The Marxist Theory Of Art: An Introductory Survey 9781000303230, 1000303233

Citation preview

Studies in Ancient Monarchies

Christopher Bendle

The Office of Magister Militum in the 4th Century CE A Study into the Impact of Political and Military Leadership on the Later Roman Empire

Ancient History Franz Steiner Verlag

Studies in Ancient Monarchies Herausgegeben von Ulrich Gotter (Konstanz), Matthias Haake (Bonn), Nino Luraghi (Oxford) und Kai Trampedach (Heidelberg) Volume 10 www.steiner-verlag.de/brand/Studies-in-Ancient-Monarchies

The Office of Magister Militum in the 4th Century CE A Study into the Impact of Political and Military Leadership on the Later Roman Empire Christopher Bendle

Franz Steiner Verlag

Umschlagabbildungen: Links: King Tiglath-pileser III of Assyria. Stone panel, ca. 728 BCE. From the Central Palace in Nimrud, now in the British Museum. © akg / Bible Land Pictures Mitte: Emperor Justinian. Mosaic, ca. 540 CE. Church of San Vitale, Ravenna. © akg / Bildarchiv Steffens Rechts: Alexander the Great at the Battle of Issos. Mosaic, ca. 100 BCE. From the Casa del Fauno, Pompeii, now in the Museo Archeologico Nazionale di Napoli. © akg / Nimatallah Bibliografische Information der Deutschen Nationalbibliothek: Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über dnb.d-nb.de abrufbar. Dieses Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist unzulässig und strafbar. © Franz Steiner Verlag, Stuttgart 2024 www.steiner-verlag.de Layout und Herstellung durch den Verlag Satz: DTP + TEXT Eva Burri, Stuttgart Druck: Beltz Grafische Betriebe, Bad Langensalza Gedruckt auf säurefreiem, alterungsbeständigem Papier. Printed in Germany. ISBN 978-3-515-13614-3 (Print) ISBN 978-3-515-13621-1 (E-Book) DOI 10.25162/9783515136211

Preface This monograph has grown from the thesis I submitted for my Master of Arts degree at the University of Melbourne. The original work was examined by Dr. Simon Corcoran (Newcastle University, UK) and Professor Bruno Bleckmann (Heinrich-Heine Universität Düsseldorf), who both provided excellent feedback and advice, which has been included here to create this final product. I would like to acknowledge and thank the University of Melbourne and the Australian Commonwealth Government for awarding me with the Australian Government Research Training Program Scholarship, which provided an instrumental source of funding for this book.

Acknowledgments My most profound gratitude goes to Dr Frederik Vervaet, who has provided me with years of support, advice, and guidance during my undergraduate and postgraduate studies. The same can be said of the invaluable expertise and assistance provided by Dr Hyun Jin Kim. I would also like to extend my thanks to all the staff at the University of Melbourne that have supported me, especially Dr Tim Parkin and Dr Louise Hitchcock, as well as Dr Sean Scalmer and Dr Zoe Laidlaw. Conversations with Dr David Parnell, Dr Mark Hebblewhite, Dr Benet Salway, Dr Ronald Ridley, and Commander Andrea Argirides has also proven most beneficial and I am grateful for their guidance. I would also like to thank Dr. Simon Corcoran and Professor Bruno Bleckmann for examining my Master of Arts thesis and providing extremely helpful constructive criticism and advising me to include the discussion that is now Chapter Three. My good friends Samer Ata and Matthias Neill also offered their assistance where my expertise failed. All the members of the University of Melbourne Classics and Archaeology Postgraduate Society have my gratitude for making my time at Uni so enjoyable. Much of this book was written during the COVID-19 global pandemic and its accompanying unrest, and so I have an endless debt to my family and friends both in Melbourne and in Philadelphia for keeping me safe and sane, especially my mum and dad for their never-ending love and encouragement to pursue my passion for history, and my wife, Stasha, for all the happiness she brings me.

Table of Contents List of Tables and Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 State of the Question. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Chapter One The Magistri Militum from 341 to 363 CE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 1.1.Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 1.2. The First Attested Magistri. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 1.3. Two Magisterial Usurpers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 1.4. Ammianus’ Account of Some Magistri Militum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 1.5. Julian’s Western Roman Empire. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 1.6. Julian and Jovian as Sole Emperors, 361 to 364 CE. . . . . . . . . . . . . . . . . . . . . . . . . . 67 Chapter Two The Magistri Militum from 364 to 395 CE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 2.1.Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 2.2. Valentinian’s Western Empire. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 2.3. Valens’ Eastern Empire. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 2.4. Crisis along the Danube. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 2.5. Civil War 383–388. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 2.6. Civil War 392–394: The Rise and Fall of Arbogast . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Chapter Three The Networks of the Magistri Militum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 3.1.Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 3.2. Silvanus’ Rebellion: The Strengths and Weaknesses of Complicated Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 3.3. Competitive Cliques and their Leaders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 3.4.Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

10

Table of Contents

Chapter Four The Prosopography of the Magistri Militum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 4.1.Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 4.2. The Fourth-Century Career Path and the Hierarchy of Military Offices. . . . . . 150 4.2.1.Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 4.2.2.Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 4.2.3.Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 4.3. The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 4.3.1.Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 4.3.2.Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 4.3.3.Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 Bibliography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 Abbreviations and Ancient Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 Modern Scholarship. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

List of Tables and Figures Table One Table Two Table Three Table Four

Fig. One Fig. Two Fig. Three Fig. Four Fig. Five

The Career Path of the Highest Late Roman Military Officers in Order of First Attestation as Magister Militum in the Sources.. . . . . . . . . . . . . . . . . . . . . 155 Years Spent in the Military Magisterium during the Fourth Century.. . . . . . . . . 169 The Perception of Identities of the Fourth-Century Magistri Militum in the Sources.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 A comparison of Teitler 1989, Elton 1996, and the present study’s results on the distribution of Magistri Militum by identity (barbarian and Roman).. 196

The Interactions between Valens and Three Magistri in 377. . . . . . . . . . . . . . . . . Ammianus Marcellinus’ Description of the Social Networks of the West from 368–78. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Consulships of the Magisterium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pie Graph of the Identities of the Magistri Militum. . . . . . . . . . . . . . . . . . . . . . . . . The Ethnic Identities of the Magistri Militum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

127 140 162 197 198

Introduction In early 392, emperor Valentinian II of the western Roman Empire sat upon his throne in Vienne and watched the approach of the army general named Arbogast. This man was the west’s sole magister militum, the highest-ranking officer of all the western forces, but his authority stretched much further than military affairs. For about a year, the magister had dictated the western Empire’s political and military policies, controlling where the Roman armies campaigned, filling the civil bureaucracy with his own cronies, and relegating the emperor to only ceremonial and religious duties.1 Arbogast had even recently personally executed Armonius, the son of a consul and personal friend of Valentinian, as the man cowered at the emperor’s feet.2 But Valentinian had decided the time was ripe to rid himself of this powerful menace. Relying on his authority as a Roman emperor, the supreme rulers of the Roman world for over four-hundred years, in court and in front of all his officials, Valentinian handed Arbogast a letter of dismissal, removing him from imperial service.3 Arbogast took the letter, read it once, and derisively tore it up and threw it on the ground in a stark refusal. In a rage, the emperor ran to one of his bodyguards and tried to draw the man’s sword to strike at Arbogast, but the soldier easily fended off his attempts.4 Even though an emperor had the technical authority to execute whomever he pleased, Arbogast clearly held much greater power. The magistri militum had not always overshadowed the Roman emperors. In 357, emperor Constantius II dismissed the magister militum Marcellus from military service for reasons the general thought were unfair.5 The only option available to Marcellus, however, was to complain and protest the command in the imperial capital.6 Disobedience to the order was not a possibility. Similarly, in 359, the magister Ursicinus was found guilty of not adequately safeguarding the city of Amida (Diyarbakır, Turkey),

1 Greg. Tur. HF 2.9. 2 Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev); Greg. Tur. HF 2.9; Paschoud 2006, 336. 3 Zos. 4.53.2–3; Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). 4 Philost. HE 11.1. 5 AM 16.7.1–3, 10.21; Lib. Or. 18.48; Hunt 1997b, 50; Drinkwater 2007, 227–8. 6 AM 16.7.1–2.

14

Introduction

even though he had tried to do exactly that but was refused the necessary soldiers.7 In anger, he insulted the same emperor, Constantius II, and was immediately cashiered as punishment. Once the order for dismissal came from the emperor, there was nothing Ursicinus could do. Marcellus’ and Ursicinus’ examples are emblematic of the magistri during the 340s to 370s, a period in which the impact of the generals on the wider political landscape was deftly controlled by the emperors. Those who were pushed too far only had the traditional path of a dissatisfied Roman general: usurpation. Defying the emperor’s authority without challenging his position was not yet a possibility. Thus, an extreme change had occurred between this period and the time when Arbogast could bluntly refuse an order from the emperor. Arbogast is an extreme example of the power of the magistri militum, but he is not an isolated case. Multiple magistri gained similar levels of hegemony over the western Empire, a trend that would continue into the fifth century. One aspect of this phenomenon is found in the person who occupied the position of emperor. Constantius II had had military commands since 336, and by the 350s he was an adept ruler with over a decade of experience as an independent Augustus, and he had developed a robust and broad patronage network.8 Valentinian II, on the other hand, had been made emperor as a four-year old child by civil and military bureaucrats who used him as a pawn to secure their own power. He had not gained independence from his half-brother until 383, and even then, he had not commanded soldiers in any military operations, nor had he won the loyalty of the western court. Valentinian is one of the child-emperors of the late imperial period that represents a dramatic weakening of the imperial office, and this development has been examined by the likes of Meaghan McEvoy.9 Powerful emperors like Constantius, or Julian, Valentinian, Theodosius I, and others, mostly had firm control over their generals, and this has led to many scholars attributing the developments in the magisterium to top-down, institutional reforms and legal directives. We must remember, however, that Constantius had also been a child-emperor, as he was raised in 324 at just seven years old, yet he was able to grow into a position of power.10 The key difference between Constantius’ development and Valentinian’s stagnation was the men who served in the civil and military bureaucracy and how their vision of the imperial office changed it into something they could annex for their own benefit. These men prohibited the child-emperors from gaining an active role in governing the Empire. It is thus my contention that significant changes occurred to the power balance of the fourth-century Empire, and the inversion of the traditional emperor-general power balance is the result of a bot-

7 8 9 10

AM 19.3.1–2, however this story is told by Ammianus Marcellinus, an officer who served Ursicinus and was not entirely objective when it came to his patron. Cons. Const. s. a. 324; Jul. Or. 1.13b. McEvoy 2013. Vanderspoel 2020.

Introduction

15

tom-up process, attributable to the magistri militum using successive moments of crisis to adapt the dynamic to suit their desires. The political and civil realms are intertwined in Roman power dynamics. Throughout this book we will be examining the nature of the magistri’s involvement in leading soldiers and conducting war, but also in affairs outside the traditional scope of a general’s purview. These realms had long been combined in the Roman world. The political success of emperors was intrinsically tied to their military capabilities, real or perceived. Although the civil and military career paths had been separated during the fourth century, the consulship remained as the shared peak of both career paths. Generals also regularly sat as judges in legal disputes. They were also deeply involved in the changing religious landscape of the Empire. Early in the fourth century, Christianity endured extreme persecution by the state, only to replace Greco-Roman Polytheism as the Empire’s dominant religion.11 This change ushered in a decline of polytheism, or as it is known by the common Christian pejorative ‘paganism’, and by 415 pagans would be formally excluded from military service.12 Christianity experienced multiple schisms, heresies, and theological disputes during this period of change, and we will see how the military leadership became involved in these developments. Additionally, the religious role of the emperors became significantly more pronounced, and the importance of their military role declined. Although many bureaucrats and aristocrats were involved in spurring on and taking advantage of this process, the magistri stood to gain the most if the emperors were reduced to ceremonial duties and military command was permanently shifted away from them and given entirely to the magistri. Because of the close intertwining of the military, political, judicial, and religious realms of the late Empire, we will not just be looking at when and how the magistri acted outside the military sphere, but the process through which they did, or did not, dominate areas other than the military. It is important to define exactly who these generals were and how they were referred to. The circumstances of the creation of this office will be discussed in Chapter One, and their variegated roles and duties will be explored throughout every section. We have already been using the term magister militum (‘master of the soldiers’) to refer to these highest-ranking generals, but this is a matter of convention. At first, the evidence indicates that the generals were called magister equitum (‘master of the cavalry’) and magister peditum (‘master of the infantry’). It was only later that some generals began to have the title magister equitum et peditum, and this was simplified into magister militum, which has become, along with the term magisterium, the standard form of referring to the office in modern literature. Other titles in use by our sources include 11 12

On the persecutions, see: Barnes 1996, 542–52; Clarke 2005, 650–1. On the end of the persecutions and Christianity’s eventual dominance, see: Corcoran 2012, 52; Drake 2005, 121–23; 2009, 216; Lenski 2017, 27; Barnes 1995; 2014, 74–80, 93–97; Chadwick 1997. Stoll 2007, 471–73.

16

Introduction

magister militiae (‘master of the military’), magister armorum (‘master of weapons’), magister utriusque militiae (‘master of both militaries’), as well as the more generic titles of dux, rector, and comes, all various forms of military leadership. Where the sources are clear and consistent on an individual’s title, I will use it in the body of this text. Where the sources are not so transparent, and when referring to the office as a whole, magister militum will be used. Titles that will not be used at all are magister militum praesens and praesentalis. These titles are often used by scholars to describe a general that spent the majority of their career serving ‘in the presence’ of an emperor or in the imperial court, in contrast to other generals that had postings in the provinces. These terms, however, first appear in the Notitia Dignitatum, and only rarely in later sources.13 Given the uncertain reliability of the Notitia, we should not use the terms praesens or praesentalis as a magisterial title earlier than the close of the fourth century without proper acknowledgment of its un-technical nature. To explore the role of the magistri militum in the changing later Roman Empire, this monograph will synthesize chronological narrative, prosopographical investigation, and the network analysis methodology. Chapter One will use the concepts of archontology – that is, the study of historical offices and the people who held them in a narrative, progressive form, to describe the magistri from their first appearances under the heirs of Constantine I through the reigns of emperor Jovian, covering the years 341 to 363. The depth of source material available for this period we will allow us to deduce the nature of the magisterial office and what the scope of its duties were intended to be. By exploring in-depth the usurpations of Vetranio and Silvanus we will better understand the methods available to the early magistri to express their ambition and power and how these rebellions will influence later generals. How the magistri were used by Julian to strengthen his regime will also be examined. It will be argued that in these years the magistri generally obeyed the emperors, and their powers did not overshadow imperial authority. Chapter Two will then continue the chronological perspective from Valentinian I’s accession in 364 and proceed until the death of emperor Theodosius I in 395. This chapter will endeavor to describe how the power of the magistri began to diverge between the east and the west. The Gothic war along the Danube frontier (376–382) prompted an enlargement of the eastern magisterium, which caused the command over the armies to be partitioned amongst multiple generals. This allowed officials in the civil administration to counter and overshadow the influence of the magistri. Conversely, in the west, the magister Merobaudes was able to take full advantage of a succession crisis that occurred in 375 to dramatically increase his own position by installing the figurehead child-emperor, Valentinian II. Merobaudes was followed by Bauto, Arbogast, and Stilicho, non-Roman – or ‘barbarian’ – magistri who continued to dominate the west-

13

This was even noted as rare by Mommsen 1901, 532 n. 4, but the term has continued to be used.

Introduction

17

ern emperors and all western affairs. These first two chapters will also delineate the responsibilities and powers of the magistri in this period. The magistri militum of this period formed complex relationships with subordinate officers, civilian officials, and other important members of the Roman world, and the systems of patronage that were developed played a critical role in the failure or success of different generals. Chapter Three will thus use network analysis to assess these connections. This methodology has only recently begun to be applied to studies in Late Antiquity, and so this chapter will explain the principles and theories that will be utilized to explore how networks of relationships affected the magistri. Through two case studies, it will be shown that those who were able to utilize networks of relationships effectively had much greater success than those who did not. The final chapter will contain two prosopographical studies on the fourth-century magistri militum. Rather than solely focusing on the most well-documented and well-studied individuals, this chapter will examine all the magistri and draw conclusions on them as a collective. The career paths that brought a man to the magisterium will be reconstructed and compared to the earlier and later career paths to determine if there are any distinguishing features that further illuminate the nature of the magisterial offices, and why particular magistri came to such heights of power. The second section will examine the origins and identification of the magistri. This analysis will be framed in light of the ongoing debate over the barbarization of the late Roman military. Accompanied by graphs and tables, these prosopographical analyses will focus on delineating the important characteristics shared by the magistri. Three diverse but symbiotic approaches to historical investigation are thus synthesized in this work. By combining the statistical results of the prosopography with the context gained from the narrative and network analyses, it will be identified how certain decisive moments in the fourth century’s military history were taken advantage of by the magistri to impact greatly the nature of their role in politics. These successive events compounded the power of the magistri until they became the most influential individuals in the western Empire, overshadowing the emperors themselves. Furthermore, unique traits can be identified for the magistri who gained the greatest degrees of power, suggesting that they were somewhat predisposed to take advantage of these decisive moments. Identifying these traits can help us better understand why these developments occurred.

State of the Question Peter Brown’s 1971 work on the religious and cultural developments of Late Antiquity invigorated interest in this era and reframed the way scholars envisioned it, yet the last major study on the magistri militum predates Brown by a year.1 The intervening decades have produced excellent works on the fifth-century magistri, as well as many works on the late Roman military in general, and even on individual magistri, but the study of the impact of fourth-century military leadership has slowed significantly. The absence of recent broad and critical coverage of the magisterium has made it commonplace for works not directly addressing military issues to leave the magistri on the sidelines of Late Antiquity. This shows itself in cases such as the civil war of the early 390s, which is often framed as a contest between emperors Eugenius and Theodosius even though the former was only a puppet controlled by the magister militum Arbogast.2 This general should be seen as holding as much control over the western Empire as we are more often inclined to see with figures like Ricimer in the fifth century. It has even been said that there is no benefit from considering the motivations of military leaders.3 The role of civilian bureaucrats in the nomination of new emperors has been emphasized on the basis that military officials rarely tried to put themselves on the throne, so they must not have held much sway.4 This shows a lack of understanding of the motivations of the magistri militum and how they reshaped the imperial office into one that they could control without assuming the risks of the imperial position themselves. This topic thus dearly needs a revisit to bring our understanding of the magistri up to the modern standard. The historical importance of the magistri ensured they featured in the earliest works on the end of the Roman Empire, but it was Theodor Mommsen who first conducted an investigation of the military magisterium itself.5 He approached the generals with a macro, institutional perspective, where changes in the office were seen as legal man-

1 2 3 4 5

Brown 1971; Demandt 1970. Lee 2007, 9; Christie 2013, 940; Elton 2018, 114, 117; Szidat 2010, 239. Cf. Kulikowski 2019, 92; Meier 2020, 44. See Chapter 2.6 for more. Lee 2007, 67. Szidat 2010, 107, 140. Mommsen 1889. Some of these early works include Tillemont 1701–1704; Gibbon 1781.

State of the Question

19

dates coming down from the emperors, and we will see how this has dictated the orientation of almost all ensuing studies. From the evidence in Zosimus and John Lydus, Mommsen deduces that the positions of magister peditum and magister equitum were created by emperor Constantine I (306–337) as a continuation of emperor Diocletian’s (284–305) removal of military authority from the powerful Praetorian Prefects to reduce their ability to usurp imperial power.6 Mommsen continued his description of the magisterial office by drawing heavily on the Notitia Dignitatum. Because the western list places magister peditum praesentalis first, Mommsen concludes that the magister peditum outranked the magister equitum in the west, and he assumes that this was the original hierarchy created by Constantine.7 He further argues that emperor Theodosius I (379–395) reformed the system in the east and made all his magistri equal to one another, and he believes this explains the differences both between the different magisterial titles and the different number of magistri between the east and the west.8 Reliance on the Notitia by Mommsen is rather unfortunate given the challenges that come with this source. It is increasingly believed that the Notitia was created with an ideological, rather than administrative purpose. Furthermore, at best it only offers a single frozen snapshot in time, while the military magisterium was constantly changing in practice. Thus, while not deliberately misleading, the intention of the Notitia appears to be neither accuracy nor technicality, and thus the order different offices are listed in is slim evidence for deducing a bureaucratic hierarchy. A later work of Mommsen’s seems to have attempted to find a new, non-institutional perspective by constructing a biography of the fifth-century magister Aetius, but half the essay became a revision of his earlier work on the institution as a whole.9 Otto Seeck, a protégé of Mommsen, continued the same macro-approach to the magisterium in an attempt to refine the creation date of the magistri militum.10 He postulates that Constantine’s incentive was protecting his sons from usurpation as they became emperors. The office thus may have appeared when Crispus was appointed Caesar in Gaul in 318. A few years later, Arthur Boak examined the developments in the differing titulature of the magistri, as well as the chronology of the creation of regional commands.11 Boak heavily incorporated the Codices Theodosianus and Justinianus, but like Mommsen, he also relied upon extrapolating information from the Notitia. Research into the magistri was continued by Ernst Nischer, who summarized the roles of the different positions and added to the discussion of how each transformed, while Ernst Stein made further deductions on the creation of different magisterial of-

6 7 8 9 10 11

Mommsen 1889, 260, citing Zos. 2.33.3 and Joh. Lyd. de mag. 2.10. Mommsen 1889, 262–64. Mommsen 1889, 265. Mommsen 1901. Seeck 1894. Boak 1915, 118–37.

20

State of the Question

fices.12 Wilhelm Enßlin challenged Mommsen’s conclusion that the magister peditum began as the highest ranked office, instead arguing that the magister equitum was superior on the basis that earlier sources almost always list equitum generals before their peditum colleagues.13 He also added that we do not know of any magistri equitum who became magistri peditum, but he believes we can identify two magistri peditum, Flavius Sallustius Bonosus and Victor, who became magistri equitum. On the basis that these men had done nothing to be demoted, he concluded that this change must have been a promotion upwards, and therefore the equitum position was the higher one. However, it is now believed that the man identified as Flavius Sallustius Bonosus is actually the conglomeration of two distinct people, Flavius Iulius Sallustius and Flavius Bonosus.14 Furthermore, it will be argued below that we cannot be certain that Victor was ever a magister peditum.15 Thus, Enßlin’s arguments were based on flawed information. A few years later, André Hoepffner authored a response to Enßlin, and he argues that rather than indicating a hierarchy, equitum preceding the word peditum might be a result of something as simple as alphabetic order, and he cautioned against some of Enßlin’s other arguments.16 This led him to the conclusion that the magister equitum and peditum were probably identical in authority, which would make sense in regards to both the Roman preference for collegiality, and Constantine’s intention of limiting the power of the generals – making some generals more powerful than others would not effectively reduce their power. Hoepffner also argues that the equalization of the magisterial hierarchy occurred in the reign of Valentinian I (364–375), rather than Mommsen’s placement with Theodosius I. Although the magistri were not the primary focus in the works of Denis van Berchem and Dietrich Hoffmann, they further developed our understanding of the wider military reforms of the period, although some of their conclusions have since been updated.17 Herbert Nesselhauf continued to rely heavily upon the Notitia in his study of the western Empire’s administration, and Wilhelm Heil further described the military reforms of Constantine.18 Scholarship on the late Roman military was then benefitted by Alexander Demandt’s 1970 contribution to one of the standard reference works for the study of Greco-Roman history in Late Antiquity, the Realencyclopädie der classischen Altertumswissenschaft, which remains the authoritative work par excellence.19 Demandt collates and critiques previous scholarship, and acknowledges the problems in heavy 12 13 14 15 16 17 18 19

Nischer 1928; Stein 1928, 186–88, 366–368. Enßlin 1930, 312–13. Also see: Enßlin 1931a, 1931b. See page 34. See pages 69–70. Hoepffner 1936, 487–95. Berchem 1952; Hoffmann 1969–70. For developments on Hoffmann’s ideas on the seniores and iuniores unis, see: Drew-Bear 1977; Scharf 1991; Nicasie 1998, 24–35. Nesselhauf 1938; Heil 1966. Demandt 1970.

State of the Question

21

reliance upon the Notitia. The issues of previous historians are remedied by affording equal weight to other sources. The comprehensiveness and breadth of this work created an excellent platform for future authors to incorporate the magistri into many works on varied aspects of the late Roman world. Nonetheless, problems persist. Sources in languages other than Latin and Greek were not incorporated, and he seems to have remained limited to previous foci. Demandt surmises his most significant conclusions to be further answers to the original questions Mommsen posed: the titles, hierarchies, and the creation of new offices. Since Demandt, there has been no dedicated study with the same scope. Hoffmann provides another discussion of the different offices that held regional commands in the late Roman military, including how the regional magisterial commands developed over the fourth century and why new titles appeared.20 However, despite Demandt’s expression of reservations, Hoffmann still heavily relies on the Notitia, asserting that the military administration in the western Empire remained essentially unchanged from Constantine to Theodosius I and that there were always two magistri in every imperial court.21 Throughout this book it will become clear that the military magisterium was actually employed in an ad hoc, ever-changing manner, with the placement and number of magistri continually shifting. Manfred Waas and Hans Teitler both studied specific subsets of the magistri, and the generals feature prominently in studies such as Raban von Haeling’s work on the religious associations of late officials.22 Giovanni Cecconi looks at the magister militum and the magister officiorum and how the separation of military and civil authority, as well as that of palatine and regional postings, was not actually as starkly divided as previously thought.23 Instead, they were fluid and changeable, being more or less separated in different periods depending on the actions of the individuals. These are ideas that will be explored further in the body of this book. Doug Lee has recently emphasized the importance of military generals to the political stability of the late Empire.24 Their control over the armies meant that securing their loyalty was paramount to the emperors, and they tried to win the magistri and their soldiers through donatives, tax privileges, the acclamatio, and use of fellowship language. There were also several forms of reprimand and punishments that emperors utilized to keep their officers in line. Marc Landelle has recently argued that the magistri may have been created later than previously thought, around 328, to relieve the Praetorian Prefects of their military duties as they were becoming overwhelmed with the broad scope of their positions.25

20 21 22 23 24 25

Hoffmann 1974. Hoffmann 1974, 387–94. Waas 1971; Teitler 1989; Haeling 1978. Cecconi 1999. Lee 2015. Landelle 2016.

22

State of the Question

It is unlikely that this date, or any of proposed alternatives, will be conclusively settled on, as they all have minimal, if any, solid evidence, but Landelle’s idea is compelling. Landelle additionally takes a different view than most historians to contend that the magister peditum and equitum had different duties, namely that when not actively campaigning, the magister peditum oversaw the infantry, while the equitum general oversaw the cavalry.26 He makes this argument on the basis that the logistics of cavalry units, incorporating both men and animals, are significantly more complex than the logistics of infantry, and therefore they needed a dedicated commander. However, we must remember that cavalry was only a small percentage of the Roman military, and the logistics for managing tens of thousands of infantry probably matched, if not exceeded, the difficulty of managing a much smaller number of cavalry. Furthermore, magistri with either equitum or peditum titles often operated alone away from the emperors or a counterpart magister, and therefore must have managed the logistics for both cavalry and infantry simultaneously. This argument is therefore unconvincing. Studies of the late Roman military as a whole bear mentioning here, as they have remained popular with scholars and non-expert audiences, and have contributed to our understanding of the military-political landscape.27 David Potter and Michael Kulikowski have offered fantastic insights on the growing autonomy of the civilian and military bureaucracies in the later fourth century, a framework that is drawn upon in this book to illustrate the increasing power of the magistri.28 Lee also examines the social impact of late Roman warfare.29 These are just some of the highlights of the many works on the wider role of the military in the late Roman world. Overall, the fourth-century magistri have mostly been approached with macro-analyses of the institution and its organization. The changes that occurred during the fourth century have been concluded to be products of legislation and intentional reform by the emperors. Power and influence have been attributed to a person’s position as magister equitum or peditum, whichever the author believed to be of higher rank, or to the status as a praesentalis or regional general. A number of articles and smaller works have taken a different tack and concerned themselves with issues of individual magistri. Bruno Bleckmann, John Drinkwater, and Alan Dearn have successively reassessed the portrayal of the magister Vetranio and his rebellion.30 E. A. Thompson devotes a chapter of his monograph to revealing the problems in Ammianus’ portrayal of the magister Ursicinus, while Drinkwater reconsiders this general’s role in the rebellion of the magister Silvanus.31 David Woods 26 27 28 29 30 31

Landelle 2014. Ie. Dixon and Southern 1996; Elton 1996; Nicasie 1998; Goldsworthy 2003; Hebblewhite 2017; Elton 2018. Potter 2004; Kulikowski 2019. Lee 2007. Bleckmann 1994; Drinkwater 2000; Dearn 2003. Thompson 1947; Drinkwater 1994.

State of the Question

23

has also published numerous articles examining the accuracy of Ammianus’ portrayal of military officers.32 Authors such as Ian Hughes, Timo Stickler, Penny MacGeorge, Jeroen Wijnendaele, and Michael O’Flynn have also made excellent contributions with monographs on the most famous and well-documented fifth-century magistri.33 These works made great progress by dispensing with the institutional approach and establishing the agency of individuals. By generally focusing on the most prominent magistri, however, the lesser-known generals and the information they can provide about the office is ignored. Nor is the accumulative effect of influential individuals tracked across a broad timespan. The role of personal relationships with the emperors, colleagues, civil bureaucrats, religious leaders, foreign leaders, and other powerful individuals has also been poorly understood. Some of these problems have been addressed by scholars, although none are entirely satisfactory for the fourth century. For example, the topic of personal relationships was undertaken by David Parnell for the generals of the fifth- and sixth-century Byzantine Empire.34 Meaghan McEvoy’s discussion of the late Roman child-emperors also helps describe the developments that occurred in this timeframe, although the magistri themselves are not the focus of the text.35 The same can be said about Mischa Meier’s 2020 reassessment of the so-called Migration Period, which, while not focused on the magistri, does give due credit to the magistri militum and their critical role in the developments of the fourth and fifth centuries.36 This current work will try to avoid repeating the same institutional, top-down perspective of the military magisterium. Only limited space will be devoted to contributing more to questions such as the hierarchy of individual offices and postings, or the date of the creation of the magisterium. It is not my intention to assert that there is nothing left to say on these matters, but I believe we will be best served by concerning ourselves with a different focus. Although the emperors did have the legal authority to make appointments and changes to the military ministry, both to its members and the powers of the different offices, at some point this shifted, as demonstrated by Valentinian II in the introduction, and the emperors no longer made those changes but instead the magistri determined their own fortunes. The role of individuals in this long process is so far poorly understood. Employing the frameworks and methodology explained in the introduction to the magistri will modernize the research of the office and address the prevailing lacunae in the scholarship. By doing so, new considerations will be offered for the impact of the magister militum on the late Roman world.

32 33 34 35 36

Woods 1995; Woods 1997; Woods 1999; Woods 2001; Woods 2010; Woods 2016. O’Flynn 1983; Stickler 2002; MacGeorge 2002; Hughes 2010; 2012; Wijnendaele 2015. Parnell 2017. McEvoy 2013. Meier 2020.

Sources The sources for the fourth-century magistri are mostly literary, and often leave much to be desired. Ammianus Marcellinus’ work, the Res Gestae, is by far our best source. Designed as a continuation of Tacitus’ Annales and Historiae in what is called the ‘classicizing’ style, Ammianus began with the reign of Nerva (96–98) and continued down to the Battle of Adrianople in 378. Unfortunately, the first thirteen books of the Res Gestae have not survived into modern times, and the remaining sixteen books begin in the early 350s, approximately beginning with the author’s professional career. Ammianus’ service in the protectores allowed him to meet personally and serve under multiple magistri, and he has unparalleled insights into the military, especially when it comes to campaigns, battles, and the distribution of forces.1 For example, Ammianus was present during the Persian siege of the city of Amida in 359, and he provides a detailed account of the operations and actions of his commanding officer, the magister militum Ursicinus. Furthermore, even though he spent the majority of his time in the east, he clearly utilized knowledgeable sources of information on occurrences in the west.2 Although aspersions have been cast regarding his accuracy, Ammianus’ work contains two-thirds of all unambiguous original Latin titles for magistri in our period.3 Unfortunately, because he wrote during the time of emperor Theodosius I, he is sensitive towards topics that might have offended the emperor.4 Furthermore, Ammianus is guilty of ignoring some details, especially when it comes to Christians and religious affairs as they were deemed inappropriate to the conventions of the genre in which he was writing, and he is considered to be overly favorable to those he admired, such as Ursicinus and emperor Julian.5 In recent decades, several works have appeared that criticized Ammianus’ replacement of fact with literary devices, and some scholars consider him to be more manipulative than is commonly believed.6 Thus, throughout this 1 2 3

Barnes 1998, 1; Thompson 1947, 2–12; Matthews 1989; Trombley 1999, 20–21. Thompson 1947, 44. Out of 92 total references to formal and unambiguous Latin titles, 59 come from Ammianus, 24 from the Codex Theodosianus, seven from other literary sources, and two from inscriptions. 4 See note 137n99. 5 Lenski 2002a, 4. 6 Barnes 1998; Kelly 2008.

Sources

25

book, we will be utilizing Ammianus and discussing the validity of his testimonies in light of these factors. The next best narrative history covers the entire period under consideration, but is a significant step down in reliability. In the sixth century, the financial officer Zosimus composed a summarized copy of the now only fragmentary fourth-century history written in Greek by the polemically anti-Christian Eunapius.7 It is often muddled and inaccurate, but it remains a useful comparison with Ammianus, and is sometimes the better source. The fourth-century Latin histories of Aurelius Victor and Eutropius, both of whom probably used the now lost so-called Kaisergeschichte (‘History of the Emperors’) are also useful for the early part of our period. The twelfth-century Greek history of the retired imperial secretary and monk Ioannes Zonaras, although written a centuries after our period, used both Christian and secular contemporary sources that are now lost to us, and so represents a useful source of information.8 Several technical documents offer important, if difficult, information on the magistri. The most important to the present book is the Codex Theodosianus. This compilation of laws covering the period of 313 to 437 was created on the orders of emperor Theodosius II (402–450), and it was promulgated in 438.9 It contains a quarter of all attested magisterial titles in our period.10 However, interpreting the Codex can be difficult. There is no context to the entries, so it is often complicated to decide to what extent the magistri were requesting the emperor to make a particular law, or how much they were simply being directed to enforce them. Furthermore, as it was composed in the fifth century, its compilers were more familiar with contemporary magisterial titles, and this arguably engenders some mistakes and overrepresentation of more generic titles for the fourth century.11 Another technical document that presents serious difficulties is the Notitia Dignitatum. A. H. M. Jones called it a “most valuable document”, while in more recent times it has been termed simply “peculiar”.12 This unique document lists the civil and military offices, along with their subordinate officers and military units, of the entire Roman Empire. The eastern portion is generally understood to be the original creation, and dates to the early 390s, and then at some point it was brought to the west where the western portion was added and updated into the 420s – unfortunately, the full extent of the updates is unclear.13 Furthermore, when one military unit was transferred from one place to another, it is clear that they were 7 8 9

Ridley 1972; Cichocka 1990. Bird 1994; Bleckmann 1992; 2012, 26. Smith 1875, 302–3; McEvoy 2013, 17. For a more extensive discussion of the Codex Theodosianus in the wider context of imperial pronouncements and missives, see: Corcoran 2000; 2014. 10 See note 24n3. 11 Demandt 1970, 568. For some mistakes, see notes 49n146, 80n29. 12 Jones 1964, 347; Kulikowski 2000, 358. 13 Jones 1964, 347; Ward 1974, 408; Mann 1991, 215–19; Brennan 1996, 164–165; 1998, 35; Kulikowski 2000, 358–77; Scharf 2005, 3; cf. Zuckerman 1998, 143–147.

26

Sources

not always removed from their first posting, leading to an inflated sense of the size of the army. Peter Brennan argues that, when considering the other manuscripts in the Codex Spirensis in which the Notitia Dignitatum survived into modern times, it must have been a singular ideological text intended to emphasize the cohesion of the Empire in response to a time that it was fragmenting and growing further apart; historians are divided on whether the ideology is attributable more to Theodosius I, Valentinian III, or even the court of Charlemagne.14 Ultimately, it seems the Notitia is barely useful for the period it covers, with increasing difficulties the further outside those years one strays. Thus, it will only come into relevance for this present study towards the end of the period covered. The Libelus de Dilectu atque Exercitio Tironum, better known as Vegetius’ Epitome of Military Science, is another technical document that offers valuable, if opinionated, insight into the Roman military. Addressed to emperor Theodosius I as a handbook of advice regarding the Roman military, in reality it is more of a political and strategic manifesto.15 Vegetius believed that the army of his day was weak and failing, and wrote his text to explain to the emperor the need to implement his ideas of rigorous standards, trainings, and practices used by the legions of the Republic and early Empire to regain Rome’s military might. Additional valuable sources of information are found in the many ecclesiastical histories that cover the period, including the Latin translation and continuation of Eusebius’ original Historia Ecclesiastica (‘History of the Church’) by Rufinus (c. 403), and then those in Greek by Socrates (c. 446), Sozomen (448/49), and Theodoret (449/50).16 While these sources preserve the Nicene perspective, Philostorgius (c. 425), whose work survives in the ninth-century Byzantine patriarch Photius’ epitome, offers the Arian view.17 Although not strictly an ecclesiastical history, Orosius’ early fifth-century Christian-apologetic work, Historiae Adversus Paganos (‘History Against the Pagans’), also offers important insights. These works provide significant value to this project when they refer to the actions of the magistri in religious affairs, including participation in Church councils, personal worship, and the enforcement of the policies of the emperors. Although there are cases of partiality, the Church histories may be considered generally accurate and can also provide useful information on military matters.18 The many letters, orations, and panegyrics produced during this period, such as those in Latin by Latinius Pacatus Drepanius, Saint Aurelius Ambrose, Quintus Aurelius Symmachus, and Claudius Claudianus, or those in Greek by Saint Basil of Cae14 15 16

17 18

Mann 1991, 219; Brennan 1996, 169; Kulikowski 2000, 360. Milner 1993, xxix, xi; Reeve 2004, v–vi. On Rufinus, see: Thelamon 1981. On Socrates, see: Urbainczyk 1997; Leppin 1996. On Sozomen, see: Urbainczyk 1997; 2002. On Theodoret, see: Leppin 1996, 91–104; Wallraff 1997. On all these dates, see: Lenski 2002a, 5; cf. Leppin 1996, 273–82. On the interdependence of these sources, see: Lenski 2002a, 5. Bidez and Winkelmann 1972, cvi–cxlii. Urbainczyk 1997. See notes 91n111 and 99n177 for some of the problems with Theodoret.

Sources

27

sarea, Libanius, Themistius, and Saint Gregory of Nazianzus, are another source of important information.19 These works often provide extensive details on a particular event or person, and some letters and orations were even addressed directly to individual magistri. Unfortunately, no works written or dictated by the magistri themselves survive, so the works addressed to them are the closest we can get to personal correspondence. However, we must remain cautious, as these works were generally written with some ulterior motive. Furthermore, although the emperors probably did not directly control or dictate the contents of imperial panegyrics, the future prospects of the orator were intrinsically tied to the favorable reception of their speech, and so they were generally designed to please.20 Despite the difficulties, these allow us the opportunity to understand the perspective of contemporaries.21 The Palestinian Talmud offers some recordings of the magister Ursicinus but has rarely been consulted in works on the fourth-century magisterium.22 This work contains the oral teaching of the tannaim, the rabbinic scholars. It was heavily edited multiple times, the last edition being hastily completed after the Jewish patriarchate was dismantled in the early fifth century, so it does not follow a particularly clear system and can be difficult to understand.23 Nevertheless, it preserves information that no Latin or Greek source contains, such as Ursicinus’ involvement in the Jewish Revolt of 351. Hence, the Talmud remains an important complementary source for understanding the role of the magistri. Finally, to supplement this mostly literary foundation, archaeological data, as well as documentary (numismatic, epigraphic, and papyral) evidence will be used to develop a more complete understanding of the magistri, and will be especially useful in providing additional titles and evidence for regional postings. This is not an exhaustive description of all the sources used in this project, but it should be enough of an introduction to the intricacies and problematic nature of the most important source materials.

19 On Libanius, see: Petit 1955; Liebeschuetz 1972; Norman 1969–77; 1992. On Pacatus, see: Lunn-Rockliffe 2010. On Symmachus, see: Matthews 1975; Pabst 1989. 20 Nixon and Rodgers 1994, 26–35. 21 Bleckmann 2012, 24. 22 Often called the Jerusalem Talmud, although it is not associated with Jerusalem. See: Schiffman 1991, 227; Jacobs 2008, 3. 23 Schiffman 1991, 227–30.

Chapter One The Magistri Militum from 341 to 363 CE 1.1 Introduction During the middle of the third century, for roughly the five decades between 235–284, in a period known as the Crisis of the Third Century, the Empire was beset by plague, depopulation, a decline of the prevailing slave economy, coinage debasement, and an increase in brigandage, all calamities which exacerbated the damage caused by successive foreign invasions and civil wars.1 This resulted in the greatest military crisis the Empire had yet faced. The Roman frontiers grew porous, and attacks came across the Rhine and Danube Rivers, the Black Sea, and from the new Sassanid Persian Empire. By necessity rather than any strategic choice, the Roman army fell back from the frontiers and formed temporary centralized cores around the emperors, which could only provide security in their immediate locale.2 A outburst of regionalization led to secession of territories and a division of the empire into three parts, and a redoubt of Roman territory north of the Danube was permanently abandoned. All these factors interlocked and compounded one another, and restoration of cohesion and security required the massive efforts of a series of militarily active emperors who reunified and restored the stability of the Empire during the 260s and 270s. The pressures exerted on the Empire and the efforts to restore it led to extensive overhauls and transformations in imperial governance, from the emperor himself, to taxation, the military, religion, social order, and provincial administration. Emperor Diocletian formulated an effective new defensive system based on fortified border defenses and the establishment of a comitatus army for each of the four vigorously active emperors of the new political system he created, the Tetrarchy.3 From their capital cities in important frontier provinces, the Tetrarchs were able to reassert Roman control of the limes and defend

1 2 3

For general works on the period, see: Potter 2004; Bowman, et al. 2005; Corcoran 2012, 35–39. In defense of the use of the term ‘crisis’, see: Liebeschuetz 2007. Luttwak 2016, 146–216. Petit 1974; Watson 1999; Williams 1985; Barnes 1981; Corcoran 2008; Elton 2018, 90, 97. For an excellent summary of Tetrarchic period, see Corcoran 2000, 5–9.

Introduction

29

themselves, at least for a time, from the incessant civil wars that had plagued the Roman world.4 These changes occurred concurrently with another development in the military’s administration. During the Principate, the regular upper echelons of military command had been reserved for men of the hereditary senatorial rank. Senators with the title legatus commanded individual legions, and senatorial governors were used to command field armies.5 However, sometimes particular situations called for the creation of extraordinary commands, such as cases like the senator Gnaeus Domitius Corbulo in 63 CE, or Gaius Avidius Cassius in 169, who were both given special jurisdiction over large parts of the eastern Empire. These ad hoc military commands indicate that the military administration was not unwaveringly strict. Rather, the emperors could act with flexibility, and they often favored “pragmatism over traditionalism”.6 This elasticity enabled a slow but steady change that replaced senators with men from the equites Romanus, a non-hereditary class based on one’s wealth, and the ad hoc commanders were often Praetorian Prefects or the commanders of the legions I, II, or III Parthica.7 Although the fourth-century historian Aurelius Victor attributes Gallienus with finally removing military commands from senators, the flexibility of the military administration allowed the emperors to place trusted non-senators into formal military positions from as early as 87/88 CE.8 As the Roman Empire swelled across the Mediterranean, more officials and civil servants were required to manage the growing bureaucracy, many in positions of insufficient prestige for senators. These equites, or equestrians, were thus increasingly relied upon to take the majority of positions in the Roman government, leading to a gradual but major increase in their importance.9 During the Tetrarchic period, the bureaucracy was further expanded with the intention to increase the ability of the central government to harvest tax from the provinces. Moreover, despite the rampant fiscal inflation of the third century, the required property assessment was not raised and therefore the barrier to entry into the equestrian class became much easier to meet.10 These developments greatly expanded the number and importance of the equestrians. Eventually, veterans from regular soldiery were promoted directly into the equestrian military career path known as the militiae equestres.11 In the 250s, a new corps of junior

4 5 6 7 8 9 10 11

Kulikowski 2015, 140–3. Syme 1958, 239, 242–43, 645–48; Saxer 1967, 25–27; Bennett 2001, 194–95, 200–1; Birley 1993, 155–57. Davenport 2019, 506–8; Vervaet 2007. Howe 1942, 22; Speidel 1994, 99–100; Bingham 2013, 41; Davenport 2019, 521–23. CIL V 875 = ILS 1374; Davenport 2019, 488–502. Gallienus’ law: Aur. Vict. 37. See also: Keyes 1915, 49–54; de Blois 1958, 37–44; Le Bohec 2004, 124; Cosme 2007. Kulikowski 2016, 245–64. Dillon 2015, 45; Davenport 2019, 558–60. See Chapter 4.2.

30

The Magistri Militum from 341 to 363 CE

military officers appeared, called the protectores.12 These soldiers seem to have been in part bodyguards, and part staff officers, and this institution granted equestrian status upon entry. Thus, through promotion into the protectores, a path was created for theoretically any soldier to advance to the upper ranks.13 These developments sounded the end of traditional, hereditary senatorial control of high military command in the 260s, shortly after the creation of the protectores.14 The highest military commanders of the fourth century did receive the senatorial status of vir clarissimus, but it was a reward of, rather than a barrier to, their service.15 Thus it seems that the protectores established a wider precedent for military promotions to also grant social status. Constantine I is thought to have expanded this equestrian access to senatorial status as a way to bind the bureaucracy more tightly to him personally.16 By flooding the senate with many newly minted members, the body began to owe its loyalty to him rather than to their hereditary status. The highest military commanders also received the status of comite primi ordinis (‘companion of the first class’), indicating their closeness to the emperor.17 Concurrently, the senate also lost its role in the elevation of a new emperor. Their acknowledgment was no longer needed, and recognition by the army replaced it.18 The army itself, however, did not choose the candidate and was not able to refuse a candidate presented to them.19 The real power to choose an emperor was in the hands of any incumbent emperors, specifically the senior Augustus, and lacking them, we will see how the leading civilian and military officers had the power to decide the succession. In time, emperor Constantine I overthrew the other Tetrarchic rulers in a series of civil wars, ultimately uniting the Empire under his sole reign.20 New military structures emerged from this series of civil wars (306–324). These arrangements are attributed to Constantine by Zosimus in a polemical, anti-Christian attack, but he was probably not the sole progenitor of the final system.21 The new army was divided into three classes. The first, variously called ripenses, burgarii, or most commonly in modern literature, limitanei, was in charge of protecting the border that they garrisoned. These soldiers had lower physical standards, less service length, and reduced tax benefits compared to the field units, however there is no discernable lack of fighting capabilities.22 They 12 13 14 15 16 17 18 19 20 21 22

Speidel 1978, 130–33; 1986 451–52; Davenport 2019, 533. Davenport 2019, 537–38, 544. Davenport 2019, 534–36. It is unclear exactly when the magistri militum obtained senatorial status, but it may have been from their first appearances: Lee 2015, 107–8. Dillon 2015. Demandt 1970, 565. Szidat 2010, 44. Szidat 2010, 77. Lenski 2008; Barnes 2014. Zos. 2.33.3. Isaac 1988; Berchem 1952; MacMullen 1963, 153; Elton 2007, 273–74.

Introduction

31

were commanded by duces, and there is some indication they were, or became, parttime farmers in some areas.23 The remaining two parts of the new military composed the field armies. They were the comitatenses, and the palatini, the latter having the greater privileges and remuneration.24 At least in the period of the Notitia Dignitatum, the palatini were not strictly attached to the emperors or their palaces, but were deployed alongside the comitatenses in the militarily important frontier regions.25 In line with the greater prestige of the field armies, their commanders, the comites rei militaris, were of higher rank than the duces.26 Vegetius writes that a comes rei militaris commanded roughly ten-thousand infantry and two-thousand cavalry, although we should be careful with such numbers, as this may be only the on-paper, official numbers under a comes’ command.27 This system created a new, deeper defensive style, which combined border defenses with reserve armies stationed deeper within the Empire.28 Along with a continued build-up of frontier fortifications, the new military structure afforded the Roman emperors greater flexibility to respond to foreign invasion, create expeditionary campaign forces, and deal with civil war. Because the Praetorian guards had supported Constantine’s enemy Maxentius, and likely also because of their long history of unreliability, he completely disbanded them as a force.29 Constantine kept the Praetorian Prefects, but he replaced the previously equestrian holders of the office with senators, and completely stripped them of military authority, leaving them as the highest civil servants in the administration.30 The armies still required senior commanders, and during this period a new office emerges, the creation of which is generally attributed to Constantine, but much like the new tripartite division of the army, the new office might have had a direct precedent under his precursors and contemporaries.31 Constantine certainly had a fondness for Republican titles, as he had already revived those of censor and patricius, and this may have inspired him to revitalize the old office of magister equitum, the second-in-command of a Roman Republican dictator.32 In light of the Roman tradition of collegiality, his creation of a second office, the magister peditum, may have been intended to spread control of the armies between two officers to limit further the ability of generals to usurp the imperial position.33 These generals could only be outranked by the emperor and,

23 Jones 1964, 649–53; Elton 2007, 274. 24 Jones 1964, 608–9; Elton 1996, 94. 25 Eg. Not. Dig. Occ. 5–6, Or. 8–9. 26 See Chapter 4.2 for more. 27 Veg. 3.1. 28 CTh 7.20.4; Pan. Lat. 12.3; Jones 1964, 97–98; Luttwak 2016; Brennan 2007. 29 Zos. 2.33.3. 30 Elton 2018, 65. 31 See pages 28–32. 32 Demandt 1970, 560; Cecconi 1999, 73. 33 Dixon and Southern 1996, 148.

32

The Magistri Militum from 341 to 363 CE

at times, each other. Both the duces and comites rei militaris reported to the magister militum in their local region. The only soldiers outside their purview were the scholae, the new cavalry imperial guard units which replaced the Praetorian guards, who were under the authority of the magister officiorum, a civil servant.34 Thus, the magistri were like the ad hoc commanders of the Principate, with command over many thousands of soldiers – Vegetius records they could command double that of a comes rei militaris, up to twenty-thousand infantry and four-thousand cavalry.35 Once again, we should not take these numbers as absolute fact, but as illustrative of their great authority, double that of the next highest rank. The new magisterial generals had no authority in the civil realm of logistics, an intentional design intended to hinder their ability to organize the logistics of a rebellion, although we will see that this did not entirely prevent usurpations, and over time they gained greater powers, especially in the judiciary and politics.36 Throughout the following chapters, more will be stated on the developing roles of the military magistri. 1.2 The First Attested Magistri It is an unfortunate result of our poor records of the early fourth century that, despite the confidence with which it is believed emperor Constantine I created the office, not a single magister militum who served him is known with certainty. He is said to have raised barbarians to the highest distinction in the Roman administration, the ordinary consulship, and the most common path for a barbarian to be raised so high was the magisterium.37 The consuls of 338, Polemius and Ursus, the latter of whom was possibly identified as a στρατηλάτης, potentially fill this lacuna.38 Στρατηλάτης, or στρατηγός, were Greek generic, non-technical term for a high-ranking military officer which often, but not always, designates someone to be of magisterial rank. This situation occurs for several men under examination here, and we will use other evidence to acquire more certainty about their positions. For example, magistri militum were the only fourth-century military officers to be consuls. Thus, if Ursus was a στρατηλάτης and received the consulship, then he was most likely a magister militum, although it is possible that he was only promoted to the magisterium after the death of Constantine in 337, and thus there are still difficulties in making this assumption. Because of the uncertainties of the links between this evidence, Polemius and Ursus are left out of the list of magistri militum.

34 Woods 1997; Barlow and Brennan 2001. 35 Veg. 2.9, 3.1; cf. Cecconi 1999, 80. 36 Lee 2015, 103–4; Jones 1964, 447–62. 37 Lee 2015, 107. See pages 160–64 on the importance of the consulship. 38 Polemius: PLRE I, 710. Ursus: PLRE I, 989.

The First Attested Magistri

33

After the death of Constantine I on 22 May 337, the Empire had four Caesars spread across its breadth – three sons and one nephew of the expired Augustus. However, shortly afterwards, Constantius II orchestrated a massacre that included the nephew, Dalmatius, and several other male family members that might have threatened the reigns of the sons through usurpation or by being overbearing advisors.39 The remaining three Caesars then came together in Sirmium (Sremska Mitrovica, Serbia) to discuss and coordinate their next steps. They were faced with a crisis – precedent had established that the promotion of a Caesar to Augustus required the presence of an Augustus or the approval of the reigning senior Augustus. If not, the resulting promotion could be regarded as illegitimate. Thus, the three brothers spent months in negotiations amongst themselves and certainly other powerful individuals and groups, before being proclaimed as Augusti by the army. They had formulated a tripartite division of the Empire: Constantine II controlled Britain, Gaul, and Spain from the frontier city of Trier; Constans ruled Italy, Africa, and Illyricum from Milan; and Constantius II reigned over Thrace, Anatolia, Egypt, and the eastern frontier with the Sassanid Empire mostly from Antioch (Antakya, Turkey) but also his father’s capital city, Constantinople (Istanbul, Turkey). This division of the Empire remained the dominant one through the period. For much of the period, the regions of Britain-Gaul-Spain would be joined with that of Italy-Africa-Illyricum, heralding the eventual division of the Western and Eastern Roman Empires, but it was not a formal separation by any means in the fourth century. The first identifiable magistri that emerge under these sons of Constantine were involved in the religious affairs of the Empire. When Constantinople’s Arian bishop, Eusebius of Nicomedia, died in 341, his followers ordained Macedonius I as his successor, while his Nicene opponents instead ordained Paul I.40 This led to a bitter religious conflict in the city, forcing the Arian emperor Constantius II to command his magister equitum Hermogenes to quell the unrest.41 Hermogenes was ordered to expel Paul and support the ordination of Macedonius; however, the populace resisted. When the magister attempted to use his soldiers to oust Paul with force, the people torched Hermogenes’ residence, executed him, and then dragged his corpse through the streets.42 Salia then appears in the service of emperor Constans. Salia is not explicitly stated to be a magister militum, as he is predominantly recorded by Greek sources. Like the case of Ursus, these represent him as a στρατηγός. Salia was given a consulship in 348, 39 40 41

42

Burgess, 2008; Humphries 2020, 157. Soc. 2.12. Soc. 2.13; Soz. 3.7. Jer. Chron. s. a. 342 calls Hermogenes “magister militiae”, but AM 14.10.2 gives him the title “magister equitum”. Balkanska 1980 identifies him as the same man who restored a building in northern Thrace, which would indicate he had already taken up his position there, but this has been adequately refuted by Krawczyk 2016. Hermogenes owned a house in the city of Tyre, which may indicate he was native to that city (Lib. Ep. 828). Soc. 2.13; Soz. 3.7; Jer. Chron. s. a. 342; AM 14.10.2.

