¯es , , ,  Menas the pagarch , , –, ,  Milgram, Stanley ,  Montaillou  Mounkrekin  Mueller, Katja , ,  multiplex ties , , , , –, , , , , –, , , , , , , , , , –, , , , ,  Negev  Nessana , ,  NetDraw , , ,  network analysis and the ancient world – and Egypt  history of  tutorial  nodes  one-mode networks , , ,  Ouesotos  Oxyrhynchite nome centralization of , , , , , , , , , , , , , , ,  estate formation , ,  great estates , , –,  network analysis of  network characteristics of , , , , , , , , ,  population , 

Subject index population pyramid , , , ,  topography , , , , , ,  Padgett’s Florentine families ,  Pajek , , , , , ,  Pakerke ,  Panolbios, Flavius  Pela –, ,  pendants  Peto – Petra , , , , , , , ,  Pharatopos  Phennis , ,  Pherko ,  Phoibammon son of Triadelphos , , , , , , , , –,  and Aphrodito’s network –, ,  archive of ,  business affairs –, ,  dispute settlement of  life of ,  network characteristics of  and strong ties , , ,  Phoibammon, comes , , – Phthla , , –, , , , , , ,  Pilatos the grammateus  plague , –, ,  power-law distribution  Praiecta ,  Prosdokios, Flavius , ,  pr¯otok¯om¯et¯es , , , , , , , , , , , , , , , , , , , ,  Pruneti , , , , –, , , , , , , , , , , , –, , , –, , ,  Psinabla , ,  rank-size rule  Samuel, comes , , , ,  Sarapammon –, ,  Sardinia  Senouthes son of Apollos , , –, –, , –, , ,  Serenos the prono¯et¯es , , ,  Serenos, grandson of Eulogios – Serila , 



shepherds , , , , , , –, , , , , , –, –, –, , , , , ,  signed graph ,  Sophia , , , , ,  Spania , , , , , , , – Strategios I ,  Strategios II , , , ,  Strategios III  Strategios IV  Strategios Paneuphemos  strength of weak ties , ,  structural equivalence  structuralism ,  suntelest¯es –, , , , ,  Takona , , , , , , , , ,  Tanuaithis , ,  Tarouthinos , , ,  Tausiris  Temseu Skordon  Teruthis , , ,  Thebaid , , , , , , , , , , ,  Theodora estates of  petition to , , , , , ,  and protection of Aphrodito , ,  Theodoros of Petra , ,  Theodoros the prono¯et¯es , ,  Theon, estates of , –,  Theoteknos, Flavius –, , , , , , , , ,  Thinis , ,  Thmonachthe , , , , ,  Thmounameris  Tholthis  Timagenes descendants of , ,  estates of , –, ,  tracing  two-mode networks , , , ,  UCINET , , , –, , , , , , , , –, , , ,  valued graph  Zuckerman, Constantin , , , , , , , , , , , , , , , 

Index locorum

P.Aphrod.Lit.   P.Aphrod.Lit.   P.Aphrod.Lit.   P.Bingen   P.Cair.Masp. . , ,  P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. . P.Cair.Masp. .

, , , , , , , , –,    ,   , ,       , ,   ,  ,    ,   ,   ,    ,  ,  ,  , ,    ,  , , ,  ,   , , 

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





Index locorum P.Mich. . , ,  P.Mich. . – P.Mich. . – P.Mich. .  P.Mich. .  P.Mich. .  P.Mich. .  P.Mich. .  P.Mich.Aphrod. ,  P.Michael.  , , , ,  P.Michael.  ,  P.Michael.  ,  P.Michael.   P.Michael.  ,  P.Michael.  ,  P.Michael.  , ,  P.Michael.   P.Ness. .  P.Ness. .  P.Ness. .  P.Ness. .  P.Oxy. .  P.Oxy. .  P.Oxy. . , , , , ,  P.Oxy. . ,  P.Oxy. . , , ,  P.Oxy. .  P.Oxy. . ,  P.Oxy. . , ,  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. . , , ,  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. . ,  P.Oxy. .  P.Oxy. . ,  P.Oxy. .  P.Oxy. .  P.Oxy. . –

P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . ,  P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. . P.Oxy. .

,   ,   ,  , ,   , ,  , , –,  , ,  ,   ,  ,  ,       ,     ,  ,     ,        ,    ,  , , , , –, , ,     , –,  , , ,  , , , – ,    , , , ,  , ,  , , , 

 P.Oxy. . ,  P.Oxy. . , a P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. . ,  P.Oxy. .  P.Oxy. . ,  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. . , , –, ,  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. . , , , ,  P.Oxy. ..  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Oxy. .  P.Princ. .  P.Ross.Georg. .  P.Ross.Georg. .  P.Thomas   P.Turner   P.Warr.  

Index locorum PSI . ,  PSI .a  PSI .  PSI . , ,  PSI .  PSI .  PSI . ,  PSI .  PSI .  PSI .  PSI .  SB .  SB . ,  SB .  SB .  SB . ,  SB .  SB .  SB .  SB .  SB .  SB .  SB .  SB .  SB . , , , , , ,  SB .  SB . ,  SB .  SB .  SB .  SB . ,  SB .  SB .  SB . 