34

The Magistri Militum from 341 to 363 CE

and so we can be fairly confident he was of magisterial rank.43 In 344, the Nicene Constans ordered Salia to escort two western bishops to Constantius’ court in the east.44 The bishops were to demand that Constantius acknowledge the crimes of the Arian ex-bishop of Antioch, Stephanus, and to agree to restore the Nicene bishop Athanasius to the See of Alexandria. The bishops were to threaten war if Constantius did not acquiesce, so we can deduce that Salia, and the soldiers and junior officers that accompanied a general of his rank, probably escorted the bishops to protect them and give weight to the threat.45 This was effective, and Constantius gave in.46 The activities of Hermogenes and Salia are informative about the important role the magistri played in the Empire’s religious affairs, but they should not be interpreted as suggesting that the magistri were initially intended to mostly have a religious role. These early events are recorded in the ecclesiastic accounts that were concerned with documenting the history of the Church, and they thus recorded the generals only in this context. Both Hermogenes and Salia, and the other magistri lost to time, would have commanded armies against Rome’s enemies, but the ecclesiastic historians had little interest in recording purely secular events. The year 344 was also one of the stranger years for recording of the Roman consuls, specifically in regards to two magistri militum. Flavius Domitius Leontius, the Praetorian Prefect of the east, was recognized as consul across the Empire for the entire year, and Flavius Iulius Sallustius, a magister peditum, was recognized for the entire year in the east, but only from 28 July in the west.47 For the first part of the year, until at least 3 May, Flavius Bonosus, a magister equitum, was recognized as consul in the western half of the Empire.48 Unfortunately, nothing more is known of Sallustius, so there is little else that can be said about him.49 More is known of Bonosus, and he presents an intriguing case. In addition to the short consulship, a law was issued with his name and the title of magister equitum in Syria in 347, and some bricks from Pannonia appear with the stamps “Bono(sus) Mag(ister)” and “Bono(sus) p. v.[…].”50 This has presented the curious mystery of how one man could be a consul for part of the year in only part of the Empire, a magister militum in both the west and the east, and have For Salia’s consulship in 348: CTh 10.1.6, 10.14.2; Athan. Fest. Ep. 20.3.4; BGU 2 405.1; ICUR 1 97 = ILCV 1267; Bagnall, et al. 1987, 230. 44 Theod. HE 2.8.54. 45 For other example of magistri being accompanied by an entourage of soldiers and aides, see: Philost. HE 8.1; AM 14.9.1, 11.5; cf. Them. Or. 16.208d. Also cf. the journey of Theophanes in the early fourth century, as discussed in Roberts and Turner 1952, 104–56. 46 Barnes 1993, 87–93. 47 Bagnall, et al. 1987, 222–23. For these dates, see: Salway 2008, 302–3. For Sallustius as magister peditum, see: P. Abinn. 2. 48 For Bonosus’ magister equitum title, see: CTh 5.6.1. 49 It has been thought that Sallustius and Bonosus were the same person, but this has been disproven by the publishing of a new papyrus (Enßlin 1931a, 102, 122; Hoepffner 1936, 488; Salway 2008, 304). 50 CIL III 4669a-b, 11376a-g, and CIL III 143603, ad n. 11376 (p. 232843). The law of 347: CTh 5.6.1. 43

The First Attested Magistri

35

the equestrian title vir perfectissimus even though a consulship grants senatorial rank. While it was not unheard of for a consul to be stripped of the office partway through the year, it had always previously been accompanied by damnatio memoriae, which would indicate a major loss of favor by the consul.51 Such was the situation only a few years earlier in 337 with Flavius Felicianus.52 However there is no evidence that this happened to Bonosus, removing this as a simple explanation and further clouding the conundrum.53 Benet Salway conducted the most thorough investigation of this issue, and he concludes that the mystery is most likely to be explained as a scribal error.54 He argues that Constantius actually appointed the magister peditum Sallustius as consul, but somewhere in the line of communication between east and west his name was confused with that of the magister equitum Bonosus, and Constans’ court received this incorrect information. Thus, when the mistake was realized it was rectified without need to shame Bonosus’ name with damnatio memoriae for a mere clerical error. Salway’s theory would also mean that senatorial rank was never conferred on Bonosus, and so the title vir pefectissimus on the Pannonian brick stamps could come from any point of his career. Salway offers that they may date to when Constantius gained control over Pannonia after the battle of Mursa (Osijek, Croatia) in 351, which would mean Bonosus’ entire career took place within the eastern army, a more likely scenario than a transferal from the west.55 David Woods rightfully points out that a scribal error like this should have been noticed and corrected by the imperial bureaucracy, however his counter proposal that Constantius initially wished for Bonosus to assume the consulship and communicated this to Constans but then changed his mind at the last moment to Sallustius, is even more doubtful.56 Woods’ proposal that Sallustius’ forces might have captured a city, perhaps in the year 340, and this could have changed Constantius’ mind for the consulship, stretches beyond the limits of conjecture. Furthermore, an eleventh-hour removal of consulship for no fault would certainly cause great upset to Bonosus and potentially even drive him to rebellion. In either case, we should discount the assigning of a consulship to the magister equitum Bonosus. The final magister militum to appear in this period was Eusebius. He is named as exmagister equitum et peditum in a law from 360.57 The inclusion of ex- indicates that this is a reference to a previously held title.58 It is likely that Eusebius was magister

51 52 53 54 55 56 57 58

Bagnall, et al. 1987, 17, 222. Salway 2008, 308. Salway 2008, 307–8. Salway 2008, 306–9. For other possible but weaker theories, see: PLRE I, 222; Barnes 1993, 313 n. 22. Salway 2008, 308. Woods 2012, 896–99. CTh 11.1.1. Enßlin 1930, 318; 1931a, 112. The same method is used by AM 22.11.2.

36

The Magistri Militum from 341 to 363 CE

sometime in the 340s prior to his consulship in 347.59 Because Bonosus was magister equitum, then Eusebius may have been only a magister peditum, with the compilers of the Codex Theodosianus making a mistake or perhaps not knowing which title to give him and so recording both, although as we will see later, having one magister of each distinction was not necessary.60 We do not know much of Eusebius himself, but we do know of his family because his daughter, Eusebia, married Constantius, and Julian wrote a panegyric for her.61 Thus, the first five magistri to appear in the sources emerge in a disjointed fashion. Without a comprehensive narrative source for the period, we remain unaware of any magistri involved in the civil war between emperors Constantine II and Constans, or in Constans’ campaigns in Britain and north of the Rhine, or Constantius’ battles with Persia.62 However, despite their small imprint in the sources, these five early magistri were awarded two consulships, one died in a dramatic city riot, and they were heavily involved in enforcing the religious policies of the emperors. They could also be sent on special missions at the behest of the emperors. This indicates that, from their inception, the magistri were extremely important actors in Roman affairs with broad roles and influences. 1.3 Two Magisterial Usurpers During the third century, dozens of military generals and high-ranking officers rebelled against the reigning emperors and claimed imperial authority for themselves. Usually coming from the highest military offices, such as the commander of the important cavalry-centered army that emerged in northern Italy during the third-century upheavals, those who succeeded in securing lasting authority and the acknowledgment of their imperial colleagues survived as emperors, while those who did not were tarnished with the brand of usurper and faced civil war. Such was the nature of the Roman succession. The prevalence of these rebels was reduced during the Tetrarchy as the new political system incorporated a greater number of individuals into the legitimate ruling imperial college, and the spread of the emperors throughout frontier provinces, the regular hotbeds of usurpation, helped meet the needs of the provincials for imperial contact and patronage. The headquarters of the armies also became centered on the emperors’ provincial residences, which placed the strongest armies at their disposal and helped ICUR n. s. 1 3164 = ILCV 3831; CIL X 477; CTh 5.6.1, 11.36.8; P. Oxy 9 1190.15; Soc. 2.20; Bagnall, et al. 1987, 228–9. 60 Demandt 1970, 565. 61 Jul. Or. 3. 62 Cecconi 1999, 76. For the civil war: Epit. 41.21; Zon. 13.5.7–16; Bleckmann 2003. Constans’ campaigns can be found in Barnes 1993, 224–25. Constantius’ eastern campaigns can be found in Barnes 1980, 162–66. 59

Two Magisterial Usurpers

37

secure the loyalty of the soldiers to the emperor himself, rather than with the generals. The de facto requirement of being raised by a reigning Augustus, or of having the approval of the senior Augustus in absentia, also helped limit the possibility of usurpers. Although lessened, usurpations did not entirely go away, and usurpers continued to rise from the ranks of the military throughout the late Empire. It is thus somewhat surprising that only two usurpers arose from the ranks of the magistri militum during the fourth-century period under review here.63 Vetranio and Silvanus each made their attempts early in our timespan, and within five years of one another. Neither found lasting success. In disagreement with a recent assertion that there is “little point” in analyzing the motivations of the leaders of usurpations in the fourth century, the debate on whether Vetranio was loyal to the dynasty of Constantine or whether he was a regular, self-interested usurper seeking to raise his station, is extremely relevant to our understanding of how the magistri could express their power and ambition.64 This question has been the subject of much debate, and a new historiographical investigation will hopefully shed further illumination on this issue. Following on, the mysterious nature of Silvanus’ rebellion also represents either the act of a desperate general in fear for his life, or the invention of ancient authors. By examining the nature of these uprisings and understanding what caused the two fourth-century magisterial usurpers to fail, we can elucidate the nature of military-political power dynamics and the position of the magistri in the middle fourth century. In 340, Constantine II had invaded from Gaul the territory of his younger brother, Constans. In Italy, Constantine was killed, leaving Constans the victorious ruler of the western portion of the Empire, from Britain, Gaul and Spain to Italy, Illyricum, and Africa.65 Over the ensuing decade, Constans failed to adequately manage the vast and complicated network of relationships in the west, and alienated key parts of his realm. In Gaul in early 350, the army officer Magnus Magnentius overthrew emperor Constans and declared himself to be the new emperor.66 It is not stated exactly what rank in the army Magnentius had achieved, but as a commander of the Ioviani and Herculiani units, he was probably a tribunus or comes rei militaris posted to the imperial court; both ranks sat below the magisterium.67 It is unclear whether there was a magister militum in Gaul at this time. The sources for Magnentius’ generals are almost all in Greek, and consequently do not provide the title of magister militum with certainty. As with Salia, other information can help us be more confident.68 A man named Gaiso com63 It has been theorized that Theodosius was also a usurper, but the evidence is slim. See note 99n182. 64 Lee 2007, 67. 65 Lewis 2020. 66 Omissi 2018, 163–64. 67 Zos. 2.42.2. See Chapter 4.2 for the military hierarchy. 68 The PLRE lists Marcellinus 9 as another potential magister militum of Magnentius. The PLRE also lists Marcellinus 8 as Magnentius’ magister officiorum. Demandt 1970, 563, points out that, just like it is unlikely that there were two Bonosi serving in the upper echelons of the Empire in the 340s,

38

The Magistri Militum from 341 to 363 CE

manded the soldiers who chased down and executed Constans, and he was also given a consulship in 351.69 The consulship is firm evidence that Gaiso was a magister, but it is unknown whether he was one of Constans’ who transferred loyalty to Magnentius, or was a new promotion. If he had been a magister of Constans, we would probably expect him to not have given up his dominant position to a lower-ranked officer easily, and seeing as we have no reason to believe Magnentius was simply a figurehead for a powerful magister, we can be confident that he was promoted by Magnentius. Another known officer of Magnentius’ is Romulus. He did not receive a consulship, but he was Magnentius’ only officer that we know of at the Battle of Mursa in 351, and thus perhaps the most senior commander present.70 There is thus good indirect evidence that favors assigning the title of magister militum to Gaiso and Romulus.71 Magnentius quickly asserted control over Constans’ former territories up to the Italian border of Illyricum. Three months later, Vetranio, the magister militum appointed by Constans to Illyricum, was also declared as a new emperor by his soldiers.72 This is the first time we are made aware of Vetranio, but many of the sources emphasize his advanced age and extensive military experience, so he had presumably been a magister for a number of years already.73 The archetypal story of his rebellion was developed in the early twentieth-century by Otto Seeck, who relied mostly on the account of Philostorgius.74 This source records that Constantina, the daughter of Constantine I and sister of emperors Constans and Constantius, pushed Vetranio to seize power in order to ensure Illyricum remained out of Magnentius’ control and secured for the Constantinian, or as it is sometimes known, the Neo-Flavian dynasty. As a member of the imperial family and potentially an Augusta herself, Constantina could have potentially acted as the auctor of imperial authority and legitimately given Vetranio the imperi-

it is unlikely that two Marcellini would be serving under Magnentius without any of the sources indicating a distinction. Peter the Patrician’s use of στρατηλάτης (fr. 16) to refer to Marcellinus is probably a reference to the magister officiorum’s capacity to command the scholae, as he did against Nepotianus in Rome (Zos. 2.43.4). Therefore I do not include Magnentius’ officer Marcellinus as a magister militum. 69 Epit. 41.22–23; Zos. 2.42.5; Chron. Min. 1.69. 70 Zos. 2.52.2, most likely also Jul. Or. 2.57d. Salway 2008, 6–7 theorized that Magnentius’ Romulus may be the same man who received a consulship in 343. This is tenuous but possible, and if true would make it more certain that he was a magister militum. 71 It is likely that these two were magister equitum and peditum, but there is no way to discern who had which title (Demandt 1970, 563). 72 The sources variously say this elevation occurred in Sirmium, Naissus, or Mursa: Chron. Pasch. s. a. 349, 350; Aur. Vict. 41.26; Epit. 41.25; Eutr. 10.10.2; Jer. Chron. s. a. 350; Philost. HE 3.22; Soc. 2.25; Soz. 4.1.1; Zos. 2.43.1; Joh. Ant. fr. 173 (Müller) = fr. 199 (Mariev); Theoph. 5849; Zon. 13.7; Jul. Or. 1.26c; Or. 7.29.9–10; Demandt 1970, 562–63. 73 Jul. Or. 1.30b, 1.33a, 2.76c; Eutr. 10.10; Oros. 7.29.9–10; Soc. 2.28; Them. Or. 3.45b. For a discussion on what his role in the 340s may have been, see: Drinkwater 2000, 149. 74 Philost. HE 3.22; Seeck 1922, 98–104.

Two Magisterial Usurpers

39

al position.75 Constantius initially rewarded Vetranio’s loyalty by acknowledging his claim and sending him military support, but the threat of Magnentius’ Gallic armies drove Vetranio into seeking an alliance with the other usurper. Constantius could not tolerate this, and he extracted his armies from the wars with Persia and marched into Pannonia to confront Vetranio. The old general’s loyalty remained strong, and the two met under a truce. Constantius then addressed Vetranio’s army and, in conjunction with strategic bribery, convinced them to abandon the emperor they had created and join him instead. When they shouted their agreement, Vetranio threw off his imperial regalia and begged mercy, which was somewhat uncharacteristically granted by Constantius.76 In this account, Vetranio is ultimately only a half-senile old man that was useful to others as a pawn. Because of the jumbled and contradictory versions of different sources, Seeck’s account of this rebellion has been revised a number of times.77 John Drinkwater draws a number of illuminating conclusions about this affair, including that Magnentius extended his territorial control only as far as northern Italy and ceased expanding well before Vetranio was declared emperor.78 It is unclear whether Magnentius hoped the Balkan troops would come to him willingly, or if he decided to consolidate his hold on the western provinces before making any moves further east. What is clear is that Vetranio’s rebellion was not done out of an immediate pressure to keep Magnentius out of Illyricum, and was therefore as much a rebellion against Constantius as it was against Magnentius. Drinkwater still argues that Vetranio’s “heart” remained loyal to Constantius, and his rebellion was forced on him by the Illyrian soldiers who had developed a hatred for the Neo-Flavian dynasty.79 Vetranio was essentially forced to either support his troops in rebelling against Constantius, or resist them and be killed for it. However, the disdain for Constantius that Drinkwater sees in the Pannonian soldiers is incongruous with their admiration for Constantine I, the motifs on Vetranio’s coins that portrayed the Neo-Flavian dynasty in a positive light throughout his rebellion, and their eventual acceptance of Constantius as their emperor.80 Furthermore, the idea that the soldiery was uniformly opposed to Constantius while the army’s upper leadership was just as homogeneously supportive of the eastern emperor but kept this from their soldiers implies a vast conspiracy that stretches the limits of plausibility. The numismatic evidence for Vetranio’s regime has prompted historians to argue both the loyalist-theory, as well as the idea of Vetranio as a regular, independent, and ambitious usurper. It also seems to indicate that he was trying to find a balance be75 76 77 78 79 80

Szidat 2010, 246. Szidat 2010, 73. Recently, Humphries 2020 offers a more guarded and cautious account than many other opinions on the episode. Drinkwater 2000, 148–49. Drinkwater 2000, 149–56; cf. Bird 1994, 198 n. 28. Ie. Zos. 2.44.3–4, 46.3.

40

The Magistri Militum from 341 to 363 CE

tween associating himself with the incumbent Neo-Flavian dynasty, please his subjects by evoking previous rulers of Illyricum, and also establishing that he intended to stand alone as emperor. Alan Dearn argues that Vetranio’s association with and depictions of Constantius indicate that he never wanted imperial power and from the very beginning planned to abdicate to the incumbent ruler.81 This is most strongly evidenced by Vetranio’s depiction of himself with the imperial wreath headgear, while Constantius has an imperial diadem, perhaps implying that Vetranio was portraying himself as a subordinate Caesar to Constantius as Augustus. Dearn acknowledges the problem with this is that on the very same ‘subordinating’ coins Vetranio uses the title ‘AUG[USTUS]’ for himself.82 Dearn’s proposal that this was done to elevate Vetranio above Magnentius but still below Constantius is difficult to prove. Furthermore, coinage was more often used as propaganda for a ruler’s own populace, rather than as signals to other regimes.83 It thus seems likely that Vetranio’s pro-Constantinian motifs were instead a method of legitimizing his regime by associating him with the incumbent dynasty, a common tactic of imperial claimants.84 Dearn also writes that, because neither Magnentius nor Vetranio minted coins featuring the other’s image, they never had an alliance, but this is a difficult position to hold against the weight of the literary evidence that confirms an agreement between the two usurpers, although their alliance was given a lower priority than the relationships they sought with Constantius.85 Constantine I had issued the novel HOC SIGNO VICTOR ERIS legend on his coins from the mint at Siscia in Illyricum (Sisak, Croatia), recalling his Christian victory at the Milvian Bridge in 312, and thus Vetranio’s issuance of the same motif would seem to indicate that he was trying to associate himself with the Christian religion increasingly shaping imperial politics, as well as the Neo-Flavian dynasty, and previous rulers of Illyricum.86 Once Constantius retook Vetranio’s territory, he would also issue coins with the same legend from Siscia, indicating its importance to the people of the region. At the same time as Vetranio was associating himself with Constantius on his coins, he also emphasized his independence. His depicted himself with a beard, unlike the shaven faces of the Neo-Flavians, probably to differentiate himself from the incumbent dynasty just enough to ensure his subjects knew he intended to stand on his own.87 This bearded image also recalled that of the Illyrian soldier emperors of the late third century. Ultimately, portraiture on coinage was a primary tool of self-representation and is an effective method for understanding how Vetranio wished his regime to be perceived. It is difficult, however, to know to what extent Vetranio’s imagery were 81 82 83 84 85 86 87

Dearn 2003; cf. Watts 2015, 250 n. 5. Dearn 2003, 180–3. Szidat 2010, 283. Szidat 2010, 260–1. Dearn 2003, 178. Bleckmann 1994, 47–48. Dearn 2003, 179–78; Sanchez 2003, 37.

Two Magisterial Usurpers

41

actual depictions of subordination to Constantius, or simply appeasing messages to try to gain recognition, or unifying ways to bind his territory, or a combination of all the above. Adrastos Omissi also addresses Vetranio’s rebellion in the context of the larger phenomenon of usurpation in the later Roman Empire.88 Omissi perpetuates the idea that Vetranio’s rebellion halted Magnentius’ expansion, but as discussed above this was not the case.89 He argues, based on the panegyrical evidence, that Vetranio’s usurpation was a loyalist movement to defend Constantius. The loyalist-theory, as detailed by Drinkwater, Dearn, or Omissi, is ultimately unable to adequately explain why Vetranio needed to be usurp imperial authority at all. Although he was not actually a magister militum under Constantius, but was a general of Constans, it still would have been a much safer course of action for himself, and everyone else involved in the undertaking, if he simply remained a magister militum and declared his loyalty for Constantius. Such was the choice in 361 for Lucillianus, the magister militum in Sirmium, potentially the exact same city in which Vetranio was raised to the purple.90 Facing down the approaching armies of the usurper Julian, Lucillianus remained a loyal general of Constantius. No counter-rebellion was necessary. Julius Nepotianus could be used as an example of a man becoming emperor to face down an encroaching usurper, but his rebellion was of a fundamentally different nature. He was declared an emperor in Rome a few months after Vetranio’s own declaration.91 However, as the son of Eutropia, Constantine I’s half-sister, Nepotianus was a member of the imperial dynasty, perhaps making his declaration as emperor more palatable and less surprising than Vetranio’s. Furthermore, Rome was in Magnentius’ direct line of advance and only had limited troops available for its defense, and so Nepotianus was in extreme danger and any small advantage he could create had to be taken. The authority of an emperor might have also been a necessary step to call upon the city’s gladiators to join in its defense, as Nepotianus did.92 As established, Vetranio was not under direct threat from either Magnentius or Constantius at the time his rebellion was declared, and thus there is no good reason for his usurpation other than ambition. Because of the difficulties with the loyalist-theory, Bruno Bleckmann argues for giving Vetranio significantly greater agency in his rebellion, and views him as a traditional usurper who seized the opportunity provided by Magnentius’ rebellion to gain a greater position for himself.93 When Constantius marched westwards, Bleckmann posits that his forces were hopelessly outmatched by a potential alliance between Vetranio 88 89 90 91 92 93

Omissi 2018, 164–68. Cf. Szidat 2010, 219. See note 99n182. Aur. Vict. 42.6–8; Epit. 42.3; Zos. 2.43.2–4; AM 28.1.1; Eutr. 10.11; Philost. HE 7.24; Soc. 2.25.10; Soz. 4.1.2; Oros. 7.29.11; Joh. Ant. fr. 174 (Müller) = fr. 200 (Mariev). Eutr. 10.11. Bleckmann 1994, 44, followed by Kovacs 2017.

42

The Magistri Militum from 341 to 363 CE

and Magnentius.94 He thus entertained the possibility of recognizing Vetranio, and used this pretense of peace as a cover for inducing key members of Vetranio’s party to defect. Frightened by these developments and his deteriorating position, Vetranio agreed to meet Constantius under a truce, where Constantius addressed their combined armies. His defensive advantage having been entirely squandered, Vetranio meekly surrendered his claim.95 Because he agreed to play the part of regretful simpleton, Constantius gave Vetranio clemency and allowed him to retire. Further light can be shed on Vetranio’s rebellion through careful Quellenforschung, the exploration of the origin and transmission of our literary sources. Historians agree that many of the accounts of this event are tied to Constantius’ propaganda, and the official ‘party line’ dictates their record.96 Other views are clouded by anti-Christian polemic. Dearn has already provided a summary and binary categorization of the literary evidence, but his division into positive and neutral, or negative sources, is perhaps too simplistic.97 Instead, we should focus on perspectives, intentions, and importantly for the sources written much later than the fourth century, where did their information originate. Some sources, such as Jerome, the Consularia Constantinopolitana, and the ecclesiastical histories of Socrates and Sozomen, have only the most limited reference to Vetranio’s rebellion, and generally lump it in together with Magnentius’.98 The panegyrics of Julian (composed between the mid and the late 350s) and Themistius (composed between 357 and 364), on the other hand, are a more useful genre of literature to assess. They both denigrate Vetranio as a feeble old man who was initially loyal to Constantius, but then treasonously betrayed his trust – typical motifs of imperially sanctioned stories propagated to downplay the seriousness of a rebellion.99 They praise Constantius for giving Vetranio aid, and the verbal eloquence that allowed him to win over the rebel soldiers and gain a bloodless victory.100 The intention of these works was to celebrate and extol Constantius: Julian hoped to please Constantius through his panegyrics, and the future career prospects of Themistius depended on the positive reception of his works. Thus, no dissent or insults were included, and so the shared elements of these works delineate the features of the officially supported account of events – that is, the recasting of these events to appear like a loyalist uprising. Another important literary tradition is exemplified in Philostorgius, the Chronicon Paschale, and Theophanes, which are all believed to draw upon the ‘Arian History’, 94 Bleckmann 1994, 52, 56. 95 Bleckmann 1994, 53–54. 96 Bleckmann 1994, 44; Drinkwater 2000, 149, 158–59; Dearn 2003, 170. 97 Dearn 2003, 170–76. 98 Jer. Chron. s. a. 350, 351; Cons. Const. s. a. 350; Soc. 2.25, 2.28; Soz. 4.1.4. 99 Jul. Or. 1.26c, 1.30b, 1.30–2, 1.33a, 2.76c-77, 2.77c-d; Them. Or. 3.45b-c, 4.56a-b, 6.80c; Szidat 2010, 260–1. 100 Jul. Or. 1.30–2; Them. Or. 3.45b-c, 6.80c.

Two Magisterial Usurpers

43

the author of which is known as the ‘anonymous Homoian historian’, who wrote during the time of emperor Valens (364–378).101 The Homoians, also known as Acacians after their leader, Bishop Acacius of Caesarea, were one of the sects involved in the fourth-century Arian controversy. They were associated with the Arians and therefore had a mutually supportive relationship with the Arian emperor Constantius – he favored them, and the anonymous Homoian historian recorded Constantius’ reign in a positive and sympathetic light.102 Thus the historian includes the same line of Vetranio’s loyalty as the panegyrics, but because he was writing after Constantius’ death and a new dynasty had control over the Empire, the Arian writer is able to also discuss a wider range of information than the panegyrists, and thus he includes the role of Constantina, and instead of only shaming and insulting Vetranio he affords some praise for his religious and charity work after his forced retirement to Bithynia.103 The missing Kaisergeschichte, identified by Alexander Enmann in the late nineteenth century as the missing link between several ancient sources, is the basis for a further tradition.104 The works that relied upon the Kaisergeschichte include Eutropius, Aurelius Victor, the Epitome de Caesaribus, and John of Antioch.105 This source tradition still describes Vetranio as an old man, but instead of being idiotic, he has old-fashioned virtues and was successful and experienced in warfare, although he was uneducated. Differing quite significantly from the two strains of courtly propaganda already identified, the Kaisergeschichte represents an important source of comparative information. Dearn nonetheless places these works in his ‘negative traditions’ category along with the panegyrics, even though he writes that Eutropius has a “generally positive view” of Vetranio.106 The apparent misclassification of these sources stems from Victor, who owed his professional career to Constantius and thus modified his sources to closely mirror the orthodox imperial propaganda line wherever possible – he thus calls Vetranio illiterate and rustically stupid.107 Importantly, none of these sources indicate Vetranio remained loyal to Constantius, and in fact they record that Vetranio seized power much like a regular usurper.108 This indicates that literature generated outside the eastern court’s influence did not record any unique loyalist characteristics for Vetranio, and even when Victor wanted to portray Constantius positively he did not include this trope, perhaps suggesting the idea of Vetranio as a loyalist rapidly lost credence. The remaining sources are more complicated, and historians argue for varying networks of reliance. None of the surviving fragments of Eunapius’ Universal History, or 101 Bleckmann and Stein 2015, 253. 102 Bidez 1913, clii. 103 Philost. HE 3.22, 3.24; Theoph. 5849. 104 Enmann 1884. 105 Bird 2011, xiviii; Dearn 2003, 173. 106 Dearn 2003, 173. 107 Aur. Vict. 41.26; Nixon 1991, 124. 108 Epit. 41.25; Aur. Vict. 41.26; Oros. 7.29.9–10; Joh. Ant. Fr. 173 (Müller) = fr. 199 (Mariev).

44

The Magistri Militum from 341 to 363 CE

the Chronicle After Dexippus as it is called by Photius, the Byzantine patriarch who preserved the greatest portion of it, cover Vetranio’s rebellion. There is no indication, however, that the sixth-century financial officer Zosimus drifted from his habit of faithfully reproducing Eunapius’ work, and thus what is contained in his New History is most likely a reliable copy of Eunapius. Dearn correctly identifies that Eunapius read and used the writings of Julian and so would have been influenced by the panegyric tradition.109 Zosimus’ account of Vetranio, however, bears no resemblance to Julian’s encomium, and Eunapius’ description of this episode must have come from elsewhere. It has been proposed that Eunapius used the anonymous work Against the Christians, written in the west as anti-Christian propaganda.110 This identification has met resistance, and other proposals include Ammianus’ Res Gestae.111 In his other major work, the Lives of the Sophists, Eunapius names Tuscianus as a source.112 Tuscianus served as an assessor to Anatolius, the Praetorian Prefect of Illyricum from 357–360.113 In this role he would have had contact with a great number of individuals who had lived through Vetranio’s rebellion as well as access to records and documentation of what transpired, and thus potentially represents a major original source of information for Eunapius’ account of Vetranio. Whatever the origin of his information, Eunapius clearly saw the exaggerations and untruths of Julian’s panegyric and decided to produce a different account. Eunapius may have been motivated by an opposition to the Christianization of the Empire to more negatively portray Constantius. Eunapius, through Zosimus, emphasizes that Vetranio was overcome with the same ambition as Magnentius, and that he was proclaimed by his troops, the most important step in the acclamation of a new emperor.114 Nothing is recorded of ultimate loyalty to Constantius, nor assistance given to Vetranio – Constantius was busy dealing with Persian attacks in the east. When able to extract his armies to deal with the rebels, Constantius decided that it is better to fight one usurper than two, and so wanted to win Vetranio to his side. Meanwhile, Magnentius wished for the same and also sought to make an ally of the Illyrian emperor. Vetranio only then chose to side with Constantius, and Zosimus implies he received recognition only at that point. At a meeting that was framed as a time to plan the upcoming war with Magnentius, Constantius tricked Vetranio by addressing his soldiers and convincing them to come to his side.115 Combined with bribes, the armies proclaimed their support for Constantius, and they stripped Vetranio of his imperial vestments,

109 Paschoud 1985; Blockley 1983, 24–25; Dearn 2003, 175. 110 Paschoud 1975, 150. 111 Blockley 1981, 23–25. 112 Eun. Lives, 9.2.15. 113 Tuscianus: PLRE I, 926; Anatolius: PLRE I, 60. 114 Zos. 2.43.1; Szidat 2010, 44. 115 Zos. 2.44.1–2.

Two Magisterial Usurpers

45

with only Constantius’ mercy allowing him to survive and retire.116 Thus the Eunapian version of Vetranio’s rebellion is a criticism of Constantius’ trick on Vetranio, and portrays Vetranio as a usurper of regular characteristic, who followed the established path of acclamation, and sought recognition from the incumbent emperor in an attempt to secure lasting legitimacy. In his categorization, Dearn writes that the Kaisergeschichte was also the source for Peter the Patrician and Zonaras.117 These two accounts are similar to one another and clearly share a relationship, but are contain significant differences when compared to other sources, including the Kaisergeschichte descendants, and thus represent the final literary tradition for us to assess. The origins of this tradition and the path by which it reached Peter and Zonaras are the subject of heated polemic that need not be outlined in detail here. It suffices to say that Bruno Bleckmann has demonstrated that a source covered the third and fourth centuries, and it was not Eunapius nor Ammianus.118 This now-lost source was probably written in Greek, and covered events up to at least the death of Jovian. This source is the reason why the Byzantine historians Peter the Patrician and Zonaras contain genuine evidence that differs significantly from other sources. Peter and Zonaras mirror each other closely in language and content, most prominently in their description of a joint embassy from Vetranio and Magnentius to Constantius, and in the story of a vision Constantius had of his father and brother.119 The account of the vision naturally had to have come from Constantius himself. It may have been an attempt by Constantius to use familial vengeance to justify why he allowed Vetranio to live, but waged brutal war on Magnentius and the west. This is compatible with the inscription on the base of the monument we now call the Lateran Obelisk. Taken from the Karnak temple in Egypt, the thirty-two-meter obelisk was set up in the Circus Maximus at Rome by Constantius to commemorate his victories. The inscribed base commemorated Constantius’ father, as well as his own salvation of the entire world (totus orbis).120 It also references the loathsome tyrant (taeter tyrannus) laying waste to Rome, a reference to Magnentius’ magister officiorum Marcellinus, who put down Nepotianus’ rebellion there. No mention is made of Vetranio, indicating that Constantius tried to downplay him in the Urbs Aeterna. Why the propaganda strain of Vetranio’s stupidity but ultimate loyalty did not make it into the Peter-Zonaras tradition is unclear. Perhaps this idea could not gain traction in the west, if that is indeed the origin of the source, because Vetranio’s western contemporaries had a greater degree of personal understanding of the high-ranking and presumably well-known general. They would have been aware he was not idiotic. Zo-

116 117 118 119 120

Zos. 2.44.3–4. Dearn 2003, 175. Bleckmann 1992. Petr. Patr. fr. 213; Zon. 13.7; de Boor 1892, 21; Banchich and Lane 2009, 214–15. CIL VI 1163.

46

The Magistri Militum from 341 to 363 CE

naras does retain a reference to a letter from Vetranio to Constantius, urging them to work together against Magnentius, but there is no hint that Vetranio was only serving Constantius, or intended to abdicate to him.121 With this summary and evaluation of the available literature it becomes clear that the loyalist idea originates from and was propagated by the panegyrical sources that wanted to spread Constantius’ official propaganda line. In reality, Vetranio most likely fit the mold of almost all other fourth-century magistri and was military experienced and capable.122 Many of the sources attest to this.123 He perhaps had not received the typical Roman aristocratic education, and thence stemmed the theme of his idiocy, but there is no indication he was uniquely militarily incompetent. As such, Vetranio knew his position was untenable if he faced off alone against either Magnentius or Constantius. Thus, while engaging in high-level diplomatic negotiations with both other emperors to gain their allegiance, he made defensive measures against encroachment.124 Against Magnentius, he probably secured the fortifications known as the Claustra Alpium Iuliarum, as Ammianus mentions that Magnentius seized these fortifications from an officer named Acacius.125 It is unclear whether this was one of Vetranio’s men captured before his abdication, or one of Constantius’ during his subsequent war with Magnentius, but either way these are probably the same fortifications that Philostorgius says were garrisoned by Vetranio.126 In the east against Constantius, Vetranio occupied the critical Succi pass fortifications, located on the Via Militaris, the major east-west highway stretching from Singidunum (Belgrade, Serbia) to Constantinople. Succi itself was located between Serdica (Sofia, Bulgaria) and Philippopolis (Plovdiv, Bulgaria), on the Via Militaris as it passes between the Rila-Rhodope Massif to the south and the Sredna Gora mountain range to the north. These mountain ranges made Succi a critical juncture, because if it was blocked the alternate routes would involve travelling all the way north to the Danube and proceeding along the river, or returning almost all the way to Constantinople and then transiting through Greece to the Adriatic Coast and proceeding from there. Importantly, a force that avoids Succi for any other route would be at risk of encirclement. Vetranio’s occupation of Succi was thus a militarily strategic move that forced Constantius into confronting him. It is difficult to see the occupation of these defenses as anything other than indicating Vetranio desired to keep his reign and territory secure and independent from both other emperors.

121 Zon. 13.7. 122 See Chapter 1.3. 123 Joh. Ant. fr. 173 (Müller) = fr. 199 (Mariev); Eutr. 10.10; Petr. Petr. fr. 213; Zon. 13.7; Jul. Or. 1.33a. 124 Philost. HE 3.24; Jul. Or. 3.76c. 125 AM 31.11.3; Šašel and Petru 1971. 126 Philost. HE 3.24.

Two Magisterial Usurpers

47

Facing two usurpers as well as Persian attacks on the eastern frontier, Constantius initially recognized Vetranio as a legitimate emperor out of the need to buy time and attempt to stave off an attack on Constantinople rather than as a reward for loyalty.127 Still requiring protection against the threat of an invasion by Magnentius, Vetranio also cultivated an alliance with the western usurper.128 Late in 350, Constantius extricated his army from the frontier and marched westwards. At Heraclea in Thrace (probably Marmara Ereğlisi, Turkey), according to Peter the Patrician and Zonaras, diplomats from both Magnentius and Vetranio met with Constantius.129 They offered him the first rank in the imperial college, but requested acknowledgment of their own imperial claims. Bleckmann argues that Constantius was at an extreme disadvantage against the two opposing emperors, and was thus forced to consider their proposals.130 There is no indication, however, that Magnentius and Vetranio had actually developed a cooperative military alliance, and so even though Constantius faced two enemies, he faced them individually. The eastern army would defeat the western in two civil wars later in the fourth century, indicating that the eastern army was the superior force.131 Moreover, Julian highlights Vetranio’s preparedness, but also attests that Constantius did not feel daunted.132 His eastern soldiers were veterans from over a decade of hard fighting against the Persians. It thus seems doubtful that Constantius’ position was entirely hopeless, and his entertaining of the ambassadors was polite political practice rather than desperation. According to Peter and Zonaras, that night Constantius had a vision of his father and murdered brother who urged him to seek vengeance.133 As stated above, this seems like a justification for his brutal war against Magnentius but gentle treatment of Vetranio – the latter had done nothing more offensive to Constantius than claim the purple, while the former had killed his brother. In reality, more underhanded events had been taking place. Bleckmann rightly emphasizes that Vulcacius Rufinus, Vetranio’s Praetorian Prefect and envoy to Constantius, had clearly orchestrated a deal with the eastern emperor.134 When Constantius awoke, he arrested all envoys other than Rufinus, and proceeded rapidly through the Succi pass to Serdica. Rufinus had clearly agreed to join Constantius and aid in his passage through Vetranio’s territory. Ammianus attests that the officer Gomoarius, who we will later meet as a magister militum, had also betrayed

Zon. 13.7.1–12; Them. Or. 6.80c; Philost. HE 3.22; Jul. Or. 1.27a–28d, 30b. Zon. 13.7; Them. Or. 6.80c; Philost. HE 3.22, 3.24; Petr. Patr. fr. 213; Jul. Or. 1.30–32. Petr. Patr. fr. 213; Zon. 13.7; cf. Zos. 2.44. Bleckmann 1994, 50. Drinkwater’s (2000, 155) claim that Constantius must have been in the weakest position of the three emperors must be disregarded, as Vetranio controlled by far the smallest territory and army. 131 See Chapters 2.5–6. 132 Jul. Or. 1.31a-b. 133 Petr. Patr. fr. 213; Zon. 13.7. 134 Bleckmann 1994, 52, 56.

127 128 129 130

48

The Magistri Militum from 341 to 363 CE

Vetranio, and so it seems likely that he had been in charge of the forces occupying Succi and had betrayed the defenses to Constantius.135 This echoes similar cases of officers betraying their leaders and switching sides throughout the fourth century.136 Vetranio knew his position had become untenable once Constantius’ force had moved through the pass. His only military advantage had been lost, and he decided to meet Constantius under a truce in Serdica to negotiate and plead for clemency. His soldiers also now faced the possibility of fighting a doomed battle, or taking the undoubtedly large bribes and donatives that would come with accepting Constantius. Although Vetranio had clearly gained the necessary recognition for a usurpation from the various acceptance groups present in Illyricum, such as the army, civilian bureaucrats, the senators who had fled from the west, and imperial figures like Constantina, he had apparently not managed to secure their lasting faithfulness.137 The two forces moved to Naissus (Niš, Serbia), where, after the necessary bribes had been distributed and the outcome assured, Vetranio’s soldiers hailed Constantius as their Augustus, and Vetranio meekly surrendered and offered his imperial vestments to Constantius to demonstrate his total renunciation of imperial claims.138 In return for his submission, he was allowed to retire with a pension to Prusa (Bursa, Turkey), rather than be executed, the normal punishment for usurpers.139 It is worth addressing this apparent out-of-character decision by the normally brutal Constantius as it is an argument often used in favor of the loyalist-theory. If he is seen as a devotee, it is easy to understand why Constantius allowed Vetranio to live. Our opinion of Constantius as ruthless comes mostly from the hostile Ammianus, which will be discussed further in the following section. But our view of Constantius in this manner may not be entirely accurate. Even others that record his brutality were biased against Constantius.140 Thus the emperor might not have been as brutal as the sources impress upon us. His decision of clemency might have instead been a decision born of practicality. Constantius knew that he had a major battle against Magnentius in the near future, and it would be easier to incorporate Vetranio’s army into his own through a display of clemency. Demonstrating to Magnentius’ forces that betrayal was a quick path to forgiveness and even promotion was also a prudent measure. Rufinus became Praetorian Prefect of Gaul, and his role may have prompted Constantius

135 AM 21.8.1. Drinkwater argues that Rufinus’ conspiracy was actually sanctioned by Vetranio, who wished to submit to Constantius but wanted to hide this from his soldiers. The problems with this idea have been dealt with above. 136 See pages 47–49, 89–90, 106–107. 137 Szidat 2010 is the preeminent discourse on the various acceptance groups involved in the raising of an emperor, their roles in a new regime, and how their loyalty can be won or lost. 138 Jul. Or. 2.76c–77d; Epit. 41.25; Philost. HE 3.24; Soc. 2.28; Zon. 13.7; Zos. 1.44. 139 Philost. HE 3.22; Soc. 2.28; Zon. 13.7; Zos. 1.44. 140 Ie. Eutr. 10.15.

Two Magisterial Usurpers

49

to promote his nephew, Gallus, as Caesar.141 Constantina also had her station raised through marriage to Gallus.142 It may have been these measures that encouraged men like Silvanus, who we will meet below, to join Constantius, as well as to smooth the transition of power as Constantius took over the west. Why Constantius did not have Vetranio assassinated later in life to avoid a possible resurgent rebellion is also explainable. Theophanes indicates that Vetranio became a pillar of the Christian community in Prusa, and he presented no threat in a demilitarized area far away from his western allies.143 Furthermore, clemency was considered a commendable imperial virtue that Constantius wished to emblemize.144 Thus it is unusual, but not particularly surprising that Vetranio lived through his failed rebellion. Ultimately, it seems that all interpretations of Vetranio as a loyalist are unable to explain why he took such a course of action against Constantius. As was typical of victorious emperors, Constantius wished to downplay the significance of those who rebelled against him, and thus Vetranio was retroactively portrayed as a loyal subject. Viewing Vetranio instead as a regular, self-interested usurper without special loyal characteristics is more congruous with the weight of the reliable evidence. This interpretation indicates that from their early appearances in the historical record, some magistri militum retained the ambitious character that had dominated the military hierarchy for the previous century. Although Tetrarchic developments had slowed the rate of usurpation from its heights in the third century, the reforms of the civil and military commands had not entirely eliminated the possibility of military usurpation. Thus, when another magister faced the possibility of imperial persecution, he pursued this same path. This second magisterial usurper was named Silvanus. Although there are reasons to doubt whether he was truly a usurper, chiefly that no coins have been recovered bearing his name, the weight of evidence suggests he was another usurper from the ranks of the magistri. He first appears as a tribunus in Magnentius’ army, but just prior to the Battle of Mursa in September 351, he deserted with his soldiers to Constantius.145 Sometime after, possibly in the winter of 351–352 or after the final defeat of Magnentius in August 353, Constantius rewarded Silvanus for his timely defection by promoting him to magister peditum in Gaul.146 During this time, Silvanus was able to operate with

141 142 143 144 145

Bleckmann 1994, 57–58. AM 14.7.4, 11.22; Zos. 2.45.1; Philost. HE 3.22, 3.28; Zon. 13.8.4; Epit. 42.1. Theoph. 5849. Szidat 2010, 329. Zon. 13.8 relates that Silvanus went over to Constantius with a regiment of infantry, while Jul. Or. 1.48b, 2.97c and AM 15.5.33 seem to agree that it was actually cavalry. Aur. Vict. 42.14 leaves the details of the desertion out. 146 Ammianus called him “pedestris militiae rector” (15.5.2) as well as “magister peditum” (16.2.4), while Aurelius Victor calls him “pedestre ad magisterium” (42.14). CTh 7.1.2 is recorded from the year 349 and gives Silvanus the title of comes et magister equitum et peditum. However, the manu-

50

The Magistri Militum from 341 to 363 CE

a large degree of independence to defeat barbarians that Constantius had induced to cross the Rhine to destabilize Magnentius during their civil war.147 The sources then report that Silvanus fell afoul of a plot orchestrated by a low-ranking official named Dynamius.148 Letters were forged to make it appear as though Silvanus was plotting to usurp power, and even though this was eventually revealed as false, the magister felt his life to be in such danger that he ultimately did declare himself emperor on 11 August 355.149 Constantius was quick to respond, and sent another magister militum, Ursicinus, to go to Silvanus under the false pretense of being a co-conspirator.150 This was a success, and members of Ursicinus’ entourage were able to bribe a group of soldiers into assassinating Silvanus only twenty-eight days after his rebellion started.151 There are multiple important points to discuss in this episode, and it will be revisited multiple times in this text. The following section will analyze Ammianus’ personal account of how Silvanus was executed only 28 days after his declaration, and Chapter 3.2 will reconstruct the networks of relationships involved in the plot and subsequent events. To avoid repetition of the same accounts, in this section the focus will remain on Silvanus as a usurper and what this teaches us about the military magisterium. It first bears discussing that John Drinkwater doubts that Silvanus’ usurpation occurred at all because of the complete lack of coins, and the inconsistencies of Ammianus’ account, the most detailed version of these events.152 This has been followed by Michael Kulikowski, who writes that it is “certain” the rebellion did not happen and instead Silvanus was executed while still a magister militum.153 Elements of this stance have been countered in several attempts, most extensively by David Hunt.154 He points out that, even if Ammianus presents difficulties, only a few years later Julian was also referring to the usurpation, as was Aurelius Victor.155 Zonaras also makes mention of

script seems to be faulty, because it records Constantine’s name rather than that of Constantius. Thus, we can presume this law was particularly poorly created or transmitted, and therefore we should not doubt the chronology of Silvanus’ career that is offered by the other sources. Furthermore, because this particular form of a magisterial title is not repeated until the late 360s, we can also doubt that it was given to Silvanus (CIL III 10596 = ILS 762). CTh 8.7.3 is also dated to 349 and gives Silvanus the different title of comes et magister militum, but it is probably similarly mistaken, because it is issued by Constantius alone, not also with Constans, who was still alive in 349. For more discussion, see: Demandt 1970, 568. 147 AM 16.2.4. 148 The story is best told in AM 15.5, and also: Aur. Vict. 42.15–16; Epit. 42.10–11; Eutr. 10.13; Oros. 7.29– 14; Jul. Or. 1.48s, 2.98c–99a; Soc. 2.32.11; Soz. 4.7.4; Theod. HE 2.16.21; Joh. Ant. fr. 174 (Müller) = fr. 200 (Mariev); Zon. 13.9; Lib. Or. 18.31; Pan. Lat. 11.13.3, Ep. ad Ath. 273d; Jer. Chron. s. a. 354. 149 Boer 1960. 150 AM 15.5.18–22. 151 AM 15.5.29–32. 152 Drinkwater 1994, 574–5. 153 Kulikowski 2016, 298. 154 Hunt 1999; Gracia 2014. 155 Jul. Or. 1.48c, 2.98c–99a; Ep. ad. Ath. 273d; Aur. Vict. 42.14–16.

Two Magisterial Usurpers

51

Silvanus being forced into rebellion because of intrigues.156 The weight of the literary evidence is thus impossible to dismiss. Hunt further argues that the rebellion was simply smaller than Ammianus would have us believe, and Silvanus had not yet secured the mint at Trier to begin distributing coins.157 Ammianus’ significance of the rebellion is exaggerated because of his personal involvement: initially, before the historian was involved, Ammianus wishes to portray Silvanus and a poor innocent man falling afoul of the brutal conspiracies plaguing Constantius’ court. Then when Ammianus does become involved in Silvanus’ downfall, the usurper is shifted to be more akin to a regular rebel that deserves to be executed, in order to make Ammianus’ role in such an end more palatable. Other purported issues with Ammianus’ account have their own answers. For example, Doug Lee and Drinkwater both point out the impossibility of how Ursicinus, the magister militum in command of the forces that brought about the execution of Silvanus, could have pretended to be unaware of the rebellion at all.158 Ammianus provides the answer when he writes that this ruse failed, so instead they pretended to be supporters of Silvanus and join his rebellion instead.159 It is also worthwhile to consider John Weisweiler’s argument that Ammianus is offering a meta-lesson to his audience, educating them on how to read and use literature through a motif of the misuse of literature (Dynamius’ forged letters).160 Thus the potential rhetorical elements in Ammianus’ account muddy the waters of what really occurred, however not to a fatal extent. There might also be an answer as to why Silvanus issued no coins. Kulikowski points out that even the shortest-lived usurpers of this period, such as Julius Nepotianus who was usurper for about the same length of time as Silvanus, managed to distribute coins from Rome.161 However, Joachim Szidat describes in his extensive coverage of the technical aspects of a usurpation how a period of extensive planning occurs before rebellion.162 This planning is naturally secretive, and the act of usurpation is often portrayed as spontaneous. But furtive connections have been established between the usurper and other leading men in his rebellion with other important groups: military leaders and soldiers, senators, the Church, and civilian bureaucrats, including those who ran the imperial fabricae and mints. We can assume that these connections are what Vetranio established during the three months between the murder of Constans and his declaration. Nepotianus had several months to plan before his own declaration. Magnentius had presumably also spent time securing the loyalty of the important groups in the west, such as Constans’ Praetorian Prefect, Fabius Titianus, who became Magnentius’ 156 157 158 159 160 161 162

Zon. 13.9. Hunt 1999, 52. Lee 2015, 116; Drinkwater 1994. AM 15.5.25. Weisweiler 2014, 111–12. Kulikowski 2016, 298. Szidat 2010, 232–35, 273.

52

The Magistri Militum from 341 to 363 CE

praefectus urbi of Rome.163 In Chapter 1.5, we will see how the Caesar Julian spent years preparing and planning for his usurpation. Silvanus, however, did not have the luxury of a planning period. Dynamius’ conspiracy must have happened over a relatively short span of time, with Silvanus only making his decision to usurp suddenly. Prior to his declaration, he was considering other options such as fleeing across the border to the Franks instead of planning a rebellion.164 Furthermore, it is surprising that Silvanus gave a donative to his soldiers on Constantius’ birthday, effectively boosting their loyalty to the emperor, only four days before he would declare himself a rebel emperor.165 This could in fact indicate that on that date Silvanus had still not decided on his course of action, and thus potentially only had three or fewer days to prepare, assuredly not enough time to connect with the commanders in charge of the mints in Trier. Although that mint was not far from Cologne, it was a very different situation to Nepotianus, who had a mint directly in Rome. Even after Silvanus’ declaration and him reaching out to contact the mint, the commander would want to verify messages and assess his own position in regard to the likely ensuing civil war before committing to Silvanus’ side. Silvanus would have to decide on the motifs and symbols he wished to use, and have dies created. Then there would be the matter of distribution. Thus, although it can still strike as surprising that Vetranio issued no coins, it is not certain evidence that his rebellion never occurred. Furthermore, the lack of coins pales in comparison to the weight of contemporary literature in favor of the rebellion. Although there has been extensive literature published with doubting theories on both fourth-century magisterial usurpers, it seems that they both followed the moreor-less regular path for military generals faced with opportunity or crisis. Following the destabilization of the murder of emperor Constans and Constantius’ inability to quickly respond, Vetranio seized the chance to elevate his station and secure his territory against both other emperors. A few years later, feeling his life was in danger after being targeted by courtly plotting, Silvanus took the same step of elevating himself in an attempt to secure his safety. He was probably also ambitious – other magistri, such as Ursicinus or Barbatio that we will discuss below, were also unhappy at their circumstances and fell afoul of plotting, but they did not similarly claim the purple. This tells us that some magistri were ambitious and would seek to increase their positions, but this was not universally true for them all. Thus we will see over the rest of this book how some individual magistri stand out as particularly motivated to seek more power. Compared to their contemporary usurper Magnentius, who over the course of his rebellion established control over most of the west, suppressed a resistance in Rome, appointed a Caesar, and fought multiple bloody battles against Constantius, both mag163 PLRE I, 918. 164 AM 15.5.16. 165 AM 15.6.3; Nutt 1973, 85–86.

Ammianus’ Account of Some Magistri Militum

53

isterial usurpers were lackluster rebels without significant successes to their names.166 It has been pointed out that maybe the magisterium was working as Constantine intended, and the magistri had difficulty creating the necessary relationships with acceptance groups and portraying themselves as viable imperial candidates.167 Both Vetranio and Silvanus were brought down by betrayal, and some future magistri would also suffer at the hands of civil officials. Other magistri, however, were able to form positive, working relationships with other powerful individuals and groups, which will be explored further in Chapter Three, indicating this was not a comprehensive impediment to all magistri who might seek the purple.168 It might also be argued that there were unique, unrepeated circumstances that led to these rebellions. However, while no magister fell victim to quite the same plot as Silvanus, the conditions of Vetranio’s rebellion were closely mirrored in 365 during Procopius’ rebellion, but the magister in Sirmium still declined to start his own revolt.169 Moreover, the comparative frequency with which military officers of lower ranks rebelled indicates there was nothing stopping military usurpers in totality. Instead, later magistri may have learned a lesson from the failures of Vetranio and Silvanus. We do not have any insight into what extent magistri looked back at the failures and successes of recent history for guidance, but it would be deeply surprising if they did not know the recent military history of the Empire. Desiring to not repeat the failures of their predecessors, the magistri and other officials would explore other methods of raising their station, especially during the succession crises of the fourth century. 1.4 Ammianus’ Account of Some Magistri Militum Ammianus Marcellinus has long been considered the most reliable source for understanding the military history of the fourth century.170 His eye-witness accounts of battles, campaigns, politics, as well as a critical use of additional informants, provides incredible insights into the activities of the magistri. However, he has recently come under increased scrutiny. David Woods has claimed that “the myth of the accuracy of Ammianus as a military historian deserves to be exploded in the same fashion that T. D. Barnes (…) has exploded the myth of his objectivity in religious matters.”171 This is a task that has so far not been attempted other than for few, disparate issues, probably because Woods is exaggerating Ammianus’ flaws. For the purposes of this

166 167 168 169 170 171

For Magnentius’ rebellion, see: Hunt 1997a, 10–11, 14–22; Barnes 1993, 101–8. Lee 2015, 103 n. 17. Compare the success of Merobaudes to the failure of Flavius Theodosius in Chapter 2.2. See pages 79–80, 89–90. Dixon and Southern 1996, 110. Woods 2001, 303 n. 12, referencing Barnes 1998.

54

The Magistri Militum from 341 to 363 CE

book, Ammianus proves to be generally accurate on matters such as titles, disposition of forces, and general narrative context. The sixth-century historian Procopius, like Ammianus, personally witnessed many of the campaigns he later included in his writings. When assessing his reliability, David Parnell wrote that Procopius “should generally be trusted unless there is particular reason to be suspicious of his motives.”172 This same statement can be made about Ammianus. E. A. Thompson sensibly argues that we should be suspicious of Ammianus’ motives when he records events in which he had a personal stake.173 During his time as a protector in the army, Ammianus served on the staff of the magister militum Ursicinus.174 Ammianus’ own career and reputation would have been tied into the perception of his commander: if the magister was remembered as honorable and accomplished, then so would Ammianus. For this reason, he did not record this part of history with the same truthfulness as others. Despite this problem, we can use Thompson’s critique to recognize Ammianus’ biases and continue using this source to understand important details that we could not otherwise obtain on the nature of the magistri militum. Comparing the differences between the roles of Ursicinus, Sabinianus, and Arbitio – the three most contemporaneous magistri to Ammianus’ military career – will help to further illuminate the roles of the early magistri. Ursicinus is first identified as a soldier under Constantine I, and he might have been a dux in Pannonia Superior and Noricum.175 He is known to be a magister militum in the early 350s, but it is possible that he received his promotion to the magisterium in the wake of the major Roman defeat at Singara (Sinjar, Iraq).176 This event has been dated to either 344 or 348, when Constantius was still in the east.177 The Romans routed the Persian army in the field, and then the emperor ordered them to fall back and regroup, but the soldiers pursued the Persians to their camp.178 The Romans looted the camp, but the Persian shahanshah (‘King of Kings’), Shapur II, brought up his reserve archers and took advantage of nightfall to counter-attack, killing many Romans.179 The sources do not mention any Roman officers, magistri or otherwise, being slain during this engagement, but the Persian missiles could have easily struck and killed one if he had been present with the soldiers. Moreover, even if the commanding officer was not killed in the ambush, it is possible that he would have received the blame for not having adequate command over 172 Parnell 2017, 6. 173 Thompson 1947, 54. 174 See page 24. 175 As a soldier under Constantine, see: AM 15.5.19. For the brick stamps, see: CIL III 11853c, 4656; AE 1934, 272a-b. For the title of magister equitum, see: AM 14.9.3. 176 Festus 27; Eutr. 10.10.1; Lib. Or. 18.208; Soc. 2.25; Jul. Or. 1.26b. 177 For 348, see: Seeck 1919, 196; Barnes 1980, 163–4, n. 13. For 344, see: Portmann 1989. 178 Lib. Or. 59.99–120; Jul. Or. 1.22d–25b; Festus 27.1–3; Eutr. 10.10.1; AM 18.5.7; Jer. Chron. s. a. 348; Oros. 7.29.6; Soc. 2.25.5; Zon. 13.5. 179 This project will refer to the Persian shahanshah in the common Anglicized manner of ‘Shapur’, but in later chapters we will refer to a magister militum with the same name as ‘Sapores’, both to avoid confusion and follow prevailing parlance.

Ammianus’ Account of Some Magistri Militum

55

the battlefield, and may have been punished with demotion or dismissal. In either of these scenarios, Ursicinus could have been promoted in his place. Additionally, Ammianus says that, in 359, Ursicinus had held command in the east for ten years, indicating that his promotion could have occurred after this battle, if it took place in 348.180 Ursicinus’ first recorded command was during the ‘Gallus’ or ‘Jewish’ revolt of 351. He is not documented by the Latin and Greek accounts, which state that the eastern Caesar Gallus took command, and violently razed three cities during the suppression of the unrest.181 For this reason, Ursicinus’ involvement is omitted from major academic works on the magistri. However, the Palestinian Talmud recounts that it was Ursicinus who was in command of the Roman forces, and it downplays the severity of the revolt, even indicating friendly relations between the local rabbis and the Roman commander.182 Furthermore, the archaeological record does not corroborate the violent destruction of cities in this area, indicating the Latin and Greek sources may be inaccurate.183 This has led historians to reinterpret this mutiny as a series of local, decentralized, and non-elite disturbances, rather than a revolt led by the elites and involving large numbers of participants.184 Ursicinus’ involvement in the Jewish revolt can be compared with the Isaurian raids of 353 to further understand the responsibilities of a magister militum.185 The Isaurians in southern Anatolia attacked travelers and plundered trade ships along the coast, but despite Ammianus describing it as a “troublesome war” (bella gravia) it seems to have been more or less a minor affair, as passing cohorts of cavalry and local garrisons were able to fend off the attackers from most towns.186 The most difficult part seemed to have been getting the Isaurians to offer battle, because they would disperse into the mountains when the Roman soldiers approached.187 Gallus eventually ordered Nebridius, a comes, to redress the situation, and Ammianus tells us that he was only sent because Ursicinus was too far away.188 This implies that, had he been closer, the magister would have been sent to deal with the insurgency. Therefore, we can conclude that the magistri did not only take command during major campaigns, but could also be charged

180 181 182 183 184 185 186 187 188

AM 18.6.1–2. Aur. Vict. 42.9–12; Theoph. 5843; Soz. 4.7; Soc. 2.33. j. Shebi 4.2; cf. j. Ber 5.1. Schäfer 1986, 184–201, showed that the archaeological record does not reflect the story of a Roman destruction of these cities, but perhaps the story comes from destruction suffered during an earthquake in 363. j. Yeb 16.3 clearly indicates a troubled relationship with the general population. For more details, see: Mor 1989, 339–341; Nathanson 1986, 34; Geiger 1979; 1979/80; Avi-Yonah 1962, 178–9; Lieberman 1946, 336–7. See Matthews 1989, 362–4, for a more extensive examination of this revolt, and Lenski 1999 for the Isaurians in general. AM 14.2.1. For the attacks on travelers and trade ships, see: AM 14.2.2–4. AM 14.2.6–7. AM 14.2.20.

56

The Magistri Militum from 341 to 363 CE

with managing less important operations and battles, which further evidences that the Jewish Revolt might have only been a small disturbance. Ursicinus then became involved in the trials orchestrated by Gallus to remove his opponents and secure his independence from Constantius. Held in Antioch, Ursicinus was summoned to give the trials an air of legitimacy.189 This illustrates the diverse roles that the magistri could play across the Roman government, and not just in military matters, especially during special periods of terror or purges.190 We will see these powers increase over the period. Ursicinus apparently realized the trials were a sham, and wrote secret letters to inform Constantius, although this is possibly just Ammianus trying to portray him in a better light and smooth over his role in passing judgement against innocents.191 Ammianus would have us believe that informing Constantius actually led to the emperor being suspicious of Ursicinus, mostly due to the insidious courtiers whispering in his ears.192 Ursicinus then became involved in the usurpation of Silvanus in August 355 that we discussed in the previous section. As outlined, a plot tarnished Silvanus’ reputation so badly that, even though he was innocent, he feared Constantius’ retribution. In an attempt to ensure his safety, he had himself declared emperor. This caused Constantius a great deal of consternation. His armies had been greatly weakened against Magnentius, and another civil war against Silvanus would only deplete his resources further. In order to avoid this outcome, he sent Ursicinus to Cologne to defeat Silvanus through guile.193 Ursicinus came to Silvanus as a friend, and ingratiated himself into Silvanus’ inner-circle. While he was doing this, the officers that accompanied Ursicinus – including Ammianus – incited a small group of Silvanus’ soldiers to betray and assassinate the usurper.194 This story as recorded by Ammianus hints strongly at significant bias. On the basis that furtively duping and assassinating Silvanus was a very dishonorable task, Thompson argues that Ammianus would not have been happy being involved in the deceit, and tried to save his reputation by misleading his audience.195 Ammianus claims that, just before Silvanus’ usurpation, Constantius and a cabal of sycophants almost had Ursicinus executed, but that he was saved and pardoned so he could be sent on this mission.196 Thus, Ursicinus was forced to agree to dupe Silvanus because of fear for his own life, thereby justifying both his and Ammianus’ participation. Ammianus strengthens this tale by comparing Ursicinus to Domitius Corbulo, a “steadfast and loyal defender of the provinces,” who had been ordered to commit suicide by Nero, the archetype 189 190 191 192 193 194 195 196

AM 14.9.1. Cecconi 1999, 81. Blockley 1972, 443. AM 14.11.2. AM 15.5.18–22. AM 15.5.30–31. Thompson 1947, 3, 43, 45, 52. AM 15.5.5–6.

Ammianus’ Account of Some Magistri Militum

57

of a paranoid and evil emperor.197 Furthermore, Ammianus does not specifically implicate himself or Ursicinus as the ones who recruited the assassins.198 Nonetheless, there seems to be no reason to doubt the general course of events, and we can use this episode to understand that magistri did not always have to command armies, but could be ordered to carry out surreptitious and irregular tasks.199 Ursicinus spent the next few years in Gaul before being transferred back to the east in late 357.200 By late 359 he was summoned back to Constantius’ court and the eastern command was taken over by a man named Sabinianus.201 This new magister made for an interesting choice because he was not particularly militarily experienced. Ammianus describes him as cultured but inbellis.202 He was a pious Christian, something that might have turned Ammianus against him, and led him to use Sabinianus as something of a foil for Ursicinus; where Ursicinus was adept and experienced, Sabinianus was untried and incompetent.203 Nonetheless, there still is no evidence to indicate that Sabinianus was actually particularly proficient in war, and Constantius may have promoted him for some other reason. One possible explanation is that Constantius may have wished to remove Ursicinus from the eastern frontier to prevent him gaining the glory of any victories there. The magister already had a renowned military reputation, which increased the possibility of him challenging Constantius’ power. If Constantius planned to return to the eastern frontier the following year it would also ensure the glory would go to him, helping him counter the growing fame of the new Caesar of Gaul, Julian, who will be discussed further in the following section. Another potential explanation is that Sabinianus was appointed to further Constantius’ new approach to the eastern frontier.204 Constantius had needed to extract a significant number of eastern troops for the civil war against Magnentius, and many had died at the bloody Battle of Mursa in 351.205 Constantius had then been occupied fighting the Alamanni and Sarmatians, and had not been able to return the remaining troops.206 Thus the eastern frontier was under-manned, and would struggle to repel an invasion. To try to delay any Persian offensive, Constantius had been trying to negotiate with Shapur and come to some sort of armistice.207 Sabinianus may have been appointed as part of this approach.208 197 198 199 200 201 202 203 204 205 206 207 208

“provinciarum fidus defensor et cautus” AM 15.2.5. See: Vervaet 2002, 189–90. AM 15.5.31. Doubts have been expressed by Drinkwater 1994, 574–5, and Barnes 1998, 18. AM 15.13.3, 16.2.8, 10.21; Thompson 1947, 45. AM 18.5.4–5. AM 18.5.5, 6.5; Corcoran 2014, 204–5. Szidat 1991; 1992, 110–11; Barnes 1998, 85–86. Thompson 1947, 50. Eutr. 10.12.1; Epit. 42.2. AM 15.4.1, 16.10.20. AM 16.9; Blockley 1992, 17–22. Thompson 1947, 48–49.

58

The Magistri Militum from 341 to 363 CE

However, when Shapur learned that Ursicinus had left the east and been replaced by someone less competent, he immediately launched a surprise invasion. Ursicinus was hastily directed to return to the east and aid Sabinianus in defending the frontier.209 Ammianus now provides insights into the day-to-day activities of a Roman magister militum during the crisis of an invasion. Ursicinus and Ammianus aid fleeing civilians, dodge Persian patrols, confront spies and traitors, and prepare for the inevitable sieges.210 Ammianus ended up inside Amida when Shapur besieged it, while Ursicinus managed to escape and re-join Sabinianus and the rest of the eastern army, probably in Antioch.211 There, Sabinianus retained the primary position and authority even though Ursicinus was the more experienced general.212 Ammianus tells us that Ursicinus wished to relieve Amida by sallying against the Persians, but was stopped by Sabinianus, who was under orders from Constantius to not risk the lives of his soldiers.213 Ursicinus was only permitted to send a few scouts to Amida.214 In the end, the Persians successfully breached the walls of the city after a seventy-three day siege.215 Ammianus only managed to escape by fleeing out of a postern gate after nightfall.216 Afterwards, an investigatory commission would find Ursicinus guilty of not saving Amida, and he verbally lashed out and insulted the emperor.217 Constantius did not tolerate this insubordination, and he dismissed Ursicinus.218 The magister militum Arbitio provides a highly interesting counterpart to Ursicinus, further complementing our understanding of the different roles of the magistri. Arbitio is said to have begun his career as a common soldier, the only magister attested as doing so, and then rose as high as the office of dux under Constantine I, the same as Ursicinus.219 His career during the 340s and early 350s is unknown, but Ammianus indicates he excelled during the civil wars against Vetranio and Magnentius.220 This caused him to gain a highly privileged position under Constantius, confirmed by the fact that he was the only magister militum during reign as sole emperor to receive a consulship.221 Whereas Ursicinus was moved all across the Empire at Constantius’ discretion, Arbitio remained in a stable position with the emperor and commanded the primary

209 210 211 212 213 214 215 216 217 218 219

AM 16.10.21. AM 18.6.10–16, 18.8.4–11. AM 19.3. AM 19.3.1. AM 19.3.1–2. AM 19.3.3. Matthews 1989, 58. AM 19.8.5. AM 20.2.4. Thompson 1947, 51–52. On his common soldier beginnings, see: AM 15.2.4, 16.6.1. See page 154 for further discussion. As a dux under Constantine I, see: AM 16.9.4. 220 AM 21.13.16. 221 Athan. Hist. Ar. 81; AM 15.8.17.

Ammianus’ Account of Some Magistri Militum

59

field army. He operated in campaigns against the Alamanni in Gaul, and probably also against the Sarmatians across the Danube.222 Such a placement with the emperor could afford Arbitio the privileged title of magister militum praesentalis, but we must remember that that epithet was not in use during this period, nor were the regional command titles such as per Gallias or per Illyricum. It is thus difficult, perhaps impossible, to compare the court magistri with the regional appointments. It is more accurate to the period to simply describe the fact that Arbitio’s closeness to the emperor allowed him to build a stronger personal relationship with Constantius than those generals posted elsewhere in the Empire. Arbitio clearly benefitted from Constantius’ patronage, as not only did he receive the consulship, but he amassed a fortune and a great deal of property.223 Arbitio was also able to develop his own network of patronage: when he was accused of aiming to make himself emperor, his powerful allies at court were able to protect him from any repercussions.224 According to Ammianus, Arbitio’s proximity to the emperor also allowed him to defame and malign his political opponents to further his own position. Ammianus records him favoring his friends, sabotaging other magistri, and executing those he disliked.225 Ammianus even compares him to a snake waiting to strike at potential victims.226 However, if he was actually doing these things, Arbitio was not very successful. For all Ammianus paints Arbitio as working against him, Ursicinus’ career progressed perfectly normally.227 In the following section, we will see how Arbitio was involved in the fall of the magister Barbatio in 358, but it was not a plot of his origination.228 Part of this contradictory depiction is the result of an exaggeratedly negative portrayal by Ammianus because he harbored ill-will over Arbitio’s role in Ursicinus’ downfall.229 Therefore we should doubt whether Arbitio was truly a mastermind behind all of these intrigues, and if he was attempting to manipulate Constantius, he was unable to do so with significant effect. The emperor was able to keep his magistri under control. The accounts of the magistri that Ammianus personally met that are preserved in his work are unparalleled for the details and depth of information they provide. While his work can be subject to significant biases, there are still important kernels of truth that are highly illuminating. We can learn that the magistri had incredibly varied roles: they could command campaign armies in large operations or be tasked with the suppression of minor revolts and civil disturbances. They could even be sent on missions without an army at all. Some magistri had well-known reputations for military experience, 222 223 224 225 226 227 228 229

AM 15.4.1–13. AM 16.8.13, 26.8.13. AM 16.6.1–3. AM 15.2.4, 3.11, 5.1–2, 5.8, 18.3.1–4, 20.2.3. AM 15.12.4. Thompson 1947, 51–52. AM 18.3.1–4. Barnes 1998, 108.

60

The Magistri Militum from 341 to 363 CE

and yet Sabinianus was appointed as a magister to de-escalate the military situation and avoid conflict. Magistri could remain in a stable position, or be regularly transferred and relocated at the emperor’s discretion. It seems that the functions of the magistri militum varied greatly according to circumstances. 1.5 Julian’s Western Roman Empire Returning to the western Empire, after the death of Silvanus in September 355, there was a period of rapid turnover for the magistri militum, and it will be posited here that Julian, the Caesar of the west, was responsible for causing many of the changes.230 Constantius had only appointed Julian to be a subservient and junior imperial presence, with real authority resting in the generals and bureaucrats.231 However, Julian was not content with this, and it is clear that he began working towards rebellion against Constantius almost as soon as he arrived in the west.232 It has been argued by John Drinkwater that a crucial part of this preparation was using campaigns against the Alamanni to gain the political capital and loyalty of the soldiers that was necessary for a rebellion.233 By leading his soldiers directly from the frontlines in numerous victorious battles, Julian demonstrated his competency and willingness to place his own life at risk alongside the lives of his soldiers, further gaining their admiration. In this section, it will be proposed that sidelining and removing the magistri was an essential part of ensuring the renown from these victorious campaigns and daring operations went to Julian alone. Once again, Ammianus is the most detailed source for this period, but he idolized Julian as a fellow polytheist and a learned man.234 Furthermore, he was only present in the western Empire for the first years of Julian’s time as a Caesar, so the majority of his information is second-hand.235 Thus, we must work to demystify some of the obscurities about the western magistri in these years. In 356, Julian commanded his first operation along the Roman side of the Middle Rhine where he dealt with barbarians who had settled along the riverbank.236 After returning to winter quarters, he was besieged by Alamanni in the city of Sens, and Mar-

230 For biographies of Julian, see: Browning 1975; Bowersock 1978. 231 Jul. Ep. ad Ath. 277d. Constantius even personally wrote a directory for things such as what Julian should eat (AM 16.5.3). 232 Jul. Ep. ad Ath. 281b-c. 233 Drinkwater 2007. This has been identified as a tactic also employed by other emperors. See: Bleckmann 2009, 193, cf. 211. 234 Thompson 1947, 72–73, 84–85. 235 His source may have been the eunuch Eutherius. See: Thompson 1947, 4–5, 20, 46. 236 AM 16.2.12; Jul. Ep. ad Ath. 279A; Lib. Or. 12.44, 12.48, 18.36–37; cf. Eun. Hist. fr. 16.2. Also see: Drinkwater 2007, 219–20.

Julian’s Western Roman Empire

61

cellus, Julian’s nearby magister equitum, did not send reinforcements to his aid.237 Ammianus does not record a reason why Marcellus did not relieve his Caesar. A possible explanation is that he was aware the besiegers did not have the necessary supplies for a long winter investment, nor the knowledge of siegecraft to successfully assault Sens. If so, he was correct: the Alamanni retreated after one month having done no harm.238 However, Ammianus records Marcellus’ actions as shameful, and this is perhaps the same language that Julian used to report the events to Constantius.239 Such a report could have been a ploy to defame the magister militum, and if so it was a success because Constantius dismissed Marcellus for his inaction.240 Resentment over this unfair dismissal would also explain why Marcellus travelled to Milan to protest his termination.241 Marcellus’ son would also be executed for conspiring against Julian a few years later, which is further evidence that there was a sour relationship between the Caesar and the magister.242 Thus it seems possible that Julian took advantage of the circumstances and exaggerated Marcellus’ actions in order to have him removed.243 David Hunt connects Ursicinus’ transferal from the west with the downfall of Marcellus, arguing that Ursicinus was also blamed for not sending reinforcements to Sens.244 A closer inspection of the chronology, however, reveals that the two generals were separately removed from Gaul. Marcellus was first dismissed and then travelled to Milan to protest to Constantius, which must have occurred before Constantius went to Rome in late April 357.245 It was only after Constantius travelled to Illyricum at the end of May that he recalled and replaced Ursicinus.246 Ursicinus had probably been left in Gaul so that at least one experienced magister was in the province. There is no indication he faced repercussions for not relieving Sens, which could be an intentional omission by Ammianus, but given the fact that Ursicinus seems to have only been a subordinate or auxiliary magister at this time, it may have been outside his authority to relieve Sens, similar to how he was only able to send a small number of scouts to Amida.247 To replace the removed magistri, Constantius appointed an experienced general named Severus as magister equitum in Gaul, and Barbatio as magister peditum in charge of the imperial comitatus, at the time stationed in Italy.248 Barbatio would take this Ital237 238 239 240 241 242 243 244 245 246 247 248

AM 16.2.8, 4.3. AM 16.4.2. AM 16.4.3. AM 16.7.1–3, 10.21; Lib. Or. 18.48; Hunt 1997b, 50; Drinkwater 2007, 227–8. AM 16.7.1–2. AM 22.11.2; Eun. Hist. fr. 25.5. For a theory on what may have happened to Marcellus later in life, see: Woods 1995. Hunt 1997b, 49. AM 16.7.2, 10.1–13. The timing is calculated from AM 16.10.20. AM 16.10.20–21, 11.1. On his replacement, see below. See page 58. This general, Severus 8 in the PLRE, was magister equitum from 357–8 in the west, and he shares the name of the PLRE’s Severus 10, who was magister peditum from 367–72, also in the west. Fortunate-

62

The Magistri Militum from 341 to 363 CE

ian army on campaigns to the upper and middle Rhine River as well as to the provinces of Raetia and Noricum. Barbatio’s actions during one such campaign in 357 may be the first clear evidence of a magister treasonously subverting their imperial commander.249 After some initial skirmishing, the main Alamanni army fled to some islands in the middle of the Rhine.250 Barbatio had boats and other materiel to bridge the river and get to them, so Julian requested these supplies. Barbatio instead destroyed it all and turned his army back to Italy.251 Afterwards, he travelled to Constantius as if he were a victor, and slandered Julian at court.252 Drinkwater suggests that there were tense relations between the Caesar and Barbatio from the beginning of his appointment, and Barbatio may have also feared suffering a similar fate to Marcellus if he came too close to Julian.253 This could explain his sabotage. John Matthews, however, points out the Julian diverted from the original plan ordered by Constantius, forcing Barbatio to return to garrison rather than also betray the senior emperor’s orders.254 In this case, the destruction of the supplies may have actually been to allow him a swift retreat and keep it out of the hands of the Alamanni, rather than direct sabotage of the Caesar. In either case, Julian was happy to let the magister leave so he could conduct his own campaign, which led to the major Battle of Strasbourg, in which the magister Severus played a key role and Julian’s reputation as a commander gained a significant boost.255 The next year, Severus and Julian gathered the army from its winter garrisons and first moved against the Franks in northern Gaul. Severus performed well, cutting off the enemy’s retreat and forcing them to surrender.256 Julian then led his army across the Rhine, and Ammianus records that Severus suddenly lost his courage, and intimidated the scouts into delivering false reports in order to stall the army’s advance.257 This is rather out of character, because Ammianus wrote multiple times that Severus was experienced and capable.258 Ammianus explains that he may have become afraid due to a premonition of his own impending death.259 It seems possible, however, that Constantius had ordered Severus to assist Julian in defending Roman territory, but

249 250 251 252 253 254 255 256 257 258 259

ly, Severus 8 exists entirely in this chapter, while Severus 10 is in the following chapter. It should therefore be easy to distinguish them. In Chapter Four, where they appear together in multiple places, their PLRE numbers will be used to distinguish them. Obviously discounting the actions of the usurpers, and the less direct inaction of someone like Marcellus. AM 16.11.8. AM 16.11.14. Woods 2010 proposes an alternate and unconvincing idea that Barbatio’s boats were not intended to be used for a bridge, despite all the evidence explicitly stating so. AM 16.11.15. Drinkwater 2007, 231–35. Also see: Rosen 1970, 94, contra Woods 2010. Matthews 1989, 299–300. AM 16.12.1–27. AM 17.8.1–4; Drinkwater 2007, 241–46. AM 17.10.1–2. AM 16.10.21 11.1, 17.10.1, as did Lib. Or. 18.48. AM 17.10.2.

Julian’s Western Roman Empire

63

not in aggressive campaigns across the frontiers. This would explain why Severus conducted himself competently during the Battle of Strasbourg and in northern Gaul, but not when Julian took his army across the Rhine. Constantius’ motivation would have been to ensure the provinces were secure but minimize the glory Julian could win. We cannot be certain if Severus actually died on this campaign after his pusillanimous behavior, but he was not in command the following year.260 There were no major engagements during this campaign, hence there is no easy spot to place Severus’ death in a battle.261 It is possible he died of natural causes, but also possible that Julian grew suspicious of his actions and had him executed, either publicly or privately. Ammianus may be preserving an incorrect story because he was unaware of Julian’s schemes, or he was covering them up because of his admiration for the Caesar. Ammianus does not record whether Barbatio again joined Julian in 359, just that he spent the year on a campaign.262 Ammianus reports that while he was away, a plot against him was hatched by one of the ancilla in his household. This servant had previously served Silvanus, and so must have been given to Barbatio by Constantius as some sort of gift or reward. Although it was not Barbatio who had been responsible for overthrowing him, she seemed to have harbored ill-will against him for taking over Silvanus’ position, or perhaps he was just a poor master.263 Thus, after Barbatio returned from campaigning, she brought some incriminating letters to Arbitio, who then, likely wishing to gain more favor with the emperor, showed them to Constantius.264 Barbatio and the ancilla were both executed. If we could be certain that Barbatio had joined Julian across the Rhine we would be able to infer Julian’s hand in the conspiracy more confidently, but otherwise we can only implicate Arbitio and the ancilla. No magister peditum replaced Barbatio as commander of the Italian army, and until the 380s it would be commanded by lower officers.265 Thus a third magister was removed from office in Julian’s western Empire. In 359 Constantius appointed Lupicinus to be Severus’ replacement.266 The following year, the Scots and Picts raided Britain. While attacks there could reasonably warrant the response of a magister, it is a conspicuously opportune coincidence that Julian claimed it was imperative that he remain in Gaul, so Lupicinus had to be sent to distant

260 261 262 263 264 265

AM 18.2.7; cf. Demandt 1970, 574. AM 17.5–10. AM 18.3.2. AM 18.3.2. AM 18.3.2–4. As discussed above, Ursicinus was moved into Barbatio’s position, but he never took it up. Lucillianus would be ordered to take up the position in 363, but was diverted to Gaul (see pages 72–73 for further discussion). 266 AM 18.2.1–19; Drinkwater 2007, 247.

64

The Magistri Militum from 341 to 363 CE

Britain where news would only intermittently reach him.267 This once again ensured all the renown for the campaigns in 360 went solely to Julian, and it allowed him and his supporters to work towards the goal of usurping power without the risk of Lupicinus reporting to Constantius.268 Before news of Lupicinus being sent to Britain had reached him, Constantius sent a summons for the magister to bring a significant number of western troops to the Persian frontier.269 This comports with a law from the same year contained in the Codex Theodosianus which states that the magister militum equitum et peditum must be consulted on any transferal of soldiers so as to ensure the security of the Empire.270 Thus strategic planning was a key duty of the military magisterium. The soldiers requisitioned from Julian were purportedly to help shore up the east’s defenses in the wake of the defeat at Amida, and while this is a logical defensive decision, it is believed that Constantius’ true intent was to weaken Julian’s position by reducing the size of his army.271 The Gallic soldiers refused to leave their homes in the west because Julian had promised them they would never be forced to fight across the Alps, and to cement their resistance to Constantius’ orders they proclaimed Julian to be their rightful emperor.272 It is no coincidence that this was done far away from both the magister militum Lupicinus and the Praetorian Prefect Florentius, who were still loyal to Constantius.273 Constantius had also sent Gomoarius to replace Lupicinus.274 Gomoarius was known to be untrustworthy because he had betrayed Vetranio in 350, so his appointment helps to demonstrate that Constantius’ policy had changed from appointing beneficial veteran commanders to actively undermining Julian’s position. However, Gomoarius is not recorded as participating in any operations in the west during the time he was there, and it may be that Julian sent him back to Constantius as early as mid-360.275 Thus, it seems that Julian continued to ensure the magistri remained sidelined so he could exert total control over his half of the Empire and its armies. Julian spent some time in Gaul organizing and legitimizing his new government, but eventually decided military action was required to effectuate his rebellion. In preparation, he ousted those who retained loyalty to Constantius and replaced them with men

267 AM 20.1.2–3. Drinkwater 1983, 370; Matthews 1989, 95–96. See pages 80–82 for more attacks on Britain that warranted the response of a magister militum. 268 For the campaigns that year, see: AM 20.10.1–2. 269 AM 20.4.1–6; Zos. 3.8.3–4. 270 CTh 1.7.1. 271 Drinkwater 1983, 383–87. 272 AM 20.4.4–22. 273 AM 20.4.6–9. Although Julian sent a summons to Florentius, Browning 1975, 100, theorizes that this was probably not an earnest summons, and Julian may have been hoping the Prefect would continue to keep away. 274 AM 20.9.5. 275 Woods 2016, 183–84.

Julian’s Western Roman Empire

65

he trusted. This included promoting Nevitta to the office of magister equitum.276 Julian also halted all traffic crossing to Britain, which prevented Lupicinus from hearing of the rebellion and returning to resist him.277 He would ultimately be arrested upon his return to the continent.278 Once these changes were completed, Julian mobilized his forces to march eastwards. Although the soldiers had previously refused to cross the Alps for Constantius, they were willing to do whatever was asked of them for Julian, demonstrating that he had masterfully secured their loyalty.279 The manner in which he conducted his army on this campaign is particularly useful for understanding the roles of the magistri. During the initial march, he split the army into three groups. He assigned one to himself, one to Nevitta, and the third he assigned to two other senior officers, Jovinus and Jovius.280 Ammianus does not specify a rank for these two at this time, but he expressly differentiates them from the rank of the magister militum Nevitta.281 Additionally, he indicates that Jovinus was only promoted to the magisterium later in the year.282 Therefore it can be safely assumed that Jovinus and Jovius were of a lower rank and were probably less experienced, and that is why their portion of the army required two commanding officers. Julian first targeted the city of Sirmium. Travelling with a small force by boat rapidly down the Danube, they disembarked at Bononia (Banoštor, Serbia), about nineteen miles away from Sirmium.283 He then sent a small party under the command of the comes domesticorum Dagalaifus to capture Lucillianus, Constantius’ magister equitum.284 Lucillianus had likely performed well and been noticed for his successful diplomatic mission to Persia in 358.285 He may have taken over the Danube theatre once Constantius and Arbitio left the region in early 360. David Woods offers a reinterpreted chronology of this episode which would have Arbitio as the magister equitum in Sirmium at this time.286 Woods, however, seems to be ignorant of a passage in Ammianus that places Arbitio on the eastern front at this time.287 Ammianus’ accounts of this campaign is unfortunately broad and lack detail, as opposed to his accounts of the Battle of Strasbourg or Julian’s Persian campaign. Ammianus’ style is also to provide us with the entire story of Julian’s advance through the Balkans, and then discuss all of Con276 AM 21.8.1. Nevitta is called magister armorum here, but is called magister equitum at AM 21.8.3. 277 AM 20.9.9. 278 Jul. Ep. ad Ath. 281a-b. 279 AM 21.5.10. 280 AM 21.8.2–3; cf. Zos. 3.10.2. 281 AM 21.8.1. 282 AM 21.12.2. 283 Dušanić 2004, 56. 284 AM 21.9.5–8, 10.1. 285 AM 17.14.3, 18.6.17; Demandt 1970, 575–76. 286 Woods 2016. 287 AM 21.13.3. Woods 2016 further proposes that Constantius promoted Jovinus to the magisterium despite Julian doing this in AM 22.3.1 as well as the marching order of 361 as described above.

66

The Magistri Militum from 341 to 363 CE

stantius’ responses but without any chronological references to Julian’s action, so we are left to try to match up the timing. Woods’ best demonstration is that it is possible to reconstruct these events in a plethora of ways and interpretations. Ultimately, Dagalaifus successfully captured the magister equitum, and Julian entered the city. As infiltrating a fortified city would be a difficult task on its own, and including breaching the military garrison and sneaking up on the commander of the city makes it nigh impossible, it is more likely that the magister surrendered or defected to Julian. Ammianus might have been ignorant of the exact occurrences at Sirmium, or he may have wanted to preserve the reputations of those involved, or even just vaunt the derring-do of Julian and his officers. Afterwards, Julian appointed Nevitta and Mamertinus, the Praetorian Prefect of Illyricum, to the consulship of 362.288 Julian doubted the loyalty of the two legions stationed in Sirmium, so he sent them back to Gaul where they could not cause trouble. Unfortunately for Julian, they were persuaded by a man named Nigrinus, a commander of a turma of cavalry, to divert to Italy and seize the city of Aquileia in Constantius’ name.289 This was a serious problem, because Aquileia had strong walls and could be easily defended, and the soldiers there could potentially cut Julian off from easy routes back to his western powerbase. Hence it was critical that he retake the city. Nevitta had already been tasked with occupying the important Succi pass, so Julian promoted Jovinus to the rank of magister equitum and sent him to encircle Aquileia.290 After a period of time, Julian determined that the siege was going to be long rather than difficult, and so sent a different commander to take over, freeing up Jovinus for more important duties. Thus, while earlier with Ursicinus we established that magistri could be assigned minor tasks, this indicates that, when multiple missions were presented, the most important went to the magistri. Furthermore, if a mission were significant enough to require a magister militum but none were available, one could be promoted. Lesser or more furtive tasks, such as sneaking into Sirmium to capture its commander, fell to lower ranked officers. Constantius used his officers in a similar manner. Agilo, as eastern magister peditum, had filled the space left by Sabinianus’ dismissal, and along with Arbitio, they commanded Constantius’ armies against the Persians and led the forces that moved to oppose Julian.291 Gomoarius had lost his rank of magister militum after being dismissed by Julian, and seemingly was not reinstated in the east, perhaps because Constantius already had two magistri. However he still played an important role, taking troops to oppose Nevitta in the Succi pass.292 Before the armies could meet in a major battle,

288 289 290 291 292

CTh 7.4.7, 8.1.6; CIL VI 753; P. Flor. 1 30.25; AM 21.10.8, 12.25, 22.7.1; Bagnall, et al. 1987, 258–59. AM 11.2–3. AM 21.12.3. For their operation against the Persians, see: AM 21.13.3. On moving against Julian, see: AM 21.13.16. AM 21.13.16.

Julian and Jovian as Sole Emperors, 361 to 364 CE

67

Constantius died from a fever on 3 November 361.293 His civilian bureaucrats and military officers assembled to discuss whether to raise a new emperor and continue prosecuting the war, but they decided that was not in their best interests and they allowed Julian to take control of the entire Empire as the sole sovereign.294 This chapter has attempted to demonstrate that marginalizing and removing the magistri militum that Constantius appointed in the west was part of Julian’s plan to gain the necessary military distinction and loyalty for a successful rebellion. It is undeniable that this argument is somewhat circumstantial – Ammianus would never level such an accusation against Julian. Thus, mystery remains regarding the fates of some magistri and these events could be interpreted in a different light. We cannot be certain that Julian had a hand in the downfall of Barbatio or Severus, although the circumstances around their disappearances are conspicuous. Yet, the reaction of Marcellus to his dismissal is best explained on the grounds that he was unfairly accused of cowardice by Julian. Furthermore, there can be little doubt that keeping the officials who were loyal to Constantius generally away from the centers of power in the west was a top priority of Julian because of his treatment of Lupicinus, Gomoarius, and Florentius, and therefore it should not be a surprise if Julian also tried to sideline the magistri as a whole. If this was truly the cause for the high turnover of western magistri during these years, it strongly suggests that the magistri were still serving at the whims of their imperial rulers and enjoyed limited personal agency in political affairs. 1.6 Julian and Jovian as Sole Emperors, 361 to 364 CE The successive reigns of Julian and Jovian as sole emperors contribute fascinating insights into how the magistri began to push back and exert their own will upon the emperors, as well as how the military administration could be used by an emperor to establish both long- and short-term security. This section seeks to demonstrate how the magistri initially held a strong position vis-à-vis emperor Julian, and how he was able to regain control over them in the lead up to his major campaign against Persia. After his unexpected death in the midst of this campaign, the officers and officials of the army were presented with the first of multiple important succession crises that occurred during the period under review here. The magistri were some of the important individuals in the search for a new emperor and their indecisiveness resulted in the soldiery promoting a candidate of their own choosing, which took the opportunity away from the magistri to install an emperor. Finally, in this section we will also see

293 Jer. Chron. s. a. 361; Cons. Const. s. a. 361; AM 21.15.3; Epit. 42.17; Soc. 2.47.4, 3.1.1; Soz. 5.1.6; Zon. 13.11. 294 AM 21.15.4; Szidat 2010, 103–4.

68

The Magistri Militum from 341 to 363 CE

how there was a surge of magisterial appointments during Jovian’s brief reign as he tried to bolster the precarious security of his position. After learning that Constantius had died en route to confront him, Julian moved to Constantinople and was acclaimed by its Senate and people as the rightful ruler over the entire Empire.295 Not long after, the trials known as the Commission of Chalcedon began, and it seems that they were orchestrated by the military officials from both halves of the Empire.296 Only two of the appointed judges were civil officers, while the others were all four of the presently-serving magistri: Nevitta, Arbitio, Jovinus, and Agilo.297 Moreover, the Commission took place under the protection of two of the most important units in the army, the Ioviani and the Herculiani, and even though the Praetorian Prefect was supposed to be the chief official, Arbitio steered the judgements.298 Furthermore, all the accused came from the civil bureaucracy, including even close friends and allies of Julian, such as Ursulus.299 He had helped established a close relationship with Julian by securing pay for his soldiers while he was Caesar in Gaul, and there is no indication that their relationship had soured.300 However, the eastern military officers resented Ursulus, because when he saw the ruins of Amida, he bemoaned that the soldiers did not have the spirit to defend the Empire even though they were paid so much it was bankrupting the treasury.301 Additionally, some who had opposed Julian, like the spy Gaudentius, were never tried by the Commission.302 Thompson therefore surmises that the Commission was not created by Julian, but formed to give release to the resentment of the army and to help conciliate the eastern magistri

295 AM 22.2. 296 Cf. Kulikowski 2019, 27–28. 297 AM 21.12.3, 13.3, 22.3.1. It is perhaps the right place here to address a 1999 article by David Woods, who theorizes that the former dux Aegypti Artemius was promoted to the magisterium by Constantius and continued in this position under Julian before his execution in late 362. Woods makes a plausible argument that may very well be correct. The Artemii Passio (AP 35) indicates that Artemius had authority over Egypt and Syria, which must mean he was in a position higher than only a dux. Furthermore, it is plausible that Ammianus remained silent on Artemius’ promotion because of his anti-Christian biases. However, Woods does not adequately prove that Artemius could not have been promoted as comes rei militaris instead. He assumes that a comes would not be able to intervene in Egypt, as Artemius is said to have been able to do, but this is not exactly clear; after all, a comes outranked the dux Aegypti. Furthermore, two known duces Aegypti, Sebastianus and Traianus, were promoted as comites rei militaris before reaching the magisterium, indicating this was the regular career path, as will be discussed further in Chapter 4.2. Finally, without the same supporting evidence we have used for other potential magistri, such as the granting of a consulship, it is my opinion that assigning a magisterial promotion to Artemius is too flimsy for this prosopographical project. 298 AM 22.3.1–2, 3.9. 299 Thompson 1947, 73–75. 300 AM 22.3.7. 301 AM 22.3.8. 302 Gaudentius had spied on Julian’s doings while he was Caesar, and reported them to Constantius (AM 17.9.7, 21.7.2). He was only executed later (AM 22.11.1).

Julian and Jovian as Sole Emperors, 361 to 364 CE

69

with Julian’s new regime.303 Julian’s claim that he had no knowledge of the sentences the Commission delivered could have been a lie to save his reputation, or it could support Thompson’s conclusion that he either gave the army free reign so he could win their favor, or was unable to stop them doing as they pleased because he feared their power.304 Once Julian had established the short-term security of his reign through the Commission and other measures, he began formulating plans to establish long-term legitimacy. Constantius’ death had enabled him to combine the western and eastern field armies without a significant loss of life. Such a rare concentration of force presented an excellent opportunity for a major campaign against the Persian Empire, which could rectify the territorial losses Constantius had suffered, prove to the recently incorporated eastern armies that Julian was a militarily competent emperor who deserved their support, and demonstrate the righteousness of Julian’s polytheism.305 This last matter was extremely important to his reign, as his actions had been alienating the Christians in his government and army.306 He spent a year-and-a-half preparing, and during this time he consolidated his position vis-à-vis the military. He dismissed a significant portion of the protectores, who had been responsible for several attacks on him and his supporters.307 He also dismissed Agilo and Arbitio. Our two main sources for the period, however, disagree on whether new magistri were appointed in their place. Zosimus wrote in Greek, and therefore cannot provide the Latin magisterial titles, but his phraseology indicates that Victor commanded infantry and therefore may have been magister peditum, while Hormisdas and Arinthaeus commanded cavalry and so could have been magistri equitum.308 This has been generally accepted in most studies on the magistri.309 However, Ammianus does not record any new promotions, leaving only Nevitta and Jovinus as magistri. A careful critique of each source can help resolve this apparent discrepancy. As was discussed earlier, Zosimus reproduced the now only fragmentary history of Eunapius, who had used the writings of Oribasius, Julian’s private physician, and Magnus of Carrhae, who was an officer in Julian’s army, for his information on the Persian campaign.310 On the other hand, Ammianus himself participated in the campaign and used a combination of his own memories and notes along with interviews with other par-

303 304 305 306 307 308

AM 22.3.8. See Thompson 1947, 75. AM 22.3.8. Harries 2012, 312–13. Kulikowski 2019, 26–30. Lenski 2000, 504–5. Zos. 3.13.3. Those who follow this translation are: Demandt 1970, 581, 583; Ridley 1972, 57. Enßlin 1931, 119, believed Arinthaeus was a magister equitum per Illyricum. 309 Enßlin 1931a, 119; Demandt 1970, 581–82, 584. 310 Eun. Hist. fr. 15; Seeck 1906a, 531. Mendelsohn 1887, xxxix, first proposed the idea of Magnus of Carrhae as a source for Zosimus.

70

The Magistri Militum from 341 to 363 CE

ticipants to form his edition.311 While we have seen in the previous section that Ammianus’ aversions and biases render parts of his account untrustworthy, there seems to be no display of the same favoritism for particular magistri in this section. All three officers in question are included in what appears to be a fair and accurate manner.312 This has led some modern scholars to doubt this rendering of Zosimus’ passage.313 A comparison of the tasks and duties assigned to the different officers will help to further elucidate this issue. As the army set out to invade Persia in 363, Julian split his army into multiple parts, like he had when marching against Constantius in 361. He again personally commanded the primary force in the center, while Nevitta commanded the right flank with several legions closest to the banks of the Euphrates.314 The left flank of cavalry was assigned to Arinthaeus and Hormisdas, and a fourth part, the rear-guard, was commanded jointly by Victor and Dagalaifus. This reflects the marching order Julian established in 361, where a magister militum could command a section of the army alone, but lower officers were required to share command. Furthermore, Zosimus corroborates this configuration, and this would also explain his reference to Arinthaeus and Hormisdas commanding cavalry; rather than indicating they were magistri equitum, it suggests that they commanded the outriders covering the flank.315 Furthermore, during the various battles and sieges of the invasion, Nevitta received the more important tasks. For example, during the siege of Maiozamalcha, he oversaw the construction of a tunnel under the walls, which eventually allowed the Romans to capture the city.316 Conversely, Hormisdas was utilized as a negotiator and a guide because he was a Persian himself, while Victor and Arinthaeus mostly commanded scouting, raiding, and path-finding missions.317 Thus, we can see a distinct difference in the responsibilities of Julian’s senior officer which, in light of ambiguous titulature, can help confirm the hypothesis that Nevitta was the only magister militum on the campaign. After the Roman army began withdrawing from Persian territory, Julian’s predilection for leading from the frontlines, a key tactic in gaining his support base, eventually 311 312

313 314 315 316 317

AM 15.1.1. See also: Thompson 1947, 20, 32–33. This has led to a number of historians insisting that if there is a clash between the two of them, Ammianus is to be preferred (Chalmers 1960, 160; Ridley 1970, 101–2). He does not give Arinthaeus or Victor a single title during the campaign, and only calls Victor a dux (AM 24.4.13, 6.13) or comes (AM 24.4.31, 6.4) during the campaign, however, these titles are often used by Ammianus when referring to a confirmed magister militum, even for his hero Ursicinus (AM 15.5.26), so they cannot be used as evidence that Victor was not of magisterial rank. Ammianus never gives Hormisdas a Roman military title at any point. See: Woods 1997, n. 57, 290. Eg. Woods 1997, n. 44. AM 24.1.2. Zos. 3.14.1. AM 24.4.1–13, 4.22–25; Zos. 3.21.4. Zos. 3.22.1 says that Julian replaced those who were in charge of the mines because of their laziness, but this could be a reference to the sappers themselves and not Nevitta and Dagalaifus. Hormisdas: AM 24.1.8, 2.20, 24.2.4; Zos. 3.15.4–6, and also probably 3.29.2–4. Victor: AM 24.4.13, 4.31, 6.4–7; Zos. 3.16.3–17.1, 21.5, cf. 25.1–7. Arinthaeus: AM 24.7.2; Zos. 3.24.1.

Julian and Jovian as Sole Emperors, 361 to 364 CE

71

proved to be his undoing when he was mortally wounded in a skirmish, and died a few days later.318 As there was now no Augustus in the Empire to raise a new emperor, the selection and election of a new leader fell back onto the acclamation of the army. To choose a successor, the officers initially split into factions based on their past loyalties: Arinthaeus and Victor led the faction that had served in the east, while Nevitta and Dagalaifus headed those from the west.319 Arinthaeus and Victor were also both Christians, whereas Nevitta’s and Dagalaifus’ religious affiliations are unrecorded, but there is the chance that they were both polytheists, which suggests that religious preferences, either to maintain Julian’s status quo or to return it to Christianity, may also have played into the division.320 It is interesting to note that no individual, even the senior officer Nevitta, was able to assert themselves to make a unilateral decision, something we will see occur later in the fourth century. Each faction sought a candidate from their own group, but they ultimately compromised and agreed upon raising Saturninus Salutius Secundus, the Praetorian Prefect. This decision probably reflects the reality that none of the magistri militum could accept raising one of their own, as that individual would subsequently greatly overshadow the other generals because of their own military reputation, significantly reducing the influence and power of the remaining magistri. By raising a civilian Praetorian Prefect, the magistri would all retain their positions of influence. However, Salutius declined the offer on account of his old age. Before they could arrange another candidate, a group of soldiers, discontent with being marooned deep in the enemy’s territory without a leader, raised an officer of the protectores, a domesticorum ordinis primus by the name of Jovian, as emperor.321 These soldiers have been identified by Noel Lenski as a group of Christian imperial guardsmen who had been marginalized under Julian, and this helps explain why they raised Jovian, a man of small repute but a fellow Christian and guard.322 Although it seems he was not the choice of the magistri, the generals probably could have quashed the acclamation of Jovian if they had so chosen. But, he was of low enough rank in the military

318 AM 25.3.23, 5.1; Soc. 2.21.17; Cons. Const. s. a. 362; Eutr. 10.16.2; Zos. 3.29.1; Epit. 43.4. 319 AM 25.5.2. 320 On Arinthaeus’ religious affiliation: Bas. Ep. 269; Theod. HE 4.33.3. On Victor’s religious affiliation: Theod. HE 4.33.3; Vita Isaacii 2.7–8. On Nevitta’s and Dagalaifus’ potential religious affiliation, see: Haeling 1978, 239, 253. Von Haeling presumes that because Julian was a polytheist, he probably promoted similarly polytheist generals, although this assumption might not be valid considering Julian also promoted to high office known Christians like Victor and Arinthaeus, and that the polytheist Richomeres was one of the Christian emperor Theodosius I’s most privileged generals. All these officers were also barbarians: Waas 1971, 12. For further discussion on the identities of the magistri, see Chapter 4.3. 321 AM 25.5.4. Zos. 3.30.1 gives a different account of events, one that states Jovian was chosen by the succession council’s common consent, although Lenski 2000, 495–96 lays out the reason we should not believe this account over Ammianus’. 322 AM 25.5.4; Lenski 2000. The church histories tell that Jovian was a reluctant emperor until the army professed itself Christian (Theod. HE 4.1; Soc. 3.22; Soz. 6.1.1).

72

The Magistri Militum from 341 to 363 CE

that even as emperor he would find it difficult to challenge the power of the magistri, in a similar way to the civil bureaucrat Salutius. Thus, although the opportunity to directly choose a candidate was seemingly taken away from the magistri, the new emperor was still one who would be favorable to their interests, and this can be seen in the flurry of magisterial appointments that occurred in Jovian’s short reign. He quickly secured the loyalty of both succession factions by promoting Victor and Dagalaifus, one from each faction, as magistri equitum.323 Shortly after, he requested that his father-in-law, Lucillianus (the same man who had surrendered to Julian) come out of retirement and take up the position of magister equitum et peditum in Italy.324 Jovian was thus quickly securing the loyalty of both the campaign army and the rest of the Empire by promoting important officers to the magisterium. If Jovian was promoted with the tacit support of the magistri, it is surprising that Nevitta disappears from the sources as soon as Jovian became emperor.325 It is possible he died or retired, although neither Ammianus nor Zosimus specify this.326 A more likely explanation might be that Jovian had him removed, because he attempted to do the same to Julian’s other incumbent magister, Jovinus. He was still in Gaul, and Jovian feared that he could use his influence with the soldiers to usurp power, so he chose to replace him with the retired tribunus Malarichus. He is said to have been a man of few prospects, so Jovian thought that he could easily secure his loyalty by suddenly raising him up.327 Malarichus, however, bluntly refused the offer. This is the only known time this occurred in the fourth century. Malarichus thus did not ever become a magister militum, yet he was capable of being such. His career and life are therefore useful to understanding the individuals that could be magistri. In Milan, Lucillianus learned that Malarichus had turned his promotion down, so he forwent his appointment in Italy and travelled to Rheims to take command of the Gallic legions instead. Ammianus criticizes him because he did not secure the province nor win the loyalty of the soldiers, but instead chose to immediately put an administrative official on trial for corruption.328 This official fled to the army camp and told the soldiers that Julian was still alive and Jovian was only a usurper, which incited the garrison to riot and kill Lucillianus.329 Jovinus, who must have remained in Gaul even though he had been dismissed, managed to calm the soldiers and convince the garri-

323 324 325 326 327

AM 26.5.2. On Victor’s title, see: AM 27.5.1. On Dagalaifus’ title, see: AM 26.1.6. AM 25.8.9–10. Demandt 1970, 585. Cf. Waas 1971, 13. AM 25.8.11, 10.6. This incident seems to be mistakenly attributed to Valentinian I’s reign by Potter 2004, 534. 328 AM 25.10.7. 329 AM 25.10.7.

Julian and Jovian as Sole Emperors, 361 to 364 CE

73

son to support Jovian.330 In return, Jovian reaffirmed him as magister equitum in Gaul.331 Jovian’s final appointment was to make Lupicinus magister equitum in the east, and then he would die in early 364.332 Because of Nevitta’s disappearance and the attempt to remove Jovinus, it is clear that even though Jovian was raised without apparent resistance by the magistri, he still wished to win the loyalty of the armies by removing the most influential generals of the previous regime and replacing them with new generals with loyalty to himself. Thus, much like Julian as described previously, Jovian was still able to use and manipulate the magisterium for his own gain. The brief but well-documented reigns of Julian and Jovian as sole emperors are important for our understanding of the office of magister militum. Most significantly, we can contrast the actions taken during the planning of a major military campaign and those taken after a military disaster. When preparing for the invasion of Persia, Julian reduced the number of magistri by dismissing those who had served under Constantius and perhaps had questionable loyalties. It seems he did not replace them, and only took a single magister militum with him on campaign, leaving another to oversee Gaul. Julian had pacified the Rhine frontier, Constantius had pacified the Danube, and the Persians were soon to be preoccupied with his invasion, so he must have felt confident that the Empire was secure. Furthermore, as Julian was a militarily proven and capable emperor, he was able to keep the magistri in line and obedient, something later emperors would struggle with. Conversely, when Jovian inherited the Empire, he was in a vastly inferior position. The great invasion force was exhausted and hungry after marching through Persian farmland scoured of food by saboteurs, and they were being whittled down by the constant raiding of Persian cavalry.333 Moreover, the stability that Julian had established was not guaranteed to survive a change of regime and dynasty. In this context, we can understand Jovian’s rapid removal of the two incumbent magistri and the promotion of five new generals as an attempt to secure short-term stability during a crisis, a phenomenon that will be repeated over the remainder of the century. * In totality, the emperors we have seen in this chapter generally retained a dominant position over their magistri. The generals functioned as useful military ministers that furthered the plans and enforced the policies of the sovereigns rather than pursuing their own agendas. The forms this took varied, and included religious duties, the sup330 AM 25.10.8. 331 Arinthaeus carried this message to him (AM 25.10.9–10). For his title, see: CTh 7.1.7, 8.1.10. 332 AM 26.5.2. Jovian also gave a promotion to both Victor and Arinthaeus but Ammianus does not define the rank they were promoted to until they were serving under Valens (AM 27.5.1–9). On Jovian’s death: AM 25.10.13; Soc. 3.26.5, 4.1.1; Eutr. 10.18.2; Jer. Chron. s. a. 364; Epit. 4.44; Zos. 3.35.3; Theod. HE 4.5. 333 AM 25.1.1–19, 2.1, 3.1–14.

74

The Magistri Militum from 341 to 363 CE

pression of rebellions, and taking command during campaigns and battles. What inhibited the agency of these magistri was the adept use of containment strategies by the capable and experienced emperors. Constantius had competently faced multiple Persian invasions in a brutal series of campaigns during the 340s.334 He also had a reputation for proficiency in civil war – not an honorable thing to the Romans, but still a demonstration of capability. These actions afforded him a venerable martial respect that influenced his subjects into obedience. He was able to cleverly turn important men in Vetranio’s rebellion to his side and overcome the usurper without bloodshed, saving his manpower for what would be a brutal but eventually victorious clash with Magnentius. He stalled the ability of his magistri to develop regional alliances by regularly reassigning them across the Empire, and limiting the time in office for many of them. Arbitio enjoyed a long career, but Constantius retained his loyalty with him and others through a robust network of rewarding and patronage.335 When other methods of control failed him, Constantius did not hesitate to cashier Ursicinus, or order the execution of Silvanus and Barbatio. Emperor Julian utilized similar methods to ensure the magistri militum did not endanger his ultimate goals. As only Caesar, he did not have the legal authority to transfer, remove, or execute his magistri, and in fact the officials, including the generals, were instituted as his overseers. Despite this, he skillfully manipulated Constantius to sideline the western magistri by removing Marcellus. Although the historical record is opaque, it seems possible that Julian also masterminded the death of Severus while on campaign. By the time his resistance to Constantius became overt, Julian had gained the necessary respect and loyalty of the powerful groups and individuals in the west, and was able to send Lupicinus to Britain and isolate him there. A crucial part of gaining this control over the west was the exclusion and sidelining of the magistri. By doing so he ensured that the prestige of successive military campaigns was afforded to himself alone. This is not to say that either Constantius or Julian always fully dominated and controlled their magistri. Vetranio took advantage of the destabilization of Constans’ murder to raise his station and usurp imperial power. Fearing for his life but likely also harboring ambitious desires, Silvanus also tried to usurp Constantius’ position. Other magistri, like Arbitio, used their powerful positions to advise and influence the emperors, yet were unable to seriously impact major events. Julian’s authority was challenged regularly by Constantius’ appointees, but also during the Commission of Chalcedon. This power balance did not last long, however, and Julian asserted dominance over his military administration once again. Overall, the impact that the magistri militum had on the Roman Empire was not overly significant, and they could not operate with

334 Barnes 1980, 162–66. 335 AM 18.3.3, 18.5.5, 26.8.13.

Julian and Jovian as Sole Emperors, 361 to 364 CE

75

a great deal of personal agency. It seems, instead of the military magisterium, it was the leaders of the palatine units of soldiers that were successful during this period of the fourth century. Magnentius commanded the Ioviani and Herculiani palatine units, while Jovian was a leading officer of the protectores, and Valentinian, soon to be elected as Jovian’s successor, was an officer in the scholae. All these units were based in the emperor’s comitatus, and seemingly allowed their officers to form the necessary interpersonal connections to usurp or be elected to the imperial position. Although some magistri were also stationed at the court and would have been the supreme commanders over these units, and therefore would have likely had the same opportunities to establish a powerbase, efforts by the emperors to retain their loyalty, remove them from power, or simply execute them seem to have been effective at stopping magisterial disloyalty. It may be that the emperors were failing to adequately manage the ambitions and influence of the middling-rank palatine officers. It may also be that promotion of these men as emperors, as opposed to the highest-ranking men of the military and civil administrations, was more appealing to the powerful acceptance groups. No one desired an emperor with too much experience and dominating influence. During this period of muted influence in which some magistri explored the traditional methods of expressing their power, others were also beginning to explore new avenues. It was realized that imperial succession crises offered the opportunity to promote an emperor that would be favorable to their interests, although the magistri failed to fully exploit the chance after the death of Julian. This was a chance not lost on future magistri, and we will see in the following section that they became increasingly capable of furthering their own interests. Thus, while the early magistri were powerful, they were still inhibited and overshadowed by the emperors, and they would need to refine their approach to events if they were to reach the full potential of their position.

Chapter Two The Magistri Militum from 364 to 395 CE 2.1 Introduction The selection of Jovian’s successor in February 364 was a calmer affair than his own election. The soldiers were now safely back within Roman territory, so they did not feel the same level of urgency that caused them to elect an emperor on their own eight months prior – although they would not be content to wait too long, so the officials in command quickly began looking for a successor. No source states clearly who formed the search group when the army stopped in the city of Nicaea (İznik, Turkey). Ammianus only writes that the highest civil and military leaders were involved.1 Dagalaifus was in the vicinity, and thus probably closely involved in the search for a new emperor, as were several other military officers.2 Joachim Szidat has emphasized the role of civilian bureaucrats in this nomination process, but there is no true indication they had a role that outstripped that of the military officers.3 The potential imperial nominations are important for understanding the manner in which the imperial office was envisioned at this time. Equitius, tribunus of the prestigious schola prima Scutariorum, was considered first, but he was then rejected for being too rustic and unrefined.4 They next considered Januarius, a relative of Jovian and an officer in Illyricum, but he was rejected because he was far away and it was unlikely he would arrive before the army once again promoted someone of their choosing.5 Succession may have been offered once more to Salutius, but he again declined.6 Zosimus records the

1 2 3 4 5

6

AM 26.1.3. AM 26.1.6, 4.1–2; cf. Waas 1971, 13, who also believes Arinthaeus was involved, although AM 25.10.9 indicates he was in Gaul. Szidat 2010, 107. AM 26.1.4. There is some possibility that Januarius was a magister militum at this time, however this is unlikely because in this period Ammianus is detailed with the titulature but does not give Januarius a magisterial rank, nor does Januarius participate in any campaigns, battles, or other events to warrant assigning that rank to him. See: Demandt 1970, 586; PLRE I, 454. Zos. 3.36.1–2; Zon. 13.14. See: Lenski 2002a, 20 n. 43.

Introduction

77

candidacy was also offered to Salutius’ son, but Salutius declined on behalf of the boy because of his youth; unfortunately, we do not know his exact age.7 Seeing as Salutius made the decision for the boy, he must have been quite young, under the age one attained the toga virilis, the indicator of Roman manhood.8 Because Salutius declined his own offer on account of his advanced age, we can infer that the offer to the young boy was an attempt at compromise: Salutius himself would be able to watch over and guide his son at least for a while, while the boy’s youth would ensure he would have a long reign. Thus, the idea of a child-emperor existed in the Romans’ minds, but they understood the need for an older and wiser guardian. Interestingly, as far as the sources indicate, none of the officials considered Jovian’s infant son to be a viable successor at the time, perhaps because he had no father to oversee him. Nonetheless, it was clearly feared that his connection to an imperial dynasty might allow him to threaten the chosen emperor in the future, because he was cruelly blinded to eliminate any succession possibility.9 Finally, the powerful officials looked to Flavius Valentinianus (Valentinian), tribunus of the schola secunda Scutariorum, who had been left in Ancyra (Ankara, Turkey) to follow behind the main force.10 We can see thus see that there were several factors considered for a new emperor, including their cultural refinedness and logistical matters, like proximity to the succession committee. More importantly was that no potential imperial candidate was from the highest echelons of the military, much like during the succession that led to Jovian. The magistri declined to raise one of their own to power, presumably because they could not face the possibility of raising a man with a strong military support base that would weaken their own and mean that the new emperor would not have to rely on them for military support. Civilian bureaucrats, young children, or middle-ranking officers were all considered acceptable, as an emperor from these backgrounds would not overshadow the magistri, and would need to continue paying for their support to ensure the soldiers remained loyal and their reign secure. During the ten days it took for Valentinian to arrive to Nicaea, some of his supporters calmed the army and convinced them to not resist his election or look for their own candidate.11 Finally, Valentinian appeared and was officially accepted by the soldiers as emperor.12 Upon his accession, the soldiers’ only demand was for Valentinian to appoint a co-emperor, so if he died they would not be left without a sovereign.13 The new

7 8 9 10 11 12 13

Zos. 3.36.1–2. Although we do not know the son’s age, if his father declined on his behalf, he was presumably quite young. McEvoy 2013, 3–13. Lenski 2002a, 20. AM 26.1.5. AM 26.1.6. AM 26.2.2; Zos. 3.36.2–3; Epit. 45.3; Oros. 7.32.1; Jord. Rom. 307; Theod. HE 4.6; Soz. 6.6; Soc. 4.1; Eun. Hist. fr. 31; Cons. Const. s. a. 364. Zos. 4.1.2; Soz. 6.6.9. AM 26.4.3, Soc. 4.1, and Theod. HE 4.5 imply that Valentinian made the decision independently.

78

The Magistri Militum from 364 to 395 CE

emperor sought advice on who to appoint, but only the magister equitum Dagalaifus was bold enough to tell him that if he loves his family, he should promote his brother, Flavius Valens, but if he loves the Empire, he should find someone else.14 Valentinian must have felt he needed a trustworthy and reliable, albeit untested co-emperor, because despite the warning of his commander, he soon raised Valens. This smoothly settled the second imperial succession crisis of our period, and ushered in two new regimes that would see drastic changes in regards to the office of magister militum. Valentinian, as both the elder and the first appointed, was the senior of the two Augusti, and so he chose to assign the east to Valens and take the western Empire for himself. This seniority also affected the magistri: many of Valentinian’s generals were named in laws created when the two emperors were together, yet none of Valens’ generals were so named.15 When it came to leaving to their respective halves of the Empire, they divided the army between them, and so Jovinus and Dagalaifus returned to the west with Valentinian, while Valens took Lupicinus and Victor to the east.16 All four of these men were magistri equitum.17 This division indicates that experience was considered: Dagalaifus and Victor had been recently appointed to magisterial rank, but Lupicinus and Jovinus had been magistri since 359 and 360, respectively, so the east and the west both had a seasoned veteran and a new appointee. It also meant that all of Jovian’s magistri kept their positions, a stark difference to both Julian’s and Jovian’s treatment of their predecessors’ generals. Valentinian I’s election, and the appointment of his brother, thus represents a successful attempt by the Empire’s leading men to promote an emperor that would maintain their political positions and influence. 2.2 Valentinian’s Western Empire The reign of emperor Valentinian was a watershed for the magistri militum in the west. Valentinian significantly increased the prestige of the magistri by granting them an increasing number of consulships, and he raised their social rank to equal the highest civil officials.18 His reign also engendered an important change in how the imperial office was conceived, one that had lasting ramifications on the military-political landscape. Meaghan McEvoy argues that when Valentinian raised Gratian, his eight-year old son, 14 AM 26.4.1–2. 15 Demandt 1970, 709. 16 AM 26.5.1–3; Lee 2008, 222. 17 Jovinus: CTh 8.1.10; cf. CTh 7.1.7. Lupicinus: AM 26.5.2. Victor: AM 27.5.1. The PLRE suggests that Dagalaifus was perhaps in fact a magister peditum, as his successor, Severus, was given that title. However, AM 26.1.6 calls him magister equitum. 18 In 366 Dagalaifus was the first to receive a consulship that ushered in his retirement: CTh 11.1.13, 4.12.6; CIL V 8606, X 4487, XI 4328, XIV 1945; P. Flor. 1 84.1; AM 26.9.1; Bagnall, et al. 1987, 266–67; contra Demandt 1970, 591.

Valentinian’s Western Empire

79

to the position of co-emperor in 367, it created a new imperial precedent.19 While there had been young emperors in the past, most had already attained the toga virilis, and they were often only made a Caesar.20 The fact that Gratian, who had not yet received the toga virilis, was made a co-Augustus, legitimized the idea of children as the head of the state. As discussed above, the powerful fourth-century officials had already been considering the possibility of raising Salutius’ young son as emperor with his father to oversee him, and Valentinian’s choice made this idea a reality. Following Valentinian’s death in 375, it was Gratian’s imperial title that set the precedent for the magister militum Merobaudes to promote a child-emperor of his own. The phenomenon of child-emperors incapable of commanding the armies shifted the political landscape by allowing the magistri to take over the ‘active’ duties of an emperor – campaigning, negotiating with foreign entities, and winning the loyalty of the soldiers. All these factors combined during Valentinian’s reign to expand the power of the magistri to unprecedented levels. This section will track these developments as they occurred, as well as cast light on some of the questions that have persisted about Valentinian’s magistri. After splitting the army with Valens at Naissus in 364, Valentinian travelled first to Italy and then into Gaul.21 Jovinus and Dagalaifus were initially his only two magistri. There had not been an imperial presence in Gaul since 361, and the Alamanni had taken the opportunity to expand the scale of their raiding.22 Dagalaifus commanded two counter-attacks against the Alamanni, but his repeated failures forced Valentinian to withdraw him and send Jovinus in his stead.23 Perhaps to not alienate his veteran commander, Valentinian granted Dagalaifus the consulship for 366 along with Valentinian’s son, Gratian.24 He does not appear again after this year, so it seems that he was the first of many magistri to have received a consulship with his retirement.25 In the following section on Valens, more will be explained about the usurpation of Procopius in Constantinople, but it is relevant here to explain the significance of the rebellion in regards to the western magistri. When Valentinian heard that Procopius had declared himself to be an emperor and was waging a war against Valens, he decided that the concerns of the west were more important, and he could not personally go and help his brother. Instead, he promoted the commanding officer of the western armies

19 20 21 22 23 24 25

McEvoy 2013, 48–70. McEvoy 2013, 3–13. His journey can be tracked through the laws issued during this time: Seeck 1919, 216, 218; Lenski 2002a, 27. AM 26.4.5. AM 26.5.9, 27.1.1–6, 2.1–10.; Zos. 4.3.4–5, 9.1, 9.2 ff. CTh 11.1.13, 4.12.6; CIL V 8606, X 4487, XI 4328, XIV 1945; P. Flor. 1 84.1; AM 26.9.1; Bagnall, et al. 1987, 266–67. CTh 7.20.9. Contra Demandt 1970, 591.

80

The Magistri Militum from 364 to 395 CE

in Illyricum, the comes per Illyricum Equitius, to the magisterium.26 While we do not know of any specific increases in official prerogatives or powers which came with such a promotion, Valentinian must have had some motivating reason for doing so. Hence, the most likely explanations seem to be that it actually did increase the resources Equitius could draw upon, perhaps allowing him to furnish a larger number of soldiers. Another explanation might be that Valentinian hoped the greater remuneration and honor of the higher position would ensure Equitius’ continued loyalty, and avoid a reoccurrence of the rebellion of Vetranio. Whatever the reason, Equitius was able to arrest the agents Procopius had sent to the west, and then he blockaded the main roads through the province, effectively bottling Procopius up in the east.27 The titles that Ammianus assigns to an officer named Severus during the following period present some difficulties and require close scrutiny to untangle the apparent contradictions.28 Whilst in Milan in 365, Severus was Valentinian’s comes domesticorum.29 He was also present when Valentinian fell extremely ill in Ambianum (Amiens, France) in 367. In the event that the emperor’s sickness proved fatal, the bureaucrats looked for a potential successor. The Gallic faction at court looked to Sextius Rusticus Julianus, the magister memoriae, while others looked to Severus, who Ammianus writes was “at that time magister peditum.”30 It is important to note that here in 367, a magister militum was a desirable emperor, while in 363 and 364 only lower-ranked officers were considered. This could be explained by the predominance of the Gallic faction in the west which may have desired strong military leadership, whereas the previous successions had occurred in the east with different priorities. Whatever the cause, it is clear that successions could play out very differently depending on the circumstances in which they occurred – not all factions and parts of the Empire had uniform desires and intentions. About Severus’ rank, as will be detailed in Chapter 4.2, the comes domesticorum was one step below magister militum, so Severus’ promotion upwards was typical. Valentinian eventually recovered, and then in order to both placate his officials and safeguard his dynasty in the event of his death, he promoted his eight-year old son, Gratian, as co-Augustus.31 An imperial Caesar still had to go through an acclamation process to become a senior emperor, but as an Augustus, Gratian did not have to go through any

AM 26.5.8–9. Equitius as a comes per Illyricum: AM 26.5.3. As magister militum: CTh 7.1.8; AM 26.5.11, 7.11. 27 AM 26.7.11–12. 28 See note 61n248. This is Severus 10. 29 CTh 6.24.2, 6.24.3. The MSS record 364, Seeck 1919, 71, argues they cannot have been issued until 365. 30 “magistri tunc peditum” AM 27.6.3. We are not told explicitly who supported Severus, but if it was not the Gallic military faction then it was possibly a senatorial group. 31 AM 27.6.10–11; Symm. Or. 1.3, 2.31–2, 3.1–6; Soz. 6.10.1; Soc. 4.11.3; Epit. 45.4; Zos. 4.12.2; Jer. Chron. s. a. 367; Prosp. Tiro s. a. 384; Philost. HE 8.8; Theoph. 5857; Zon. 13.15; Cons. Const. s. a. 364; McEvoy 2013, 48–54. 26

Valentinian’s Western Empire

81

such process to become the senior emperor in the event of his father’s death.32 Thus, to Valentinian, the position of co-Augustus represented a better way to secure his dynasty. According to Ammianus, Valentinian then received a report that Britain and the Channel coast had come under a massive, coordinated attack by the Picts, Attacotti, Scots, Franks, and Saxons.33 Valentinian’s first response was to send Severus to stem the attacks, and Ammianus states that he was “still comes domesticorum.”34 This would seem to describe an unprecedented demotion. This was never reported for any other magister, and it is made more implausible by the fact that Ammianus again records Severus as a magister militum by 368.35 This apparent contradiction has been resolved by rearranging Ammianus’ chronology to have Severus sent to Britain as comes domesticorum, then return and promoted to the magisterium, all before Valentinian fell ill.36 Walter Heering points out that, because Valentinian’s primary focus during this period was the Alamanni across the Rhine, the only reason he would travel away from them and towards Ambianum is because it was along the most direct route to Bononia (Boulogne-sur-Mer, France), the largest port through which one would cross to Britain.37 Therefore, the raiding and Severus’ mission probably began before he arrived there and fell ill. While this theory resolves Ammianus’ confused chronology, a question that has so far not received an adequate answer is why Valentinian first sent Severus to deal with the crisis, and then soon replaced him with Jovinus, who was in turn quickly replaced by the officer Flavius Theodosius.38 Explanations by Alexander Demandt and Roger Tomlin remain unsatisfactory.39 A possible answer lies in remembering that Jovinus had not spent the winter in Rheims with Valentinian, but had returned to Paris with the large campaign army with which he had fought the Alamanni the previous year.40 32 33 34 35 36 37

Szidat 2010, 170. AM 27.8.1, 8.5. “tum domesticorum comitem” AM 27.8.2. AM 27.10.6. Seeck 1906b, 274–75; cf. 1906a 519–20; Heering 1927, 70; Enßlin 1931a, 124. Heering 1927, 51. Tomlin 1974, 306 added that Ammianus’ mistake may have resulted from the second-hand information he was forced to rely upon from his residence in Antioch, which has been followed by Blockley 1980; Frere 1987, 340; Drinkwater 2007, 269; Alvarez-Jimenez 2013. 38 AM 27.8.2–3. It is my choice to use the cognomen of ‘Flavius’ in the body of this text to distinguish the father, Flavius Theodosius 3 in the PLRE, from his son, the homonymous emperor Theodosius I, Flavius Theodosius 4 in the PLRE. ‘Flavius’, used by those who served in the imperial bureaucracy under the Neo-Flavian, Constantinian dynasty, had become so commonplace that it was rarely used by contemporaries to refer to each other in anything but the most formal contexts (Bagnall, et al. 1987, 36–40). As such, it has generally not been included when referring to individuals in this text, but it will be useful to distinguish the two Theodosii. In the tables and figures of Chapters Three and Four, however, the PLRE entry numbers will be used, just like for Severus 8 and 10. 39 Demandt 1970, 562; 1972, 95–96; Tomlin 1974, 305 n. 19. 40 AM 27.2.10; contra Tomlin 1974, n. 10, who has been followed by others, including Drinkwater 2007, 269. We must remember that it had become common for the emperor to hold court in one city while a magister militum commanded a second army somewhere else. Such was the practice

82

The Magistri Militum from 364 to 395 CE

Paris and Rheims are both separated from Ambianum by around 100 miles of road. After Valentinian received word of attacks on Britain and northern Gaul, he must have sent word to Paris for Jovinus to bring his army and join him in Ambianum. After mobilizing the soldiers and gathering the necessary supplies, a Roman army could cross such a distance in a few days, but the quicker the response, the more Roman citizens in Britain could be safeguarded. Thus, Valentinian seems to have sent comes domesticorum Severus, the highest-ranking officer in Rheims, ahead with a small rapid force to avert the trouble besetting Britain “if chance should allow the desired opportunity.”41 This implies that Valentinian only sent Severus to do what was possible quickly, and once both he and Jovinus had “shortly after” arrived in Ambianum, Severus was recalled and Jovinus sent out with a more substantial force.42 Upon his return, having fulfilled his mission, Severus received a promotion to the magisterium. This chronology is more likely than Demandt’s explanation that there were two Severi, one the comes domesticorum and one the magister militum, so close together in the upper ranks of the military and Ammianus did not think to distinguish between them.43 The description of Jovinus’ mission in Britain is unfortunately hampered by a lacuna in Ammianus’ text, but it appears that he quickly reported that the situation was even worse than Valentinian thought, and more soldiers were needed.44 This would explain why Valentinian then sent four senior units of the field army under the command of Flavius Theodosius, who was probably a comes rei militaris or similar, to take over from Jovinus.45 It seems somewhat contradictory that Valentinian would recall the magister militum and replace him with only a comes when the situation was increasingly dire, but Valentinian was probably looking ahead to a campaign against the Alamanni, the enemies of “the entire Roman world,” and he must have wanted his most senior commanders with him.46 By then, it was too late in the campaign season to begin a new push against the Alamanni, so Valentinian moved to Trier, close to the Rhine, to prepare for the following year. It was at this time that Valentinian increased the official social status of the magistri militum from vir clarissimus to vir illustris, giving them equal status to the Praetorian

under Julian from 355 to 360, while Constantius and later Gratian would go even further, residing in Italy and leaving the Gallic armies to magistri. Valentinian seems to have done the same when he brought his army from Milan to Paris with Dagalaifus while Jovinus commanded the army at Rheims. Therefore, we can be confident that Valentinian went to Rheims and Jovinus to Paris. 41 “si fors casum dedisset optatum” AM 27.8.2. 42 “paulo postea” AM 27.8.2. The idea that Severus was sent to the area to reconnoiter the situation and report back to Valentinian has persisted in the English literature, i. e. Tomlin 1974, 305, when in reality AM 27.8.2 shows that he was sent there to redress the situation and was then recalled back. 43 Demandt 1970, 562; 1972, 95–96. 44 AM 27.8.2. 45 Tomlin 1974, 303 n. 6, 306 n. 26. 46 “totius orbis” AM 26.5.13.

Valentinian’s Western Empire

83

and Urban Prefects.47 This was done as part of a wider reformation of the governmental hierarchy.48 Valentinian also specified that, upon retirement, precedence was to be given to those who had served the longest. The additional privileges included being able to consult with judges first, and they were given more distinguished and prominent seats in the Senate. This promotion of military officers has been interpreted as a reward to the powerbase that raised Valentinian to the purple, as well a result of general favoritism stemming from the fact that Valentinian was a military officer himself. John Noël Dillon has also recently postulated that the rewarding of increased ranks to senators was done to reward valuable servants and extend the important ties of patronage to individuals who might threaten the emperor’s reign.49 While the increased standing did not carry with it a tangible increase in real powers, it is a significant acknowledgment of the increasing importance of the magistri under Valentinian’s government. We will see in Chapter 3.3 that Valentinian’s general lack of interest in internal squabbling between his ministers allowed them to wield great power and conduct brutal attacks against each other, and political cliques formed with sharp divides between them. The magistri were important members of these cliques, and were able to use networks of allies to pursue their personal agendas and secure their powerful positions. Overall, the decision to raise a lower-ranking officer as an emperor that would favor their interests was clearly paying off for the magistri. In 374 the Quadi and Sarmatians invaded across the upper Danube river.50 After they found some initial successes, the Sarmatians were driven back, and then Valentinian arrived in early 375 to finish off the Quadi.51 Equitius had had multiple recent failures, including procrastination on a building project and a shameful retreat from the Quadi, and these did not endear him to Valentinian, who removed him from active command.52 Equitius seems to have remained a magister militum, but he might have been relegated to a vacans position, although there is no evidence that points to this explicitly.53 Valentinian instead promoted Merobaudes, an experienced soldier who

47 48 49 50 51 52 53

CTh 6.7.1–2. CTh 6.22.4 also specifies that the magistri outranked those who had achieved a proconsulate. The law, however, only specifies magistri equitum, and the wording of the law has caused some debate. See: Demandt 1970, 595. Some of the other laws include CTh 6.9.1, 14.1, 22.4. Also see: Jones 1964, 142–43, 528; Näf 1995, 20; Heather 1998, 188–91; Dillon 2015. Dillon 2015, 60. This was prompted by an aggressive fortification-building project: AM 29.6.1 ff. AM 29.6.12–16, 30.5.1; Drinkwater 2007, 284–85. AM 29.6.3, 6.12. Magistri militum vacans were generals without a fixed area of responsibility but they still could command soldiers (Demandt 1970, 647). This could mean they were in charge of an expeditionary campaign army. Vacantes officers were allowed the vestments and prestige of an office, although it was not as prestigious as the regular office (CJ 12.8.2, 440–441; Boak 1915, 141). It seems possible that Equitius was taken off regular command duties and made a magister militum vacans as it was not appropriate to cashier him while his name was still being used for postconsular dating.

84

The Magistri Militum from 364 to 395 CE

had served as an officer since at least Julian’s Persian campaign.54 Along with the comes rei militaris Sebastianus, Merobaudes took an expeditionary force into Quadi territory to destroy their villages while Valentinian unsuccessfully sought to bring their main forces to a decisive battle.55 After the armies had returned to their winter garrisons, the Quadi sent envoys to Valentinian’s residence in the city of Brigetio, Pannonia (Szőny, Hungary) to negotiate a truce.56 At this fateful meeting, the Quadi envoys tried to justify their invasion and make demands of Valentinian, and this angered him so greatly that he burst a blood vessel and fatally hemorrhaged on 17 November 375.57 We have labelled the periods after the deaths of Julian and Jovian as succession crises, because the entire Empire was without a sovereign. In these situations, there was no way for a reigning Augustus to approve the accession of a new imperial ruler. But in 375, the Empire had two reigning rulers, and Gratian was even in the west as the designated heir. Thus, there was no imperative need to organize a succession and raise a new emperor on the acclamation of the army – the incumbents could simply take over. Therefore, the events that followed have been identified by Gavin Kelly as a political crisis, which Roman officials took advantage of to further their own desires through the distinctly new solution of a figurehead emperor.58 In Brigetio, Equitius and Petronius Probus, the Praetorian Prefect of Italy, Africa, and Illyricum, took immediate action.59 Whether they thought of this as an opportunity to raise either of themselves as emperor is unclear, but it mattered little while Merobaudes commanded the army at nearby Aquincum (Budapest, Hungary).60 Not only was acclamation in front of the soldiers a necessary step in the election process, but any emperor not supported by the army would be easily overthrown.61 Since Equitius’ dismissal from active service, Merobaudes was the only magister who had been campaigning across the Danube with the soldiers. Whereas all the other officials had gone with Valentinian to spend the winter at court in Brigetio, Merobaudes had remained with the soldiers in their military camp, undoubtedly increasing the esprit de corps between soldiers and commander. Thus, despite Equitius having more prestige and social standing than

54 Philost. HE 8.1. Zos. 4.17.1 says that Merobaudes was in charge of the entire army, but this could only mean he was the supreme general of Valentinian’s main comitatus. At this point, he is most likely a magister peditum, but the evidence is not specific. AM 30.5.13 says he is in command of “militari peditum” and he replaces Equitius, who was an undefined magister militum. 55 AM 30.5.13. Sebastianus had fought with Valentinian previously (AM 27.10.6–11). 56 AM 30.6. 57 AM 30.6.3–6; Soc. 4.31; Zos 4.17.1–2; McEvoy 2013, 54. 58 Kelly 2013. 59 Equitius: Zos. 4.19.1–2; Epit 46.10. Rolfe 1939, 349, identifies this Equitius with another man who was later a cura palatii under Valens, but a transferal like this seems unlikely. Probus: Ruf. HE 11.12; AM 30.5.10. Philost. HE 9.16 also adds in Justina, Valentinian’s wife. They must have acted with extreme rapidity: Potter 2004, 543. 60 AM 30.5.14. 61 Szidat 2010, 44.

Valentinian’s Western Empire

85

Merobaudes because of his consulship shared with Gratian in 374, and Probus having greater still because of his earlier consulship with Gratian in 371, Merobaudes had the practical, real power of the armies. Although Merobaudes probably could not have acted completely unilaterally because the support of wide acceptance groups was critical in the Roman Empire, he was the individual with the greatest leverage. Working with him was therefore necessary for the Equitius and Probus. They forged a letter under Valentinian’s name summoning the general to Brigetio, but the magister uncovered the truth of what had transpired.62 Whether the messenger told him the truth, or he was sufficiently suspicious of this strange summons, he took immediate action to safeguard his position of power over the coming electoral process. His first action was to send Sebastianus away to a distant post.63 Ammianus writes that Sebastianus was not an ambitious man, implying he would not seek to be emperor, but he was experienced and popular, which meant he was a likely candidate for the army to raise on their own.64 This action also makes it clear that Merobaudes was not seeking to raise another middle-level officer that would continue being favorable to the military, but instead the magister sought a different solution. From his previous experience, Merobaudes would have also been aware that Julian’s favorite magister, Nevitta, had been sidelined by Jovian. Thus, in my view, sending Sebastianus away indicates that Merobaudes was not willing to hand the position to another military man who could potentially exclude him from power, and was therefore an ambitious general intent on securing his own position in the next regime. Merobaudes had been present at Jovian’s election, and he may have also been at Valentinian’s.65 He thus had a good understanding that time was of the essence if he were to control the promotion of the next emperor. Although the army was not in hostile enemy territory like it had been in 363, the situation was still precarious, and several pressures weighed on Merobaudes and the other officials still in Brigetio. The Quadi had sued for peace, but the fact that they had still tried to make demands of Valentinian indicates that they had not been annihilated and still posed a threat. Zosimus records that it was feared they would seize the opportunity to attack, and Ammianus corroborates this by writing that the bridge across the Danube was ordered to be destroyed by the officials in Brigetio in order to delay any Quadi incursions.66 They and Merobaudes were probably less afraid of an actual invasion than of the army raising a candidate of its own due to their fear of facing a Quadi attack without imperial leadership. Ammianus writes that the Gallic soldiers, meaning those who had accompanied Valentinian to the

AM 30.6.2–4, 10.2–3; Zos. 4.19.1; Epit. 46.10. Probably Italy: AM 31.11.1; Zos. 4.22.4; Errington 1996a, 441–43. AM 30.10.3. Eun. Hist. fr. 44.3 also praises Sebastianus’ virtues and describes the reason he was so popular with the soldiers. 65 Philost. HE 8.1. 66 Zos. 4.19.1; AM 31.10.2. 62 63 64

86

The Magistri Militum from 364 to 395 CE

Danube and not those still in Gaul, were particularly unrestrainable and likely to raise an emperor.67 The officials must have feared that if the soldiers were to raise their own candidate, he would be likely to sideline or execute the present elites, or at least fail to protect them from the other powerful men in the west, and worst of all, risk civil war with Gratian and Valens. Kelly has postulated that fear of being exposed to attacks from other powerful blocs of officials in the west motivated the officers along the Danube to safeguard themselves through raising a favorable emperor.68 There is not much to oppose this theory; Kelly, however, unfortunately follows a recent trend in scholarship to downplay the role of the military, and especially military officials, in political affairs.69 When describing the strengths of the officials on the Danube against the other power blocs, Kelly lists a plethora of factors, including the support of the majority of the western armies, but does not mention the presence of the highest-ranking general in the entire west – Merobaudes.70 In fact, Kelly calls it “problematic” that Ammianus’ account of the affair has Merobaudes as the dominant figure.71 Why this would be considered inherently problematic is elusive, unless one is trying to downplay the role of military leadership in favor of social and political factors. Kelly compares Ammianus’ apparently more detailed and inclusive description of the election of Jovian, where four military commanders, including two magistri, are listed.72 This is hardly a significant comparison, however, because between Merobaudes, Equitius, and Sebastianus, Ammianus names three military commanders, including the same amount of magistri, following the death of Valentinian. One magister militum dominates Ammianus’ account of 375 because, after the removal of Equitius from active duties, Merobaudes was the only man of magisterial rank operating with and commanding the soldiers, and he was the only official that could call upon the personal loyalty of them. He is the main actor in Ammianus’ account because he had the most power and influence, and therefore the greatest agency in the matter. This should not detract from Kelly’s overall argument that, as well as seeking to avoid civil war and being sidelined by a potential candidate of the soldiers, the officials in Brigetio were also seeking to avoid persecution from powerful men in other parts of the Empire. Kelly also rightfully identifies that their ultimate solution was essentially a coup against the reigning emperors. Regarding the two incumbent emperors, they were of no assistance to the situation: Valens was preoccupied and far too distant on the eastern frontier to take command, and Gratian also could not arrive quickly while winter made communication and trav-

67 68 69 70 71 72

AM 30.10.1–3, cf. 19.5.2, 6.3–4, 22.12.6. Kelly 2013, 401. Cf. Szidat 2010, 140; Kulikowski 2019, 69–70. Kelly 2013, 379. Kelly 2013, 364. AM 25.5.2.

Valentinian’s Western Empire

87

el prohibitively slow. Nor did he have the personal respect to command the loyalty of the troops. Thus, a suitable successor had to be found from nearby, and quickly. After sending Sebastianus away, Merobaudes went to Brigetio and joined Equitius and Probus. Despite being the most powerful man present, Merobaudes did not seek to elect himself. Onomastics indicate that Merobaudes may have had Frankish heritage that precluded him from becoming emperor, or it may be that because he did not belong to the prevailing dynasty his election would not have prevented civil war.73 Merobaudes must have also known of the recent failures of Vetranio and Silvanus, and decided that it was a safer path to remain a magister militum and instead promote someone who would follow his orders and see to his interests.74 Thus, Merobaudes and the other officials chose Valentinianus, the late-emperor’s four-year old son and Gratian’s half-brother.75 He was a good choice for many reasons. As a member of the royal family, he did not threaten the existing dynasty, and his youth ensured he would surpass neither Gratian’s nor Valens’ seniority. Thus, he was unlikely to provoke civil war. He had been living only one-hundred miles away with his mother, Justina, so it was possible for his uncle, Cerealis, a tribunus stabuli and brother-in-law to Valentinian I, to be sent from Brigetio and bring him back with rapidity.76 Additionally, those raising him knew he could be used as leverage to improve their position vis-à-vis Gratian’s court and the powerful officials in it.77 Therefore, the boy was raised with all the usual ceremony as emperor Valentinian II on 22 November 375, only five days after the death of his father.78 Thus, whereas in 364 the succession had gone to the experienced Valentinian, in 375 the powerful officials chose the option of a young and untried boy who could be overseen and protected. Recognition from Gratian and Valens was not immediate. It seems that they communicated with one another before officially accepting Valentinian II as a co-Augustus, and their final assent took months.79 Fortunately, they did eventually accept what Merobaudes and the officials had done, and civil war was avoided.80 Kelly theorizes that their assent came only with the punishment of removal of all the high officials who had caused Valentinian II to be raised, and this is largely true.81 This will be discussed in full in Chapter Three, but it bears noting that only Merobaudes’ career survived this polit73 74 75 76 77 78 79 80 81

Merobaudes’ ethnic and cultural background is not explicitly recorded in the evidence, but his name implies a Germanic, probably Frankish, origin (Schönfeld 1911, 167). For a discussion on the prohibition of barbarians becoming emperors, see pages 184–88. Szidat 2010, 370. McEvoy 2013, 56. AM 30.10.5. McEvoy 2013, 59; McLynn 1994, 84. AM 30.10.4–5; Cons. Const. s. a. 375; Soc. 4.31.7; Soz. 6.36.5; Amb. De ob. Val. Iun. 59; Epit. 45.10; Zos. 4.19.1; Philost. HE 9.16; Ruf. HE 11.12; cf. Aus. Grat. Act. 2.7, 10.48; Hebblewhite 2017, 149. Girardet 2004, 124–27; Kelly 2013, 366–72, 402. AM 30.10.6. Kelly 2013, 402.

88

The Magistri Militum from 364 to 395 CE

ical crisis and the coup that was its solution – in fact, Merobaudes did not just survive the events of 375–6, but thrived, receiving a consulship in 377.82 This is further testament to the fact that his influence and power outstripped all others in the western Empire. In summation, the reign of Valentinian I represents a major turning point for the role of the magistri in the Roman world. As a military man himself, Valentinian was an ardent promoter of the army and its magistri, which helped to ensure strength in the western provinces. Valentinian favored the generals by raising their social status, giving them greater priority in court protocols than they had previously held, and placing them on equal standing with the civil administration officials who had traditionally outranked them. He also granted them a higher number of consulships. This favoritism distributed sufficient patronage to retain the military’s loyalty, and Valentinian’s personal vigor and dynamism on campaign, despite his illnesses, managed to ensure the powerful men in his government generally worked in his interest, but his unexpected death left the western Empire in dire straits. Even though he had raised Gratian as an Augustus, he had kept the young emperor on the sidelines, and he had not yet been tested in war. Thus, the power vacuum that followed his demise offered Merobaudes and the other court officials the opportunity to use the imperial office to secure and improve their own positions. This had also been a priority during the succession crises of 363 and 364, but in 375, the precedent set by the raising of the eight-year-old Gratian as a full Augustus allowed Merobaudes and his allies to raise the four-year old Valentinian II as another full Augustus of the western Empire. When they did so, they took the first step towards a century of puppet emperors in the west. The raising of Valentinian II by Merobaudes and his alliance shows that they were sharply re-envisioning what it meant to be emperor, as well as what it meant to be a powerful official or general. Merobaudes realized that he no longer had to only serve and influence the emperors and receive what benefits he was given, but he now could control the office of emperor to further his own interests. However, it seems Merobaudes did not see Valentinian II as a permanent source of power. Whether he initially thought his place would remain with the child-emperor in unclear, but he would soon be found in Gratian’s court as the west’s sole magister militum while Valentinian was kept on the margins. 2.3 Valens’ Eastern Empire Emperor Valens lacked his brother’s proficiency in military matters, and he was indecisive and inadequate both as a defender of imperial territory and as a campaigner against Rome’s enemies. One modern biographer drolly describes him as “imperially

82

See page 95.

Valens’ Eastern Empire

89

challenged,” a description which is particularly apt when it comes to the army.83 Valens’ problems in these matters stemmed from his lack of military experience. Noel Lenski, the aforementioned biographer, believes he had been a farmer before becoming a member of Jovian’s army as a protector domesticus from roughly 359, but there is no evidence he did anything of note during this time.84 However, an alternative career path proposed by David Woods would have Valens take some of the positions that have been attributed to Valentinian, including the rank of tribunus under Constantius II.85 Eventually, Valentinian promoted Valens to the position of tribunus stabuli shortly before raising him as emperor, but this was only a logistical office in charge of supplying horses.86 Consequently, he was unfamiliar with taking command during campaigns and battles. To rectify his deficiency, Valens relied heavily on his magistri, and they were kept nearby at all times. Thus, wherever the emperor was present would be adequately protected, but other frontiers were at risk of mismanagement. This was a cause for significant instability in Valens’ half of the Empire. Because Valens’ first priority was securing the Persian frontier after the embarrassing peace treaty Jovian had conceded, he left Valentinian at Naissus and passed through Constantinople to go to the east, sending much of his army ahead under the command of Victor and Lupicinus.87 He was, however, forced to turn around when Procopius, who had been an officer in Julian’s army, claimed he had been designated as heir and was therefore the rightful emperor, not Valens.88 Procopius’ usurpation quickly gained control of Constantinople, and he recruited Gomoarius and Agilo to lead his army.89 These two are not specifically given a rank by the sources, but based on their previous positions under Constantius, they were probably both also magistri under Procopius. The usurper also tried to recruit Arbitio, but he declined on account of his age and ill-health.90 In retaliation, and perhaps to help fund his rebellion, Procopius seized Arbitio’s house and property, a move which angered the former magister and drove him to join Valens instead.91 Victor and Lupicinus returned with reinforcements from the Persian frontier, and the two armies maneuvered towards their first battle in southern Anatolia.92 On the eve 83 84 85 86 87 88 89 90 91 92

Lenski 2002a, 1. Soc. 3.13.3–4, 4.1.8; Joh. Ant. fr. 179 (Müller) = fr. 204 (Mariev); Lenski 2002a, 51–54. Woods 1998; cf. Lenski 2002b. AM 26.4.2; Jones 1964, 372–73. For how this position fit into the military hierarchy, see Chapter 4.2. AM 26.6.11; Zos. 4.4.1, 13.1–2; Soc. 4.2.4; Philost. HE 9.5; CTh 7.22.7; Lib. Ep. 1499, 1505. Soc. 4.2.6– 3.1, 5.2 and Soz. 6.7.10–8.2 even state that Valens himself was in Antioch when he heard of Procopius’ revolt, but he was most likely still in Cappadocia as AM 26.7.2 reports. AM 23.3.2–3, 26.6.2–3, 6.12; Cons. Const. s. a. 365; Them. Or. 7.91a; Zos. 4.4.2–3, 5.5, 6.3; Lib. Or. 19.15; Soc. 4.3.1; Lenski 2002a, 68–73. AM 26.7.4. AM 26.8.13. AM 26.9.4. AM 26.8.4, 9.1

90

The Magistri Militum from 364 to 395 CE

of battle, Arbitio was able to utilize his venerable reputation to denounce Procopius and call on Gomoarius to disavow the usurper.93 This worked, and first Gomoarius, then later Agilo, came over to Valens.94 With his armies having deserted him, Procopius was easily captured and executed. Valens made Lupicinus consul in 367 and with this award he retired.95 Gomoarius, Agilo, and Arbitio are not heard of again, so they presumably returned to retirement. During this time, the Goths, a Germanic people who lived north of the lower Danube River, had been raiding the Roman province of Thrace, and had sent three-thousand troops to support Procopius.96 This was both insulting and threatening to Valens, so he decided to forgo the Persian frontier in favor of supporting Thrace. He first sent an embassy to the Goths to demand an explanation for why they supported the usurper.97 They answered that they had been misled by Procopius into thinking he was the real emperor, but it probably did not matter what they said. Valens needed to prove that he was a legitimate and capable ruler, and the best method for this was to gain victories over Rome’s foreign enemies. He had set his mind on war with the Goths. It is a good comparison to note that when Constantius conducted a similar war against the Sarmatians in the 350s, he retained the magister Arbitio as commander on the Danube, while the Persian frontier was overseen by another magister, first Ursicinus and then Sabinianus.98 Valens, on the other hand, kept both magistri with him on the Danube, and he left the Persian front in command of lower ranked officers. Valens’ lack of military experience may have encouraged him to arrange his generals in this way, but it may have also been prompted by the magister militum Victor and his apparent inability to take command on the battlefield. Victor first appears as a competent soldier under Constantius, and later took part in multiple actions under Julian on the Persian campaign.99 These included commanding a daring night-time river crossing while under heavy fire, during which it seems he was wounded in the shoulder by an arrow.100 Zosimus has a slightly different version of events which might indicate the damage was even more severe.101 It would seem that this injury had a lasting physical or mental impact on the general, because as far as we can tell, Victor never again directly commanded soldiers on the battlefield.102 93 AM 26.9.5–7; cf. Zos. 4.7.3–4, 8.3; Them. Or. 7.87b. 94 AM 26.9.7; Zos. 4.8.3. 95 CTh 10.19.4, 12.18.1; CIG 4 9842 = ILCV 2878; Soc. 4.11; Bagnall, et al. 1987, 268–9. 96 AM 26.4.5, 6.11, 10.3, 31.3.4; Eun. Hist. fr. 37; Zos. 4.7.1–2, 10.1–2. 97 AM 27.5.1. 98 See page 57. 99 AM 25.5.2. 100 AM 24.6.4–6, .13; cf. Zos. 3.16.3, 25.7. 101 Zos. 3.25.7. 102 Although Victor was present at the Battle of Adrianople, he was clearly removed enough from the frontlines to first seek out the reserves, then save himself, and finally ride rapidly to bring word to Gratian, whereas two of the other magistri and Valens himself were closely involved enough to be

Valens’ Eastern Empire

91

If this incapability was real and not just a failure of our sources, it may have led Valens to consider Victor unable to administer a frontier alone, and thus both magistri were kept together. An obvious solution would have been to replace Victor with a new magister, or even appoint a third. There are several possible explanations for why Valens did not do this. He may have feared creating more potential usurpers in the upper ranks of the military, or he chose to not divert from the two-magister system established by his predecessors, even though Valentinian had already promoted a third.103 The explanation might also be as simple as Valens failed to recognize the potential for disaster his disposition of magistri was going to cause. Whatever the reason, Victor’s specialty became diplomacy and negotiation rather than military command, and he was in charge of the embassy that was sent to the Goths in 366.104 To gain the glory he desired, Valens launched punitive attacks against the Goths in 367 and 369. Arinthaeus, who had been promoted to replace Lupicinus as a magister, pillaged the Gothic countryside, but there was no decisive battle.105 In the east, Shapur had spent 363 to 366 cementing his control over the Roman territories that had been ceded to him, and in 367 he extended Persian control even further by imprisoning the king of Armenia, and replacing the king of Iberia with a puppet.106 Valens felt he had to counter these moves, so in 369 he sent the comes Terentius to escort Pap, the heir to Armenia, back to the country to begin asserting his claim as ruler.107 He also sent Victor and Arinthaeus to negotiate a peace treaty with the Gothic leader Athanaric during the winter of 369–370, despite not having achieved the resounding success he desired.108 The Goths, hurting from the Roman incursions, readily agreed. The following year, Arinthaeus took the field army to Armenia to enact Pap’s full restoration.109 His arrival interrupted a planned invasion by Shapur, and most of Armenia fell under Roman influence again.110 Arinthaeus was rewarded the consulship of 372 and then he disappears, probably into retirement.111 Epigraphic evidence indicates that killed. His role during Theodosius I’s defeat in Macedonia is also unclear. See page 101 for more. This lack of command ability has cases led to some historians believing he retired sooner than in reality, e. g. Errington 1996b, 2. 103 See pages 79–80. 104 AM 26.10.3, 27.5.1. 105 AM 27.5.4–6, 5.9; Zos. 4.11.2–3. 106 Lenski 2002a, 167–8; 2007, 123–4. 107 AM 27.12.10; Mos. Chor. 3.36–37. 108 AM 27.5.9; Zos. 4.11. 109 AM 27.12.13; Faust. Byz. 5.1; Mos. Chor. 3.37 110 Lenski 2002a, 173–4. 111 CTh 7.22.8, 6.4.19; ICUR n. s. 8 23412 = ILCV 2795B adn.; P. Lips. 47.17; Bagnall, et al. 1987, 278–79. It is disputed whether Arinthaeus remained a magister militum after his consulship in 372. P. Col. VII 184.1 indicates that Arinthaeus survived his consular year, but Theod. HE 4.33 is the only dateable evidence remaining after this which indicates Arinthaeus was alive and in Constantinople in June 378. Woods 2001 is plagued with some unfortunate inaccuracies surrounding the magisterium in general, but he nevertheless demonstrates that Theodoret is unreliable and this story was probably

92

The Magistri Militum from 364 to 395 CE

in the early 370s a man named Julius was promoted to comes et magister equitum et peditum per Orientem.112 Comes used in this manner looks to have been an honorary title that appears increasingly through the later period of this century and into the next.113 Valens thus maintained the same complement of two magistri. Although he did not personally campaign in Armenia, from 371 the emperor was headquartered mostly at Antioch, but also in Hierapolis (Manbij, Syria) and Edessa (Şanlıurfa, Turkey).114 Victor was also in the east at this time, so Valens also continued to employ both magistri in the same theatre as himself.115 This left the Danube provinces of the eastern Empire to be managed, or more accurately mismanaged, by lower-ranking officers.116 In late 376, Valens sent Victor in an embassy to Persia to demand that Rome be allowed to maintain its influence in the Caucasus. Tensions escalated after two territories in Persian-controlled Armenia offered to surrender themselves to Victor, and Valens aggressively threatened Persia with war.117 His confidence was boosted because he had received word that a large group of Goths, different from those he had fought, had arrived on the Danube seeking permission to enter and settle in the Roman Empire.118 In exchange, they offered to provide many recruits for the army, which Valens hoped could bolster his forces and potentially even allow him to launch his own invasion of Persia. However, the officials overseeing the settlement of the Goths severely mistreated them and drove them into open rebellion.119 By spring 377 Valens sent Victor to Shapur once again, this time to negotiate an urgent truce so the main Roman army could be extricated from the east and sent to the Danube.120 In terms of military operations, Valens’ reign up to 376 was a muted period. Although he engaged in both civil and foreign war, he did not manage to fight any major battles. It is important to note, however, that despite his unremarkable military career,

112 113 114 115 116 117 118 119 120

fabricated to defame Valens’ religion. Additionally, AM 30.2.3, which Woods misses, might indicate Arinthaeus was dead by 374. Therefore, I agree that Theod. HE 4.33 should not be considered an accurate recording of history, although it still demonstrates the kind of supportive relationships that magistri could have, and Arinthaeus most likely died in 373. The letters addressed to him from Basil of Caesarea (Bas. Ep. 179, 269) which are often dated to 374 and 378, must be dated earlier. CIL III 88. Elton 2018, 66. Lenski 2007. AM 30.2.4–5. Although the western province of Illyricum was controlled by the magister militum Equitius, he held no authority in the eastern provinces. AM 30.2.5–7; Blockley 1992, 36. Lenski 2002a, 183. On the developments that led to the appearance of these Goths, see: Kulikowski 2016, ch. 8. AM 31.5.1–8; Zos. 4.20.6–7. AM 31.7.1; Blockley 1992, 37; Lenski 2002a, 184. The details of the truce between the Romans and Persians are not recorded, but the Armenian sources relate that they defended themselves against a Persian attempt at occupation and then ruled themselves independently for seven years (Faust. Byz. 5.37–8; Mos. Chor. 3.40).

Crisis along the Danube

93

and his reliance on his magistri, Valens kept up the expected image of a military capable emperor. He was vigorously active with his army, and it appears he was always planning greater campaigns when possible, whether against the Goths or Persians. This activity seems to have been enough to keep his magistri in line and serving his agenda rather than their own. 2.4 Crisis along the Danube The Gothic crisis of 376 to 382 was the greatest challenge to Roman sovereignty during the fourth century. The Goths ravaged the Thracian countryside, won numerous battles against the Romans, and even killed emperor Valens. Worst of all, the Romans never truly overcame them, but only abated the Gothic devastation by negotiating a peace treaty which gave the Goths a large swathe of Roman territory. Being a military affair, the magistri militum were critically important to these events. Unfortunately, Ammianus’ history ends in 378, just after the death of Valens, and the evidence that remains is disjointed and less coherent, which results in a somewhat stuttering account, where magistri appear, disappear, and then reappear a number of years later. This makes concise narration difficult. Although narrative will not be avoided, the emphasis of this section will be placed on the dramatic expansion in the number of magistri that occurred during the Gothic crisis. At its beginning, there were three magistri militum in the entire Roman Empire, but over the course of the crisis, eight new magistri would be promoted, and in 378 there were six in the Danube region alone. Scholars have often interpreted this as the inception for the larger eastern magisterium that persisted well into the Byzantine Empire.121 This section will especially focus on the effect of so many magistri on the political dynamic of the late fourth century, as well as whether the numbers helped or hindered the effectiveness of the army during this period. What is more, shortly after the crisis abated, almost all these magistri disappeared. We will explore why this may have happened, and what it meant for the future of the magisterium and the Empire. In early 377 the Goths were in open revolt in Thrace, and Victor had negotiated a hasty peace with Shapur so the army could be extricated from the east. Before he was able to go to the Danube, he was forced to deal with a revolt by the Saracens and their queen, Mavia.122 They were rebelling because an Arian bishop had been ordained to them instead of the Orthodox hermit by the name of Moses whom they wanted. The

121 Mommsen 1889, 265; Enßlin 1930, 317; 1931a, 144; Hoepffner 1936, 485; Hoffmann 1974, 393; Demandt 1970, 557. 122 Ruf. HE 11.6; Soc. 4.36; Soz. 6.38; Theod. HE 4.23; Theoph. 5869; Nic. Call. HE 11.47; Bowersock 1980; Shahid 1984, 138–202; Lenski 2002a, 204 n. 284. For a more cautious reading of the sources, see: Mayerson 1980. For dating, see Lenski 2007, 127 n. 107.

94

The Magistri Militum from 364 to 395 CE

Saracens raided from Phoenice to the Sinai, and the local commander of the limitanei requested assistance from an unnamed magister militum, who has been provisionally identified as Julius.123 In the first major confrontation, the Saracens defeated the Romans and the magister barely escaped with his life.124 The Romans then sued for peace, and it was likely Valens’ go-to negotiator, Victor, recently returned from negotiating peace with the Persians, who was sent to Mavia to reach a settlement.125 Victor negotiated a satisfactory deal in which Valens received peace and a detachment of Saracen cavalry to serve in his army, and Mavia received Moses as bishop.126 Mavia’s daughter, who perhaps was the princess Chasidat, was also married to Victor.127 Irfan Shahid argues that, in light of the later Islamic principle of kafā’a, which probably has its origins in earlier Arabic traditions, this marriage elucidates the high-standing of the magistri.128 This custom requires that the two partners of a marriage have the same social rank, which means Mavia felt that the magister militum Victor was of a high enough position to join her royal family. Due to the delay from these events, Valens sent ahead an army under the command of Profuturus and Traianus.129 Ammianus says that they both had high aspirations but were imbellis, and he only awards them the title of rector in this episode, so it seems as if Valens had not yet assigned magistri to the developing Gothic crisis.130 The Gothic forces had divided into small, spread-out bands to raid more efficiently. These would stand little chance of resisting the newly arrived Roman army, so they retreated and regrouped north towards the Danube, establishing a fort near a place called Ad Salices, ‘At the Willows’, whose modern location is unknown but may have been in Dobruja, the area of land contained by the Danube River when it turns north and runs parallel to the Black Sea for roughly one-hundred miles.131 123 For discussion about this magister militum’s title and identity, see: Demandt 1970, 704; Shahid 1984, 150; Woods 1998, 328–34. 124 Soz. 6.38.4. 125 Blockley 1992, 39 n. 49 indicates the chronology of events favors Victor as the negotiator of the settlement. This is a more likely scenario than the one proposed by Woods 2004, 733. 126 Lenski 2002a, 335 n. 94. 127 Soc. 4.36.12. Mavia’s daughter has been tentatively identified with the ‘Chasidat’ recorded by the Anasartha inscription from 425 (Shahid 1984, 227–38; Ball 2016, 107). 128 Shahid 1984, 160–61, although Shahid is mistaken in thinking that, as magister equitum, Victor was particularly involved in commanding cavalry. It seems Shahid is also mistaken when theorizing that Victor may have been particularly lackluster in conducting the war against Mavia because of their shared Orthodoxy, and this may explain the Roman defeats (Shahid 1984, 162–63). However, earlier Shahid acknowledges that it was Julius, not Victor, who conducted the war, meaning his Orthodoxy should not have factored into the Roman army’s competency (Shahid 1984, 151). Furthermore, Shahid makes a number of titulature mistakes, giving Julius the title of magister utriusque militiae and Victor the title magister equitum praesentalis, neither of which they receive in the sources (Shahid 1984, 161, 164, 167, 169). 129 AM 31.7.1. 130 Contra Demandt 1970, 705–6. 131 AM 31.7.2 ff.; Meier 2020, 177–78; Wanke 1990, 63, 157–58; contra Heather 1991, 144.

Crisis along the Danube

95

Because of the seriousness of the Gothic threat, Valens requested reinforcements from the western Empire. To aid Valens, the young emperor Gratian ordered some forces from Pannonian and the transalpine region to go to Thrace under the command of Frigeridus, whose precise rank is unknown.132 Gratian also intended to send a portion of his Gallic army under the command of his comes domesticorum, Richomeres. Before they could leave, however, Merobaudes incited a significant portion of the soldiers to betray their orders and remain in Gaul.133 This was done under a pretense of maintaining the safety of the region, but Merobaudes’ true motivation was most likely to ensure a copious number of soldiers remained under his direct command, rather than having them be split up, sent away, and potentially have another commander win their loyalty. Sources for the western Empire during this time are rather scarce, but this episode helps illustrate its political landscape. It is clear that Valentinian was entirely left out of military matters: Valens specifically contacted Gratian for reinforcements, and Gratian was able to give orders to soldiers from Pannonia, an area that was supposed to be under the control of Valentinian.134 It also shows that Merobaudes was now in Gaul and in command of the Gallic army. Combined with his consulship of 377 which was shared with Gratian, this indicates Merobaudes had been able to successfully leverage the elevation of Valentinian II to gain a central and powerful position in the western government.135 This position was so powerful and secure that Merobaudes was able to contravene Gratian’s direct orders to the military, and he faced no repercussions or punishment. Gratian may have sought to limit Merobaudes’ power in the future by promoting more magistri, which will be discussed below, but for the moment, Merobaudes operated on his own agenda with impunity. In the end, Frigeridus became ill and could not lead his own troops, so Richomeres was able to take command of them and still lead a force of western soldiers into Thrace.136 The Goths had also increased their numbers by gathering more of the raiding bands to the wagon-fort.137 The combined eastern and western forces fought a large battle at Ad Salices, with both the Roman and barbarian sides taking heavy casualties.138 The Goths suffered so badly that they did not emerge from the wagon-fort for seven days.139

132 AM 31.7.3–4. For Frigeridus, see: PLRE I, 373–74. 133 AM 31.7.4. 134 Zos. 4.19.1–2. Errington 1996a, 441–42, convincingly disputes whether it was even officially under Valentinian II’s command at all. 135 CTh 9.35.3, 8.7.14; ICUR n. s. 1 3188 = ILCV 4289; P. Flor. 1 95.82; AM 31.8.2; Bagnall, et al. 1987, 288–89; McEvoy 2013, 63. 136 AM 31.7.5. 137 AM 31.7.7. 138 AM 31.7.2 ff., 8.1; Wanke 1990, 63, 157–58; contra Heather 1991, 144. 139 AM 31.8.1.

96

The Magistri Militum from 364 to 395 CE

To bolster the now-weakened Roman forces, Valens gave an officer named Saturninus “temporary command of a cavalry army”.140 This may imply that he was a magister militum, perhaps a magister equitum, and his important commands and eventual consulship in 383 help us be more certain that he attained this rank.141 Saturninus joined with the Roman forces still in Thrace, where he found that the Goths had received some Hunnic and Alani reinforcements, and so the Romans retreated to the cities for the winter while the Goths renewed their raids.142 During this time, Richomeres returned to Gratian to requisition more soldiers.143 Valens finally arrived in Constantinople on 30 May 378.144 While there, or perhaps at the same time Saturninus was promoted, Traianus was promoted to the magisterium. At this point of his history, Theodoret writes that Traianus was confronted by Valens for not soundly defeating the Goths.145 Traianus retorted that it was not his fault, but rather it was Valens’ Arianism which had turned God against the Romans.146 The fellow magistri Arinthaeus and Victor supported Traianus, but Valens removed Traianus from command in retaliation. Although Ammianus seems to corroborate this story, David Woods makes the convincing argument that this story is false, and Theodoret was trying to defame Valens.147 Whatever the truth of the matter is, Sebastianus, the officer who had been sidelined by Merobaudes in 375, was also promoted to the magisterium in Valens’ army.148 Valens spent the first part of the campaign season preparing, and then come June he marched the army to the city of Adrianople (Edirne, Turkey).149 He had with him as magistri militum Traianus, Victor, Sebastianus, and Saturninus.150 Combined with Julius on the Persian frontier, this meant there were five eastern magistri, the highest number seen so far. They were also met by Richomeres, who informed them that Gratian was not far away with the western field army.151 However, Valens was jealous of

140 “equestris exercitus ad tempus cura commissa” AM 31.8.3. 141 See pages 163–64 for more on Saturninus’ consulship. 142 AM 31.8.4–6. Note that, while the Goths were still the main body of barbarians, from this point onwards it was a large coalition of other barbarian groups, workers and slaves who had joined the raiding, and even some Roman deserters. See: Lenski 2002a, 331–32. 143 Lenski 2002a, 330. 144 For Valens’ date of arrival in Constantinople, see: Cons. Const. s. a. 378; Soc. 4.38.1; contra AM 31.11.1; Eun. Hist. fr. 44.1; Zos. 4.21.1; Soz. 6.37.17, 39.2. 145 AM 31.12.1. 146 Theod. HE 4.33.3. The whereabouts of Profuturus after Ad Salices is unclear, but AM 31.8.3 indicates he survived the battle. 147 See note 91n111. 148 AM 31.11.1. 149 Cons. Const. s. a. 378; Soc. 4.38. 150 AM 31.12.1. As mentioned in note 91n111, Arinthaeus had probably died before the Gothic war began, but if he had not, it may be that he retired or died at this point and Traianus was promoted to replace him, as argued by Lenski 2002a, 363. 151 AM 31.12.4.

Crisis along the Danube

97

both Gratian’s recent victory over the Lentienses, an Alamanni group along the upper Rhine, and some successes that Sebastianus had won against the Goths.152 Valens himself had still not been able to achieve his goal of decisive victories over foreign enemies. Scouts had also reported that the nearby contingent of barbarians numbered only ten thousand, which made him overly confident.153 He therefore called a council of war where he forced his generals to seek a decisive battle against the barbarians before the western reinforcements could arrive.154 The assigned duties of the magistri during the opening phases of the Battle of Adrianople on 9 August 378 are unfortunately unrecorded. Ammianus writes that Richomeres and some other officers were in attendance when the Goths attempted to delay the battle by negotiating, and it is likely that some of the magistri were also present.155 They were certainly not competently supervising the army, because during the negotiations some soldiers prematurely rushed forward and attacked the Goths.156 This over-extended the Roman lines, and the Gothic cavalry was able to launch a devastating surprise attack on the Roman flank.157 The Goths exploited this success by attacking the main body of Roman infantry on two sides, which compressed the Romans so tightly that many could not even swing their swords.158 The magistri reappear towards the end of the battle. Traianus was most likely in the center with the emperor, because he called for aid after Valens was abandoned by his bodyguard.159 Victor received the message and sought out the reserves, but they had fled, so he, Saturninus, and Richomeres retreated off the field.160 Sebastianus may have co-commanded the center with Traianus, because the two of them died on the battlefield, along with a multitude of other tribuni and rectores, and emperor Valens himself.161 They were two of only five magistri to fall in battle in the fourth century.162 Although none of the surviving magistri were punished for fleeing, some ancient sources blame them for the defeat.163 Libanius wrote that many people accused the generals of incom152 AM 31.12.1. 153 AM 31.12.3; Nicasie 1998, 246, theories that this was probably half the Roman number, although estimates vary widely throughout scholarship. 154 AM 31.12.5–7 writes that Sebastianus supported Valens, while Victor cautioned the emperor to wait, although Zos. 4.23.6–24.1 switches it, writing that Sebastianus urged Valens to delay and the general’s enemies at court who supported the emperor. 155 AM 31.12.14–16. 156 AM 31.12.16. For the way contemporaries thought generals should act during battle, see: Veg. 3.17. 157 AM 31.12.17. 158 AM 31.13.2. 159 AM 31.13.8. 160 AM 31.13.9. 161 AM 31.13.18. 162 The others being Romulus (at the Battle of Mursa, see Zos. 2.52.2), Promotus (although this was somewhat more of an assassination, see page 113), and Quintinus (he presumably met his death on his failed expedition, see pages 106–107). 163 Them. Or. 15.189d; cf. Greg. Naz. Or. 22.2.

98

The Magistri Militum from 364 to 395 CE

petence and cowardliness, which may have stemmed from their failure to maintain order in the army, as well as their flight.164 Vegetius also dedicated a significant portion of his handbook to the desired order of battle – perhaps in the hopes that soldiers and their officers would better keep to it.165 The battle at Adrianople was one of the worst disasters in Roman military history. Ammianus, alive at the time, compared it to another infamous thrashing of the Romans, the defeat at Cannae by Hannibal in 216 BCE.166 Rufinus calls it the origin of all the Empire’s evils, and Ambrose sees in the marauding Goths the apocalyptic attack of Gog.167 It has been used by modern historians as a ‘beginning of the end’ point for the western Empire, and even though this is somewhat tendentious – it occurred in the eastern Empire and against the eastern army, after all – this is still an adequate reflection of the magnitude of the disaster.168 Although it might not have been a fatal military disaster, it does mark the end of an era of military emperors. Not until Heraclius (610–641) would an emperor lead a Roman army from the frontlines.169 Furthermore, Roman armies would cease conducting expeditionary campaigns across the limes; instead warfare would permanently shift to the internal side of the frontiers. After the defeat, Victor quickly rode to Pannonia to deliver the news to Gratian.170 The western emperor clearly had little interest in avenging Valens or continuing the Gothic war himself, because he retreated to the defensible city of Sirmium. From there, he undertook a campaign against the Sarmatians across the Danube and began rebuilding the devastated eastern administration, and in the four years it took for the Gothic crisis to be resolved, eight magistri, including five new appointees, played crucial roles in eastern affairs.171 One of Gratian’s first actions was to order the newly promoted magister Sapores to remove the Arian priests that Valens had supported and replace them with adherents to the Nicene Creed.172 Sapores managed this easily in all cities bar Antioch, where he was required to mediate a theological dispute before awarding control of the city’s churches to the Nicene Meletius.173 Because he only appears in Greek sources, Sapores is not explicitly titled as a magister militum, but only as a στρατιωτῶν. As with the other ambiguous cases, we can make a judgment based on other evidence. Both Libanius and Theodoret praised his military successes and said he was highly regarded, and he is also associated

164 Lib. Or. 24.3–5. 165 Veg. 2.15–18, 3.14–18. 166 AM 31.13.19. 167 Ruf. HE 11.13; Amb. fid. 2.16.137–138. 168 Elton 2018, 86. 169 Mitchell 2015, esp. 89. 170 Zos. 4.24.3; Errington 1996a, 440. 171 Theod. HE 5.5; Pan. Lat. 12.12.9–10. 172 Theod. HE 5.2–3. 173 Theod. HE 5.3.

Crisis along the Danube

99

with Julius and Victor, other magistri.174 Additionally, his mission is similar to the one given to the magister militum Salia in 344. Thus, we can be confident Sapores was a magister militum, and therefore we can conclude that in the later fourth century the magistri continued to play an important role in exerting the religious will of the emperors. The most prominent appointment of this period was the new eastern emperor Theodosius, the homonymous son of the magister militum Flavius Theodosius discussed previously. This new emperor may have reached the imperial position through his own promotion to the magisterium.175 It is recorded that he had been summoned from Spain after the defeat at Adrianople.176 It has been demonstrated by Malcolm Errington, however, that this timeline is impossibly short, and Theodosius must have already been present in Gratian’s army by the beginning of 378.177 It is unclear what rank he occupied during that year. One scholar theorizes he was recruited directly as a magister militum, but others have argued that he began as a dux and only later received a promotion to the magisterium.178 The reasons behind Theodosius’ initial recruitment into Gratian’s army, as well as his eventual promotion to emperor, are complex. He was probably supported by an influential clique of friends and relatives at court, indicating favoritism and nepotism played a part in achieving military ranks.179 Gratian may have also been eager to promote a second magister militum to challenge Merobaudes’ dominance, as he had been the west’s sole magister since the death of Flavius Theodosius père in 376.180 In the same way that he used his dominant position to keep his army whole, Merobaudes may have undermined any attempt to promote a second magister. None of the sources record Merobaudes’ presence on the 378 campaign, so it was only once Gratian had escaped from Merobaudes’ oversight that he was able to promote Theodosius, and then later Bauto, as magistri.181 Whatever the chronology of his promotions, Theodosius was the only fourth-century magister militum to successfully become a lasting and legitimate emperor, and his accession shows that, as long as the process was valid and filled a power vacuum without challenging anyone, a promotion from magister to emperor was possible.182

174 Lib. Ep. 957; Or. 2.9; Theod. HE 5.2. 175 Cons. Const. s. a. 379; Soc. 5.2; Theod. HE 5.6.3; Pan. Lat. 12.11.1 ff.; Epit. 48.1; Soz. 7.2.1; Oros. 7.34.2. 176 Theod. HE 5.5–6. 177 Errington 1996a, 438–40; and also Matthews 1975, 91. A similar conclusion on Theodoret’s reliability was drawn by Sivan 1993b. 178 Blanco-Perez 2013, 150; Errington 1996a, 449; Heather and Moncur 2001, 219. For the primary evidence, see: Them. Or. 15.187d, 198a, and Pan. Lat. 12.10.2–3. For his previous experience, see: Pan. Lat. 12.8.3; Zos. 4.24, 16.6; AM 29.6.14–16. 179 Matthews 1975, 91–98; 1971, 1075–76; Vanderspoel 1995, 187–95; Errington 1996a, 449. For an alternative view of where Theodosius received support, see: Sivan 1993b, 121; cf. Barnes 1990, 162. 180 See page 137. 181 Contra Errington 1996a, 451. 182 Sivan 1996 theories that Theodosius should rightfully be called a usurper on the basis that the decision was forced upon Gratian. Theodosius’ long-term legitimacy, however, differentiates him from the other magisterial usurpers, Vetranio and Silvanus.

100

The Magistri Militum from 364 to 395 CE

After Theodosius’ appointment, the magister Julius orchestrated a massacre of a large number of young Gothic men that had been taken as hostages into the eastern cities.183 It was feared that these men could act as a fifth column, so with the promise of gold they were lured into simultaneous ambushes across their host cities. Interestingly, Julius did not seek the permission of emperor Theodosius, who was busy battling the Goths and who had maybe not even confirmed that Julius was to remain in his office. He instead sought authorization from the Constantinopolitan Senate.184 This is an interesting glimpse into power dynamics at a time when the emperor’s attention was entirely elsewhere. Along the Danube, Theodosius utilized every available strategy to overcome the Goths, including adding new magistri to his army.185 The first appears after Gratian ceded control of the Prefecture of Illyricum to the eastern Empire, perhaps to distance himself from the dangers of further failure and so that Theodosius could control the entire theatre of war.186 Theodosius appointed Majorianus to take command in Illyricum, and this new magister continued Gratian’s campaigns across the Danube, probably still against the Sarmatians, and he also fought against the Goths in Pannonia.187 Unfortunately, we are unaware of more detail than this. Once Gratian retook control of Illyricum in late 380 it seems as if Majorianus lost his command, and we do not hear of him again.188 Modares may have also received an appointment to the magisterium under Theodosius. Zosimus called him a prince of the Gothic royal family, and this is often presumed to mean he was a member of the Balt dynasty.189 This would make him a relative of Athanaric, who remained north of the Danube until 381.190 Athanaric was an enemy 183 Zos. 4.26.2–9; AM 31.16.8; Kulikowski 2019, 91. Eun. Hist. fr. 42 corroborates the existence of these Goths, but his record of the massacre has been lost. Ammianus put this event just after Adrianople and therefore before Theodosius’ accession, but Zos. 4.26.4 places it later, either in 379 or early 380. Although Ammianus is typically the more preferable historian, Zosimus’ version of this event is quite comprehensive and significantly more coherent than the surrounding sections of his narrative, and it allows more time for the necessary planning for such an operation, and so he should be followed in this case (Kulikowski 2007, 147; cf. Heather 1991, 147–48). It is convincingly argued that Ammianus included this massacre in his history as a subtle criticism of emperor Theodosius’ settlement of the Goths on Roman land in 382 (Sivan 1993a, 118). Direct criticism was too risky, but by lauding Julius’ slaughtering of the Goths, Ammianus is clearly telling us which approach to the Goths he supports. 184 Zos. 4.26.5–9. Also see: Demandt 1970, 710–11. 185 Other strategies included using reinforcements brought from Egypt (Zos. 4.30.2). 186 Soz. 7.4.1; Errington 1996b, 22–27; Kulikowski 2007, 150. 187 Sid. Ap. Carm. 5.107–15. 188 Demandt 1970, 602–3, believes that Majorianus’ successor, Vitalianus, may have been a magister militum at this time, but because he had previously only reached the rank of protector domesticus, and there does not seem to be a magister militum in the area either before or after, I do not think this is likely. 189 Zos. 4.25.2; Wolfram 1988, 131; Heather 1991, 189–90. 190 Them. Or. 15 190d ff.; Soc. 5.10; Jord. Get. 142 ff.; Cons. Const. s. a. 381; Zos. 4.34.4–5; Oros. 7.34.7.

Crisis along the Danube

101

of Fritigern, the leader of the Goths marauding Thrace and Illyricum, and it may have been this rivalry that gave Modares cause to join the Roman side.191 In Thrace, Modares’ scouts discovered a group of intoxicated and sleeping barbarians, a trope common to Zosimus that should be read skeptically.192 Zosimus also writes that Modares ordered his men to take off their heavy armor and attack with only their swords and shields so they could be silent. In a fragment which does not mention Modares’ name but might be describing the same event, Eunapius recounts an attack during which the Roman soldiers advanced silently to surprise the barbarians.193 Whatever the true circumstances, Modares reportedly won a great victory over the enemy, which could have been the justification for the announcement in Constantinople of a victory over the Goths, Alans, and Huns.194 This is the only record of Modares’ military commands, but he was later in Constantinople where he received two letters from Gregory of Nazianzus.195 Both are full of praise for the commander, but Epistle 136 in particular lauded him as having the soul of a Greek despite his barbarian body, and it praised him for ending foreign wars. While Gregory was hoping to gain Modares’ support during the First Council of Constantinople and was therefore probably exaggerating his exploits, the fact that Modares wielded enough power and influence to make a difference during the religious council indicates he probably reached the magisterium and was quite an important figure in eastern affairs. Neither Saturninus nor Victor are recorded to have undertaken any missions of note during the first few years after Adrianople, despite them being the two most experienced and veteran generals under Theodosius.196 This might be because they were in command of Theodosius’ primary field army, and the deeds of this force were attributed to the emperor rather than the generals. Theodosius spent 379 replenishing the losses suffered at Adrianople, and in 380 the Goths moved against him in Macedonia.197 The new recruits were too inexperienced, and the ensuing battle was almost as much of a disaster as Adrianople.198 Saturninus and Victor were probably present during this battle, and probably also followed Theodosius when he retreated first back to Thessalonica (Thessaloniki, Greece) and then to Sirmium.

191 Zos. 4.25.2. 192 Zos. 4.25.2–3. Another example of this trope is Zos. 4.49.1. 193 Eun. Hist. fr. 51. 194 Cons. Const. s. a. 379. Zosimus’ praise of Modares as having totally annihilated the Goths in Thrace is a good example of Zosimus confusion about this period that stems from his attempt to synthesize two sources on the same topic (Heather 1991, 77–78, 147). 195 Greg. Naz. Ep. 136, 137. 196 Demandt 1970, 717–18; Enßlin 1931a, 137 f. 197 Zos. 4.31.2–3; Heather 1991, 152. 198 Zos. 4.31.4–5.

102

The Magistri Militum from 364 to 395 CE

Theodosius then entered Constantinople for the first time in November 380.199 Saturninus and Victor were with him, and after entering the city they set about releasing a monk they had imprisoned on the orders of Valens.200 By this time, Victor had been a high-ranking officer for around twenty years and a magister militum for over seventeen under five different emperors, the longest career of the fourth century. Although there was no fanfare or note of his retirement, he already having held the consulship in 369, it seems as if the long-serving general now left military service. Richomeres, who was Gratian’s comes domesticorum up to at least the Battle of Adrianople, was at some point transferred to the eastern administration under Theodosius. He seems to have been one of the more important persons, because it was said that he had been promised the consulship of 384 as early at the beginning of the previous year, a rare occurrence.201 To the best of our knowledge, no comes domesticorum was ever made consul without first being promoted to the magisterium. In 391 Richomeres is confirmed as a magister utriusque militiae, and he was possibly promoted to this position as early as 383 to fit in before his consulship.202 Libanius wrote that Richomeres had been somewhere on the Persian frontier before 384, so he may have replaced Sapores or Julius once Theodosius took over control of eastern affairs.203 In 381, Gratian sent an army to Illyricum and Thrace as part of the new agreement to provide more assistance. This army was commanded by the Frank Bauto, who was aided by his countryman Arbogast.204 Bauto is another general who the sources do not explicitly name as a magister militum, but the importance of his mission probably required someone of magisterial rank. Furthermore, he later gained such a powerful position that he must have become a magister at some point, and it was probably at the beginning or end of this campaign to the Danube.205 The efforts of all these magistri helped wear down the Goths through attrition and push them to the negotiating table in 382.206 One of the most significant agreements of the peace treaty that resulted from these negotiations was the semi-autonomous settlement of the Goths in the provinces of Moesia II and Dacia Ripensis.207 These

199 Thessalonica: CTh 10.10.10. On entering Constantinople in November: Cons. Const. s. a. 380. 200 On the arrest of Isaac, see: Vita Isaacii 2.8. On his release, see: Vita Isaacii 3.10. For further discussion, see: Shahid 1984, 168–69; Lenski 2004, 107–113. 201 Them. Or. 16.201b. For other westerners who served in the early years of Theodosius’ reign, see: Errington 1996b, 2–3. 202 CTh 7.1.13; cf. Enßlin 1931a, 138. 203 Lib. Or. 1.219. Richomeres also returned to the east, perhaps around 391. See page 113. 204 Zos. 4.33.1–2. Soc. 5.25 says Bauto hailed from Galatia Minor, although this is contradictory both with the other evidence of his Frankish heritage and his service in the western military. On dating this campaign, see: Heather 1991, 155. 205 cf. Demandt 1970, 609. 206 Zos. 4.33.1–2; Them. Or. 16.208b–209d, 210d. 207 Cons. Const. s. a. 382. Syn. de reg. 19.43.5 records that the Goths lived under their own laws, and Jord. Get. 27.141–46 describes it as an equal agreement rather than a surrender, in the manner that Them.

Crisis along the Danube

103

lands were right on the frontier which the Goths had crossed, sandwiched between the Haemus Mountains and the Danube River, but they were still Roman territories that had been effectively ceded to invaders.208 This provision, and similar agreements made in the fifth-century, were one of the primary ways the Empire lost sovereignty over territories it had controlled for centuries. The magistri involved in the Gothic crisis seem to have left our records just as quickly as they appeared. Julius, Modares, Majorianus, Sapores, and Victor all vanished by 383. Saturninus remained present in Constantinople, but is not attested as having campaigned or commanded soldiers again.209 The only magistri that remained active were Merobaudes and Bauto in the west, and Richomeres in the east. This rapid reduction in numbers has multiple explanations. Some left the magisterium in an expected manner, such as Victor’s overdue retirement.210 Majorianus was appointed to oversee Illyricum when Gratian gave it to Theodosius, and when it was returned, Majorianus lost his command.211 Another potential explanation comes from Zosimus. While he gives credit to the generals for having saved the eastern Empire, he castigated Theodosius for having so drastically strained the imperial fisc paying them all.212 Thus, financial concerns could have also motivated Theodosius to reduce his magistri.213 Theodosius may have also been wary of the ever-present risk of usurpers. Themistius, an imperial orator and propagandist, tried to place a positive spin on the peace treaty by explaining that it was better to turn the Goths into farmers rather than continue fighting and fill the Empire with corpses.214 But Theodosius had spent nearly four years failing to destroy the Goths, and he probably felt his legitimacy as an emperor was threatened. If so, he could have taken the opportunity to remove some potential military rivals from their powerful positions.215 Another possible explanation is that

Or. 16 portrays it. The speculative nature of modern academic works on the 382 treaty is summarized in Kulikowski 2002, 77–78 n. 28. Also see: Stickler 2007, 504–5; Faber 2013. 208 Meier 2020, 186. 209 Them. Or. 16; CTh 6.2.13. 10.3.4; CIL VI 501 = ILS 4149; P. Gen 12.1. He remained in Constantinople as a prominent aristocrat until his exile was demanded by Gainas (Greg. Naz. Ep. 132; Zos. 5.18.7–9; Soc. 6.6; Soz. 8.4; Joh. Ant. fr. 190 (Müller) = fr. 215 (Mariev)). 210 Victor had been an officer for at least twenty years and a magister for seventeen, under five different emperors. This was the longest career of our period, although it is almost matched by Stilicho’s, whose career extends past our purview. 211 Soz. 7.4.1; Errington 1996b, 22–27; Kulikowski 2007, 150. 212 Zos. 4.27.1–3. Errington 1996b, 3, incorrectly writes that we do not know enough individuals to fill Zosimus’ claim of five magistri, although as demonstrated, we have more. 213 For a lengthier discussion of the vicissitudes of the late Roman economy, see: Potter 2015. 214 Them. Or. 16.211a-b. 215 Interestingly, Mommsen believed that Theodosius might have maintained the enlarged magisterium he inherited from Valens in order to counter-balance the power of the military, especially in the face of the power exhibited by Merobaudes (Mommsen 1901, 536–37). However, neither Mommsen nor any other historian has emphasized the temporary downsizing in Theodosius’ magisterium, which presents a flaw in this theory.

104

The Magistri Militum from 364 to 395 CE

the magistri may only be appearing to have vanished because of the poor state of the ancient sources. However, the sources do not leave this period entirely dark. There are references that indicate that lower ranks were in command of many of the military operations that took place after the Gothic peace treaty.216 So perhaps some more magistri existed in this gap, but it seems very unlikely that Theodosius maintained the full complement of magistri during these years. Yet, more crises would soon plague his reign, and he would once again expand his magisterium. Another assessment to make is whether the employment of so many magistri over the course of the Gothic crisis assisted or hampered the army’s efficacy. While it was Valens who made the reckless decision to engage at Adrianople without western reinforcements, the catastrophes on the day of battle can be attributed to the operational and tactical failures of the magistri. They failed to prevent units from prematurely engaging, they did not locate the Gothic cavalry prior to the surprise assault, and they left the reserves unsupervised which allowed them to leave the battlefield. Additionally, Saturninus and Victor were later unable to prevent Theodosius’ newly recruited army from being destroyed in Macedonia, although we are aware of fewer details in this engagement. However, the magistri were redeemed when they operated independently. Julius was praised for saving the eastern cities by massacring the hostages, Modares won a great victory in Thrace, and Majorianus and Bauto both successively drove the Goths out of Illyricum.217 Thus, it seems that many magistri together were not successful in the largest engagements of the war, but when operating with smaller armies and independent of the emperors, they were able to wear down the barbarians and push them towards the peace agreement. 2.5 Civil War 383–388 Eastern and western affairs have generally been separated in the previous sections, but the final years under review in this project revolved around two major civil wars between the halves of the Empire. This means that a separated examination no longer best serves our purposes, and so this section will combine the perspectives of the east and west and instead proceed chronologically. These civil wars were both symptoms of, and responsible for, major changes in the magisterium and the wider Empire. We will see how a small number of western magistri concentrated great power into their hands and were able to dominate the western emperors, while at the same time the eastern magisterium again expanded to a larger number of generals. This resulted in the dilution of eastern military power, and no individual magister was able to eclipse the 216 Them. Or. 18 220d–221a; Claud. de cons. Stil. 1.51–68; Zos. 4.40. 217 Julius’ praise: Zos. 4.26.9. Modares’ victory: Zos. 4.25.2–3. Majorianus’ role: Sid. Ap. Carm. 5.107– 15. Bauto’s role: Zos. 4.33.1–2.

Civil War 383–388

105

power of the primary civil officials.218 This difference had a significant impact on the divergent paths of the two halves of the Empire. Gratian had shifted his court and administration from Trier to Milan to be nearer the Gothic crisis, and after its end, Bauto and Arbogast returned there with the army, most likely in late 382 or early 383.219 This move to northern Italy changed the established power dynamics of the western Empire by placing the emperor closer to the influence of the Roman Senate and other important individuals, such as Ambrose, the bishop of Milan. It also took Gratian further away from the Gallic and British armies, which had, for quite a while, enjoyed proximity to an emperor favorable to their interests. Gratian might have been aware that this risked alienating these armies, because he tried to placate Merobaudes, who remained in Gaul as magister peditum, by awarding him the consulship of 383.220 This made him the first non-royal family member since the time of Constantine I to be granted a second consulship. Furthermore, the social hierarchy codified by Valentinian I in 371 was reaffirmed by Gratian and Theodosius, along with Valentinian II.221 In addition, Gratian officially ordained that the consulship was another rung in the administrative hierarchy.222 This constitution confirms the priority of the consulship as dictating the senior honor within senatorial meetings, but clarifies that prefects and magistri who have also attained the consulship outrank men who have held only the consulship but not a prefecture or a magisterial position. The need for such a ruling most likely arose because the official hierarchy was not adequately reflecting the changing power dynamics that had arisen since Valentinian I’s upgrading of the military magistracy’s status, as well as of the practical power these highest prefects and generals commanded. Because these laws appear towards the end of Gratian’s reign, when we know Merobaudes had already been exerting a great deal of influence over western affairs and taking actions that supported his own agenda, this should not be seen as some sort of institutional reform that gave Merobaudes’ office of magister peditum a hierarchical status above all other magisterial offices in the west.223 Merobaudes had already obtained these powers during his years as the sole magister

218 Mommsen 1889, 264–65; Boak 1915, 124; Grosse 1920, 186; Jones 1964, 161, 177–78; Demandt 1970, 727. Even though Rufinus would eventually be killed on the orders of the magister militum Stilicho, the eastern generals were still unable to take control of the eastern government, and another civilian official, the praepositus sacri cubiculi Eutropius, gained the preeminent position in Constantinople. 219 It is unclear whether Valentinian had remained with Gratian this entire time. Ammianus said that Gratian acquiesced to overseeing the young boy’s education in 375, which might mean the young boy travelled to Trier to stay with his half-brother (McEvoy 2013, 62; cf. Errington 2006, 31). 220 CTh 6.2.13; ICUR n. s. 2 5996 = ILCV 4623; CIL VI 501; P. Gen. 12.1; Bagnall, et al. 1987, 300–1, 650–52. 221 CTh 6.5.1–2. These laws are dated to after Gratian’s death, but this must be a mistake in the manuscripts (Pharr et al. 1952, 127). 222 CTh 6.6.1; Demandt 1970, 604–5. 223 Demandt 1970, 605.

106

The Magistri Militum from 364 to 395 CE

militum with two untried and weak emperors in the west. The legal reforms only reflected this power. Despite these attempts at patronage and placation, Gratian had a difficult time shedding a perceived weakness that stemmed from his promotion as a child-emperor, and he angered the Gallic armies through an apparent favoritism of Alan mercenaries.224 These mercenaries may have been an attempt to create a military powerbase loyal to him rather than Merobaudes, but before any results could be realized from this, the soldiers rebelled under the banner of an officer in Britain named Magnus Maximus in early 383.225 Maximus crossed onto the continent in the summer, and Gratian moved north to confront him.226 They skirmished near Paris for five days, after which Gratian’s entire army defected to Maximus.227 Gratian fled but was soon captured and executed on 25 August 383.228 By summer 384 an unofficial armistice with a tripartite division of the Empire began to form. The magistri militum of each emperor poses unique questions and analyses that will be discussed below. The most difficult problems arise from Magnus Maximus’ magistri. Because his reign was de-legitimized and denounced after his death, none of his laws are explicitly contained in the Codex Theodosianus, although two may have been included under the names of the other emperors.229 Furthermore, none of the histories treat him with the detail the other emperors received. It is thus difficult ascertain much information about his magistri. Andragathius was the officer who captured and executed Gratian, and he would continue to be one of the only named generals from Maximus’ reign, so he was probably a magister militum. In 388, we encounter two other generals in Gaul when the Franks attacked near Cologne.230 Nannienus and Quintinus were both given the title militaris magister by Sulpicius Alexander, and their role in Gaul supports interpreting these two as magistri. It is unclear why Maximus decided Gaul required two magistri when it had been under the control of one since 373, and after he was overthrown, Gaul was placed once again under the control of a single general. The role of Merobaudes after the defeat of Gratian is also unclear. Because of his position in command of the Gallic soldiers, he must have been present at Paris. Prosper of Aquitaine records that he betrayed Gratian and if so, given his position, he was probably instru-

224 Zos. 4.35.2–3; McEvoy 2013, 84–85. 225 Pan. Lat. 12.23.3, 38.2; Prosp. Tiro s. a. 384; Soc. 5.2.2; Soz. 7.13.1; Philost. HE 10.5.8–9; Zos. 4.35.3–4; Epit. 47.7; Eun. Hist. fr. 55; Claud. de IV cons. Hon. 73; Ruf. HE 11.14; Oros. 7.34.9. 226 Errington 2006, 32; McLynn 1994, 154. 227 Zos. 4.35.5. 228 Zos. 4.35.5–6; Ruf. HE 2.14; Soz. 7.13.8 ff.; Soc. 5.11.7; Zon. 13.17. 229 Honoré 1998, 187. 230 This entire episode is found in Greg. Tur. HF 2.9.

Civil War 383–388

107

mental in prompting the desertion of the army.231 If we follow this account, then we can assume that Merobaudes was content enough with his powerful position in Gratian’s regime, but once it seemed that Gratian might lose to the usurper, Merobaudes was quick to abandon the young emperor, again placing his own interests first. However, Pacatus testifies that Maximus ordered the murder of the officer Vallio, and that he also forced Merobaudes to commit suicide.232 Some historians have taken this version to mean Merobaudes did not join Maximus, but was captured like Gratian.233 However, it is possible to reconcile Pacatus with Prosper by theorizing that Merobaudes deserted Gratian for Maximus, who then forced the magister to commit suicide at a later date.234 If such an interpretation is true, then Maximus seems to have been able to exert the same level of control over his magistri that earlier emperors had done, indicating that it was not inevitable that western magistri would become extremely powerful, and strong emperors could still keep their ambition in check. Merobaudes was clearly unable to transfer his dominance over Gratian to also dominate the more military experienced Maximus, even though Maximus was only a middle-ranking officer at the time of his usurpation. While Gratian was still alive, Valentinian had only been a ‘phantom emperor’ without any real power, but after Gratian’s death he underwent a ‘real’ accession and became technically the senior emperor, having been raised more than three years before Theodosius.235 However, he faced difficulties in truly gaining the imperial authority of someone who came to the throne as an accomplished adult. In his government in Italy, the magister Bauto quickly became one of the leading men. Historians, however,

231 Prosp. Tiro s. a. 384. For historians who take Prosper’s version, see: Enßlin 1931a, 134; Vetter 1960; Matthews 1975, 173; Rodgers 1981; Nixon and Rodgers 1994, 477–78 n. 79; Bagnall, et al. 1987, 650– 52; Potter 2004. 232 Pan. Lat. 12.28.4–5. Vallio is probably the most difficult potential magister militum to include or exclude in this study. Pacatus, in Pan. Lat. 12.28.4–5, gives Vallio high praise, saying he had celebrated a Triumph. However, Triumphs were a privilege reserved for the emperors alone, and therefore it is unclear exactly what Pacatus means by this and it leaves Vallio’s rank in the military uncertain. Amb. Ep. 24.11 emphasizes that Vallio was a warrior of high repute and loyalty, but he only calls him a comes. Thus, the sources are not clear, and we once again must rely on contextual evidence. It was uncommon for magistri to be stationed in Britain during the fourth century. Some had been sent there during the unrest of 360 and 367, and it is possible that Gratian had heard the stirrings of the rebellion which would raise Maximus to the purple, and so he may have sent a magister there to deal with it in the early 380s. However, a comes, the title Ambrose gives Vallio, is a more common rank found in Britain (Not. Dig. Occ. 18, 19), and furthermore, we have seen that Merobaudes tended to restrict the magisterium to himself alone during this period. Thus, we should conclude that it is unlikely that Vallio was a magister militum (Demandt 1970, 603). 233 For proponents of this theory, see: Demandt 1970, 598–99, 606; Waas 1971, 97; Barnes 1975, 159–60. Barnes is the only academic to propose a reason to dismiss Prosper, on the basis that he confused Merobaudes with Andragathius. For the problems and dismissal of this argument, see: Rodgers 1981, 94–95; Waas 1971, 98; Bagnall, et al. 1987, 650–2. 234 PLRE I, 599; Rodgers 1981, 104; O’Flynn 1983, 4. 235 McEvoy 2013, 50.

108

The Magistri Militum from 364 to 395 CE

have had mixed interpretations of his actual importance. Some have said that he was the “mainstay” of the regime, and he dominated all affairs, even those unrelated to military matters, while other works on the period have ignored him almost entirely.236 The belief that Bauto was so powerful possibly stems from the works of Ambrose, who describes Bauto as a man of the “most distinguished dignity as chief of the military,” and he also writes that Maximus complained Bauto was trying to form his own personal kingdom by controlling Valentinian.237 Additionally, Bauto was the only non-imperial family member of Valentinian’s government to be made a consul.238 These details would seem to indicate that Bauto was quite powerful, but deeper inquiry reveals that his power was challenged by other ministers and grandees in Italy. Valentinian’s first nominee for the consulship in 385 was actually the Praetorian Prefect Vettius Agorius Praetextatus, but he died before taking the office.239 Bauto was thus only a second choice. Additionally, his name was not attached to any of the laws that came out of Valentinian’s court during this time. This weakness may have arisen from the fact that a magister militum’s power stemmed from the army, but the Italian garrison had become extremely weak.240 Most of the army had defected to Maximus in 383, leaving only light garrisons and whatever forces Bauto could recruit, which was a difficult task in the late fourth century.241 Thus Bauto was forced to rely upon barbarian soldiers recruited from the Huns and Alans to repulse a Iuthungi attack on Raetia.242 He later used the same soldiers to harass Maximus in Gaul, and might have used them to fortify the Alpine passes.243 The ongoing struggle over the Altar of Victory and other religious privileges also highlights the limits of Bauto’s power. This Altar, placed in the curia, the Senate meeting chamber, by emperor Augustus (27 BCE–14 CE), had been removed by Gratian in 382 as part of his pro-Christian agenda.244 In 384, the polytheist praefectus urbanus Romae, Quintus Aurelius Symmachus, petitioned for the restoration of the Altar on behalf of

236 For those that prioritize Bauto, see: O’Flynn 1983, 6–7; Demandt 1970, 607–8; McEvoy 2013, 67, 129; McLynn 1994, 159; Bloch 1945, 214; Bleckmann 2009, 219. For those that ignore him, see: Matthews 1975, 188; Williams and Friell 1994, 39; Kulikowski 2019, 92. The situation might be more clear if Augustine’s panegyric to Bauto had survived, which is referenced in Aug. c. Lit. Petil. 25.30. 237 “amplissimus honore magisterii militaris” Amb. Ep. 57.3, 24.4. 238 In 385: CTh 1.23.5; ICUR n. s. 7 17489 = ILCV 4258B; P. Lips. 62 2.17; Bagnall, et al. 1987, 304–5. 239 Symm. Rel. 3.12; Jer. Ep. 23.23. 240 Maximus was aware that Valentinian was in a much weaker position than he (Amb. Ep. 24.3). 241 For the most expansive work on the late Empire’s endemic recruitment problems, see Boak 1955, although the theory of manpower shortage has been doubted by Elton 1996, 154–56. 242 Amb. Ep. 18.21, 24.8. Amb. Ep. 20.12 also indicates that Justina employed soldiers recruited from amongst the Goths as a personal retinue. 243 Amb. Ep. 24.7–8. 244 According to Symm. Rel. 3.7, Constantius had already removed the Altar, but it had evidently been restored, possibly under Julian (Cameron 2011, 33–34). Gratian also declined the traditional imperial title of pontifex maximus, the chief priest of Roman Polytheism (Zos. 4.36.5, cf. Cameron 2011), and he defunded the polytheist cults and Vestal Virgins, which may have been the most important issue for the Roman Polytheists (Symm. Rel. 3.15, Cameron 2011, 98 ff.).

Civil War 383–388

109

the city of Rome.245 His petition was stridently opposed by Ambrose, who composed two counter-letters and delivered them at court in the presence of Bauto and Rumoridus, another magister militum in Italy, although one of significantly less importance than Bauto.246 Ambrose convinced the emperor to deny Symmachus, and the two generals were forced to acquiesce despite their own polytheist beliefs.247 This may have been a calculated move by Bauto in order to win favor with the Christians in government and allow the emperor to have control over these ceremonial affairs, or it may have been a decision he could not contravene.248 In the end, while Bauto was undoubtedly one of the most important men in Valentinian’s government, and possibly one of the few men able to actively preserve the reign from an invasion by Maximus, his power was not unconditional and seems to have remained mostly within the bounds of the military. Bauto died before summer 387, and he was replaced by his long-time deputy, Arbogast. Eunapius records that Arbogast’s bravery and contempt for wealth had earned him the loyalty of the soldiers, and he was able to assume the position of dominant magister militum without the assent of Valentinian.249 However, while an important demonstration of the power of the military, the sources do not specify if Valentinian vehemently resisted the appointment of such an experienced general. Thus, this might not be quite as an important moment as it has been made out to be.250

245 Symm. Rel. 3; Amb. Ep. 57. 246 Amb. Ep. 17, 18. Amb. Ep. 57.3 records that Rumoridus was of the “same rank” (“eiusdem”) as Bauto, which implies he was also a magister militum, although it would seem he was not as important as Bauto as he had no further recordings in the fourth-century, other than maybe some undated brick-stamps from Thrace (Bull. Comm. 1942, app. p. 140). He reappears as consul in 403 (Bagnall, et al. 1987, 340–41). 247 There is some uncertainty regarding Bauto’s religious affiliation. Amb. Ep. 57.3 is discussing Bauto and Rumoridus, but uses only singular terminology that only explicitly identifies Rumoridus as a polytheist. Demandt (1970, 607–8), Rapisarda (1972, 27 n. 22), and Ruggini (1979, 10 n. 13) believe that Bauto was also a polytheist, and that Ambrose’s phrasing is either a mistake in the manuscripts or a particular choice to emphasize Rumoridus’ religion for an unknown reason. Alan Cameron (2011, 85, 104 n. 20) however, argues that we should not correct Ambrose in this fashion and instead consider Bauto to have been a Christian, because he believes that Arbogast was a Christian. Arbogast is recorded as a polytheist by both Orosius (7.35.12) and Paulinus (V. Amb. 26); however, because the former used the latter as a source, we only need to consider Paulinus. He writes that it was Arbogast and the Praetorian Prefect Flavianus who urged emperor Eugenius to restore the Altar of Victory. Although Paulinus is generally considered trustworthy, Cameron disputes him on the basis that Arbogast is said to have been friends with the bishop Ambrose and therefore must have been Christian (Paul. V. Amb. 30). However, throughout his work, Cameron argues that Christians and polytheists intermingled freely in the western court of the late Roman Empire, and that there were no sharp divisions. Thus, by his own logic, it should not come as a surprise that Arbogast, even if he was a polytheist, was friends with Ambrose. Therefore, if we doubt Cameron’s assertion that Arbogast was a Christian, we have no reason to follow his argument that Bauto was a Christian. 248 See pages 114–116 for more on the developing religious role of the emperors. 249 Eun. Hist. fr. 58; Zos. 4.53.1; Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). 250 Ie. Demandt 1970, 608–10.

110

The Magistri Militum from 364 to 395 CE

The final magistri to be examined during this period of détente are those under Theodosius in the east. After a brief period of stability following the Gothic peace, the eastern Empire faced renewed military pressure, including negotiations with a new Persian king, tensions in Armenia, another Arab revolt, a barbarian invasion, and an expedition to retrieve Gratian’s body from Maximus.251 In the face of these concerns, Theodosius again expanded his magisterium. The first addition was Ellebichus, who seems to have appeared in the east from 384.252 He was an important magister militum and aristocrat in the east who maintained regular correspondence with notable easterners and built a mansion and baths in Antioch.253 He was also invited to give his opinion in the Constantinopolitan Senate, the first barbarian to do so.254 In 386 a law carries the name Timasius with the title comes et magister equitum, and this man travelled to Valentinian’s regime in Italy where he presided over two legal cases on behalf of Theodosius.255 The third new appointee, Promotus, fended-off an invasion by the Greuthungi in a major river battle in 386.256 He is never explicitly given a Latin title, but he was one of the most important eastern generals, and we can be confident he was a magister militum. One more magister, Butherichus, appears in the east around this time.257 Although we do not know his exact title, when he was killed in a riot in Thessalonica, emperor Theodosius ordered an unprecedented massacre of the city’s civilians. The information on this event is imprecise and even contradictory at times. Stanislav Doležal lays out a lot of the discrepancies and plain misstatements offered by many modern scholars on the story of Butherichus’ death.258 Doležal’s argument that the massacre was a public execution gone awry is mostly based on conjecture and is not entirely convincing, but without more to go on it might be as likely as any other conclusion. In the end, such an extreme reaction to an individual’s death suggests that person was of particular importance to Theodosius, and in addition to the fact that Butherichus was probably assigned the important task of suppressing the raiding occurring in the area, lends itself to the assumption he was high-ranking, probably a magister.259 Finally, Theodosius retained Richomeres, who was awarded the consulship of 384, indicating he was probably the most privileged general, which brings the east’s total to five mag251 Persian negotiations: Oros. 8.34.8; Them. Or. 16.212d–213a; Epit. 48.5; Marc. com. s. a. 384; Pan. Lat. 12.22.4–5. Armenia: Blockley 1987, 230–34; 1992, 42–44; Gutmann 1991, 229–32; Garsoian 1999, 45–46. The Arab revolt: Ball 2016, 105–7. The Greuthungi invasion: Zos. 4.35.1, 4.38–39; cf. Ridley 2006, 88 n. 109. The expedition: Them. Or. 18 220d–221a; Matthews 1975, 178. 252 CTh 9.39.1. 253 Lib. Or. 1.232, Ep. 898; Greg. Naz. Ep. 225; cf. Demandt 1970, 711, 718. 254 Lib. Ep. 925; Waas 1971, 33. On his ethnicity, see the Chapter 4.3. 255 CTh 4.17.5; Symm. Ep. 3.72–73. 256 Zos. 4.35.1, 4.38–39. Petronius Probus was mistaken for Promotus in this incident by Williams and Friell 1994, 43. 257 Soz 7.25.3; Ruf. HE 2.18. 258 Doležal 2014. 259 Zos. 4.45.3; Heather 1991, 184; McLynn 1994, 315–30.

Civil War 392–394: The Rise and Fall of Arbogast

111

istri.260 This is the first time we can confidently place the correct number of magistri in the appropriate locations as contained in the eastern list of the Notitia Dignitatum, and thus that source is best used during this period for the eastern Empire. Eventually, Maximus broke the stalemate and invaded Italy in 387, but Valentinian had enough warning to flee to Thessalonica and plead with Theodosius to assist him retake his territories.261 He agreed, and in 388 Theodosius marched on the west in order to restore Valentinian to the western throne, bringing the magistri Timasius, Promotus, Richomeres, and Arbogast as his commanders.262 Valentinian did not accompany the army, but travelled back to Italy via the Adriatic Sea, so Maximus created a naval fleet and assigned Andragathius to stop him.263 This is the only fourth century example of a magister militum being given a specific seafaring command.264 On land, Theodosius drove Maximus back into Aquileia where he was eventually captured and executed.265 After learning that Maximus had been defeated, Andragathius is reported to have leapt overboard and drowned himself rather than suffer the same fate as his emperor.266 The roles of the magistri militum during the civil war of the 380s makes it clear that the nature of the office of magister militum diverged under different emperors. Merobaudes had established a powerful position in Gratian’s regime, but the record that he was forced to commit suicide by Magnus Maximus indicates this militarily capable and experienced emperor was able to control his magistri effectively. Similarly, Theodosius seemed to face no resistance from his magistri, and was able to easily increase the number of eastern generals, something that Gratian had failed to do while under the control of Merobaudes. Conversely, Bauto and then Arbogast, as well as other important individuals like Ambrose, were able to establish firm control over Valentinian II, even though he was rightfully the senior legitimate emperor. 2.6 Civil War 392–394: The Rise and Fall of Arbogast The next major civil war under examination here was preceded by a new height of magisterial power. Although Valentinian was restored as ruler of the west, the magister militum Arbogast was able to exploit the young emperor’s weaknesses to concentrate a

260 CTh 13.1.12; ICUR n. s. VII 17486; P. Lips. 62 i.24; Soc. 5.12. 261 Zos. 4.42.1. 262 Zos. 4.45.2; Philost. HE 10.8. Interestingly, Zosimus says that Timasius was in charge of the infantry, even though the Codex Theodosianus lists him as magister equitum (CTh 4.17.5). This is one of the few examples of contradiction between sources. 263 Zos. 4.46.1–2; Amb. Ep. 40.23. 264 Demandt 1970, 606. 265 Pan. Lat. 12.34–36; Amb. Ep. 40.22; Zos. 4.46.3–4. 266 Zos. 4.47.1; Soc. 5.14.2; Soz. 7.14.6; Oros. 7.35.3–5; Marc. com. s. a. 388; Claud. de IV cons. Hon. 91–93; Amb. Ep. 40.22; cf. Joh. Ant. fr. 186 (Müller) = fr. 211 (Mariev); contra Zon. 8.18.

112

The Magistri Militum from 364 to 395 CE

massive degree of power into his own hands. He dominated the western government, which was made clear to the imperial court when he publicly refused a direct order of dismissal from Valentinian. After Valentinian’s suspicious death, Arbogast held onto his preeminent position by raising Eugenius as an emperor. However, this bold move would trigger the civil war that would ultimately bring about Arbogast’s death. At the same time, it is possible to discern a distinctly different development in the eastern Empire, as command of the armies became divided amongst five separate magistri, and the civil official Rufinus was able to overshadow them all and prevent them from gaining too much power. After he defeated Maximus, Theodosius resided in Milan for the next three years to consolidate the western Empire. Valentinian was sent to the distant frontier city of Trier, while Theodosius remained in Milan to appoint a new western government. This might not seem unusual, since Trier had been a regular imperial residence since 293. Yet, Valentinian was seventeen years old by 388, an appropriate age to begin having independent agency in imperial affairs. If Theodosius was truly acting as a de facto guardian over his technically senior, but in reality vastly junior imperial colleague, he would have encouraged Valentinian to conduct a campaign across the limes against the Alamanni or other barbarian enemies of Rome in order to gain a strong martial reputation and win the loyalty of the soldiers. Yet no such campaign happened. Instead, Valentinian existed meekly and uneventfully on the outskirts of his territory, while Theodosius formed a government that was ultimately tied to him. He gave many important positions, such as some of the Praetorian prefectures, to easterners.267 He additionally won over many westerners, including by giving Virius Nicomachus Flavianus an imperial quaestorship and eventually the prefecture of Italy, Africa, and Illyricum, and Theodosius also pardoned Symmachus for his panegyric of Magnus Maximus and made him consul in 391.268 This vast distribution of patronage indicates that Theodosius was tying the west to himself, rather than the actual western emperor. He also depicted the imperial college in a way that ignored Valentinian’s technical seniority and relegated him to a junior, subservient role.269 On his coinage, Theodosius depicted Valentinian in the same manner as his eldest son, Arcadius, who had only been raised as Augustus in 383.270 Furthermore, the Madrid missorium, a large silver decorative plate created for the tenth anniversary of Theodosius’ reign, depicts Theodosius in the center and

267 McEvoy 2013, 93–5. Trifolius became Praetorian Prefect of Italy (CTh 16.5.15) and Constantianus of Gaul (CTh 15.14.8). 268 Flavianus: Symm. Ep. 3.81, 90; CTh 9.40.13. Symmachus: Symm. Ep. 2.13, 28, 30, 31, 32; Soc. 5.14.6; Lib. Ep. 1004.8; CIL VI 32018. For more loyalty winning favoritism towards westerners, see: Errington 2006, 134–35, 138–39. 269 Matthews 1975; Croke 1976, 236; McLynn 1994; Errington 2006, 37–38, 134–39; McEvoy 2013, 93– 95. 270 RIC 9. xx, 9. xvii.

Civil War 392–394: The Rise and Fall of Arbogast

113

much larger than the other emperors, who are both in junior, subservient positions.271 Theodosius undoubtedly intended that the eastern Empire was to be inherited by Arcadius. Although he had not yet raised his youngest son, Honorius, as even a Caesar, it seems plausible that Theodosius envisioned that the west could also be inherited into his dynasty one day, and so he was minimizing and weakening Valentinian while building support for himself. In 391, Theodosius returned to Constantinople, and we can deduce that a shift in power dynamics occurred in the final decade of the fourth century. During his time with the campaign army in Italy, the magister officiorum Rufinus began increasing his influence by expanding the powers of his office.272 He remained formidable in Constantinople, and he emerged as the preeminent eastern official in late 391. Rufinus insulted the magister Promotus, who struck him in retaliation.273 In revenge, Rufinus convinced Theodosius to transfer Promotus to a position in Thrace, and then arranged for him to be assassinated by a band of mercenaries.274 This episode is reminiscent of the plots Ammianus accused Arbitio and Eusebius of hatching against Ursicinus, with the difference being that Rufinus was actually successful. Afterwards, Rufinus continued his rise through a promotion to Praetorian Prefect, and he was made consul for 392.275 This powerful position allowed Rufinus to exert power over other officials and bureaucrats, including the magistri. The deeds of those magistri during this period are often difficult to reconstruct, but we can still learn some useful details. We saw above that after the Gothic crisis, Theodosius reduced his accompaniment of magistri. This time, however, Theodosius chose to continue employing a high number of magistri throughout the remaining years of his reign. To restore magisterial numbers after the deaths of Butherichus and Promotus, as well as the apparent retirement of Ellebichus, Theodosius added Stilicho, Abundantius, and Addaeus to his veterans Timasius and Richomeres, keeping the total of generals at five.276 Little is known of the actions of Abundantius or Addaeus, but Stilicho’s future fame allows us to know that he commanded soldiers in Thrace and avenged Promotus’ death.277 As the magistri gained the greatest part of their power from the military, dividing the armies amongst five separate generals diluted the power any one of them could hold. Furthermore, the central government was clearly trying to keep the extension of magisterial powers curtailed, as a ruling affirms that magistri were not to judge in cases 271 MacCormack 1981, pl. 55. 272 CTh 10.22.3. 273 Zos. 4.51.1–2. 274 Zos. 4.51.3; Claud. de cons. Stil. 1.94–6, in Ruf. 1.316–17. 275 PLRE I, 778–79. 276 Stilicho: CTh 7.4.18, 7.9.3. Addaeus: CTh 16.8.9, 1.5.10, 7.2. Abundantius: CTh 12.1.128. Timasius: Zos. 4.51.1. Richomeres: Zos. 4.54.2; Lib. Ep. 1024. 277 Claud. de cons. Stil. 1.94–115. Also: Claud. in Ruf. 1.308–22; Laus Serenae 207–9.

114

The Magistri Militum from 364 to 395 CE

involving provincial governors, but authority in such cases was the realm of the Praetorian Prefects.278 Combined with Rufinus’ dominant position, as well as Theodosius’ own military competency, it seems that the power of the eastern magistri was severely limited in the final decade of the fourth century. They were not able to achieve the same level of influence and personal agency which was exerted by western magistri, a trend which would continue into the fifth century. This, combined with other assets like Constantinople’s defensive position and the discovery of a virgin source of gold somewhere in the Caucasus that helped tie the provincial bureaucracy to the imperial center, maintained the eastern Empire’s strength during the fifth and sixth centuries.279 After Theodosius left the west, the now twenty-year old emperor Valentinian nominally held power from Trier, but in reality, Theodosius had left the west in the hands of Arbogast. Theodosius had made the acquaintance of Arbogast during the campaign against Magnus Maximus, and clearly thought he had won over the magister, and that he would be a good, loyal overseer.280 Events quickly showed, however, that Arbogast’s allegiance was to himself and his own position as he moved to establish control over all ‘active’ parts of governing.281 During this time we see the culmination of a fundamental change of the traditional, military image of the Roman emperors to a more religious, ceremonial role. This was done in order to make the reigns of the new child-emperors more palatable. The previous acceptable depiction of an emperor is emblemized in Menander Rhetor’s late third-century writing, in which he identified the important elements that were to comprise the portrayal of a good emperor.282 Having a respected family and an education, as well as being just, pious, and moral were all important, but perhaps the single most crucial aspect was military prowess and strength.283 Of course, not all emperors were exceptional warriors or strategists, but the imperial image was built upon the idea that they could be.284 We can also see the portrayal of some fourth-century adult emperors in this manner. For example, Ammianus praises Constantius II’s use of javelin and bow, and he writes that Julian inspired his soldiers through his brave deeds.285 Aurelius Victor offers his opinion that an emperor should be educated, refined, intelligent, and successful in war.286 Eutropius also idolized Trajan as the perfect emperor because of his military and civic virtues.287 Even when an

278 CTh 1.7.2. 279 Kulikowski 2019, 54. 280 Eun. Hist. fr. 58.1. 281 McEvoy 2013, 117–27. 282 Russell and Wilson 1981, xxxiv–xl. For a discussion of the difficulties that stem from the nature of the kind of works Menander wrote, see: McEvoy 2013, 24–26. 283 Menander 371.29–372.2, 375.5–13, 375.24–376.9. 284 McEvoy 2013, 29. 285 AM, 16.2.11–13, 12.39–41, 21.16.7. 286 Bird 1994, 28. 287 Eutr. 8.2–5.

Civil War 392–394: The Rise and Fall of Arbogast

115

emperor was not militarily experienced or capable, they would often still accompany the army and take nominal command. For example, Elagabalus (218–222) was declared emperor at only fourteen years of age and had had the career of a priest, not a soldier.288 Yet he was still present and seemingly leading his soldiers as they fought for primacy against Macrinus.289 Closer to our time period, we have seen that even though he did not have a military education, emperor Valens actively led his soldiers at every opportunity.290 No one could be fooled, however, into believing that a child-emperor, let alone an infant-emperor, could take to the battlefield and fulfill the military duties of his office. Instead, the idea of youthful promise was emphasized: it was stated that the young Gratian and Valentinian II would be excellent commanders some day in the future.291 In the interim, however, the role of the emperor was redefined to have a new emphasis on Christian piety. Since Constantine I had offered his support to Christianity, Roman emperors had developed a new religious role by convening ecumenical councils, ordaining and removing bishops, and weighing in on theological disputes.292 This role became the one Gratian, and especially Valentinian II, could best exemplify. Personal piety, religious commitment, and the promotion of a united faith became their most valued characteristics, with the likes of bishop Ambrose lauding the imperial virtues of piety and gentleness, and praising Valentinian for being more mild than a lamb.293 The imagery of St. Peter and St. Paul was also invoked to reinforce the “corporate image” of the Theodosian dynasty, stabilizing their monopolization of power through the ideal of concordia.294 Military depictions did not entirely disappear for the emperors, as can be seen on the diptych of Probus, which shows Honorius in armor holding a spear.295 However, Ambrose also wrote that Valentinian’s self-control and meekness protected Italy from the depredations of Magnus Maximus.296 Thus, in quite a stark reversal, imperial passivity became a virtue. The bishop Ambrose is most often identified as one of the primary drivers behind the development of emperors into figureheads with an emphasis on Christian piety, along with other officials and advisors, such as Symmachus or Anicius Auchenius Bassus, or even imperial family members, such as Justina, Valentinian II’s mother.297 It is

288 Arrizabalaga y Prado 2010, 162–205. 289 Herod. 5.4.5, 4.8. 290 See Chapter 2.3. 291 McEvoy 2013, 109–10. 292 McEvoy 2013, 40–43. 293 Amb. De ob. Val. Iun. 74, 79; McEvoy 2013, 117–27. 294 Dal Santo 2015. 295 McEvoy 2013, 18. 296 McEvoy 2013, 128. 297 McEvoy 2013, 119–25; Errington 2006, 200; Testa 2007, 262; McLynn 1994, 151; Potter 2004, 560–61.

116

The Magistri Militum from 364 to 395 CE

also credited to many unnamed “emperor’s advisers”.298 It is not always exactly clear who these advisers are; late Roman courts were densely populated with eunuchs, bureaucrats, and guardsmen who could exert influence on the emperor in a myriad of forms. One group that has often been dismissed or excluded from these discussions are the magistri militum, even though they were some of the most powerful individuals in the late fourth century.299 Significantly, the magistri probably also had the most to gain if the emperors were reduced to ceremonial duties and military command was permanently shifted out of their control. We have already seen how Merobaudes treasonously contravened Gratian’s orders to the Gallic army.300 It thus seems likely that the magistri were deeply involved in removing the emperors from their traditional military role. Although Hugh Elton assumes there was a second magister in the west at this time, there is no evidence that Arbogast shared the military magisterium with anyone else, and instead power was concentrated solely in his hands.301 As we have seen occur during Merobaudes’ career, being the only magister militum allows one to use the full leverage of the military in one’s own favor. This enabled Arbogast to fill the western bureaucracy with his lackeys and loyalists.302 He, not Valentinian, campaigned across the Rhine against the Franks, and the magister even executed a personal friend of Valentinian.303 In an ultimate display of preeminent authority, Arbogast publicly and blatantly refused to step down when Valentinian tried to oust him.304 Unlike Merobaudes, Arbogast never received a consulship to place him ahead of other officials, nor is there evidence of significant legal reforms to the magisterium to grant him more power. Thus his influence did not come from legal or social standing, but from his utilization of the leverage of the armies to win loyalties and distribute patronage. As we have seen throughout the preceding decades, the magistri had served at the whim of the emperors, who had dismissed or had them executed as they saw fit.305 Now it was made 298 McEvoy 2013, 125. 299 For example, in Cameron’s 2011 reassessment of fourth-century polytheism he claims that it was chiefly three civilian aristocrats and the emperor Eugenius who were the strident polytheists during the late fourth century. Cameron 2011, 59, argues that Arbogast was actually a Christian, and that there is “no doubt” he was the denier of the Altar of Victory restoration petition that came to Vienne in 392. For the flaws in Cameron’s argument on Arbogast’s religious affiliations, see note 109n247. 300 See page 95. 301 Elton 2007, 308. 302 Szidat 2010, 203. 303 Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev); Greg. Tur. HF 2.9; Paul. V. Amb. 30; ILS 790. 304 Zos. 4.53.2–3; Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). According to Zosimus, he told Valentinian that because he had not appointed him then he could not remove him. Demandt 1970, 609–10 thinks this is linked with Zos. 4.53.1 and Arbogast was referring to his appointment by the soldiers, but Bleckmann and Stein 2015, 525 point out that it is much more likely a reference to when Theodosius appointed him to oversee Valentinian. 305 i. e. Ursicinus’ dismissal in 360.

Civil War 392–394: The Rise and Fall of Arbogast

117

clear that emperor Valentinian held no such power. Interestingly, as Bauto before him, Arbogast still did not override the western Empire’s Christian convictions: another petition from the Roman Senate to restore state funding of Roman Polytheism came to Valentinian in 391 but it failed to gain any respite.306 As it may have been for Bauto, Arbogast could have purposefully allowed Valentinian to decide religious matters to encourage the emperor’s ceremonial role. In the winter of 391–92, some barbarian raiders appeared in the Alps, and the Italians panicked at their apparent complete lack of defense.307 This may have occurred because Theodosius had commandeered much of the western army into his own eastern forces, and the remainder may have travelled with Arbogast into Gaul.308 The Praetorian Prefect, Virius Nicomachus Flavianus, convinced Ambrose to travel to the imperial court which had moved to Vienne and request military aid, but the bishop hesitated because he did not want to incite conflict with either Theodosius or Arbogast by giving Valentinian an opportunity to command a military operation and win the loyalty of the soldiers. Valentinian, however, learned of this mission and became desperate for the opportunity. The young emperor may have been developing some political acumen, because he told Ambrose that he wished to be baptized by him and him alone, an honor the bishop would not refuse. Around this time, Valentinian also sent letters to Theodosius begging him for aid against Arbogast, and he might have been planning to flee to him, maybe with Ambrose’s help.309 Ambrose eventually departed for Gaul in early 392, but was only halfway there when word came that Valentinian was dead. The sources on the young emperor’s death vary. Some declare that he had committed suicide, while others believe he had been assassinated by order of Arbogast.310 The inconclusiveness of the sources has led to speculation by modern academics, but it is generally believed that Arbogast did not orchestrate Valentinian’s death because he was already in a position of power and he had nothing to gain through the emperor’s death, and he seemingly had no succession plan.311 However, Valentinian was growing increasingly resistant to Arbogast’s control. Both Zosimus and John of Antioch record the vitriolic and hateful exchange that stemmed from Valentinian’s attempt to dismiss 306 Amb. Ep. 57.5; Paul. V. Amb. 26; McLynn 1994, 335; cf. Waas 1971, 30. On Arbogast’s polytheism, see: Paul. V. Amb. 26; Oros. 7.35.12. 307 Amb. De ob. Val. Iun. 24; McLynn 1994, 335–37; cf. Hoffmann 1974, 389. 308 Zos. 4.47.2; Hoffmann 1969–70, 469–519. 309 Amb. De ob. Val. Iun. 24; Zos. 4.53.4; Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). 310 For the verdict of suicide, see: Prosp. Tiro s. a. 392; Cass. s. a. 392. For murder by Arbogast, see: Soc. 5.25; Zos. 4.54.1–4; Cons. Const. s. a. 392; Marc. com. s. a. 392. Philost. HE 10.1 and Oros. 7.35.2 attribute Valentinian’s death to murder but state that it was made to look like suicide (for more on Philostorgius, see Stickler 2011). Ruf. HE 11.31, the most contemporary ancient source, ascribes the death to reasons unknown. The unknown author of the Epitome de Caesaribus reports that Eugenius was responsible (Epit. 48), and finally, Soz. 7.22 records both tales without judgment. 311 Croke 1976, 244; Matthews 1975, 238; Szidat 1979, 490; McLynn 1994, 336–37; Barnes 1990, 165; McEvoy 2013, 97; Kulikowski 2019, 119.

118

The Magistri Militum from 364 to 395 CE

Arbogast, and Philostorgius even records that the emperor tried but failed to draw a sword to strike at the magister.312 Arbogast may have also feared the emperor might escape and return with Theodosius’ eastern army, as he had in 387.313 Thus, while powerful, it hardly seems that Arbogast’s position was comfortable. Furthermore, it seems that the report of suicide originated from Arbogast, and so was perhaps a cover for his actions.314 Therefore he could not quickly implement a succession plan, as this would have ruined his subterfuge. Thus, we cannot confidently rule out murder. While the sources make it very clear that Arbogast dominated Valentinian, they are not so transparent about Arbogast’s relationship with the new emperor he raised, Eugenius. For example, Eugenius himself received the customary oaths of fealty from the Franks, and he would march with the armies in the later civil war as their leader, which is a departure from Arbogast’s overt control over military affairs under Valentinian.315 This has led to some historians assigning Eugenius the larger share of agency and almost ignoring Arbogast’s role.316 However, it is impossible to believe that Arbogast so quickly changed from being a tyrannical and even cruel de facto ruler to an obedient and loyal general under the Augustus he personally chose. Zosimus records that, since they were introduced by Richomeres sometime between 388 and 391, Arbogast and Eugenius had become close and trusted friends.317 Thus a more likely explanation is that Eugenius, because he became emperor as an adult, maintained the appearance of an active role, but they ruled in close concert. Because he had been a grammarian, Eugenius would not be expected to have the strategic mind to command the armies, but as emperor, it was expected for him to be present and nominally lead, and this is the role he duly played, while Arbogast exerted control from behind the scenes.318 In the east, Theodosius prepared for another military expedition, this time to place Honorius on the western throne. One aspect of preparation involved investing a Mauri prince named Gildo with the new office of comes et magister utriusque militiae per Africam in 393.319 Because of its status as a key tax-base for the imperial coffers and a major exporter of food to the rest of the Empire, securing Africa had been a priority for both Constantius II during the civil war against Julian, and Valentinian I against Procopius.320 Theodosius clearly felt the same pressures to secure Africa, and looked for a trusted ally to secure it. Gildo was the son of the Mauri king Nubel, and had fought

312 Philost. HE 11.1. 313 Zos. 4.53.2–3; Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). 314 Bleckmann and Stein 2015, 529–31. 315 Greg. Tur. HF 2.9. On the importance of the oath, see: Lee 2007, 52–53. 316 Ie. Cameron 2011, 5; Szidat 1979, 506. 317 Zos. 4.54. 318 Szidat 1979, 492. 319 CTh 9.7.9. 320 Szidat 2010, 280.

Civil War 392–394: The Rise and Fall of Arbogast

119

with Flavius Theodosius against his own brother, Firmus.321 It seems that he may have been appointed comes in charge of the African garrison in 386, and then as he prepared to campaign against Arbogast and Eugenius, emperor Theodosius embarked upon a strategy to win his lasting loyalty, part of which was done by creating the novel position of comes et magister utriusque militiae per Africam.322 This was not a temporary or vacans position, as those magistri without specific posting were labelled, but the creation of a new regional magisterial appointment.323 This did not work out the way Theodosius hoped, as Gildo sent no assistance to the eastern emperor during the upcoming war, although nor did he help Arbogast and Eugenius.324 As he departed the east to campaign against Arbogast and Eugenius in 394, Theodosius brought Timasius, a veteran from fighting Maximus, along with Stilicho, who had risen to a position of eminence in the east.325 He had married Theodosius’ niece, Serena, and therefore was now a part of the royal family.326 Theodosius planned for Richomeres to also participate, but he died of disease before the campaign began.327 It does not seem he was immediately replaced, which left Addaeus and Abundantius in charge of the eastern Empire’s military affairs.328 Even though Arcadius was not only a full Augustus but an adult, Theodosius still left Rufinus to watch over and protect him. The two armies met at the Frigidus River in early September 394. On the first day, Theodosius purposefully sent his Gothic auxiliaries, recruited from those settled in 384, in a reckless frontal attack that brutally cut their numbers and led to a solid eastern defeat in the initial engagement. Rallying for a second day of battle, a powerful bora wind 321 AM 29.5.2, 5.6. 322 Demandt 1970, 719; Meier 2020, 248–49. 323 For vacans appointments, see note 83n53. 324 Claud. de bello Gild. 246–47, de VI cons. Hon. 104–5, 108–10. 325 Zos. 4.57.2 also says that in command of Theodosius’ barbarian allies were Gainas, Saul, and Bacurius. Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev) only records Gainas and Saul as barbarian commanders. These men are potentially magistri, but unlikely to be so. Of Saul, nothing more is known, although Zos. 4.58.2–3 indicates he survived the campaign. Bacurius was identified as a comes domesticorum by Ruf. HE 1.11, and it would be entirely possible for him to be brought on the campaign and be praised by Lib. Ep. 1060 in that rank rather than as a magister militum. The best method of deduction is an understanding of Gainas’ career. He is ranked as a comes by Jord. Rom. 319, 320, as well as Marc. com. s. a. 395 and 399, and in 395 he took orders from the magister militum Stilicho, indicating he was subservient and a lower rank (Zos. 5.7.4–6; Philost. HE 11.8; Joh. Ant. fr. 190 (Müller) = fr. 215 (Mariev), Marc. com. s. a. 395; Jord. Rom. 319). Additionally, in 399 he was frustrated at not receiving more honors (Zos. 5.13.1–2, 17.4; Eun. Hist. fr. 67.10; Marc. com. s. a. 399), and he was promoted to the magisterium later that year (Soz. 8.4.5; cf. Theod. HE 5.32.1). All these things indicate that Gainas was not a magister militum in 394, and if he was equivalent in rank to Saul and Bacurius, then, in the absence of any evidence to the contrary, we can assume that none of these barbarian commanders were magistri during the 394 campaign. 326 Zos. 4.57.2. 327 Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev); Zos. 4.55.2–3. 328 Seeck and Demandt theorized that Moderatus may have also been a magister militum in the east around this time, but it is more likely that he was only a tribunus (Seeck 1906b, 213; Demandt 1970, 711; PLRE I, 605).

120

The Magistri Militum from 364 to 395 CE

blew into the western army’s lines, both blinding them and causing their missiles to be ineffectual.329 At the same time, it would have carried the eastern force’s missiles further and faster, and given them a greater tactical maneuverability. These advantages allowed the east to crushingly defeat the western army, and Eugenius was captured and executed.330 Arbogast fled to the mountains where he committed suicide two days later.331 Theodosius once again proceeded to Italy to assert his control over the west. He intended to leave the ten-year old Honorius as Augustus there when he returned to Constantinople, and to oversee the young emperor he officially appointed Stilicho as a guardian and the controlling authority.332 It has been theorized that it was at this point that Stilicho convinced Theodosius to leave him as the only western magister.333 However, as we have seen, the west had been dominated by one powerful magister since Merobaudes, so Theodosius’ decision was not new, and instead it was just a maintenance of the status quo. Thus, despite Arbogast dominating and perhaps murdering Valentinian from a similar position, Theodosius still felt that it was magistri who could safeguard the western Empire. If Theodosius had doubts about whether this was the best course of action, they may have been assuaged by Stilicho’s marriage into the imperial family, that he was personally well known to Theodosius himself, and he may have been prohibited from becoming emperor because of his barbarian heritage.334 Not long after he made this decision, however, and well before he could ensure the stability of his youngest son’s reign, Theodosius fell fatally ill.335 After his death, ten-year old Honorius and eighteen-year old Arcadius were left as emperors, but the Empire was truly in the hands of their controllers, Stilicho and Rufinus. * In this chapter, we have witnessed an increasing concentration of power into the hands of the magistri. After the disaster of Julian’s invasion of Persia, emperor Valens initially

329 Oros. 7.35.17; Soz. 7.24; Ruf. HE 2.33. Zos. 4.58.4 once again chalks this victory up to the over-eating and laziness of the losing side. 330 Zos. 4.58.5; Oros. 7.35.19; Soz. 7.24; Ruf. HE 2.33; Philost. HE 11.2. 331 Zos. 4.58.6; Oros. 7.35.19; Soz. 7.24; Philost. HE 11.2; Claud. III cons. Hon. 102–5, de IV cons. Hon. 91–93; Sid. Ap. Carm. 5.354–56. 332 Zos. 4.59.1, 5.1.1–4; Oros. 7.37.1; Joh. Ant. fr. 188, 190 (Müller) = fr. 213, 215 (Mariev); Eun. Hist. fr. 62.1–2; Philost. HE 11.3; Claud. III cons. Hon. 142–62. This was of course a convenient retrospective claim for Stilicho to make after Theodosius’ death, but it seems as if he might be telling the truth and have been given this position over the western Empire before the emperor’s demise: Cameron 1969, 269; McEvoy 2013, 142. 333 Mommsen 1889, 265. 334 The prohibition on barbarians or those with barbarian heritage becoming emperors will be discussed below. 335 Cons. Const. s. a. 379; Chron. Edess. 39; Soc. 6.1.1; Theod. HE 5.25.2.

Civil War 392–394: The Rise and Fall of Arbogast

121

struggled to manage the military pressures in the east while being restricted, or restricting himself, to only two magistri. At the same time, Valentinian raised the official status of the magistri militum and gave them greater license by allowing them to manage the armies and conduct operations at times without his direct oversight. His death in 375 afforded the chance for the magister militum Merobaudes to manipulate the imperial office for his own benefit, and this was the first time the magistri became veritable kingmakers. Soon after, the outbreak of the Gothic crisis saw a massive expansion in the magisterium, with eight magistri operating along the Danube within only four years. The nature of the eastern and western magisterium diverged over the following decade. Theodosius at first reduced the number of eastern magistri, but then expanded their numbers in the face of increasing pressure, including two civil wars with the west. A strong emperor and powerful Praetorian Prefect were able to inhibit the power of these eastern magistri, who had the armies divided between them. Conversely, in the west, a series of magistri gained increasing degrees of power by dominating the weak emperor Valentinian II, culminating in Arbogast’s total control and eventual appointment of a puppet-emperor, Eugenius. The importance of these men was so great that Theodosius appointed Stilicho to serve as a guardian over the young emperor Honorius, setting a strong precedent for the continued domination of the magistri.

Chapter Three The Networks of the Magistri Militum 3.1 Introduction Through globalization, the rise of social media, and ideas like ‘six degrees of separation’, social and professional networks have become an integral part of the modern world. Social prestige is measured by the size of one’s social media following, we constantly observe the prevalence and importance of interpersonal connections all around us, but this is not a modern development. The Roman imperial government was, above all else, run on a system of “kin and correspondence, debts and favors”.1 Managing one’s social network was key to political, legal, and social success in the Greco-Roman World.2 Personally petitioning provincial governors was important, but even more so was pleading a case directly to the emperor. Litigants would travel the breadth of the Empire for their day in court.3 The celebrations devoted to this indicate how important such matters were: for example, three separate towns in Tripolitania (Sabratha, Gigthis, and Lepcis Magna) set up statues of Lucius Aemilius Quintus, permanently memorializing his successful petitioning of the emperor for aid.4 Personal relationships were also the deciding factor in one’s career progression. The structured Republic-era practice of clientela (‘patronage’) between a patronus (‘patron’) and cliens (‘client’) no longer existed in the same form, but commendationes (‘letters of recommendation’) were the equivalent of the modern cover letter, curriculum vitae, and list of references all in one during the late imperial period. The giving of gifts and money played an important part in obtaining office for those with wealth, but this was only secondary to personal connections.5 All administrative and military officials thus had robust networks of contacts and connections that had aided in their professional

1 2 3 4 5

MacMullen 1988, 99. Ruffini 2020, 333. Kelly 2006, 121–31. IRT 111, 588; CIL VIII 27. Kelly 2006, 132–37.

Introduction

123

careers, and who in turn they were obligated to assist. As such, this Chapter will be investigating the impact of military and political leadership through studying the complex networks of relationships the magistri and others formed in the fourth century. At the peak of these networks, for the majority of the fourth century, were the emperors. They were the ultimate distributors of justice, legislation, offices, and wealth. Connections to the emperor were the most sought after, and the emperors were also obligated to listen to leading aristocrats.6 This accrued honor to both the emperor, through the chance to display his power and benevolence, and to the petitioner, who was able to demonstrate that they had the sovereign’s trust. Generosity was thus an imperial virtue.7 Even when not directly petitioned, an emperor might make an appointment on the expectation that it would please a leading aristocrat.8 The child-emperors that rose in the later parts of the period under review here were not able to distribute imperial patronage with the same effectiveness. As individuals like Arbogast subsumed the active parts of governing, the child-emperors could not distribute offices in the same way as their predecessors, and thus the reciprocal relationships were not formed. Instead, the domineering officials created networks of patronage and personal connections in competition with each other. For the magistri militum, they developed diverse networks with emperors, subordinates, bureaucrats of the civil administration, as well as each other. The complex webs of associations, alliances, and enemies that were formed by the magistri played critical roles in many of the events we have seen in previous chapters. Those who were able to establish relationships with other important figures in the imperial administration were able to further their own positions and that of their associates, while those who did not create such coalitions found themselves excluded from power, or even killed. To understand and analyze this phenomenon, we can employ the framework of social network analysis. This system was developed during the mid-twentieth century by psychologists, social scientists, and mathematicians, with a breakthrough occurring from the block model analysis established at Harvard University in the 1960s.9 It is a method of mathematically modelling relationships in wide-ranging areas, such as occupational mobility, web interconnectivity, innovation diffusion and adoption, group problem solving, and many others.10 These diverse applications are unified by the idea that the structure of relationships – the network – functions as an “explanatory prism” for social actions and events.11 Network analysis relies on the idea that social relationships are one of the foundational blocks of human interactions, and that most

6 7 8 9 10 11

Salzman 2002, 191–92. Pan. Lat. 12.16.2–3. AM 21.12.24. Broekaert et al. 2020, iv. Wasserman and Faust 1994, 5–6; Scott 2000, 3–37. Broekaert et al. 2020, iv.

124

The Networks of the Magistri Militum

historical events are considerably influenced by social structures.12 The methodology can measure a person’s position relative to other in the network – whether they were central or on the periphery.13 Network analysis also has a deep theoretical side which can be applied to the present study. For example, in 1962 Richard Emerson changed the way authority was defined in the social and political sciences by positing that power is not a property of an individual, but instead a function of the relationship between individuals.14 Accordingly, it is less precise to say that a person is powerful, than it is to state the conditions of a relationship that allows one power over another. This does not mean it is wrong to just call someone powerful, but we better comprehend the realities of an individual’s authority when we map their relationships and understand who was able to exert influence over whom. To do this in the present project, we will be using the subfield called sociometry. First established by Jacob Moreno in 1934, sociometry is “the study of positive and negative affective relations … [in] a social network data set consisting of people and measured affective relations between people.”15 This is the best structure for analyzing the types of relationships we will be seeing here. To paraphrase Wim Broekaert, the role of networks in coordinating virtually every aspect of society has made them a primary focal point for the social sciences for decades.16 Recently, historians have shown that network analysis is not merely applicable to sociology or modern history, but it can present all scholars with a new means by which to see long-familiar data. Most extensive dedicated network analysis studies approach a body of evidence or a specific social phenomenon. Giovanni Ruffini analyzes the social networks of Egypt through the papyrus records of Oxyrhynchus and Aphrodito.17 Adam Schor uses network analysis to describe the complex relationships in Roman Syria in the context of the fifth-century Christological debate.18 Diane and Eric Cline demonstrate the ‘Small World’ concept in the Bronze Age Mediterranean by using the Amarna letters.19 Julia Hillner also employs network analysis to investigate the role of imperial women as patrons of banished clerics.20 Additionally, it has recently been shown by David Parnell that network analysis as part of a larger study on military and political matters can help elucidate the nature of high-ranking generals.21 Furthermore, the Journal of Historical Network Research has also began regularly publishing limited forays into the more scientific application of network analysis to

12 13 14 15 16 17 18 19 20 21

Broekaert et al. 2020, v. Cline and Cline 2015, 21. Emerson 1962. Moreno 1934; Wasserman and Faust 1994, 77. Broekaert 2013, 471. Ruffini 2008. Schor 2007; 2011. Cline and Cline 2015. Hillner 2019. Parnell 2017, 103–29.

Introduction

125

history, with the 2020 edition specifically focusing on ‘Ancient Politics and Network Analysis’. It has been noted that politics lends itself particularly well to network research.22 These works have demonstrated the importance of relationships in the imperial political sphere, and they make it clear that we must also strive to understand the role of networks on the impact of the office of magister militum. Social network analysis and sociometry traditionally involve populating a dataset and then applying algebraic equations to establish predictive models – what has been called the “heavy industry” of network analysis.23 Most historians have avoided this technical approach and instead often used network analysis as only a framework, or in a conceptual manner.24 This present chapter straddles a middle-ground by employing network analysis programs and utilizing the ‘heavy’ industry, but with a retention of the overall analytical themes of the work. In the first section, sociometrical concepts and terminologies are employed as tools to better understand the importance of relationships to events of Silvanus’ rebellion in 355. The events that caused the magister militum to usurp power, and the usurpation itself, were not military affairs in the sense of involving armies and campaigning. Instead, they were heavily reliant upon relationships. Assessing the effectiveness of the networks of the different actors involved will help us better understand how relationships affected events in Late Antiquity. The second case study draws more heavily on quantitative and mathematical analyses to assess the networks of the social alliances that appears in the rough decade from the final years of the reign of Valentinian I into the first years of Gratian’s period as senior emperor in the west. These groups have typically been called factions, but we will use the network analysis term of ‘cliques’. They were extremely important in dictating affairs in the west, and recognizing their roles will help us understand the power and influence of the magistri, especially that of the most powerful general, Merobaudes. The centrality factor of different individuals will be explained and analyzed through a sociogram, and we can identify individuals who occupied powerful central positions that could control or exclude others from the network. This approach will lend itself effectively to furthering our knowledge of the developments in the magisterial office. The shortcomings of applying network analysis to the present study mostly arise from the poor state of the sources. This is a problem faced by every scholar in ancient history attempting network analysis, as data will always be fragmented at some level.25 One difficulty is that the extent of unknown relationships is itself unknown – we cannot tell how much information is missing from our sources. Perfunctory connections, relationships ignored by our sources, and the complex dynamism of social interactions

22 23 24 25

Broekaert et al. 2020, iii. Ruffini 2008, 15. Broekaert et al. 2020, ii. Broekaert 2013, 474.

126

The Networks of the Magistri Militum

are all impossible to include in the stationary perspective of network analysis.26 For his study, Schor was able to identify the ‘Antiochene network’ using specific watchwords and expressions in the personal correspondence of fifth century clerics.27 However, we have no letters penned by the fourth-century magistri, and very few directed to them. We must rely on other sources, especially Ammianus Marcellinus. His information is often second or third hand, and even when he was an eyewitness to events, he writes with significant favoritism towards his heroes, and hostility to those he dislikes.28 There will always be an element of uncertainty in historical social network analysis, but such challenges of using our sources should not stop our attempt. Michael Alexander and James Danowski, in one of the first major applications of network analysis to ancient history, similarly used a single source, the letters of Cicero, to study the dynamics between senators and equestrians.29 They concluded that the two groups were closely interlinked. While this was not revelatory, the use of network analysis helped mathematically prove the idea. They acknowledged that this conclusion is inevitably also the perspective of Cicero himself. They argued that, although the synthesis of multiple sources is preferable, if they were to discard Cicero’s pool of information because it offers unparalleled detail that cannot be effectively compared or contrasted to other sources, the most helpful material available to them would be lost.30 We are similarly receiving the perspective of Ammianus, often without the ability to compare it. When using the Res Gestae for other methods of investigation, we critically analyze the author’s biases and perspective to surmount the difficulties in the source. The same processes can be applied when conducting network analysis to assist us in using the source to draw effective conclusions. We must also keep in mind that the networks presented by scholars are not depictions of perfect reality, but are models based on the available evidence which can assist or detract from conclusions drawn through the traditional Quellenkritik methods. Thus, this network analysis should not be conceptualized as true historical reality, but as a model based on Ammianus’ description of networks which can “help us gain further insights into the intriguing societal processes lurking behind narrative and political history which cannot be gleaned from traditional close reading of sources”.31 The conclusions drawn from this chapter will also not be made to stand alone, but will be supported by or used to critique other deductions, and to clarify certain ideas. Hence the network analysis here has been conducted alongside parallel narrative and prosopographical investigations.

26 Cline 2020, 38. 27 Schor 2011, 19–39. 28 See Chapter 1.4. 29 Alexander and Danowski 1990. 30 Rollinger 2020, 12. 31 Nitschke 2020, 275.

Introduction

127

Another difficulty arises from the fact that relationships change over time, and it is often hard to know exactly when these changes occurred. Without evidence to the contrary, we must assume that a serious hostile or positive interaction between two parties is emblematic of their entire relationship, but this is not true in every situation. For example, during his rebellion in the early 350s, Vetranio received the support of many officials and officers, but eventually it seems that his defenses were betrayed, probably by the officer Gomoarius and the Praetorian Prefect Vulcacius Rufinus.32 It is unclear exactly when Vetranio lost the loyalty of these subordinates, but at one point they were inside his network of allies, and then moved outside it to work with Constantius instead. Thus, we must clearly define the time period of each study and acknowledge what changes occurred during it. Because any interaction can be rightfully considered a link between two people, it is necessary to define exactly what this project will consider to be a relationship.33 As already stated, we will be dealing with mostly Ammianus’ recording of how people interacted. The metadata of who is interacting with whom is useful, but we can take it one step further by assessing the nature of the interaction: was it an act of support or hostility, is the relationship professional or familial? For example, according to Theodoret, when Traianus was blamed by Valens for failing to defeat the Goths in 377, Arinthaeus and Victor defended their fellow magister militum and all three blamed Valens’ Arianism.34 This is easy to comprehend by understanding that emperor Valens was being hostile to Traianus (negative relationship), Arinthaeus and Victor were supporting Traianus (positive relationships), and all three magistri were hostile to the emperor (negative relationships). Such a network is displayed in Figure One.

Fig. One The Interactions between Valens and Three Magistri in 377. Own Illustration.

32 Philost. HE 3.24; Bleckmann 1994, 52, 56. 33 Ruffini 2008, 21. 34 Theod. HE 4.33.3. Although, as discussed in note 556, David Woods argues that episode is probably fabricated, it can still be utilized as an example of how magistri could work together in mutual support of one another.

128

The Networks of the Magistri Militum

From this type of graph, we can see that Valens only had hostile, negative relationships, indicating a weak position, while Traianus had the support of his fellow magistri through their supportive, positive relationships. This of course does not prove on its own that Valens was weaker than the magistri, just that, in this moment in time, he was in a less-supported position in this particular social network. Expanding this type of graph to a greater number of people, events, and relationships makes it more correctly indicative of reality, and thus can help us further understand the power and influence different magistri wielded during the fourth century. 3.2 Silvanus’ Rebellion: The Strengths and Weaknesses of Complicated Networks A strong social network is one where the members can trust and rely on each other for assistance. A weak network is one which can be easily fragmented by members betraying each other. The varying strengths of the networks involved in the usurpation of Silvanus in summer 355 were pivotal to determining the outcome of his rebellion, and the way different magistri were able to cultivate networks of allies illustrates their successes and failures. These events are therefore an excellent place for our first network case study. The tale is most thoroughly told by Ammianus Marcellinus, but it is fortunately confirmed by multiple other sources.35 The story goes that an actuarius by the name of Dynamius obtained letters of recommendation from Silvanus under the pretense of using them to make himself look well-connected.36 This was a favor often sought from the magistri as it was from all high officials.37 Instead of using the letters for their intended purpose, however, Dynamius removed all the writing except for Silvanus’ signature, and then forged new letters on the effaced spaces which gave the pretense that Silvanus was planning to declare himself emperor and overthrow Constantius II.38 Ammianus does not record why Dynamius decided to do this, as it is not as if he was in a direct position to gain from Silvanus’ downfall. It has thus been theorized that someone of a higher position, perhaps the magister militum Arbitio, was directing Dynamius.39 Dynamius, being of a relatively low rank, could not get the letters directly into the hands of the emperor, so he gave them to the Praetorian Prefect Lampadius, who passed them on to Constantius.40 Ammianus also writes that Eusebius, a former comes rei privatae, and Aedesius, a former magister officiorum, were co-conspirators in the

35 See page 50. Forgery was not an uncommon problem in Late Antiquity: Corcoran 2014, 199. 36 AM 15.5.3. 37 For example, Lib. Or. 884. 38 AM 15.5.4. 39 Hunt 1999, 48–49; Kulikowski 2016, 297. 40 AM 15.5.5.

Silvanus’ Rebellion: The Strengths and Weaknesses of Complicated Networks

129

scheme.41 Being ex-officers of the court, we can imagine that they relied upon their reputations and associates to help make the plot more believable. Either one, or both, could also be the originator of the plot instead of Dynamius or Arbitio. Whoever was ultimately behind the plot, Constantius believed the evidence of the letters. He then summoned his advisory council, the consistorium. They also thought the letters were legitimate, and orders were given to arrest Silvanus, as well as the tribuni and privati to whom the letters were addressed. But Malarichus, a Frank like Silvanus and the man who would later be offered a position in the military magisterium but would turn it down, and who was at this time a tribunus in the imperial court, spoke up to defend the suspected magister militum.42 Malarichus summoned his collegae, his associates, of which most are unnamed except Mallobaudes, also a tribunus and a Frank. These officers are thus linked to Silvanus not only by their supportive actions, but also by shared ethnicity and career. They protested against Constantius’ orders, and Malarichus requested that he be allowed to go to Silvanus and investigate the truth of the matter.43 The strength of this network and their connection to Silvanus is demonstrated when Malarichus offered his own family as hostages to guarantee he would not betray Constantius. Even so, Constantius refused him, and instead sent one of Arbitio’s associates, the agens in rebus Apodemius.44 We should remember that Ammianus says Arbitio convinced Constantius to send Silvanus to deal with the raids along the Rhine frontier in the hopes he would be too far removed from court to exert any influence, and would maybe even perish.45 Thus, Arbitio and probably Apodemius were not friends of Silvanus. Dynamius continued to thicken the plot by forging a new letter and sending it to the tribunus of the Cremona fabrica.46 This time the letter appeared to come from Silvanus and Malarichus, requesting that the tribunus hastily complete the final preparations for the rebellion. Obviously confused by the unexpected arrival of the letter, the tribunus showed Malarichus the letter and asked for clarification of what exactly he meant. A forged letter now being in his possession, Malarichus was able to once again bring together his collegae to help him confront Constantius with the new evidence. Florentinus, deputy to the magister officiorum, was appointed to investigate the matter and concluded that all the letters were forgeries. The plot uncovered, Constantius looked to punish the true conspirators. Lampadius was arrested and questioned, but was ac-

41 42 43 44 45 46

AM 15.5.4. This is the same Malarichus who would later turn down an offer from Jovian to be made a magister militum. AM 15.5.6–7. AM 15.5.8. For his office, see: AM 14.11.19–23. AM 15.5.2. AM 15.5.9–12.

130

The Networks of the Magistri Militum

quitted because of the aid of his friends. Most surprisingly, the orchestrator of it all, Dynamius, was found innocent and promoted to the rank of corrector.47 Even with his innocence revealed, Silvanus feared that he would still face punishment from the unpredictable emperor. He had been receiving word from his friends about the emperor’s suspicions, and that Apodemius had been seizing his properties in Gaul with impunity, and he did not feel confident that his innocence would protect him. He thought about escaping back to his countrymen, the Franks, but a tribunus in his army, Laniogaisus, told him that the Franks would most likely either kill him or turn him over to the Romans for a ransom. Silvanus, seeing no other viable choice, decided he would truly have to usurp imperial power, and after discussions with the chief officers of his army, he declared himself emperor.48 When this news came to the imperial court at Milan, Ammianus reports that at first there was great panic, and Constantius called a midnight meeting of his consistorium.49 They settled on the idea of using the magister militum Ursicinus to deal with Silvanus. In Chapter 1.4, we discussed the reliability of Ammianus’ account of these events.50 The motivations he puts behind the decisions are most likely fabricated, and his portrayal of Ursicinus and other actors is designed to present his particular point of view. Despite this, there is little reason to doubt the general course of events. Ursicinus was sent to Cologne with Ammianus himself, along with Verinianus, eight other protectores, and some tribuni.51 They were to deliver an official letter to Silvanus, recalling him to court as if nothing were amiss, apparently in the hope that he would simply give up the usurpation. When they arrived, however, Ursicinus figured the rebellion had progressed too far for this to succeed, and he instead decided he would pose as a fellow malcontent, unhappy with Constantius and willing to join the rebellion. This worked, and Ursicinus was taken on as one of Silvanus’ closest confidants.52 While Silvanus complained to Ursicinus about the unfairness of Constantius’ decisions, the protectores and tribuni were able to bribe the Bracchiati and Cornuti companies of soldiers into betraying Silvanus, and they murdered the usurper while he was on his way to a Christian church service.53 After Silvanus’ death, his rebellion rapidly dissipated and those involved were arrested and tried.54 Ursicinus and his subordinates, themselves a small, close-knit network, had thus been able to infiltrate Silvanus’ network furtively and destroy it from the inside by breaking the Bracchiati and Cornuti away from Silvanus. Even though Silvanus’ network was large, covering Cologne, Gaul, and 47 48 49 50 51 52 53 54

AM 15.5.13–14. AM 15.5.15–16. AM 15.5.17–18. See pages 56–57. AM 15.5.21–23. AM 15.5.27–28. AM 15.5.30–31. AM 15.6.

Silvanus’ Rebellion: The Strengths and Weaknesses of Complicated Networks

131

the imperial court, it only took one the fracturing of one small group to bring him down. Furthermore, Silvanus was, as we might expect of a usurper, the linchpin of his network – with his death, the rebellion and network collapsed. As we can see from this description of events in summer 355, relationships were of the utmost importance at all levels of this affair. Every single step of Dynamius’ conspiracy involved contacts and liaisons. He first established an ostensibly friendly connection with Silvanus and received letters of recommendation from him. He relied upon his friendship with the Praetorian Prefect to get the forged documents into the hands of the emperor. Eusebius and Aedesius must have also had their roles to play. Similarly, Silvanus had a massive network of friends, associates, and subordinates. Ammianus tells us that multiple tribuni and privati were targeted in Dynamius’ forged letters.55 Although these individuals were not actually planning a rebellion, they were clearly associated closely enough to Silvanus to be conceivably part of a secret plot. Moreover, two inscriptions have survived in Campania, Italy, which have been interpreted to be acknowledging Silvanus as rightful emperor.56 Bruno Bleckmann believes these may have been set up by some of Silvanus’ high-ranking friends in the Italian peninsula seeking to gain benefits by being quick to acknowledge him as emperor.57 from Constantius’ court, Malarichus, Mallobaudes, and the unnamed other Franks formed a powerful cabal of military officers linked by their shared careers and ethnicity to each other and Silvanus. This Frankish military network in the western court might trace its origins to the extensive recruitment from this group that occurred under Constantine I.58 They might have further shared their Nicene religious affiliations which were in opposition to Constantius’ Orthodoxy, and thus could have already been under suspicion by the emperor. The strength of the bonds in this clique is demonstrated by Malarichus’ willingness to go against the emperor’s suspicions, offer his family as hostages, and even risk being implicated in Silvanus’ fabricated usurpation, all to save Silvanus. The strength of this relationship was in its mutual support, which ultimately absolved Silvanus of the suspicion against his name – although by then it was too late. Silvanus’ network continued to be extensive and important during his usurpation. The military officers in Cologne supported Silvanus’ declaration, and his friends advised him of goings-on in the court and provinces. From Ammianus’ account it seems that Silvanus dallied in Cologne, and he certainly did not achieve much of note. But as proposed in Chapter 1.3, Silvanus seems to have not had the usual planning period that preceded a usurpation.59 What he might have been doing during the twenty-eight days of rebellion is thus establishing relationships with more important individuals in the

55 56 57 58 59

AM 15.5.5. CIL X 6945, 6946. Bleckmann 2000. Kulikowski 2012, 357–58. See page 52.

132

The Networks of the Magistri Militum

provinces, such as the commanders of the fabricae and mints, and other units of soldiers throughout Gaul – much as he thought he was doing with Ursicinus. Ammianus writes that Proculus, a domesticus, and the comites Asclepiodotus, Lutto, and Maudio, were all executed for their participation in Silvanus’ rebellion, but he does not record their specific roles.60 They may have been some of the associates with which Silvanus’ was establishing connections during his twenty-eight-day rebellion. Although Silvanus’ network was extensive, and Malarichus went to extreme lengths to help him, Silvanus’ associates were ultimately untrustworthy and led to his downfall. Silvanus was first betrayed by Dynamius, and then also by Ursicinus and his accompanying officers. They were able to bring the Bracchiati and Cornuti out of Silvanus’ network and into their own. Thus, Silvanus’ network, although large, was insecure, and indicates that he may have had greater success in a smaller, but more reliable group of associates, such as Ursicinus’. The officers under his command did not betray their mission to Silvanus, and Ursicinus was able to spend his time lulling Silvanus into complacency while his subordinates organized the rebel’s assassination. Constantius’ network was also broad and powerful during the summer of 355. We can assess who was part of Constantius’ network on the basis of trust and cooperation. In this regard, the advisors of the consistorium were a part of his network, as was his most favored magister militum, Arbitio. Dynamius, as he was not able to bring documents directly to Constantius, was outside the network. Lampadius is an interesting case, as he was initially trusted enough to bring documents directly to the emperor, but when his involvement in the conspiracy was revealed, he was ejected from the network. Some of the other cases are a little more difficult to determine. We concluded in Chapter 1.4 that, contrary to Ammianus’ portrayal, Constantius and Ursicinus had a normal, trusting relationship. Perhaps being sent to furtively deal with Silvanus was just a case of an officer following orders, but to be trusted with such an important task indicates Ursicinus was probably inside Constantius’ network of confidants. In Chapter One, we concluded that even though the magistri were pushing the limits of their power and exploring avenues of utilizing it, Constantius was able to keep their ambitions largely in check, and network analysis helps demonstrate this point. Constantius was the linchpin through which all actions passed through. Dynamius’ plot centered on convincing Constantius that Silvanus was planning to usurp, and the orders for Silvanus’ arrest and their eventual countermanding came from the emperor. Fear of the emperor’s fickle temper drove Silvanus into rebellion, and when Silvanus’ friends came to his aid, they petitioned directly to Constantius. Although someone in the consistorium proposed the idea of using Ursicinus to deal with the rebel, it was Constantius’ assent that was required. Although how easily Ursicinus formed a relationship with Silvanus demonstrates that connections between officials were abundant,

60

AM. 15.6.

Competitive Cliques and their Leaders

133

the centrality of Constantius amongst these links ensured he remained in a position which enabled him to exert power over all affairs. Although this is not a revelatory insight, much like Alexander and Danowski’s conclusions on senators and equestrians, network analysis helps further support prevailing scholarly opinion. In summation, networks of relationships could be both weakening and strengthening factors in the ancient world. As we see during Silvanus’ rebellion, Dynamius and the network of conspirators abused their relationships with Silvanus, the emperor, and others to form their plot. On the other side, after he declared himself to be an emperor, Silvanus was ultimately betrayed by Ursicinus and some of his own soldiers. But relationships were also helpful. Although it did not avert Silvanus’ fate, the lengths Malarichus was willing to go to in order to save Silvanus’ life are incredible. Constantius, and in turn Ursicinus, were also both able to trust their subordinates in carrying out their duties. Thus, relationships and cooperation were extremely important factors in determining the course of events in the summer of 355. 3.3 Competitive Cliques and their Leaders The reign of Valentinian I saw vigorous and complex maneuvering by many of the western government’s powerful officials, and they were able to manipulate the emperor in pursuit of their own agendas. Much has been written about this period spanning approximately a decade beginning in 368, yet many questions and disagreements remain. Social network analysis is well-suited to furthering our comprehension of this topic, and it will hopefully shed light on this complex period of subterfuge and machination. To this end, this section will take a similar approach to the previous. Brief narrative of the main individuals, with special focus on the magistri militum, will be followed by an analysis of the social networks. This will be supplemented by a sociogram, a useful tool for visualizing networks and calculating the importance of different individuals. Our main source for this period is, once again, Ammianus Marcellinus, with other supporting evidence coming from additional sources. Ammianus is not detailed enough for us to be aware of all the interpersonal connections; he even admits his inability to record everything that happened.61 Furthermore, it seems that he wished to exaggerate Valentinian I’s regime as one dominated by corruption and discrimination, with an eventual triumph of justice.62 Thus, we must again work to overcome the flaws in his work. Upon his election in 364, Valentinian I did not have a powerful support base in the government, so he formed one by promoting men who would support him.63 Some of 61 62 63

AM 29.3.1. Coşkun 2004. Potter 2004, 534.

134

The Networks of the Magistri Militum

these men had been instrumental in smoothing over his election, and some were from Valentinian’s native Pannonia, and they were perhaps favored for these reasons.64 It is clear that Valentinian had to constantly work to retain their loyalty, as we can see that those in the imperial court were not tightly bound to his dynasty during his illness of 367.65 Thus, perhaps hoping to ensure their fidelity, and because of his distraction with waging wars across the frontiers, Valentinian granted his officials a large degree of freedom, and they were regularly able to exert their will on him by manipulating the information he received.66 These factors combined to make Valentinian’s favored bureaucrats very influential. This power reached its zenith during the early 370s when a series of senators in Rome were investigated and tried for crimes such as adultery, poisoning, and magic. A Pannonian, the praefectus annonae Maximinus, was placed in charge of these prosecutions by Valentinian because the city prefect was ill.67 According to Ammianus, this gave Maximinus the opportunity to pour out the “innate cruelty” in his heart, an invective that makes clear Ammianus’ sympathies towards the senators who suffered.68 The prosecutions were extensive and surprising in their severity, even touching the most prestigious families and briefly overturning the ban on torturing senators.69 Maximinus, promoted to vicarius Urbis Romae, was not alone in this investigation, but had the assistance of the notarius Leo, and the new prefect of Rome, Publius Ampelius.70 Maximinus himself was eventually withdrawn to be Praetorian Prefect of Gaul, and he was replaced as vicarius by Ursicinus.71 This new appointee proved too mild in his conduct, so he was soon replaced by another Pannonian, Simplicius, who had previously worked with Maximinus.72 Simplicius remained in charge of the investigations until Maximinus arranged for Doryphorianus to be appointed in late 375, seemingly after the death of Valentinian.73 During these investigations, these men were able to use their positions of power both to favor their allies and to harm their enemies. Maximinus supported his son, Marcellianus, in gaining high office.74 He also acquitted some associates of his friend, Victorinus.75 In 374, Leo plotted against Petronius Probus, the Praetorian Prefect of 64 Ie. Leo and Equitius, but cf. Demandt 1969, 619. 65 See page 80. 66 Potter 2004, 533–52. 67 AM 28.1.9. For his office, see: CTh 14.17.6. 68 “genuinam ferociam” AM 28.1.10. 69 CTh 9.35.1, cf. 9.16.6; AM 28.1.11, 1.24–25; Matthews 1975, 57–58. 70 AM 28.1.12, 1.22. 71 AM 28.1.41, 1.44. 72 AM 28.1.45. 73 Seeing as Simplicius was still in office after 17 November 375 (Coll. Avell. 13.3) then Doryphorianus must have been appointed during Gratian’s time as senior emperor (AM 28.1.53). 74 AM 29.6.3. 75 AM 28.1.27.

Competitive Cliques and their Leaders

135

Italy, Africa, and Illyricum, with hopes of taking his position.76 Maximinus also insulted Probus on a separate occasion.77 Of Maximinus’ insult, Probus was informed by Aginatius, an opponent of Maximinus who had himself previously insulted Victorinus. However, Probus feared Maximinus’ power, and so rather than retaliate, he told him that Aginatius had informed on him. In revenge, Simplicius secured an order for Aginatius’ execution, and Doryphorianus tortured Aginatius’ slaves to extract evidence and then had Aginatius executed.78 Thus, these men participated in a complex web of associates and opponents and were able to work together to further their positions and harm others. Africa also suffered during the reign of Valentinian I, in part because of the actions of its high officials. A man named Romanus had been appointed, probably by Jovian, as comes Africae, and he continued in this position under Valentinian I.79 It was his role to see to the military protection of Africa, so when the people of Lepcis were attacked by the Austoriani, they appealed to Romanus for assistance.80 Altay Coşkun estimated these attacks occurred during the winter of 363–364.81 Romanus came to the city and demanded they provide him the supplies necessary to equip four thousand camels. When the people said they could not, Romanus refused to attack the Austoriani.82 Although this might at first appear to be dereliction of duties, B. H. Warmington points out that the Codex Theodosianus records legislature that places the burden of supply for the frontier forces on the provincials in the area.83 Most likely, such a campaign in the harsh deserts to the south would have been impossible without the camels and supplies, so Romanus had no choice but to refuse. Because of Romanus’ decision, the people of Lepcis appealed to emperor Valentinian for aid sometime during 366 or 367.84 Perhaps similar to Silvanus’ predicament, Romanus clearly felt there was a real danger he would be punished even if his actions were legal.85 He therefore asked his relative, the magister officiorum Remigius, to come to his aid. Remigius was able to ensure the investigation of the affair was entrusted to Romanus and his associate, the vicarius Dracontius.86 As he was now investigating himself, Romanus’ safety was ensured. Eventually, a tribunus and notarius named Palladius was sent to Africa to begin a new investigation. Romanus, however, duped

76 77 78 79 80 81 82 83 84 85 86

AM 30.5.10. For this entire episode, see: AM 28.1.31–33, 54–56. AM 28.1.54–56. Warmington 1956, 58. AM 28.6.2–5. Coşkun 2004, 295. AM 28.6.5–6. Warmington 1956, 56–57, citing CTh 11.1.11. Coşkun 2004, 295. Warmington 1956, 57. AM 28.6.7–8.

136

The Networks of the Magistri Militum

Palladius into embezzling some money.87 This was then used to blackmail Palladius into reporting that all was well with the people of Lepcis, and they had no cause for complaint against Romanus. Thus, Romanus’ position remained secure. Romanus next appears in 372. At an undetermined earlier year, Firmus, an heir to the Mauri kingdom, murdered his brother, Zammac, who was a friend of Romanus.88 In revenge, Romanus besmirched Firmus’ name at court, and Remigius was able to suppress the reports that came in his defense by distracting Valentinian with rumors of impending barbarian invasions.89 These actions drove Firmus to start a war in Roman Africa, and the magister militum Flavius Theodosius was sent by the emperor to suppress it in 373.90 Upon arrival, Theodosius arrested Romanus and seized his property, and also arrested his deputy, the curans Romani vicem Vincentius, for their roles in spurring the rebellion. Palladius was also stripped of his position and sent into retirement.91 After some time, more evidence was found of Palladius’ involvement in disguising the events in Africa, and rather than face his due punishment, he committed suicide. Remigius, who had been replaced at court by Leo, retired to Moguntiacum (Mainz, France).92 However, Maximinus, as Praetorian Prefect of Gaul, eventually reopened investigations into his conduct, and to gain evidence he tortured Caesarius, a notarius of Remigius. Ammianus does not explicitly tie in this investigation of Remigius into his actions involving Romanus, although it is implied to be connected.93 After learning of the investigation and his impending arrest, Remigius also ended his own life, probably in 374.94 This brings us to the death of Valentinian I on 17 November 375. As discussed in Chapter 2.2, Cerealis, Probus, and the magistri Equitius and Merobaudes took the initiative and sent for empress Justina and raised her young son as emperor Valentinian II. There is little evidence that these men had formed an alliance before 375.95 But, even if they were joined by circumstance alone, they formed a powerful group by drawing upon the military authority of Equitius and Merobaudes and the imperial family members Justina and Cerealis. This group also functioned like the others in the way they worked together to promote Valentinian II, a solution that benefitted them all, and they acted against those who might threaten them, as we see Merobaudes isolate Sebastianus.

87 88 89 90 91 92 93 94 95

AM 28.6.17–20. AM 29.5.2. AM 30.8.12. AM 29.5.3–7. AM 28.6.25–27. AM 30.2.10–11. Coşkun 2004, 302–3. AM 28.6.30, 30.2.12; Coşkun 2004, 295. Potter 2004, 535, writes that Equitius was related to Romanus, although there does not seem to be any indication of this in the source materials.

Competitive Cliques and their Leaders

137

Now the senior emperor of the west, young Gratian was initially beholden to those who held power under his father. As Symmachus puts it, the incumbent officials of his father were the “burden of his inheritance”.96 Late in 375, Doryphorianus, an agent of Maximinus, was appointed as vicarius urbis Romae to continue the persecution of the senators.97 Furthermore, in the same year, the rebel Firmus was captured by one of the African tribes allied to Rome, and rather than be turned over to Flavius Theodosius, he committed suicide.98 This successfully concluded Flavius Theodosius’ campaign again him, but the magister would soon be executed in mysterious circumstances. We only know of his death from four rather unsatisfactory sources.99 Jerome and Jordanes imply Flavius Theodosius was killed along with a large number of other aristocrats, although Jordanes makes the unlikely claim that it was because of a new religious law enacted by Valens.100 Ambrose writes that his death resulted from schemes and plots, and Orosius records that whoever killed him was jealous of his successes.101 While it seems clear that Flavius Theodosius’ death was orchestrated by a rival, the perpetrators remain unknown. It has been theorized that Merobaudes was responsible, but this is unlikely due to the congenial relationship he had with emperor Theodosius, the homonymous son of the magister militum, who probably would have rather sought revenge rather than cooperation if it was suspected Merobaudes was behind the death of his father.102 More likely perpetrators are the enemies Flavius Theodosius had made in the civil sector through his campaigns and anticorruption measures in Britain and Africa, including Maximinus.103 Thus, the incumbent power bloc remained dominant, and they were able to continue supporting one another and attacking their enemies for some time after the death of Valentinian I. Once his rule became more stable, however, perhaps once Valentinian II had travelled to his court, Gratian was able to weaken this powerful faction in the western government. Even though he had only recently been appointed, Doryphorianus was removed from office and executed, as were Simplicius and the chief architect of the

“hereditatis onera” Symm. Or. 4.10. AM 28.1.53. AM 29.5.54. Because he was writing his history during the tenure of emperor Theodosius, Ammianus chose not to preserve the story of the father’s downfall for fear of embarrassing the emperor. For further explanation, see: Barnes 1998, 184; Demandt 1969, 599. Ammianus also wrote of Flavius Theodosius in highly favorable terms, comparing him to Domitius Corbulo, just as he did with his hero, Ursicinus (AM 29.5.4). For further discussion on the way this affected Ammianus’ writing, see: Kelly 2007. 100 Jer. Chron. s. a. 376; Jord. Rom. 312; Demandt 1969, 602. 101 Amb. De ob. Theod. 53; Oros. 7.33.7. 102 For those that suspect Merobaudes, see: Tomlin 1973, 527–28; McEvoy 2013, 57; Errington 1996a, 444. For the arguments against Merobaudes’ involvement, see: Rodgers 1981, 82, 89; Demandt 1969. 103 Potter 2004, 544–45; Rodgers 1981, 83–84. 96 97 98 99

138

The Networks of the Magistri Militum

trials, Maximinus. Leo was also dismissed, and others, such as Marcellianus, disappear from the sources.104 Ammianus does not record any specific crimes that these men were charged with, only attributing it to their “intolerable brutality.”105 Thus, we would probably be correct to assume that these deaths were the result of inter-factional fighting and specific criminal charges were not important.106 As these men were cashiered and executed, they were replaced by others loyal to Gratian. Most of these men came from the family of the young emperor’s tutor, Decimius Magnus Ausonius. Clearly the years of working together had made the pupil and tutor extremely close and they had a very trusting relationship, and this endeared Gratian into privileging the Ausonii. We do have to wonder the extent of the influence Ausonius had over Gratian – it would not be surprising for Ausonius to be controlling the inexperienced Gratian from behind the scenes. Ausonius himself became Praetorian Prefect and later a consul, and his son, Decimius Hilarianus Hesperius, was made proconsul of Africa and then also Praetorian Prefect, as was Ausonius’ father, Iulius Ausonius.107 Other individuals given high office by Gratian included Eucherius, Flavius Theodosius’ brother, who became a comes sacrarum largitionum, and another relative of the Theodosian house, Claudius Antonius, became a Praetorian Prefect.108 Through these familial connections, the likelihood of hostility with Maximinus and Romanus, and because Gratian would raise Theodosius fils as emperor, Flavius Theodosius père can be confidently placed in the Ausonius-Gratian group even though he died shortly into Gratian’s reign.109 Other members of the new administration came from those who had lost their positions under Valentinian or had suffered during the senatorial persecutions, including Tarracius Bassus and Furius Maecius Gracchus.110 As a final comfort to the senators, the office from which Maximinus started the persecutions, that of praefectus annonae, was placed under the prefecture of Rome.111 Hesperius and Virius Nicomachus Flavianus were responsible for another investigation into Romanus’ abuses in Africa. They arrested and tortured Romanus’ consiliarius, Caecilius, and found out the true sequence of events.112 Romanus then took Caecilius and sought out a trial at the imperial court in Milan. Ammianus writes they wanted to accuse the judges of having overly favored the provinces, which was possibly true considering the provincial responsibility to supply the army, which was reasserted in

104 Symm. Or. 4.10. 105 “intoleranter efferens” AM 28.1.57. 106 Warmington 1956, 62. 107 Symm. Ep. 1.16; Aus. Praef. 4.35–38; CTh 1.15.8, 6.30.4, 8.5.35, 10.20.10, 15.7.3. 108 CTh 1.32.3, 9.40.12, 13.3.11. 109 Potter 2004, 544. 110 Chastagnol 1962, n. 77, 78, 79; Errington 2006, 122. 111 CTh 14.3.15; ILS 5694. 112 AM 28.6.28.

Competitive Cliques and their Leaders

139

Africa around this time.113 At Milan, they were aided by Romanus’ friend, Merobaudes. André Piganiol attributed Romanus’ acquittal to Merobaudes’ sheer power and influence, but this is not what is recorded by Ammianus.114 He writes that Merobaudes only helped ensure the trial was fair by bringing more witnesses to Milan, who were able to provide enough evidence to have the charges thrown out.115 Thus, this was not a case of Merobaudes flexing his political and military might, but simply helping his friend receive a fair trial. After their trial, Romanus and Caecilius disappear from the sources. Others, such as Probus, disappear for the early years of Gratian’s reign until reappearing later.116 This indicates Gratian removed them from office so he could promote his new powerbase. Of the important officials under Valentinian I, only the magister militum Merobaudes, and the empress Justina, remained in power. Thus, within a few years of Valentinian I’s death, Gratian and his tutor Ausonius displaced those who had held the greatest power under the previous regime, punished those who had gotten away with abusing the provinces, and secured themselves enough powerful supporters to be the only large and powerful clique in the west. The period just described involved many actors and actions across varied geographical spaces, and is a rather complex series of events. To help visualize how all these occurrences relate to each other, we can use a sociogram, such as Figure Two. This will not only help our comprehension, but we can also apply network theories to understand the importance of different individuals during this period. Firstly, we must describe the parameters of the sociogram before deducing anything from it. As described in the narrative section above, some relationships were reciprocal. For example, Maximinus supported and was supported by Publius Ampelius, Leo, Doryphorianus, and Simplicius. Similarly, Ausonius had built a trusting relationship with Gratian, which was returned with the promotion of himself and his family into high office. These ties, as well as familial ties, are denoted with a simple unadorned line. There are other cases, however, of uni-directional ties. For example, Merobaudes relegated Sebastianus to a distant post, and to the best of our knowledge, Sebastianus took no retaliation. Thus, Merobaudes had a hostile relationship with Sebastianus, but it was not provably reciprocal. To show the vector of these unilateral actions, arrowheads point at whom received the attack or support. It bears noting that this is not necessarily a person directly involved. For example, Aginatius informed Probus that Maximinus had insulted him. Aginatius’ intention was to harm Maximinus by inciting Probus against him, so this interaction is marked with an arrowhead directed from Aginatius to Maximinus. Thus, even though he was involved, this interaction is not 113 114 115 116

Inscr. Lat. Tun. 894–97; AM 28.6.29; Warmington 1956, 57. Piganiol, 1947, 184 n. 83. AM 28.6.29–30. Soz. 7.13.1; Soc. 5.11.3.

Fig. Two Ammianus Marcellinus’ Description of the Social Networks of the West from 368–78. Own Illustration.

140 The Networks of the Magistri Militum

Competitive Cliques and their Leaders

141

linked to Probus on the sociogram. Because of these directional ties, this sociogram is known as a ‘directed graph’ or digraph.117 Further information is displayed using color: red for hostile actions, green for support, and yellow for familial ties. It might be noted that Gratian is the only emperor placed in the sociogram. This is because, as portrayed above, early in his reign as senior emperor he made decisive moves to promote his loyal associates and remove those from his father’s regime. Conversely, because Valentinian I was preoccupied with attacks on Britain, the Channel coast, and the Danube, as well as with his campaigns against the Alamanni, he was not very involved in the maneuvering of his officials. Even when documents written by the proconsul of Africa, Iulius Festus Hymetius, were found that insulted Valentinian personally, he deferred the trial to Ampelius and Maximinus.118 The officials, such as Remigius and Palladius, were also able to manipulate the information Valentinian received and keep him ignorant of occurrences. Thus, even when he tried to become involved, it was on incorrect reports, and through manipulation his decrees were specifically targeted by his advisors. Therefore, he has been left off this sociogram. Valentinian II is not depicted because his relationships do not meet the same level of intentional purpose as the other shown actors. Even though he was emperor, as a young child, he could hardly be the originator of any actions. Any that came from him must have derived from his advisors. Merobaudes, Equitius, Probus, Justina, and Cerealis were all involved in Valentinian II’s accession, but as discussed in Chapter 2.2, they did this for their own benefit.119 They are thusly connected with supportive ties to each other, but cannot be shown as connected to the boy in the same manner. For these reasons, Valentinian I and II are not displayed in this sociogram. Speculative relationships have not been included in the sociogram. For example, it seems likely that Maximinus was responsible for Flavius Theodosius’ death. The magister had fought a campaign against Maximinus’ brother-in-law, Valentinus, and Maximinus’ powerful position in the government likely afforded him the opportunity to attack Flavius Theodosius.120 Furthermore, although Ammianus does not discuss the subject explicitly, it has been proposed that there are hidden meanings in some of his passages that would indicate he knew Maximinus was the orchestrator of Flavius Theodosius’ death.121 However, other theories persist and remain plausible. These include blaming Merobaudes or Romanus, or including the involvement of Probus and Marcellianus.122 Because of these other plausible theories, placing a hostile tie between Maximinus and 117 118 119 120 121 122

Distinct from the word meaning two letters making one sound: Wasserman and Faust 1994, 72. AM 28.1.19–23. See pages 84–88. AM 28.3.4–6. Rodgers 1981, 83–84. For blaming Merobaudes, see: McEvoy 2013, 57, following Tomlin 1973, 527–28 and Potter 2004, 544–45. For Romanus, see: Seeck 1913, 31; Enßlin 1934. For Probus and Marcellianus, see: Demandt 1969, 621–22.

142

The Networks of the Magistri Militum

Flavius Theodosius would only increase one speculative element of the sociogram, so it has not been done. The same applies to the proposal that Romanus knew Equitius from previous offices, and that Merobaudes and Equitius were responsible for orchestrating the downfall of Maximinus.123 These potential ties are simply too uncertain. The most significant flaw that must be acknowledged is the simplification of complex relationship down to single, colored lines. Interpersonal relationships are always complex, and they often change over time, an aspect that a sociogram cannot capture. A further difficulty arises from the fact that the evidence for the sociogram is based entirely upon Ammianus Marcellinus’ writing, as no other fourth-century source comes close to offering any similar insights into social networks. Without the ability to cross-reference the evidence from Ammianus, we cannot be entirely confident in every connection, and some might in fact be ahistorical. Although these problems are significant, they should not be fatal. They can be alleviated by not forcing the conclusions drawn here to stand on their own, but using them as additional pillars to support other assumptions and suppositions. Network analysis is not irrefutable, it is an additional tool to employ. With the analytical parameters described, we can apply network theories to the sociogram to understand better the western Roman social system of the 370s and the position of the magistri militum within it. The most evident feature of the sociogram is the appearance of three distinct cliques, which have been marked by black dotted lines.124 The first group appears on the left of the sociogram and comprises the high-ranking officials involved in the Senate trials in Rome. The second group is displayed on the right and incorporates Romanus, Remigius, their associates and aides, and the group that promoted Valentinian II, which includes Merobaudes and Equitius. The third and final clique appears at the bottom of the sociogram with Gratian and the officials that formed the administration in the early years of his regime, as well as Flavius Theodosius. Previously, these groups have been defined on some other basis and seen as opposed against their opposite faction: a senatorial alliance versus a governmental alliance situated in the imperial court, or a Gallic faction opposed to a Pannonian one.125 It has also been proposed that Maximinus received the backing of the military officers from all regions, as well as support from the provincial aristocrats, to attack the senatorial elites in Italy to disrupt their hold over the most prestigious governmental offices.126 Demandt has already pointed out that in the case of the ‘Pannonian’ faction, only a few

123 For Romanus and Equitius, see: Matthews 1975, 36. For Merobaudes and Equitius, see: Sivan 1993b, 129. 124 For other uses of network analysis to assess governmental factions, see: Nitschke 2020. 125 These theories began with Schuurmans 1949, and Alföldi 1952, who are still followed today, by for example, Errington 2006, 114; Bird 2011, xxxiii. 126 Kulikowski 2019, 63.

Competitive Cliques and their Leaders

143

were actually from the region.127 Altay Coşkun also has issues with the identification of a faction from Pannonia at this time, although he does not propose an alternative.128 Thus, even though these shared qualities perhaps endeared some members to each other, it is inaccurate to define the entire group in this way. Instead, we can identify these cliques using network principles. As seen on the sociogram, each clique is characterized by multiple internal positive associations, including acts of support, protection, assistance in securing office, and collaboration for joint goals. Furthermore, no clique contains negative connections between members, indicating exceptional internal harmony. Each of the three cliques also contains familial connections. Moreover, each clique is only connected to the others through hostile actions. It thus seems that each clique worked extremely closely together, and were staunchly opposed to their opponents, with no one crossing the lines between groups. Through network analysis we can see that the cliques, or factions, were starkly separated from the others. Most nodes only have hostility with members from one other clique; the exceptions are Maximinus and Flavius Theodosius, who had hostile relationships with members of both opposing cliques.129 This model of three cliques differs from what has been previously proposed by historians, who have mostly advocated for two factions during Valentinian’s reign. The one constant through these propositions is the Roman senatorial group attacked by Maximinus, and this is visible as the Gratian-Ausonius clique on the sociogram. We learn that the perceived governmental or military-provincial alliance, however, is quite starkly divided into two cliques when we apply network analysis. They had complex relationships with civil authorities but were not fundamentally divided from them. Every clique combined civil and military officers, although Maximinus’ faction was dominated by civil authorities; the magistri militum were only present in the other cliques. Thus, when discussing this period, we should describe a situation where three diverse, but very closely-allied cliques were in stark and brutal opposition to each other. As mentioned above, many relationships and interactions have been proposed in this period by historians, and they have not been included in the sociogram. In addition to the examples listed previously, another is the friendship between Maximinus and Merobaudes that is speculated by E. A. Thompson.130 To the best of our knowledge, they did not actually directly interact with one another and have no discernable relationship. Thus, we should not confidently assert a friendship between them, and there is accordingly no line connection them on the sociogram. There is a possibility that their friendship is concealed to us, and this can be evaluated based on the close-

127 Demandt 1969, 618–19. 128 Coşkun 2004, 298–99. 129 Firmus also had hostile interactions with members from two other cliques, but he himself did not belong to any of the three depicted cliques. 130 Thompson 1947, 99.

144

The Networks of the Magistri Militum

ness of their cliques. Merobaudes is tied to Equitius and Probus, who were both attacked by Maximinus. Maximinus also attacked Remigius, and Leo attacked Probus. Thus, there was a significant degree of hostility between Maximinus’ and Merobaudes’ cliques. As stated in the introduction to this chapter, we cannot know every interaction and relationship in the networks of Late Antiquity, and so we cannot entirely rule out a friendship between Merobaudes and Maximinus based on the sociogram alone. Considering the available evidence, however, it seems that there is a greater likelihood that Merobaudes was hostile to Maximinus, rather than friendly. Another path that we can utilize with the sociogram is mathematical. Network analysis offers multiple formulae for applying value to network nodes and calculating which are more important to the system. ‘Closeness centrality’ calculates how far a given node is from all other nodes in the network, and is a good measure of importance for the presently studied group. Part of this method has been used by previous historians, although not to its full extent.131 Thus the method will be briefly explained g here. Gert Sabidussi provides the formula Cc(ni) = [Σ j=1 d(ni, nj)]-1 where the closeness centrality (Cc) of a given node (ni) equals the sum of the number of connections in the shortest path between it and every other node.132 Put simply, one node is picked, and then the number of ties it takes to link that person to all other individuals is counted, added together, and then that number is put to the power of negative one. Murray Beauchamp suggests standardizing the indices so the result is comparable across different size networks.133 This is done by multiplying Cc(ni) by one less than the total number of nodes in the sociogram. With these calculations, a node’s centrality will have a value between 0 and 1, the higher number indicating a greater degree of connection. In a network with as many connections as ours, short paths can be found between almost any node and all others, so it is expected for the values to be close in range. We will thus be calculating the formula to three decimal points, a common choice in network analyses. Applying this to the individuals in our sociogram, the two highest values are: Maximinus = 0.516, and Romanus = 0.471. We can thus view Maximinus and Romanus as the leaders of their cliques, although we must remember that this is not a measure of absolute connectedness, but a measure of connectedness within our network parameters. Both Maximinus and Romanus are protagonists in Ammianus’ Res Gestae, and the focus upon them contributes to their high centrality. We can also apply the same formula to the magistri militum. Remembering that a higher value indicates a greater degree of connectedness, their centrality values are: Merobaudes = 0.421, Equitius = 0.416, and Flavius Theodosius = 0.405. Recall131 Eg. Schor 2011. 132 Sabidussi 1966. For a lengthier discussion of how to apply this formula, see: Wasserman and Faust 1994, 184–86. 133 Beauchamp 1965.

Competitive Cliques and their Leaders

145

ing the introductory discussion of power as influence exerted through connections, Merobaudes’ greater level of connectivity helps to explain another aspect of his strength and the robust position he occupied in the western government. Another measure of importance in social networks is the significance of a node’s connections. For example, Gratian’s four hostile interactions with members of Maximinus’ clique led to the total destruction of that faction. Thus, Gratian’s connections are extremely significant. Similarly important, Merobaudes is what is known as a ‘critical bridge’.134 If his node were to be removed, his clique would fragment into two separate and much smaller groups. Additionally, Merobaudes is the node which connects Romanus’ side of the clique with the circular subgroup of important individuals along the Danube that raised the young Valentinian II as emperor in November 375. This group comprised three members of the imperial family, two magistri militum, and the Praetorian Prefect Probus. Merobaudes also connects the group to the powerful Danubian army that Valentinian had assembled on the Danube to combat the Sarmatians and Quadi.135 Furthermore, his only hostile action was taken against Sebastianus, an outlier who was seemingly not a part of the cliques, and therefore Merobaudes did not attract anyone’s anger and he faced no repercussions for this act. Thus, even though Romanus has a higher closeness centrality value, Merobaudes’ connections might be more important in practice. Although there is not an extraordinarily large difference between the centrality values for the three magistri, it remains that Equitius and Flavius Theodosius were statistically less central than Merobaudes. Equitius shares several network characteristics with Merobaudes: he was also a magister with the Danube army, and he also helped raise Valentinian II. However, his only other network connection is the defamation directed at him by Maximinus. Similarly, Flavius Theodosius is only connected to Gratian’s clique by virtue of his familial relation to Claudius Antonius and Eucherius. Without any positive connections, his other links are his campaigns against Valentinus and Firmus, as well as his arrest of Romanus, Palladius, and Vincentius. As discussed above, by confronting members of other cliques without establishing a close network of allies, Flavius Theodosius left himself exposed to attack and ultimately his execution. Thus, network analysis shows that Merobaudes occupied the strongest position of the three western magistri, and this aids in clarifying why he survived the regime change of 375–76 when almost everyone else outside of the Gratian-Ausonius clique did not. As members of his clique were arrested, committed suicide, or met other fates, Merobaudes, along with Equitius and Probus, was able to defend himself through the raising of Valentinian II. He was then able to leverage his powerful position into dominant position as sole magister militum of the west during Gratian’s reign, while Equitius

134 Wasserman and Faust 1994, 114. 135 See pages 83–84.

146

The Networks of the Magistri Militum

disappears from the record, most likely into retirement, and Probus was out of office until the final year of Gratian’s life. His survival should not be characterized as a “concession to the old ruling group,” as David Potter puts it, but instead it was the result of utilizing his central, critical, and pro-active position in the contemporary networks.136 3.4 Conclusion Unfortunately, Ammianus’ history does not return to a western perspective after the accession of Valentinian II. Without the detail he provides, network analysis becomes decidedly more difficult to conduct for the western half of the Empire. Contemporary sources are few, and many were written in the east, and thus they can only provide limited insight into the social dynamics of the west. The extensive letters of Symmachus are one contemporary western source, and Schor has demonstrated the potential for personal correspondence to populate a network.137 Schor was, however, able to utilize the regular correspondence of multiple authors, and his study was supplemented with other sources. Most importantly, Schor’s focus was on those authors. With only Symmachus’ letters, a sociogram would inevitably have himself at the center and would offer little information on the magistri militum. Thus, it will not be possible to conduct the same depth of network analysis for generals in the years after 375, although the potential for the reconstruction of Symmachus’ network should be considered for future research endeavors. Despite the paucity of sources, the case studies presented here are by no means the only instances of networks during the timeline under consideration. As we saw in Chapters One and Two, the magistri were not only involved in military affairs, and were not always reliant on the use of military force to achieve their objectives. Networks were of particular importance during such events that did not depend on fighting, including Vetranio’s rebellion, the search for a successor after Julian’s and then Jovian’s deaths, the defeat of Procopius’ rebellion, and many more. It is also clear that the powerful magister militum Arbogast developed a strong network around himself that excluded the emperor Valentinian II.138 Networks were also not unique to the fourth century, but were important features of the Roman military in other periods.139 Network analysis is imperfect. It is difficult to account for the chronological aspect, uncertainties persist when a network cannot be fully populated, and it is heavily reliant upon source material. If we can account for these shortcomings, however, it presents an interesting new path to explore ancient history and it rightfully deserves a place in 136 137 138 139

Potter 2004, 545. Schor 2011. See page 114. Ie. Parnell 2017.

Conclusion

147

the historian’s toolkit. By using network analysis on two fourth-century case studies, we can see that the magistri militum of the fourth century were part of complex social networks that involved both military and civil officers, emperors, privati, and foreigners. Sometimes these networks worked in favor of the magistri, but sometimes they did not. Silvanus’ network, even though it contained devoted individuals such as Malarichus, ended up being the magister’s downfall, as he was betrayed first by Dynamius then by Ursicinus. Under emperor Valentinian I, we saw that three cliques formed, characterized by strong internal positivity and by stark hostile separation to each other. These cliques engaged in brutal and often deadly competition. It was Merobaudes who occupied a stronger, more central position in his clique compared to the positions of the other contemporaneous magistri, Equitius and Flavius Theodosius. It should come as no surprise then that Merobaudes was the only one of the three to survive the regime change in 375–76. Thus, networks played important roles in the careers of some magistri, and disastrous ones in others, and recognizing this phenomenon is crucial for our understanding of fourth-century developments in the magisterial office.

Chapter Four The Prosopography of the Magistri Militum 4.1 Introduction Prosopography is a useful, time-tested tool for historians. It is the inverse of a biographical approach: whereas the biographer examines one particularly unique person and details the important characteristics and events of their life, the prosopographer asks questions of a body of people in order to delineate and understand the norms and averages of the individuals within the group.1 This is not the type of prosopography that might be familiar to the reader from works such as The Prosopography of the Later Roman Empire, which is an attempt to collate and alphabetically list the names of every individual in the later Roman Empire and the ancient sources that attest to them. This will rather be an exploration of different aspects of the nature and character of the magistri by asking questions and compiling answers through statistical, quantitative analyses. Discussions of outliers and smaller groups of magistri with shared characteristics, will combine with the previously established narrative to illuminate further how decisive moments in the fourth century changed the position and power of the magisterium. Prosopography desires several elements to be successful. The first is a well-defined population to study.2 In Chapters One and Two, the evidence was proposed and discussed for each individual that we should consider a magister militum. Most magistri have their title recorded in sources like the Res Gestae or the Codex Theodosianus. The precise rank of some other individuals is more difficult to determine, often if they are recorded by Greek sources, or Latin sources that do not use technical terms. Supplementary evidence, such as the granting of a consulship, has been used to help resolve these uncertainties. Using this method, this project reached the number of fifty-two recorded magistri from their first concrete appearances under the heirs of Constantine to the death of Theodosius I in 395. Thus we have a well-defined population. This is a

1 2

On this method, see: Verboven, Carlier, and Dumolyn 2007. Verboven, Carlier, and Dumolyn 2007, 47.

Introduction

149

lower total than those used previously by scholars, as the dataset used in this study is intended to be conservative rather than widely inclusive. While this approach runs the risk of discounting some individuals who did reach the magisterium, it should not dramatically affect the prosopographical results, as many of the excluded individuals are those about whom the least information is known. The other important element prosopography requires is detailed information about the people in the examined population, and this presents the main difficulty for the present study.3 Prosopographers are often able to utilize a much deeper volume of information than is available for the fourth century. At the beginning of this project, it was my intention to include a section on the religious affiliations of the magistri to chart the impact this had on the careers of the generals, and whether the most powerful individuals tended towards particular associations. Unfortunately, there is very little information in the sources about this topic, even though the beliefs of the magistri must have been known to contemporaries, especially those who knew them personally, such as Ammianus or Libanius. Despite this, only one-third of magistri have an attested religious affiliation. Furthermore, some of these are particularly suspect, such as the denouncement of Sebastianus as a Manichaean.4 Furthermore, it remains unclear the best way to categorize religious affiliations. Rather than simply Christian and pagan, Alan Cameron proposes a breakdown into five categories: committed Christians, center Christians, middle grounders, center pagans, and committed pagans.5 This system, however, has been criticized for being a substitution, not a solution, for the simple dichotomy.6 Thus, such a prosopographical investigation would ultimately be fruitless. Fortunately, more information exists in other areas. We will first be investigating the unique fourth-century career path that led one through the military ranks of the late Roman army to the office of magister militum. Establishing this hierarchy of offices will help us understand the nature of the late Roman army administration and the offices that composed it. A comparison to earlier and later career paths will be established. We will also explore the factors that aided one in becoming a powerful magister. The second topic has been previously visited by scholars, although the two previous studies have either not been comprehensive or have used flawed methodology. We will examine the ethnic identities of the magistri militum by reviewing the previous research and discussing the improvements that can be made to obtain more accurate results. This topic will lead into a discourse regarding the effects of ethnicity in the developments of the magisterium. We will see that a small subset of the most powerful magistri had a number of characteristics in common that may have encouraged them to reach these 3 4 5 6

Verboven, Carlier, and Dumolyn 2007, 46. Dam 1985, 80–87, 101–6; Tardieu 1988. Cameron 2011, 176–77. Rebillard 2013.

150

The Prosopography of the Magistri Militum

heights of powers. The statistics for both questions will be presented with accompanying graphs, and then to conclude any final considerations that can be drawn from the results will be presented. 4.2 The Fourth-Century Career Path and the Hierarchy of Military Offices 4.2.1 Introduction To many Romans, one’s professional career was extraordinarily valued. For historians, career paths are also valued as a way of understanding the structure of bureaucracies, the character of aristocrats, and the oscillations of power. Whether civil, military, or a combination of both, academic descriptions of career structure hierarchies have contributed to a deeper understanding of the Roman world, as well as the officials that stretched from the imperial center across the provincial world. The careers of men in the Republic and early Empire were often recorded epigraphically, and that has enabled the development of extensive studies that cover hundreds of individuals. Unfortunately, the regular practice of erecting these so-called cursus-inscriptions ended in the third century, so the information for the magistri is more fragmentary.7 Furthermore, the small population of only fifty-two magistri makes drawing conclusions difficult. Nonetheless, a study on the Germanic officers of the fourth century has already been conducted by Manfred Waas.8 Expanding our view to consider all the attested magistri of the fourth century will allow us to reveal more conclusively both the hierarchy of different military offices, the normal career progression that brought an individual to the rank of magister militum, and whether these factors changed over time. Analyses of why particular offices were likely to see promotion to the magisterium, as well as why certain individuals received more rapid advancement, will further elucidate the nature of the magistri. Finally, we will be able to assess whether particular career factors influenced a magister’s power. It will be of benefit to situate the fourth-century career path in context of the career paths that predated it, and that which followed it. In Chapter 1.1, we discussed the evolution of the earlier imperial career paths – how senators held the upper echelons of military command during the Principate, and eventually lost these positions to the growing importance of the equestrians during the third century. The equestrian order itself became divided into the separate civilian administrative bureaucracy headed by the Praetorian Prefects, and the military commanded headed by the magistri militum.

7 8

Borg and Witschel 2001. Waas 1971, 5–10.

The Fourth-Century Career Path and the Hierarchy of Military Offices

151

We also saw over the course of Chapter Two that the highest echelons of each strand gained senatorial status, the civilian officers slightly earlier than those from the military. What we will explore in greater detail is the structure and nature of the senatorial and equestrians career paths, and then also the Byzantine path – that is, the eastern Roman Empire which survived the dissolution of the western Empire in the fifth century – to frame the fourth-century path. During the early Empire, an aspiring Roman senator would seek to progress through the cursus honorum. This hierarchy would see one climb through the positions of the public ordinary magistracies of the city of Rome, namely quaestor, aedile, and praetor, and although they had mostly become sinecures in themselves, the order they were to be held in was generally strictly ordained: one could not hold an office without having first held those which preceded it, and there were age restrictions upon the offices.9 Only a specific order from the emperor could allow one to circumvent these restrictions.10 Once a man had passed through the cursus honorum, he could become a provincial governor. The governors regularly commanded whole armies comprising multiple legions and units of auxiliary troops. In this way, they approached the professionalism of the magisterial generals, although their duties were still divided between military command and civilian administration. The Roman Republic had established a long precedent of extraordinary magistracies, such as the dictator, appointed to resolve a specific problem. The emperors, similarly, at times, needed to grant extraordinary commands, and senators like Marcus Vipsanius Agrippa, Germanicus, Gnaeus Domitius Corbulo, and Gaius Avidius Cassius, to name only a few, commanded combined field armies across entire theatres of war.11 Although these positions were temporary, the scope of their military commands was much like that of the magistri. These senatorial magistracies were mostly unsalaried; from Augustus onwards, however, the provincial commands were remunerated. A lack of a salary was generally not burdensome to the senators, however, as they came from the landed elite and derived their wealth from rents and revenues on their holdings. The equites Romani had two career paths separate to the senators during the early Empire. It is important for these to be understood, because although the large number of soldiers commanded by the magistri reflects the senatorial commands, the magisterium ultimately evolved from the equestrian order. The first of their paths, known as the militiae equestres, or the tres militiae for its three main offices, would see an eques begin as a prefect of a cohort of auxiliaries, followed by a posting as a military tribune, and then they would finish as the prefect of an ala of auxiliary cavalry.12 Under emperor Hadrian (117–138) command of an ala milliaria, a double-strength unit of cavalry, was 9 Talbert 1996, 340–41. 10 Birley 1981, 4–35; Talbert 1984, 16–27; Davenport 2019, 302. 11 See page 151. 12 Goldsworthy 2003, 65–66; Talbert 1996, 340; Davenport 2019, 261–62.

152

The Prosopography of the Magistri Militum

added as a fourth rung, but because of the funnel-shaped hierarchy, few made it this far.13 This career path thus ended at a much lower position than the concurrent senatorial or later magisterial paths. In an alternative equestrian career path of the early Empire, an eques could seek employment as a legionary centurion ex equite Romano. This path culminated in more significant positions, including tribunates in the vigiles, urban cohorts, Praetorian Guard, or even the equites singulares.14 For the most ambitious equestrians, the highest positions they could obtain were the prefectures of the fleets, procuratorships, and the prefectures of the Praetorian Guard, the province of Egypt, the Annona, and the vigiles.15 Unlike the senatorial magistracies, the equestrian positions were remunerated.16 These higher positions, especially that of Praetorian Prefect, at times functioned as the commander of field armies, similar to the ad hoc commands of the senators.17 There is evidence of further development of special commands in the third century. Julius Priscus first appears as a iuridicus in Egypt, and then as Praetorian Prefect, and then as governor of Mesopotamia in 244, the same year his brother, Philip, became emperor.18 He is then given the title of rector Orientis.19 Zosimus provides more context, saying that Philip entrusted the most critical miliary areas to family members, including placing Priscus in charge of the Syrian armies.20 He clearly had wide-ranging authority, as his exacting taxes eventually drove the east into rebellion. This special command given to an equestrian could have directly inspired the creation of the magistri militum. A distinction shared by the senatorial and the equestrian career paths of the early Empire is that officer served domi militiaeque – they combined civil and military service.21 Approximately 85 % of equestrian civil procurators had previously served in the military.22 However, the reforms of the late third and early fourth centuries ended domi militiaeque service. High civil and military office became the domain of the equestrians, but individuals did not move between the two structures, even though they shared the consulship as the peak of both careers, and they were combined into a unified social hierarchy enjoying three grades of senatorial status in 372 (clarissimus, spectabilis, and illuster/illustris).23 13 14 15 16 17 18 19 20 21 22 23

Davenport 2019, 261; Devijver 1999, 241. Dobson 1974, 399–403; 1978, 68–87. Davenport 2019, 303. Cheesman 1914, 35–45, 90–101; Davenport 2019, 270 n. 104, 276, 302. Davenport 2019, 524–26. ILS 1331, 8847; IGR 3.1033, 1201–2. ILS 9005. Zos. 1.19.2–20.2. For example, see: CIL III 6687 = ILS 2683; AE 1930, 121; 1954, 104. Davenport 2019, 316–17. For this term, see: Drogula 2015, 47–56. Davenport 2019, 311. CTh 6.7.1, 9.1, 11.1, 14.1, 22.4; Jones 1964, 142–44; Schmidt-Hofner 2008, 103–16.

The Fourth-Century Career Path and the Hierarchy of Military Offices

153

The later, sixth-century career path of the Byzantine Empire retained a similar structure to what we have seen over Chapter One and Two for the fourth century. There was, broadly-speaking, a similar disposition of frontier forces and internal field armies, which were still commanded by magistri militum, although, because of the predominance in our sources of Greek historians writing in a classicizing style, their positions are often preserved only with non-technical terms like στρατηγός or ἄρχων.24 The ranks below the magisterium also changed in name, but only slightly in nature. Byzantine field armies were divided into brigades, each comprising roughly five-thousand men and they were commanded by a μεράρχης.25 These brigades were further divided into legions of approximately one-thousand men commanded by a χιλίαρχος.26 These ranks seem to have commanded fewer troops than what we see in the fourth century, as if we recall from our earlier discussion, Vegetius writes that a magister militum commanded approximate twenty-four-thousand soldiers, and a comes rei militaris twelve-thousand.27 This could be emblematic of the smaller total size of the Byzantine army, but also its tendency, already visible in the fourth century, to be divided amongst a greater number of commanding officers than the west. The personal power of the magistri militum was also somewhat enhanced during the Byzantine period through the employment of bucellarii, new bodyguard units that directly served the generals.28 But overall, the early Byzantine Empire continued to utilize its army and officers in the same way developed in the late fourth-century eastern Roman Empire. This closeness indicates that, when compared to the late third and early fourth centuries, the intervening period of the fifth century was not one of extreme military innovation and reform. With the preceding and ensuing career paths thus described, we can now examine the fourth-century path that led a man to the military magisterium. 4.2.2 Survey Table Two displays the results of the research conducted on the fourth-century military career path. There seem to have been two groups of offices comprising the rungs that preceded the position of magister militum, and then the consulship made an additional final rung. We will visit each of these in successive order, but first it is pertinent to explain the ‘Unspecified Military Experience’ column. These individuals were recorded as some sort of officer in the army or as having previous military experience, but the sources failed to preserve their precise rank. These mentions appear mostly as

24 25 26 27 28

Parnell 2017, 14, 21–23. Treadgold 1995, 95. Parnell 2017, 22. See pages 31–32. Parnell 2017, 17–18.

154

The Prosopography of the Magistri Militum

references to long and extensive military careers, or a general having risen through the ranks, and are often given as part of a brief description of the background of a prominent magister militum.29 References also come from records of officers participating in Roman wars without exact mentions of their position and rank, such as Arinthaeus, Victor, and Hormisdas on the Persian expedition.30 In many of these cases, although it would be possible to assign a position to them speculatively, this kind of supposition would only serve to make this study inaccurate, and so it has not been attempted. Despite not contributing to the reconstruction of the magisterial career path, this category of generic references to military experience does confirm that at least thirty-eight magistri had military careers. When classifying the career path that led one to the magisterium in such a fashion as Table One, it is apparent that almost all the men who became magistri are first recorded in quite high ranks. Arbitio is the only individual that is said to have risen from the lowest ranks of the soldiery, and seeing as how Ammianus seems to have painted Arbitio in an exaggeratedly negative light, we should be careful in accepting this too readily.31 One reason that we are unaware of the magistri occupying low ranks is probably because the deeds of low-ranking positions were lost in the milieu of general military operations, and were either considered by our authors to be unimportant, or the authors were simply unaware of them. Service in the protectores, for example, could distinguish someone to their superiors and put them on a path to promotion, but would rarely involve actions significant enough to be recorded in the histories. Another reason may be because the career path was inherently flexible. The magisterial military career was the evolution of the flexible equestrian military careers that were entirely reliant on imperial decision-making. Thus, if an emperor so chose, it was possible to skip lower positions, although this did not go uncriticized by contemporaries if it was deemed unreasonable.32 Wealthy individuals, or individuals with fathers who distinguished themselves in the military, were able to begin their careers as officers.33 Additionally, magistri born outside the Empire may have gained experience commanding their native armies, and this could have endeared them to the emperors who allowed them to skip the lowest offices.34 The first firmly attested rung of the magisterial hierarchy is accordingly a multitude of positions that had quite important duties in the Roman army. This group includes the curae palatii, tribuni or comites sacri stabuli, praepositi, duces, and, most commonly,

29 30 31 32 33 34

E. g. Zos. 5.10.5. See pages 69–70. AM 16.6.1. AM 20.2.5. Although some believe Ammianus said this because he disliked Agilo’s barbarian background: Stroheker 1961, 128. Jones 1964, 641–42; Matthews 1989, 77–80, 270, 519 n. 33; Trombley 1999, 18. Williams and Friell 1994, 98–99.

35

Comes Domesticorum

Praepositus

Nevitta

35

361–369

361–363

361–362

360–362

367

362

367

355

351

348

344

347

Consulship (19)

Reportedly began career as a common soldier (AM 16.6.1) and therefore may have held a lower office before this, but this is not directly attested.

Unspecified

Tribunus Scholae Scutariorum

Gomoarius

Jovinus

Tribunus Stabuli, Tribunus Gentilium et Scutariorum

Agilo

360–361

Comes Domesticorum

Lucillianus

Unspecified

359–360, 364–367

357–358

357–359

353–355

350–362

Lupicinus

Severus 8

Unspecified

Tribunus Scholae Armaturarum

Silvanus

Barbatio

Dux

Arbitio

349–359

Ursicinus

Unspecified

349–353

348–350

Unspecified

Vetranio

Gaiso

344–347

Salia

34

Magister Militum, Dates Held

Sallustius

Second Rung (13) 343–34

First Rung (13)

Eusebius

Unspecified Military Experience (14)

Table One The Career Path of the Highest Late Roman Military Officers in Order of First Attestation as Magister Militum in the Sources.

The Fourth-Century Career Path and the Hierarchy of Military Offices

155

Unspecified

37 38

36

Comes Rei Militaris

Comes Rei Militaris

Comes Rei Militaris

Comes Rei Militaris

Comes Domesticorum

378–379

378

377–382

378

375–388(?)

373–378

370–375

368–372

367–373

365–375

363–380

363–366

388

363

Magister Militum, Dates Held

383

377, 383, (388)

372

374

369

366

Consulship (19)

Called a dux by AM 27.8.6, but this does not seem like a technical usage because he also uses the term to refer to one of Flavius Theodosius 3’s direct subordinates, Dulcitius. We might therefore assume that Flavius Theodosius 3 was a comes rei militaris if a dux was below him, but it is also possible that Flavius Theodosius 3 was a dux and Dulcitius was actually an even lower rank, and so we must leave Flavius Theodosius 3 in the unspecified column. Theodosius 4’s consulships as emperor are not counted as magisterial consulships. Unclear exactly what Theodosius 4’s rank was when he was recalled into service in 378, but it seems unlikely that he would have been in a position lower than his previous rank of dux.

Dux38

Theodosius 437

Unspecified

Dux

Cura Palatii

Unspecified

Saturninus

Sebastianus

Dux

Unspecified

Traianus

Merobaudes

Julius

Flavius Theodosius 3

Unspecified36

Tribunus Armaturae

Arinthaeus

Severus 10

Tribunus Schola Prima Scutariorum

Equitius

Comes Rei Militaris

Comes Domesticorum

Victor

Dagalaifus

Second Rung (13)

Comes Rei Militaris

Tribunus Gentilium

First Rung (13)

Nannienus

Malarichus

Unspecified Military Experience (14)

156 The Prosopography of the Magistri Militum

Unspecified

Abundantius Tribunus Praetorianus Militaris, Comes Sacri Stabuli39

39

392–408

391–396

391–396

386–395

387–394

386–398

386–391

393

389

389

384

385

Consulship (19)

This position probably served similar or even the same function as the tribunus sacri stabuli but was upgraded with the title comes. Perhaps this means the position should be considered to be one tier higher, or perhaps it was somewhere in between the two tiers. Without more information on the role of the position or the changes to it that came with the comes title, it is impossible to determine conclusively.

Comes Domesticorum

Unspecified

Timasius Comes Domesticorum

Unspecified

Arbogast

Stilicho

Unspecified

Gildo

Addaeus

Unspecified

Promotus

383–393

Magister Militum, Dates Held 380–386

Comes Domesticorum

Second Rung (13)

Richomeres

First Rung (13)

Bauto

Unspecified Military Experience (14)

The Fourth-Century Career Path and the Hierarchy of Military Offices

157

158

The Prosopography of the Magistri Militum

the tribuni.40 One quarter of the fourth-century magistri are attested as having attained ranks in this group. Multiple usurpers and emperors came from these ranks, most prominently Magnentius, Valentinian, and Valens, demonstrating the importance of these offices – although as described above, there were numerous reasons why men from these ranks appealed to acceptance groups as leaders.41 The tribuni commanded the legions of the field armies, which numbered approximately 1,000 men each, and they served under the magistri militum, or under the magister officiorum for the tribuni of the scholae palatini, a position higher than other tribuni.42 They could also be given special assignments, such as intelligence gathering and reconnaissance.43 Duces also likely served under the magistri militum, but commanded units of limitanei soldiers.44 It is more difficult to determine the roles of the other offices with certainty, but they all appear to have been positions of importance in the middle levels of the army. However, these positions were not typically a direct path to the magisterium.45 Of the thirteen officers in this group, only one dux and three tribuni were elevated directly to magister militum. These promotions all occurred in somewhat exceptional and dire circumstances. Silvanus was promoted for his timely defection to Constantius during the war with Magnentius, and Agilo was promoted at a time when Constantius was both reeling from the loss of Amida and had just been forced to drop two magistri.46 Similarly, in 363, emperor Jovian was trying to secure stability for his new reign and hoped to win the loyalty of the former tribunus Malarichus by suddenly promoting him to high station, although he declined the offer.47 Finally, in the wake of the disaster at Adrianople in 378, Gratian seemingly promoted the dux Theodosius directly to the magisterium.48 Of the remaining ten officers in this first tier, we know that six were promoted once more before becoming magistri.49 The other four officers have chronological gaps between their time in this group and when they are first attested as magistri, which provides the possibility that they were also promoted further up the military hierarchy, but it has gone unrecorded.50 As shown in Table One, it was the positions of comes domesticorum and comes rei militaris that comprised the next rung of the military career. Like the first rung, one 40 41

Jones 1964, 372–73, 625, 640–41. Valentinian: AM 25.10.9, 26.1.5; Oros. 7.32.2. Valens: AM 26.4.2. On the acceptance groups, see pages 48, 53. 42 Mommsen 1889, 195 ff.; Nicasie 1998, 68; Jones 1964, 640–41; Not. Dig. Occ. 9, Or. 11. 43 AM 19.11.5. 44 Nicasie 1998, 77–78. 45 Or, for that matter, a common direct path to being emperor. Jovian, as well as the usurpers Magnentius and Procopius, were all higher ranked comites. 46 AM 14.10.8, 20.2.5. 47 See page 72. 48 See pages 99. 49 Equitius, Saturninus, Sebastianus, Stilicho, and Traianus. 50 Arinthaeus, Gomoarius, Nevitta, and Silvanus.

The Fourth-Century Career Path and the Hierarchy of Military Offices

159

quarter of our magistri are attested to have occupied one of these offices, although it is not the exact same set of individuals. Seven were comites domesticorum and six were comites rei militaris. As best as can be determined, none ever occupied both positions at different times of their career. This helps confirm that these ranks were roughly militarily equivalent to one another and therefore comfortably comprise a tier of offices. From this, and the preceding rung, it seems that the Roman military was not in the habit of promoting individuals ‘sideways’ to similar positions; instead, a man would be promoted upwards or removed from military service. Although approximately equivalent in rank, these positions had different duties and stations, and probably esteemed themselves in the eyes of the emperors for different reasons. Interestingly, in contrast to this apparent martial similarity, these comites positions obtained different ranks in the unified social hierarchy. The comites domesticorum made the grade of vir illustris, the same as that of the magistri and one above the comites rei militaris.51 As they were the commanders of the protectores domestici, they were a part of the palatine troops.52 As discussed in Chapter 1.1, the palatine troops were more highly decorated and rewarded than the soldiers in the regional field armies, and this explains their greater prestige than the other military comites.53 Other palatine civil ministers were also viri illustris.54 The comites domesticorum received their orders for their broad duties directly from the emperors. For example, Dagalaifus was ordered by Julian to infiltrate Sirmium and capture its commander, the magister equitum Lucillianus, ahead of the advance of Julian’s main army in 361.55 Lucillianus himself had been promoted to comes domesticorum in 354 and it seems that he remained in this position when he travelled to Persia to undertake diplomatic negotiations with Shapur.56 The comites domesticorum also occasionally controlled small armies, such as when Severus was sent to Britain to defend against a wave of invasions in 367, or when Richomeres commanded the Gallic reinforcements sent to the Danube during the Gothic crisis.57 Being in such close proximity to the emperors and completing these important missions placed the comites domesticorum in a highly favored position. The comites rei militaris were not as close to the emperors, but their duties closely mirrored those of the magisterium, a factor which assisted in distinguishing them for promotion. The office first appears under Constans, so it may have been created by him in the 330s or 340s, or perhaps it was part of the swathe of reforms undertaken by Constantine I, or even Diocletian. The comes rei militaris commanded the smaller regional

51 52 53 54 55 56 57

Jones 1964, 528–29. Williams 1985, 25; Alföldi 1967, 412–13; Lenski 2000, 502. See page 31. Jones 1964, 528. AM 21.9.6. AM 17.14.3, 18.6.17. Severus in Britain: AM 27.8.2. Richomeres on the Danube: AM 31.7.3–5, .6–16, 12.4.

160

The Prosopography of the Magistri Militum

field armies that were not under the direct oversight of a magister militum.58 They were thus a part of the field armies, the comitatenses, and so were only viri spectabiles.59 For example, Julius and Equitius were comites rei militaris in command of the eastern and western armies along the Danube in 365, and Nannienus commanded the armies in northern Gaul during the Saxon invasion of 370.60 Similarly, Sebastianus commanded the Italian garrison under Valentinian. He played a key role as a secondary commander to the magistri militum during these campaigns, cutting the Alamanni off from a potential avenue of escape during the main battle of 368, and in 375 he assisted in burning and pillaging the villages of the Quadi.61 These duties would have demonstrated that Sebastianus was a capable commander, well-suited to the position of magister militum, and this increased his reputation sufficiently for Valens to request that specifically he come to his aid during the Gothic crisis.62 With only two exceptions, the men in this category were promoted directly into the magisterium. The anomalies, Barbatio and Julius, both have gaps in their career where it is unclear what position they held. Barbatio was replaced as comes domesticorum in 354 but continued to serve Constantius in an unknown role for at least one year.63 Perhaps he held no official position, or he was promoted to the magisterium earlier than we are aware of. As for the comes rei militaris Julius, he was captured and imprisoned by the usurper Procopius in 365, and it is not known what became of him until he reappears six years later as a magister in Arabia. Thus we can confidently conclude that these positions were a regular stepping-stone to the magisterium. The next rung was the magisterium itself. Above this existed one special position that was not a military office, but still represents the peak of a magisterial career: the consulship.64 During the Roman Republic, two consuls were elected annually in replacement of the overthrown kings, and they were the highest ordinary magistracies with broad governing powers. With the change to an Empire, these governing duties were absorbed by the emperors, and additional consular positions were created by the awarding of suffect consulships. The ordinary consulship maintained its presidency of the senate and some judicial functions, although these were a small enough burden for the office to be described as an extremely high “honor without work”, according to one fourth-century holder.65 Although a formal election process persisted, the consuls were now chosen by the princeps, and the position was often reserved for the emperors or their family members, especially during significant events such as their first year 58 59 60 61 62 63 64 65

Jones 1964, 105, 124–5. Jones 1964, 528–29. Julius: AM 26.7.5. Equitius: AM 26.5.3. Nannienus: AM 28.5.1–7. In 368: AM 27.10.6. In 375: AM 30.5.13. AM 31.11.1. AM 14.11.19–25. Lee 2015, 109–12; Bagnall, et al. 1987. “honos sine labore” Pan. Lat. 3.2.2.

The Fourth-Century Career Path and the Hierarchy of Military Offices

161

as sovereign. If not taken by the imperial family, the ordinary consulship in the fourth century was awarded by the emperors to their elite and most favored officials, initially mostly to civilian administrators, but increasingly given to military commanders.66 To be a consul along with the emperor, especially in an important anniversary year, was the greatest distinction. This highly coveted award represented the culmination of a career in the imperial administration. The changing frequency with which the magistri held the consulship helps to highlight the varying degrees of favor they received from different emperors. In some cases, such as Arbitio or Nevitta, the awardee commanded the main field army alongside the emperor and they were probably the most favored general for this reason, and this helps explain why they were chosen for consulships over other contemporaneous magistri.67 This could indicate that Gaiso, consul under the usurper Magnentius, was his most favored general, unsurprising after he successfully executed Constans.68 Others were clearly given the consulship as reward for good service, such as Saturninus in 383.69 Yet, others were given after a failure, for example Dagalaifus in 366.70 Instead, this consulship might have been granted out of respect for a long and valued career, and to usher him into retirement.71 Others, like Richomeres in 384, were awarded after a period without any known significance for the awardee.72 Thus, it seems that the inclinations of the emperors, rather than the actions of the recipients, more often determined the consulship. As can be seen in Figure Three, before 366 the consulships awarded to magistri were sparse.73 Constantius awarded at least two consulships to magistri within the first decade of his tenure as emperor of the east, but in the following fourteen, only awarded a single one to Arbitio.74 Constans, Magnentius, and Julian also only awarded one consulship each to their magistri during their shorter reigns.75 In the same period, twelve civil officials, mostly Praetorian Prefects, were awarded consulships.76 Ammianus, who wrote at a later time when the magistri were more powerful, nostalgically recalled 66 Waas 1971, 20. 67 See pages 58–59, 65–66, 70. 68 Epit. 41.22–23; Zos. 2.42.5; Chron. Min. 1.69. 69 Zos. 4.33.1–2; Them. Or. 16.208b–209d, 210d. 70 AM 26.5.9. 71 AM 26.5.9, 27.2.1. 72 See pages 110–11. 73 As discussed on page 32, Ursus and Polemius may have been the first magistri to hold the consulship, but we cannot be certain. 74 There is the possibility that some earlier consuls were magistri, but as discussed in the opening section of Chapter One, accurately identifying these men is impossible. See: Chastagnol 1992, 248; Barnes 1981, 398 n. 17. As per the same section of Chapter One, these numbers are reached by discounting the consulship of Bonosus. 75 Constans: Salia in 348. Magnentius: Gaiso in 351. Julian: Nevitta in 362. 76 Bagnall, et al. 1987, 222–65. There are quite a few consuls in the early years of this span that nothing is known about; some could have been magistri and some could have been civil officials.

Fig. Three The Consulships of the Magisterium. Own Illustration. Above the line are the names of magistri who received their consulship in the east, and below the line in the west.

162 The Prosopography of the Magistri Militum

The Fourth-Century Career Path and the Hierarchy of Military Offices

163

how they had been kept in their place under Constantius, and everyone looked up to the Praetorian Prefects as the highest officials, as he thought it should be.77 This was not to be the case under the brothers Valentinian and Valens, whose favoritism of the generals was the high-water mark for magisterial consular awards. Because he was both the elder brother and the first proclaimed emperor, Valentinian had the privilege of deciding who was to be awarded the consulships each year, although when choosing an appointee from the east he seems to have valued Valens’ recommendations.78 Out of the twenty-two total consulships Valentinian was able to distribute with his brother, six were granted to the magistri, only two to civilian officials, and the rest were taken by members of the imperial family.79 Furthermore, as we have seen, Valentinian officially raised the magistri militum to the same standing as the Praetorian and Urban Prefects.80 This favor may be attributable to Valentinian’s own military career before he was elevated to the purple: it seems that this gave him a great appreciation and fondness for the military men around him.81 He also may have felt he owed them for electing him, and needed to continue securing their loyalty through the regular distribution of patronage and rewards. Gratian had not been raised by soldiers as his father had been, but by aristocrats and civilians, and thus after he came into power as the western Empire’s senior ruler, consular favor began to swing back to the civil officials. He initially gave the 377 consulship to Merobaudes and the Urban Prefect Lucius Aurelius Avianius Symmachus, but after Symmachus died Gratian took the position for himself.82 Valens’ death made Gratian the entire Empire’s senior Augustus, and he took the opportunity to award the consulship to some of those who we have seen were in his social network: his tutor Ausonius and the Praetorian Prefect Olybrius.83 From 380 the consulship becomes more complicated. As the most recently raised member of the imperial college, Theodosius was technically the junior Augustus to Gratian and Valentinian II; however, he was their senior in both age and experience, and this allowed him to wield influence over the designation of consuls. From 381 to 383, Gratian and Theodosius each chose one consul per year, granting two consulships each to civil officials, and one each to magistri: Saturninus’ first and Merobaudes’ sec-

77 78 79 80 81 82 83

AM 21.16.1–2. Lenski 2002a, 34; contra Bagnall, et al. 1987, 14–15. Jones 1964, 142–43; Bagnall, et al. 1987, 264–91. No new consuls were appointed in 375, so Equitius was consul two years in a row, however this is not equivalent to being awarded two consulships. CTh 6.7.1, 6.9.1, 6.14.1, 6.22.4. Also see: Jones 1964, 142–43; Näf 1995, 20; Heather 1998, 188–91. AM 16.11.6–7, 25.10.8–9, 26.1.5; Soz. 6.6.3–4; Philost. HE 7.7; Oros. 7.32.2. He also seemed to have a dislike for aristocratic senators, a great number of trials occurring during his reign, e. g. the trials conducted by Maximinus. Bagnall, et al. 1987, 19, 288–9. Bagnall, et al. 1987, 292–3.

164

The Prosopography of the Magistri Militum

ond, both in 383.84 Saturninus’ award is particularly important because it occurred during Theodosius’ quinquennalia, the celebration of his fifth year as emperor, a time when he was expected to reserve the consulship for himself, but instead he awarded it to the magister.85 Additionally, the hierarchical importance of the consulship was also further increased at this point, indicating that even though the magistri were receiving fewer consulships, they were still highly valued.86 After Gratian’s death in 383, Theodosius was able to monopolize power even more firmly over Valentinian II, and this allowed him to control almost all the consular awards. Valentinian was allowed to choose one of the consuls for 385, but this was his only opportunity. He initially chose Praetextatus, the Praetorian Prefect, but he died before the consular year began and was replaced by Bauto.87 Theodosius nominated six civil officials and four magistri over this period, which further supports the conclusion that eastern civil officials eclipsed the power of the eastern magistri.88 The usurpers Magnus Maximus and Eugenius also appointed their own consuls which went mostly unrecognized by Theodosius, but they also favored their Prefects over their generals.89 In the end, twenty consulships were awarded to the magistri, with only Merobaudes receiving two. Thus, almost forty percent of the magistri received this honor in the period 344–395.90 In comparison, during the same period, thirty-seven consulships were given to civilians, almost double that of the generals. 4.2.3 Discussion There are several interesting things to note from this survey. Firstly, the distribution of known career paths reflects, to an extent, the vagueness of our source material. There are almost only entirely unknown careers during the 340s and 350s, and then to an extent again through the 380s – the parts of our period not covered by Ammianus’ Res Gestae. From what we can determine, the distribution of first rung offices is fairly well spread across the period examined, with duces and tribuni throughout. Before 365, there was a mix of comites domesticorum and comites rei militaris, and then from 365 (the

84 Bagnall, et al. 1987, 15, 296–301. 85 Them. Or. 16.203a. 86 See page 105. 87 See page 108. 88 Bagnall, et al. 1987, 302–25. 89 Even if we include Merobaudes’ potential third consulship in 388, Magnus Maximus nominated his Praetorian Prefect Euodius in 386, who was recognized by Theodosius, and Eugenius also promoted his Praetorian Prefect Virius Nicomachus Flavianus, who was not recognized by Theodosius in 394. See: Bagnall, et al. 1987, 16. 90 This is not including the inconclusive third consulship for Merobaudes, or for Stilicho’s two and Rumoridus’ one, which fall outside the time period under examination here.

The Fourth-Century Career Path and the Hierarchy of Military Offices

165

beginning of the Valentinian dynasty) until the ascendency of Theodosius after death of Gratian, the magistri almost all come through the office of comes rei militaris. This is perhaps indicative of Valentinian I’s and Valens’ favoritism of the regional armies and provincial leadership. Perhaps as a response to this favoritism, or because of his own inclination, emperor Theodosius then seems to have only promoted magistri from the position of comes domesticorum. Of course, there is a host of men with unknown careers which could entirely contradict these conclusions, but the differing career paths seem to fit what we suspect of the tendencies of different emperors. A consideration to make is whether merit or other factors dictated the promotion of men along this career path. Ammianus writes that Julian, upon his acclamation as emperor, promised his soldiers that he would end the practice of officers being promoted above their merits through favoritism.91 This indicates that it was at least perceived by disgruntled rank-and-file soldiers that they were at times led by incompetent officers. However, it seems most promotions along the magisterial career path were based on merit. For example, Flavius Theodosius was promoted to the magisterium after his successful defense of Britain in 368–69.92 Sebastianus was summoned by Valens and promoted to the magisterium because of his renowned reputation.93 Other factors may have contributed to these promotions, but the military successes of these and other magistri were the primary causes for their endorsements. Other promotions did not directly follow military successes, but still demonstrate a consideration of various merits. Sabinianus was probably promoted as magister in the east in order to continue Constantius’ diplomatic approach to managing relations with the Persian Empire.94 While not based on his military merit, Sabinianus’ promotion thus still contained a consideration of other values. Furthermore, Equitius was elevated to the magisterium so he could better defend Illyricum from Procopius in 364, but his promotion was probably also due to the support he showed Valentinian during the imperial election process because of their shared Pannonian heritage.95 These personal relationships were quite important in many promotions, and we can see it clearly in the careers of the younger Theodosius and Arbogast. As discussed in Chapter 3.3, Flavius Theodosius was not only supported and favored by Valentinian I, but was a member of the Gratian-Ausonius clique, and this group probably also offered their support to the younger Theodosius’s promotion to the magisterium and eventually as emperor.96 Additionally, Arbogast inherited Bauto’s position after the latter’s death. As discussed above, this was reportedly done without the permission of emperor Valentinian II,

91 AM 20.5.7. 92 AM 28.3.9; Pan. Lat. 12.5.2. 93 See page 96. 94 See page 57. 95 See page 95. 96 See page 99.

166

The Prosopography of the Magistri Militum

but was instead because Arbogast had the support of the soldiers.97 Eunapius records that Arbogast was Bauto’s son, and if true, this may have aided him in inheriting the position.98 However, this familial relationship is somewhat doubtful, as even though Eunapius is generally a reliable contemporary source, it seems unlikely that Arcadius would marry Bauto’s daughter and therefore, according to Eunapius, Arbogast’s sister, because at that time Arbogast was still suspected of murdering Valentinian II, Arcadius’ cousin-in-law and the senior emperor. Whether this relationship is historical or fiction, Arbogast’s promotion indicates that the loyalty of the soldiers and administrators could pass from one general to another, helping their career progress. Furthermore, high-ranking ancestors, especially fathers who had achieved renown, often favored someone for promotion.99 Arbogast was probably also supported by the many other Franks in the Roman military administration during the late fourth century.100 However, the support that Equitius, Theodosius, and Arbogast received from these different groups rested on their competency. Thus, even when nepotism or favoritism contributed to a magister’s rise, it did not necessarily preclude them from needing to be militarily experienced. Although merit was duly important, it does not explain all promotions to the magisterium. For example, Traianus was promoted in 378 after the battle of Ad Salices, which was not a failure, but was far from a success.101 Furthermore, although Malarichus was clearly capable enough to become the commander of the gentiles, Jovian tried to promote him to the magisterium because he did not have the renown to achieve the position under normal circumstances.102 Thus, while most magisterial promotions in the fourth century were based on military merit, some were based on other merits or were supported by other groups, and some were not based on merit at all. As discussed above, magisterial consulships appear at the very beginning of the period, and continued all the way through, evidencing that supreme importance of magistri during the entire fourth century. A point of interest is that Arbogast, one of the most powerful magistri of this period, never received a consulship. Western officials were not afforded any nominations during the years he supervised Valentinian, but even under Eugenius the nomination went to a Praetorian Prefect. So, while the consular awards are a useful guide to understanding the changes felt by the magistri over the fourth century it is not an exact representation of their power and influence. Furthermore, as it was an office without powers, by no means was it necessary for a magister to gain a consulship to begin or maintain their influence. The awarding of consulships

97 See page 109. 98 Eun. Hist. fr. 58; cf. Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). 99 Elton 2007, 203–4. 100 Drinkwater 2007, 158–59. 101 See page 95. 102 AM 25.5.6.

The Fourth-Century Career Path and the Hierarchy of Military Offices

167

can instead help us aid in overcoming lacunae in the historical record. For example, we are able to appreciate Merobaudes’ powerful position better when we learn that he was the first non-royal to receive two consulships in decades.103 Stilicho, the next magister to repeat this feat, similarly enjoyed a powerful position in the west, in addition to his familial connection to the imperial family through marriage.104 Thus the consular awards serve as a limited but still useful barometer of the social and political position enjoyed by some magistri during the fourth century. Other than their time as magistri, as can be seen on Table One, the careers of Merobaudes and Arbogast are relatively unknown. They were both officers in the Roman army before becoming magistri, although their ranks are unspecified. As discussed above, Arbogast served with Bauto and was able to assume the command of the armies upon Bauto’s death.105 This was more to do with personal qualities than any official inheritance policy, and so it seems that the offices held by Merobaudes and Arbogast was not the determining factor in their eventual powerful positions. Instead, it was the opportunities they exploited once they were already magistri that gained them such influence. Much more is known about the career of Stilicho, who had something of a unique career which directly contributed to his later success. Our more detailed knowledge of this magister comes from Claudian, the court poet at the time who composed multiple panegyrical pieces for Stilicho, as well as from two inscribed statue base dedications from the Roman forum.106 Although, as stated above, domi militiaeque careers had disappeared, Stilicho did have something of an irregular mixed civil and military career. He first appears as a tribunus praetorianus, an office in the branch of military administrators known as the notarii. In this position, Stilicho was a junior member of a diplomatic embassy to Persia in 384.107 At this point, historians face a conundrum of how to explain Stilicho’s subsequent marriage to Theodosius’ niece and adopted daughter, Serena. Claudian claims that Theodosius awarded Stilicho with Serena’s hand because of his outstanding achievements, but as Stilicho was not the leader of the embassy to Persia, nor had he done anything else of note, this is certainly panegyrical publicity.108 Perhaps at some point the varied duties of the notarii at some point placed Stilicho inside the imperial residence of Constantinople where he encountered Serena and they decided the match themselves, or maybe Theodosius saw a good outcome in tying a young, up-and-coming, half-barbarian general to the imperial household.109 103 See page 105. 104 Bagnall, et al. 1987, 334–35, 344–45. 105 See page 109. 106 CIL VI 1730 = VI 31913 = ILS 1277 and CIL VI 1195 = VI 1731 = VI 31913 = ILS 1278. 107 Claud. De cons. Stil. 1.51–68. 108 Claud. Laus Serenae 178–81. Joh. Lyd. de mag. 3.53 writes that the leader of the embassy was Sporacius. 109 For a discussion of Stilicho’s Roman-barbarian dichotomy, see Chapter 4.3.

168

The Prosopography of the Magistri Militum

Whatever the cause, it undoubtedly rose Stilicho’s prospects much higher than they otherwise would have been. Theodosius seems to have been aware that Stilicho’s civil career hitherto had not been quite remarkable enough for a royal family member. He thus helped Stilicho’s career skyrocket. He was first promoted as comes sacri stabuli – a similar, or perhaps a copy of, the position of tribunus stabuli that Valentinian promoted Valens into when trying to increase his prestige before raising him as co-emperor.110 Theodosius may have upgraded the position from that of tribunus to comes to further speed up Stilicho’s progression, who by 385 was promoted as comes domesticorum.111 To the best of our knowledge, Stilicho still did not achieve much of significant military note, but probably began taking major steps towards developing a network of friends and allies, including, for example, the senator Symmachus.112 Stilicho’s career had finally progressed sufficiently for him to become a magister militum in either 392 or 393, just in time for him to inherit guardianship of Honorius in 395.113 Thus, Stilicho’s eventual power and authority was based upon his early career, but rather than the offices he held or operations he was involved in, it was his marriage to the imperial family that boosted his position far above what would be realistic in other circumstances. Thus, unique factors that affect only one person could clearly have a major impact on an individual’s successes. It is revealing when we consider the length of different magisterial careers. As can be seen in Table Two, from the 340s to the 360s, most magistri only spent one, two, or three years in the position. There are some stand outs, such as Ursicinus at eleven and Arbitio at thirteen years in the office. Both are much higher than the average of only four years in office for the magistri under Constans, Constantius, Julian, and Jovian. This speaks to a relatively fast turnover of the highest-ranking generals. This changes with the Valentinian dynasty. From the accession of Valentinian I and Valens, we can now see multiple magistri having equal careers to Ursicinus and Arbitio, but also Stilicho, at seventeen years, extends significantly beyond all others bar Victor, who retained the office for an incredible eighteen years. These long careers, while standouts, are emblematic of an overall increase. The Valentiniani average was approximately six years in the magisterial office. The median also increases from three years pre-Valentinian dynasty, to six afterwards. There was thus a significant increase in time the magisterial office was held. Additionally, although it is only a small difference, the average career length in the west was five years, and in the east, it was six. While there are difficulties in applying modern military theories to ancient armies, there is perhaps a valid comparison to be made when it comes to magisterial turnover rates. It is a feature of armies in the modern western world to practice regular job

110 AM 26.4.2; Jones 1964, 372–73. 111 CTh 6.13.1 would seem to be the law upgrading this office to a comes, but Stilicho’s position predates this law by two and a half decades. 112 PLRE I, 854. 113 CTh 7.4.18, 9.3.

169

The Fourth-Century Career Path and the Hierarchy of Military Offices

Table Two Years Spent in the Military Magisterium during the Fourth Century. The lists have been divided into west and east based on where the magister spent the majority of their time in the magisterium. It should be noted that these years are only the best estimation of the author. Often, we do not know the exact year someone entered or left the magisterium. We also very rarely know the time of the year, which could easily add or subtract up to two years. Furthermore, there were cases like that of Theodosius the younger, who seems to have only been a magister for less than a year. For the sake of simplicity, such fractional amounts have not been included and those like Theodosius have been rounded up to one year. Malarichus, who was offered the position of a magister militum but declined has still been included for the reasons discussed above, but at zero years in the magisterium. Despite these hurdles, broad conclusions should not be terribly inaccurate and will still be useful. Western Empire Magister Militum Salia Vetranio Romulus Gaiso Silvanus Marcellus Severus 8 Barbatio Lupicinus Nevitta Jovinus Dagalaifus Malarichus Equitius Severus 10 Theodosius 3 Merobaudes Theodosius 4 Bauto Andragathius Rumoridus Arbogast Quintinus Nannienus Stilicho

Eastern Empire

Years in Office 3 3 1 4 2 1 2 3 5 3 9 4 0 11 5 6 13 1 7 5 4 8 1 1 17

Magister Militum Hermogenes Bonosus Eusebius Sallustius Ursicinus Arbitio Sabinianus Lucillianus Agilo Gomoarius Victor Arinthaeus Julius Saturninus Traianus Sebastianus Sapores Modares Majorianus Richomeres Ellebichus Gildo Timasius Promotus Butherichus Abundantius Addaeus

Years in Office 3 8 5 1 11 13 1 2 3 2 18 7 6 6 1 1 4 4 2 11 5 13 10 6 1 6 6

170

The Prosopography of the Magistri Militum

rotation in the upper military ranks. For example, Australian military scholars theorize that high job rotation is important for developing an officer’s adaptability and breadth of experience.114 The United States military also mandates that its upper officers must either retire after five years in a rank or a certain number total years of service, and no Chief of Staff of the Army, a four-star general position somewhat similar to a magister militum, has remained in office longer than four years since World War II.115 Furthermore, the US Defense Officer Personnel Management Act of 1980 instructs that any officer twice passed over for promotion is to be dismissed from service, and there are limits on how many generals of particular ranks can serve concurrently.116 This guarantees a high rate of turnover for upper military officers. Interestingly, Vegetius writes about similar practices during the Principate, in that men should be promoted through all ten cohorts in a legion to ensure all the men felt a sense of cohesion.117 Obviously, the same level of strategic information is not as available for all militaries. Outside observation still allows us to see examples of individuals like Oh Jin-Woo, who was an extremely decorated officer in the North Korean army of Kim Il-sung, and he remained the supreme administrator of the Democratic People’s Republic of Korea’s military for twenty-three years.118 In this position he was one of the most powerful individuals of the North Korean regime. Similarly, since 2014, Thailand has been under the rule of Prayut Chan-o-cha, who rose through the powerful Queen’s Guard units and had been a military commander for twelve years before overthrowing the government in a coup d’état.119 Of course, examples of all kinds can be found in the modern world. Muammar Gaddafi had only spent three years in the military before orchestrating a coup in Libya.120 Nonetheless, a pattern can be observed that long careers in modern militaries have allowed some individuals to form robust bases of power, whereas other militaries have been well-served by regularly rotating their highest military leaders. Thus we might see a similar correlation between lengthy military careers and securing dominant political positions in the ancient world. The fact that some of the most powerful fourth-century magistri include the likes of Stilicho, Merobaudes, Arbitio, and Gildo, those with some of the longest magisterial careers, is unlikely to be a coincidence. Arbogast’s eight years in the magisterial office, although less than some contemporaries, is still longer than the average. He also spent quite a number of years as Bauto’s deputy, during which time he could have built a strong powerbase. Furthermore, specific regulations were made to grant greater standing to those who held magisterial power for 114 115 116 117 118

Jans and Frazer-Jans 2001, 47; Masland and Radway 1957. Gardner 2005. 10 U. S. Code § 525. Veg. 2.21. Wudunn, Sheryl, ‘O Jin U, the Defense Minister of North Korea, is Dead at 77’ New York Times, 26 Feb. 1995; Martin 2004, 61–62. 119 “Gen Prayut Chan-o-cha: Fighting to Protect his Legacy”, Thai PBS World, 13 Mar. 2019. 120 St. John 2012, 138.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

171

long periods.121 We should not be too quick to place the causation of a long time in office in front of the observation of great power; undoubtedly, great power also helped one retain important office. There must have been a symbiotic relationship: if a magister was influential, he would be more likely to retain the office, and the longer he retained the office, the more powerful he could grow. This helps to explain why the majority of the magistri, thirty-two in our timespan, retained the office for five years or less, and only twenty held it for longer, with some of those being significant outliers. It was a symbiotic, exponential relationship. Interestingly, the Byzantine Empire, which survived the fifth century, continued to frequently rotate its highest-ranking generals.122 The fourth-century military career path retained some similarities, but also developed a number of differences to that of the Roman Republic and early Empire. The career paths of senators and equites incorporated both military and civil postings, while the road that led to the magisterium was almost purely military in character. The fourth-century career path was remunerated, unlike the senatorial cursus honorum but similar to that of the tres militiae, although the magistri militum occupied much higher positions than the equites did during the early Empire. Thus, the late third- and early fourth-century reforms made radical changes in the military administration. After this period of reform, however, the Roman military stayed relatively unchanged into the sixth century, indicating this period of reform was a single, intense burst of changes that eventually stabilized. 4.3 The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between? 4.3.1 Introduction The manner in which the Roman army employed foreign troops alongside native citizen soldiers has a long, complex, and dynamic history, and foreigners played an important part in Rome’s military successes and failures from the city’s earliest beginnings. At first, it was easy to tell who was Roman, and who was not. In the early fifth century BCE, the Foedus Cassianum, named after its negotiator, the first Republican-era magister equitum, Spurius Cassius, codified the use of the non-Roman soldiers that were supplied by Rome’s Italian allies, and the register known as the formula togatorum maintained an annual catalogue of these troops.123 While there are some recorded instances

121 See page 83. 122 Parnell 2012, 7–9. 123 Hoyos 2007, 70; Rich 2007, 11; Dion. Hal. Rom. Ant. 6.95.2; Livy 2.18.5, 33.9. The foedus did undergo a number of modifications during this time, e. g. Livy 7.12.7, but the intention of the agreement remained the same.

172

The Prosopography of the Magistri Militum

of betrayal by these allies, the fact that they funded their own contingents of soldiers, and that they generally fought just as viciously as the Romans themselves, greatly eased the burden of warfare on the city-state of Rome and helped its early expansion.124 The allies eventually sought greater rights and an equal standing with Rome, and the Latin War (340–338 BCE) concluded by incorporating most of the allies, as well as the Campanian and Volsci former-enemies of Rome, as Roman citizens.125 Consequently, those who were once foreigners to the city of Rome, became natives of the Roman Republic – Romans in their own right. Rome continued to extend its control and diplomatic relationships further afield in the Mediterranean, and two and a half centuries later the Social War (91–88 BCE) resulted in Roman citizenship being extended more widely to a larger number of the Republic’s allies in Italy, as well as the slightly lesser Latin Rights to the Cisalpine Gaul communities north of the Po River.126 Thus, while unplanned, Rome repeated the same system of initial military alliance, followed by territorial expansion, and eventual enfranchisement of non-Roman groups into the Roman polity. Before the Social War even began, Rome’s armies had been utilizing other forms of non-Italian allied troops, such as Numidian cavalry or Cretan archers, and this increased into the first century BCE after the Marian reforms.127 These troops became crucial elements of Roman armies.128 Estimates indicate that some armies may have been composed of twice as many foreigners as Romans, although it was more often divided into roughly equal portions.129 As Roman sovereignty grew to cover much of the Mediterranean, the expansion-­ assimilation process continued. The allied territories that had been supplying soldiers, as well as many of the territories supplying mercenaries, were absorbed into the Roman state. The Republic ultimately transitioned into the Empire, and the absorbed non-Romans were formed into the auxilia by emperor Augustus. These non-citizen residents of the Empire maintained the legal status of peregrini. Non-citizen forces amounted to approximately half the strength of the Roman army and served as a compliment to the citizen-legions.130 In time, however, Roman citizens would even serve in the auxilia alongside non-citizens.131 During the early Empire, the army also employed

124 Rawlings 2007, 52–53; Polyb. 6.21. 125 Rich 2007, 14. 126 Pallottino 2014, 157; Duncan 2017, 178–79. 127 Hoyos 2007, 68–69; McCall 2002; Dixon and Southern 1992, 20–25; Harmand 1967, 46–51. 128 Caes. B. Gal. 7.65.4; Waas 1971, 1. 129 Toynbee 1965, 2.128–35; Brunt 1971, 677–86. 130 Goldsworthy 2003, 80, 208; Heather 2005, 119; Gilliver 2007, 185–86. For some primary examples of their role in Roman warfare, see: CIL VII 218; Cichorius 1896, pl. XIX, L. 131 Gilliver 2007, 193.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

173

irregular units known as numeri, which fought in their native styles and were probably composed of entirely non-citizens.132 During the Empire, many groups of foreigners were settled within Roman territory: Augustus placed 50,000 Getae on the Danube; Tiberius 40,000 Germans along the Rhine; Nero gave land in Moesia to 100,000 transdanubians; Trajan relocated many Dacians; Marcus Aurelius many Quadi, Vandals, Iazyges, Narisatae, and Marcomanni; Constantius I settled captive foreigners on uncultivated lands; Constantine I distributed 300,000 Sarmatians throughout the Balkans; in 377 Valens placed defeated Goths and Taifals in Italy as farmers; and in 386 the entire Greuthingi nation was settled on Roman soil.133 Although some of these numbers are certainly exaggerations, it is clear that the incorporation of foreigners into the Roman system continued throughout the imperial period. These settled groups also provided soldiers to the Roman army.134 Overall, the Roman Empire continued to use foreign troops in a myriad of different ways, and they were given many opportunities and pathways to assimilate into the Empire and, within a few generations, become Romans.135 The third century CE was a turning point in the way foreign soldiers were employed. A key motivation for peregrini to join the auxilia was a grant of citizenship after completion of their service. However, in 212, the emperor Caracalla granted citizenship en masse to nearly all free persons within the Roman Empire, ending this important distinction between the recruitment pools of the legions and auxilia, as well as between citizen and non-citizen in regards to Roman identity.136 This loss of upwards social mobility through the auxilia also compounded the problems of recruitment. There was also increasing and varied pressures that hindered the recruitment of Romans, including that landowners tended to prefer to keep men working the land rather than serving the military, and would thus either make extra tax payments instead of sending recruits, or evade recruitment some other way.137 This meant that the army would find it difficult to secure Roman recruits at the same time it became necessary to expand the size of the army significantly in order to aid the recovery from the third century crises.138 To counteract these developments in the fourth century, the Empire began to rely more heavily on employing men from across the Roman frontiers, the area that

132 133 134 135 136 137 138

Gilliver 2007, 195. Mathisen 2018, 256. Cass. Dio 71.21; AE 1956, 124; AM 31.4.1; Stickler 2007, 496; Waas 1971, 2. Meier 2020, 120–30. Mathisen 2006, 1014–15; Liebeschuetz 1998, 134. Waas 1971, 4; Liebeschuetz 1993, 274; Errington 1996b, 7. The manpower shortage theory on the Roman collapse has been put forward most strongly by Boak 1955, which was been negatively received, e. g. by Finley 1958. Whether it was a decisive factor in the collapse of the western Roman Empire or not, there is a large amount of contemporary evidence which indicates the Roman army had difficulty in obtaining domestic soldiers: CTh 7.13.4, 13.5, 13.10, 22.1. For the expanded size of the army, see: Coello 1996.

174

The Prosopography of the Magistri Militum

became known as barbaricum from the 220s onwards.139 In a recent essay, Michael Kulikowski argues that we should see barbaricum as functioning in the very same way that the provinces functioned in the early Imperial period – inside the imperial frontiers was almost universally Romans, legally speaking, and outside in barbaricum were the non-Romans, those called barbarians.140 Much as foreign soldiers previously, barbarians continued to come into Roman service through a variety of ways, and the poor state of the sources do not make it easy to understand the different categories used to designate barbarians within the Oikoumene – the cultured, habitable Roman world. The most common were the dediticii, war prisoners who were given the chance to join the Roman army, and laeti, who were given lands in return for taxes and recruits.141 Others still were recruited voluntarily as individuals or small groups from external tribes existing across the Roman frontiers.142 A final important change was the aforementioned opening of a new path for any soldier, including barbarians, to be promoted to the upper ranks through the protectores.143 The degree to which the late Roman army relied upon foreign soldiers, the so-called barbarization of the late Roman military, is still avidly debated by scholars. When looking at the army as a whole, historians have considered the ratio of barbarians to Romans and whether this had a negative impact upon the army’s efficacy. The academic discussion on this topic reaches as far back as Gibbon, but even has a precedent in Vegetius’ fourth-century manual of military science.144 Some scholars, such as Wolf Liebeschuetz and Arther Ferrill, argue that barbarization deeply weakened the army and it was one of the primary causes of the collapse of the western Empire in the fifth century.145 Most historians, however, have moved away from this theory in recent years, and it is more often believed that the regular Roman army remained an effective fighting force well into the mid-fifth century.146 It has also been argued that the rankand-file soldiery of the Roman army was much too large with a distinct martial culture to become barbarized through the slow recruitment of foreigners, especially seeing as they were recruited from all the varied Germanic, Gothic, and Eastern tribal groups, which each had their own different subdivisions, further lowering the likelihood that foreign military practices specific to any group would be adopted by the whole.147

139 Strobel 2007, 278–279; Fanning 2011; Mathisen 2019, 143; Meier 2020, 120. 140 Kulikowski 2015. 141 Jones 1964, 614, 620, 665. Silvanus was the descendent of Bonitus, a laetus. For a list of all the major settlements of foreigners up to Constantine I, see: MacMullen 1963, 553. 142 Elton 1996, 134–5. Gildo was one who voluntarily joined the Romans (AM 15.5.16, 5.33; cf. Aur. Vict. 42.15; Epit. 42.11). 143 See pages 29–30. 144 Gibbon 1781, 130–33; Veg. 2.2–3. 145 Liebeschuetz 1993; Ferrill 1986. 146 Dixon and Southern 1996; Elton 2018. 147 Bleckmann 2009, 206.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

175

Ralph Mathisen even posits that barbarian integration, including in the army, was one of the reasons the end of the western Roman Empire occurred as a transformation into new successor states rather than as a violent conquest.148 Barbarization as a damaging phenomenon has become unpopular enough that in his recent monograph How Rome Fell, Adrian Goldsworthy does not once mention barbarization as a cause for imperial collapse.149 A separate but related debate in this area is the employment of standalone units of barbarian soldiers in the fifth century that were not intermixed with the regular Roman army and instead fought under their own barbarian leaders. It is much more commonly accepted that, because these groups, called foederati, were loyal to their own interests rather than Rome’s, they were a deeply destabilizing force, and they played a key role in the collapse of the western Empire. Some of the leaders of these groups, including Alaric and Gainas, became magistri militum, and at times directly competed with other magistri, such as Stilicho and Aetius, who held powerful positions controlling the child-emperors. Although much has already been published on these topics, there is more to analyze and understand, and hopefully the present study of the fourth-century barbarian magistri will contribute further to our knowledge and comprehension of barbarians in the upper Roman army command. A primary difficulty arises because, as there were no decrees to match Carcalla’s of 212, the acculturation and assimilation of Romans and barbarians became a more opaque process as it increased in speed during the fourth century, and it continued until such a time that the differentiation was meaningless.150 Romans and barbarians eventually became indistinguishable, and eventually the concept of Roman identity vanished. This was noticed by contemporaries: an unknown fifteenth-century editor of a manuscript of the Liber Historiae Francorum attributed to them all being killed during the reign of king Clovis I (481–511).151 Thus, the later a study on identities stands in Late Antiquity, the blurrier the lines between Roman and non-Roman become. In the fourth century, it is proposed here, the distinction still remained relevant, and an individual’s perceived identity had tangible effects on the Roman world. The question is how we are to identify individuals who occupied the highest offices of the imperial military administration, and what categories they should be separated into to enable meaningful conclusions. For example, there is nothing to be gained in defining identities on the basis of legality, as every magister militum was legally a Roman citizen. Furthermore, the word ‘barbarian’ itself cannot be used in the modern day without controversy. Borrowed from the Greek word βαρβαρος, meaning a non-Greek person and

148 Mathisen 2019, a view also shared by Geary 1988. For a fuller discussion of the two sides to this debate, see: Stickler 2007, 497–98. 149 Goldsworthy 2009. 150 Meier 2020, 339–40 151 Mathisen 2018, 267–68, n. 74.

176

The Prosopography of the Magistri Militum

derived from the idea that the language of non-Greeks sounded like incomprehensible ‘bah-bah’, the term is an ethnographic category with a plethora of ancient and modern interpretations. Although some may caution us away from employing the term at all, the fact remains that the word was used by ancient people themselves. Different, modern categories would be placing our own biases on the ancient sources, and moreover would be unlikely to be any less connotatively loaded. Thus, we may continue to use the word here, but only with an appropriate exploration of how the terms ‘Roman’ and ‘barbarian’ have been used by both modern and ancient writers, and an explanation of what it means for a magister militum to be a barbarian. Extensive publications have appeared on Late Antiquity identities since its inception as an area of study in the 1960s by Reinhard Wenskus. ‘Ethnicity’, and its creation and dissolution (‘ethnogenesis’), denotes a situation where a group, for various reasons (socio-economic, political, perceived kinship ties, etc.) professes belief in or claim a shared common descent, and purport to share a common historical memory. This belief in a common past and descent was often used to create ‘in’ and ‘out’ groups.152 An extensive amount of work has been produced largely by scholars from the University of Vienna, who tend to view the Late Antiquity literary works that trace the history of a particular people, or gens, as representing, at least in part, historical accounts of the movement and development of the gens.153 These accounts include Jordanes’ Getica, Gregory of Tours’ history of the Franks, Isidore of Seville’s history of the Goths, Vandals, and Suevi, and Bede’s history of the English, among others. The opponents of the Vienna School center on Walter Goffart, who instead believes that these works are not historical accounts, but are instead mostly crafted pieces of literature, created according to traditional Roman styles and intended to reflect the desires of the burgeoning new kingdoms.154 These ideas are similarly shared by Michael Kulikowski.155 Ultimately, these theories of group identity do not fit the magistri well. At one point, many of the magistri might have been a part of a foreign gens, in particular those born outside the bounds of the Empire to barbarian parents. But, as the case of Silvanus shows, even if one’s parents belonged to a foreign gens, birth within the bounds of the Empire and service in the Roman military at the level of the magisterium was enough to break ethnic ties felt by those still in the gens.156 Furthermore, we have no insight into whether individual magistri personally believed in a shared descent, but only how Roman authors portrayed them. This complexity means that it would be impossible to separate the magistri with the term ‘ethnicity’.157 The general difficulties of ethnogene-

152 153 154 155 156 157

Wenskus 1961; Amory 1997, 14; Eriksen 1993, 18. Wolfram 1990; 1993; Pohl 1994. Goffart 1988; 2006; 2009. Kulikowski 2009; 2011. AM 15.5.15–16. Noble 2006, 16; Halsall 2018, 44.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

177

sis have been recognized, with Walter Pohl, a mainstay of the Vienna School, acknowledging that ‘ethnicity’ should only be used in a way that is neither overly restricted nor too “all-encompassing.”158 ‘Nationality’ is also often used to define ancient peoples, ascribing them to a group based on belonging to a people “defined by laws, institutions, religion, language, and customs.”159 Pohl writes that direct participation in a nation’s affairs defines one as belonging to that nation; on this basis none of our magistri would be considered a part of a foreign nation as they did not engage in the internal affairs of those groups during their time in the Roman military.160 As it refers to the shared customs, ideas, and institutions of a society, ‘culture’ exists on a broader level than ethnicity. One methodology used to explore and try to identify ancient cultures is through the archaeological record, especially burials and associated grave goods. For example, Joachim Werner saw Germanic traits from north of the Rhine contained in late fourth-century burials inside the Roman Empire.161 He identified this Reihengräberzivilisation with the laeti. Horst-Wolfgang Böhme reassessed this identification, instead seeing the graves as more likely to be of members of the foederati because the lavishness of some of the graves far surpassed that of what would be expected of laeti, and that many of the graves contained weapons, whereas laeti, like regular Roman troops, did not own their weaponry.162 These interpretations engaged historians and archaeologists for decades, until Guy Halsall, along with several other scholars, in a revision spanning multiple published works, cast extensive doubt on the ability of these grave goods to actually identify shifting cultural paradigms and the migration of peoples into the Empire.163 He has shown that the graves are actually not all that different from what had already been occurring within the confines of the Empire, and that even some of the ‘Germanic’ traits from north of the Rhine were themselves borrowed or inspired by Rome. Halsall’s proposed alternative is that the different burial practices are more related to social disruption, rather than identity. Halsall’s conclusion does not disagree with the reality of Germanic migration into the Empire, but that the Reihengräberzivilisation graves are not evidence for it. In a similar way, it has been thought that it might be possible to use archaeology to track the Gothic migrations as described in Jordanes’ De origine actibusque Getarum, from the Chernyakhov culture north of the Black Sea to the Sântana de Mureș culture in former-Roman Dacia.164 However, a more critical look reveals that the two cultures are not actually as linked as it is sometimes thought, and the links that remain can be

158 159 160 161 162 163 164

Pohl 2018, 33. Kaldellis 2007, 43. Pohl 1998, 4. Werner 1950. Böhme 1974. These works and accompanying commentaries are contained in Halsall 2010. Kleeman 2005.

178

The Prosopography of the Magistri Militum

explained through diffusion and transmission rather than mass migration.165 Furthermore, as discussed above, the likes of Goffart and Kulikowski argue that the stories of these migratory peoples should not be considered factual histories. Thus, it is rather difficult to actually identify culture, and especially so at the individual level. Moreover, it was entirely possible for barbarians to become Roman through a process of acculturation, especially if one served in the military, which continued to be an important path to cultural Romanization.166 Shane Bjornlie recognizes how Ammianus seems to see Roman identity as being particularly unique to the military.167 In the Res Gestae he identifies foreign barbari, provincial civilian cives, and martial Romani. Additionally, a grave inscription from Aquincum reads: “I am a Frank according to my origin, a Roman soldier according to weapons”.168 Thus a Frankish person becomes a Roman soldier. It is unclear, however, to what extent this can be extrapolated as the opinion of the entirety of the Roman world – Ammianus and this unnamed soldier were just that: soldiers. This probably influenced their perception of what it meant to be Roman.169 Furthermore, the Roman military’s culture was decidedly different to the rest of the Roman population, as it had incorporated many of the perceived animalistic traits that were a part of the traditional Roman perception of ferocious barbarians into something of a unique subculture.170 On those barbarian traits, one of the difficulties in assessing barbarian identities is that so much of the way barbarians are described is based upon classical archetypes and established ethnography. Ancient ethnographic climate theories go back at least as far as the fifth century BCE: Herodotus describes men in the north as having the wildest character, existing without law, and regularly resorting to cannibalism.171 Conversely, the southern deserts were increasingly occupied by dangerous animals, until one went south enough to find human-dog hybrids and other unknown fantastical beasts.172 Aristotle, in the fourth century BCE, described the north of the European continent as sparsely inhabited by wild men, who had courage but insufficient reasoning to develop any sort of political order.173 Meanwhile, in the hotter south, men had craftsmanship and political reasoning, but no courage. The further away from the Greek world one went, the wilder it became.174 Of course, the Greeks, being in the middle of the world,

165 Meier 2020, 143–44. 166 Mitchell and Greatrex 2000, xi. 167 Bjornlie 2018. 168 CIL III 3576 = ILS 2814: “Francus ego cives Romanus miles in armis”. 169 AM 31.16.9. 170 Halsall 2007, 101–110; 2018, 50; Meier 2020, 59. 171 Hdt. 4.106, 18.3; Timpe 1999, 27; Steinacher 2017, 35. 172 Hdt. 4.191.4. 173 Arist. Pol. 7.7 1327b23–36; Timpe 1999, 28. 174 Halsall 2007, 46.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

179

exist with the perfect mix of all positive elements and were both ferocious enough to be effective in war but civilized enough to form the best governments. Greek ethnography was readily adopted by the Romans: in the late first century CE, Tacitus described the people living north of the Empire in Germania, all the way to the land of the Fenni, who are fierce and poor, have no buildings nor metal nor clothes.175 Yet, he maintains a certain level of admiration for these people, writing that they are happier this way than they would be working the fields, and that they desire of nothing. Cassius Dio repeated similar Greek ethnographic ideas in the early third century when describing emperor Caracalla as having inherited from his parents “the fickleness, cowardice, and recklessness of Gaul, the harshness and cruelty of Africa, and the slyness of Syria.”176 Parallel ideas were maintained in the fourth century: Vegetius writes that those closest to the sun (Persia and Africa) have less blood and therefore lack the bravery for melee fighting. Conversely, northerners have more blood on account of the cold climate, and therefore are less scared of receiving wounds – although their lack of intelligence makes them less formidable. Therefore Vegetius recommends recruits for the army come from temperate climates where it is neither too hot nor too cold; just as the Greeks exist in the perfect climate, so too do the Romans, and men from these zones comprised Vegetius’ ideal army.177 Similarly, Ammianus writes that the soldiers from Gaul and Pannonia bray like wild beasts during battle.178 The Germans were also wild beasts, while the Huns had twisted and deformed bodies, and the Saracens would swoop like hawks in battle.179 He also levels ethnographic criticisms at specific people when it suits his narrative: Leo, the ally of the persecutor Maximinus discussed above, was “a Pannonian and a grave-robber, snorting forth cruelty from the grinning jaws of a wild beast …”.180 Additionally, being fully aware of reliance on past ethnography, Zosimus invokes Herodotus when giving his account of the Huns.181 Thus, classical archetypes guided perceptions of foreign people peoples and framed the ways in which Romans conceived of barbarians. All these terms are elastic, and any definition of them will ultimately be arbitrary.182 Guy Halsall has recently argued for solving the Roman-barbarian aporia by seeing ethnicity as multi-layered, and ethnic change as a process of acquiring new layers and reshuffling the importance of old ones.183 Thus, no one is wholly one identity to the exclusion of all others. Although this might be a more nuanced way to conceptual175 Tac. Germ. 46.3–4. 176 Cass. Dio 78.6.1. 177 Veg. 1.2. 178 AM 19.5.2. 179 AM 16.5.17, 31.2.2, 14.4.1; Barnes 1998, 107–113. 180 AM 28.1.12. 181 Zos. 4.20.3–5. 182 Mitchell and Greatrex 2000, xi. 183 Halsall 2018, 41–42.

180

The Prosopography of the Magistri Militum

ize identity, and could reconcile the differences between Roman, Roman military, and barbarian identities, this results in individuals having an incredibly complex, ever-changing, and essentially an incomparably unique monolith of identity. Such an approach would render any attempt at prosopographical analysis futile. Mischa Meier, in an expansive 2020 study of the so-called ‘Migration Period’, masterfully synthesizes the advances of the scholars coming from Vienna and of others, but avoids the pitfalls engendered by some of the more rancorous polemics. Meier agrees that legal categories of citizenship made it relatively easy to identify barbarians in the early Empire, and then it becomes easy once again during the Byzantine era, but during the fourth century the distinction is much vaguer.184 However, he still argues that the fourth-century barbarian was not a rhetorical, imaginary character, entirely created by Romans to intentionally or subconsciously ‘other’ individuals they disliked. Barbarian identity could be applied or not applied to a particular individual depending on circumstances, but when it was applied it did have consequences.185 The best example of this may be the magister militum Stilicho’s downfall. Occurring outside the scope of this work, from 395 into the first years of the fifth century, Stilicho was the powerful leader of the western Empire and was generally praised as a respected hero.186 His descent from a Vandal was comfortably ignored. Stilicho, however, eventually levied a massive tax on the Roman Senate to buy an alliance with Alaric, the leader of a large contingent of Gothic soldiers. This, along with Stilicho’s great personal power, angered the powerful members of the imperial court, who orchestrated the downfall and execution of the magister militum, justifying their actions as a necessary step to end the plots of the treacherous half-barbarian that wanted to see Rome burn.187 After these events, Orosius records that Stilicho came from the Vandals, an “unwarlike, greedy, dishonest, and deceitful gens”.188 Jerome calls him a half-barbarian traitor (semibarbari proditor).189 This kind of attack for political reasons was common in the ancient world.190 This does not mean we should doubt the veracity of Stilicho’s semibarbarus status, but we should understand that before his fall from grace, when he was the savior and protector of Rome, no one cared that Stilicho’s father was a barbarian. Once he was vilified and executed, this aspect of his identity was emphasized by his opponents. It did not matter that Stilicho no longer truly belonged to a foreign culture, nation, or ethnicity, or that he was not involved in the affairs of actual Vandals. Thus, one’s bar184 Meier 2020, 51. 185 Meier 2020, 87. 186 Meier 2020 84–85. 187 Meier, 2020 86–87. 188 “inbellis avarae perfidae et dolosae gentis” Oros. 7.38.1. 189 Jer. Ep. 123.17. John of Antioch ( Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev)) records his semibarbarus identity neutrally, but also claims that Stilicho’s power rivaled that of emperor Theodosius because of Stilicho’s marriage to Serena. 190 Elton 1996, 142–45.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

181

barian status could be downplayed or emphasized depending on one person’s feelings towards another. This is seen as well in Orosius’ testimonies of Radagaisus, the leader of a Gothic army that devastated northern Italy in 405–406, who Orosius calls a “pagan, barbarian, and a true Scythian”, compared to Alaric, who Orosius describes as a Christian, more like a Roman, and through his fear of God is inclined to spare men’s lives.191 A similar sentiment is expressed by Augustine of Hippo, who believes that Radagaisus was defeated because of his polytheism, while Alaric had more success and took Rome because he, as a Christian, spared the churches.192 Thus, if one was helpful and an ally of Rome they could be more like a Roman, but if one was a hated and violent enemy, they were a ‘true’ barbarian. The study of identity thus remains important for us to understand. Without investigating identities and their practical consequences, we will forever be missing a piece of the puzzle when understanding the course of ancient events. Furthermore, despite the difficulties in conceptualizing barbarian identity in the modern day, the Roman authors of our sources regularly labelled magistri as Goths, Franks, or other groups. Ancient works do not represent a dialogue between Romans and non-Romans, but only the voices of members of the Roman elite who used their writing to distinguish themselves from the barbarian ‘Other’.193 However, because they are our only sources of information, we must seek to understand their perspective and give individuals the labels that the Romans gave them. We must understand the nuances of how the Romans applied the idea of ‘barbarian’ to the magistri militum, and how they might have identified them as such. As we have seen already, Rome had a long history with foreign soldiers, but they also had a long social history with non-Romans. The city’s etiological myths included Aeneas coming to Latium from Troy, and that the city of Rome itself had been populated by foreign criminals and wanderers in its earliest days.194 Even in the fourth century, authors acknowledged the importance that foreigners played in the strength of the state.195 Despite this favorable mythos, the Roman emperors had long portrayed one of their primary duties to be defending the Empire and its people from these barbarians. Titles such as ‘destroyer of the barbarians’ (barbarorum extinctor) and ‘subduer of the barbarian gentes’ (domitor gentium barbararum) appear in the fourth century to describe the emperors.196 At the same time, it was the duty of the emperors to extend Roman civilization to the barbarians.197

191 “paganus barbarus et vere Scytha” Oros. 7.37.9. 192 Aug. Civ. 5.23. 193 Halsall 2007 56; Heather 1999, 235–36. 194 Livy 1.8.5–6. Virgil’s Aeneid describes the life of Aeneas. 195 Demandt 1989, 75–6. Although Demandt cites AM 27.4.32, the quote is located at 28.4.32. 196 AE 1969/70, 631; 1968, 602. 197 Bleckmann 2009, 200; Ladner 1976; and also Symm. Or. 16.30.

182

The Prosopography of the Magistri Militum

This complex duality can be seen in multiple other ways. For example, the magistri were some of the most powerful individuals in Roman society, and their acquaintance was eagerly sought out by prominent aristocrats. Symmachus and Libanius fostered friendships with barbarian magistri as with Roman magistri, but Symmachus also thought the barbarians were “savage people.”198 This mirrors other references to the perceived savage, deceitful, and violent nature of barbarians that riddle Roman literature: Prudentius, a poet of the fourth and fifth centuries, wrote, “what is Roman and what is barbarian are as different from each other as the four-footed creature is distinct from the two-footed or the dumb from the speaking.”199 The relationships Roman aristocrats developed with the magistri probably also played an important role in identifying them. Simple social conversations would probably lead to a discussion of birth places, parentage, and upbringing; all things that would contribute to establishing someone as a barbarian or Roman. It would of course be possible for barbarians to lie or cover up their past, if they chose. However, it seems that such things as an inability to navigate Roman social norms would have betrayed them as barbarians. Michele Salzman identifies through the letters of Symmachus that he used the Roman custom of amicitia to reinforce his relationship with the generals Bauto and Richomeres, but at the same time reveal their failures and ignorance of its practices, which subtly portrayed them as outsiders by emphasizing their sub-par understanding of Roman culture.200 At the same time, Jonathan McLaughlin uses the letters of Libanius to the magister Ellebichus to argue that there was not a significant cultural divide between barbarians and Romans.201 Yet, Libanius’ praise of Ellebichus is only because of the political and professional assistance Ellebichus can give Libanius.202 When Ellebichus does not support Libanius’ petition for a friend to join the senate, Libanius writes a new oration emphasizing Ellebichus’ barbarian origins and calling him the son of “Gaison”, the barbarian magister militum Gaiso who killed emperor Constans, thus implying that Ellebichus was cruel, brutal, and a potential emperor-murderer – the natural state of barbarians.203 Gregory of Nazianzus’ letter that strongly praises the magister Modares similarly only contains such approval because of Modares’ support for Gregory’s religious position.204 Thus the 198 “truces populous” Symm. Or. 1.19. For their correspondence with barbarian magistri, see: Lib. Ep. 866, 898, 972, 1007, 1024, 1525; Symm. Ep. 3.54–69, 4.1–14, 4.15–16. For their correspondence with Roman magistri, see: Lib. Ep. 318, 350, 520, 596; Symm. Ep. 3.70–3, 74–80. 199 “sed tantum distant Romana et Barbara, quantum quadrupes abiuncta est bipedi vel muta loquenti.” Prud. C. Symm. 2.816–7, trans. Thomson 1961. There are truly too many references to list, but for some other fourth- and fifth-century examples, see: Pan. Lat. 2.32.4; Sid. Ap. Ep. 1.7.6, 2.1.2, 5.5.3. Also see: Woolf 2012. 200 Salzman 2006. Also see: Mathisen 1997, 148. 201 McLaughlin 2014. 202 Cf. Pellizzari 2011. 203 McLaughlin 2014, 273. 204 See page 101.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

183

perception of a person as a barbarian could change depending on the person’s usefulness, but the standard state seems to have been dislike and distrust. This aversion for barbarians would not begin to disappear until the mid-fifth century when foreign peoples had begun to settle large swathes of imperial territory.205 Even then, adherence to well-known categorizations and stereotypes persisted beyond the point they became significantly outdated.206 In one study, Pohl identifies language and method of warfare as some of the traits often used by the sources to distinguish barbarian from Roman.207 There is no indication that any of our magistri could not communicate effectively in Latin or Greek, as it was for the Armenian leader of a unit of his own countrymen in Byzantine service, Gilacius, in 547, who knew neither language according to Procopius.208 As for following the traditional Roman method of warfare, they generally did, although Modares is recorded to have ordered his men to shed their armor and ambush their enemy, a decidedly un-Roman tactic.209 Other indications of barbarism identified by Pohl included clothes and hairstyles.210 It was perhaps the red or blonde hair common to Germans that would identify men as barbarian, even if they cut it in a more Roman-style. Thus, simple physical appearance may have helped Romans identify people as barbarians, and these things rarely survive in the sources. The term ‘barbarian’ could also be applied to individuals who had very different degrees of connection to foreign peoples. For example, both Modares and Gildo were members of foreign royal houses.211 Gildo’s father may have even been a Roman officer before becoming king of the Mauri, and some of Gildo’s brothers also fought alongside the Romans.212 Thus, the Mauri royal house was closely tied to the Empire. Conversely, the Frank Silvanus was so hated by his people that they would have killed him if they received the chance.213 Some barbarian magistri therefore had close ties to their people, and some had minimal.214 The feelings of the Franks for Silvanus might have also been that way to an extent because Silvanus was a semibarbarus, a man with one foreign and one native Roman

205 Ladner 1976, 23–4. 206 Meier 2020, 361–62. 207 Pohl 1998. 208 Procop. BP 7.26.24; Parnell 2017, 44. 209 Zos. 4.25.2–3; Eun. Hist. fr. 45.1. 210 Pohl 1998, 52. 211 For Modares, see pages 100–101. For Gildo, see: AM 29.5.6; Claud. de bello Gild. 335–36, 347, 390–1. His magisterial command is first attested in 393 by CTh 9.7.9, but Claud. de bello Gild. 153–55 indicates he was promoted in 386. Also see: Demandt 1970, 719; Matthews 1975, 245. 212 AM 29.5.2; Camps 1984, 185–86. 213 See page 130. 214 Silvanus still had relations with a number of other Franks in Roman service, such as Malarichus (AM 15.5.6–11).

184

The Prosopography of the Magistri Militum

parent.215 Parentage and ancestry mattered a great deal to the Romans. The legitimacy of children was regulated by the law, and what mattered was the legality of the union between the mother and father.216 To be a valid marriage (iustae nuptiae) a man and a woman both needed to have conubium, the capacity of legal Roman marriage.217 Children from a legal marriage were in the familia of the father, and thus Silvanus, whose father was a barbarian, took on a semibarbarus identity.218 One of the most aristocratic and best documented magistri militum in our time period, Stilicho, also fell into the semibarbarus identity because of his Vandal father.219 Stilicho was an extremely important fixture in both the eastern and western courts at different times of his career. He was a member of the imperial family through his marriage to Serena, emperor Theodosius’ niece, and his two daughters consecutively married emperor Honorius.220 Additionally, though outside our time period, Stilicho would be made consul twice.221 Although no other barbarian magister militum quite reached the social heights of the royal family, many were still prominent: Merobaudes held the consulship twice, several others once, Ellebichus was noted as a wealthy land-owner, and the patronage of Silvanus was sought out.222 As the fifth- and sixth-century king of the Italian Ostrogoths, Theodoric, remarked: “the poor Roman imitates the Goth, and the rich Goth imitates the Roman.”223 The high status of these barbarian individuals makes classification, both ancient and modern, more difficult. In summation, we might be tempted to conclude that it is impossible to place the magistri into discrete categories because of the difficult body of sources. Afterall, if there was no practical difference between Roman and non-Roman, what is the point? However, it seems as if there may have actually been a distinction with practical consequences on the magistri. While it seems that there were generally no significant hurdles to a barbarian soldier’s career, there may have been a social taboo prohibiting them from obtaining the imperial office. Many historians have written about this, some with reservations, others stating there can be “no doubt” about the prohibition’s existence.224 However, the evidence for this ban is sparse and contradictory, so it must be carefully

215 AM 15.5.16, 5.33; cf. Aur. Vict. 42.15; Epit. 42.11. 216 Buckland and Stein 1975, 101–6. 217 Gaius 1.56. 218 Gaius 1.55. 219 Jer. Ep. 123.17. See also: Oros. 7.38.1; Joh. Ant. fr. 187 (Müller) = fr. 212 (Mariev). 220 Stilicho’s first daughter, Maria, married Honorius in 398 (Zos. 5.4.1, 12.1) but she died in 408 and Honorius married Stilicho’s other daughter, Thermantia (Zos. 5.28.1; Olymp. fr. 3; Philost. HE 12.2; Jord. Rom. 332, Get. 154; Zon. 13.21). 221 Bagnall, et al. 1987, 334–35, 344–45. 222 Ellebichus: Lib. Ep. 898. Silvanus: AM 15.5.3. 223 “Romanus miser imitatur Gothum et utilis Gothus imitatur Romanum” Excerp. Val. 61; Amory 1997, 27; Demandt 1989. 224 Gibbon 1781, 63; Jones 1964, 327; Cameron 1969, 274; 1970, 38; Szidat 2010, 264–65. With some more reservations: McEvoy 2013, 54, 141; Lee 2015, 108 n. 45.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

185

weighed. Who occupied the imperial throne slowly but steadily spread further from the imperial center, much as did the pattern of incorporation into the Roman world traced at the beginning of this section: the first dynasty was the Julio-Claudians of patrician origin, next came the municipal Flavian dynasty, and then the colonial emperor Trajan, before finally the provincial Severan family.225 Without any codification of the succession process, could the imperial throne be open to anyone, even non-Romans? Raban von Haeling argues that it was open to them, and there was no ban restricting the holder of the office.226 He points out that Silvanus, a semibarbarus with a Frank father, became emperor, as well did the Goth Johannes (423–425). Von Haeling also addresses the evidence that arises from an account of a Catholic synod at Rome in 501, which records that king Theodoric quoted the eastern Alanic-Gothic magister militum Aspar, who might have gained the office in 424 and held it until 471, as having refused an offer to take the eastern imperial throne because he feared that he would “give rise to a tradition (consuetudo) in royal power.”227 This quote is of particular note as it presents evidence both for and against a restriction on the imperial office. On the one hand, the offer indicates that Aspar was considered to be a viable candidate; but his response that he would “give rise” to a new tradition implies that what was being offered had never happened before and was against the prevailing customs of succession. The matter is clarified somewhat by Procopius, who writes that Aspar could have become emperor if he only renounced his Arian Christianity and converted to Orthodoxy.228 Although the fourth-century Arian emperors Constantius II and Valens might contradict Procopius and imply that Arianism was not a bar to the imperial position, by Aspar’s time Arianism had decisively lost the contest for supremacy and only existed on the margins. The strength of Aspar’s identity as a barbarian is not clear; his father was also a magister militum and he was deeply intertwined in the politics of the east his entire life.229 Therefore he may or may not have been considered a barbarian, which could have added to the apparent previous restriction on individuals like him becoming emperor. Whatever the exact sum of factors was, this record is solid proof that some limitations on who could hold the imperial office existed, even if they were not consistently and clearly articulated. Because there are no references in Roman law codices to such a ban, at most we should only infer a social taboo, not a legal restriction. Although this lack of legal codification is significant, Philostorgius seems to make a clear statement on the matter. He writes that after the death of Valentinian II, Arbogast realized that he would not be 225 Kulikowski 2015, 138. 226 Haeling 1988, 90–95. 227 “timeo, ne per me consuetudo in regno nascatur” Acta synhodorum habitarum Romae 5. Croke 2005, 150 n. 10, guesses that this occurred in 457. For the term consuetudo, see: Ranft 1957, 379–390. For Aspar’s life, see: PLRE II, 164–69. 228 Procop. DV 1.6.3. 229 For Aspar’s father, see: PLRE II, 137–38.

186

The Prosopography of the Magistri Militum

able to make himself emperor because of his barbarian ancestry – a statement which evidences the importance of parentage to the Romans.230 Von Haeling’s argument that Philostorgius fabricated this statement as a moral insult does not seem convincing; even if he had, the fact that he could present this description to his audience as a real occurrence supports the fact that there was a general consensus that barbarian origins was an impediment to exercising imperial power.231 The source of Philostorgius’ information presents a problem, as its origins are unknown. As an easterner, it is also possible that he is projecting an eastern taboo. Despite these difficulties, it is the clearest ancient statement that barbarians were restricted from becoming emperor in the fourth century and no convincing reason to discount it has so far been presented. It must be addressed that there are three cases of barbarian usurpers which might, on first glance, seem to disprove the existence of a ban on barbarians becoming emperor. Magnentius, a usurper in the western Empire from 350–53, is recorded by Themistius, Julian, Aurelius Victor, the Epitome de Caesaribus, Zosimus, and Zonaras as a barbarian.232 Drinkwater, however, argues that these claims were the result of misinformation and smear propaganda, and Magnentius was probably actually a Roman.233 Another exception is Silvanus, the Frankish magisterial usurper in Cologne, who is used as evidence by von Haeling. While his ancestry is beyond doubt, his claim to the purple never extended further than the city and his soldiers, and he never gained wide acceptance from the Roman aristocracy. Furthermore, he was born in Gaul and his father was a lauded Gallo-Roman officer of Frankish origin.234 Thus, perhaps like Aspar, this made his leadership more palatable, although according to Philostorgius, the important factor was the barbarian status of one’s father. Both Magnentius and Silvanus rose to power within a few years of each other and both in Gaul, and this may indicate they both received important support from the powerful Frankish military clique that existed at that time.235 As for Johannes, also discussed by von Haeling, he seized power in Ravenna after the death of Honorius, but was eventually defeated by an eastern army led by Aspar. Nikephoros Kallistos Xanthopoulos, writing in the fourteenth century, is the only author to record Johannes’ ethnicity as a Goth.236 We are not fully aware of his own sources, and he is generally not considered a reliable source for the fourth century, so it is difficult to employ the example of Johannes with any significance. Furthermore, all these examples of potential barbarian emperors never achieved legitimizing acknowledgment from outside of the realms they controlled, 230 Philost. HE 11.2. 231 Haeling 1988, 94–95. 232 Them. Or. 3.43a, 6.80c; Jul. Or. 1.34a, 1.33d, 1.34d, 1.42a, 1.42b, 2.56c, 2.56d, 2.95c; Aur. Vict. 41.25; Epit. 42.7; Zos. 2.54.1; Zon. 13.6. 233 Drinkwater 2000, 131–45; Kulikowski 2016, 291. 234 AM 15.5.33. 235 Drinkwater 2000, 140. 236 Nic. Call. HE 14.7.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

187

and therefore cannot conclusively disprove the existence of a social taboo on barbarians in the imperial office.237 The power of this taboo is supported by negative evidence. There seem to have been multiple occasions where barbarians, or first-generation Romans, could have made themselves emperors, but they declined the opportunity. For example, the path to the throne was theoretically open to Merobaudes in 375, or Arbogast in 392. Their decision to raise other candidates was motivated by a plethora of factors, including avoiding civil war for Merobaudes. The failures of Vetranio’s and Silvanus’ attempts to make themselves emperors may have also been a factor. It was been previously argued, that they may have also decided to remain in powerful behind-the-scenes positions where they could control the emperors rather than become emperors themselves.238 However, in both Merobaudes’ and Arbogast’s cases, a final factor that helped them reach their decisions may have been that they would never have been accepted by the Roman aristocratic population because of a prevailing disdain for barbarian rulers. Although it might have been possible for Silvanus and Johannes, and maybe Magnentius, to rule in limited areas for brief periods of time, the aristocrats wielded broad soft powers through their economic wealth, land ownership, and through their almost complete monopoly on the highest positions in the civil administration.239 Thus their support was necessary for long-term, wide-ranging rule, and if they could not support a non-Roman emperor it would have been an impossibility for such a man to gain the imperial office. If this social taboo existed for Merobaudes and Arbogast, it might also help us understand other cases of imperial ambition, or lack thereof. In 354 Ursicinus was rumored to be seeking the throne, not for himself, but for his children.240 It seems possible therefore that he was considered a barbarian, but his children were Roman enough to be possible candidates. Furthermore, in the fifth century, the barbarian generalissimo Ricimer cycled through multiple adult emperors, even leaving the throne empty for months at a time. As mentioned before, this can be explained as Ricimer perceiving that there was more power to be had in remaining a magister militum and appointing emperors at will than becoming emperor himself. While this perception is not necessarily misguided, it must be acknowledged that some magistri did continue to accept a promotion to emperor, and those who did were Romans.241 Thus, it seems that becoming emperor was still potentially desirable for some high-ranking military leaders, but only a feasible path for those considered Romans. Further study of these fifth-century

237 238 239 240 241

Meier 2012, 206–7, esp. n. 66. Waas 1971, 16–19. Matthews 1975. AM 14.11.2–3. Oros. 7.42.2; Olymp. fr. 37.

188

The Prosopography of the Magistri Militum

magistri militum, both those that became emperors and those who did not, will contribute greatly to our understanding of this issue. All told, the evidence is convincing that there was a de facto ban on barbarians becoming emperors in the fourth and fifth centuries. Subsequently, if the most powerful magistri happened to be barbarians and they could not legitimize their power by usurping or inheriting the imperial throne, their positions would never be entirely secure. Instead, there would be a greater threat of destabilizing in-fighting and civil war as others tried to take their dominant positions. Thus, understanding the changing proportion of Roman to barbarian magistri, even if it is only an exercise in speculative plausibility because of the difficult source body, is important for us to understand power dynamics in the late Empire. This social taboo on succession can also help us define what is meant by barbarian: an inability to occupy the imperial throne effectively and legitimately. Relying on the testimony of the sources will be the best method of separation – those identified as barbarians by the sources are barbarians, and if there are others that share the same described characteristics, they are also barbarians. If we are considering the support or lack thereof one might have in the imperial office, this external perception is paramount. Hopefully the preceding descriptions of barbarian and Roman magistri, make it clear that barbarian magistri did not act any differently to Roman magistri. Finding that difference is not the intention of this investigation. What is clear is that barbarians were treated differently – the external perception of a magister militum’s identity led to changed behavior by others. Thus the distinction is not internal, but external. ‘Barbarian’ is not truly an antonym of ‘Roman’. These are not two distinct genetic, cultural, or linguistic groups, and thus using them in juxtaposition will always invite criticism. Yet, presently there are no better terms that do not come without their own problems. Furthermore, these terms retained their significance to contemporaries.242 Finally, as the survey presented below hopefully shows, scholarship would not be best served by leaving the ‘barbarization’ debate in its current state. Thus, this author hopes that it is still possible to use the terms ‘barbarian’ and ‘Roman’ as long we follow recent advice to not “accept [them] too uncritically” and remove any pejorative senses to the words.243

242 Meier 2020, 362–63, 1095. 243 Conant 2014.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

189

4.3.2 Survey There have been two previous studies conducted to assess the degree of barbarization in the fourth-century magisterium.244 The first was undertaken by Hans Teitler in 1989, and looked at only the number of German magistri. Hugh Elton revisited the topic in 1996 and expanded the view to all magisterial identities. We will here discuss the methodologies of each study to assess the reliability of their results, and where it might be possible to make improvements that render the resulting statistics more accurate. This will lead into the discussion of this project’s different methodological processes and results. Rather than attempt new methods of categorization or identification, it has been found that the previous studies have left significant room for improvement, and building upon their basis will offer a useful extension of knowledge and understanding of the barbarization of the later Roman army. Teitler’s study was conducted as part of an investigation into whether the primary sources over-represent Germans in the Roman world. He states that he used the fasti of magistri militum in the Prosopography of the Later Roman Empire to find that there were sixty magistri from 342 to 395.245 However, the PLRE only contains fifty-six magistri in its appendix.246 Because of this confusion, we cannot be entirely certain who he included or excluded from his calculations. Fortunately, he does include a list of all those he considers ‘Certain Barbarians’: Silvanus, Malarichus, Bauto, Arbogast, Richomeres, Agilo, Modares, Stilicho, Nevitta, Merobaudes, Butherichus, Gildo, Victor, and Bacurius.247 There are three differences between these and my own results. First, I judged it more likely that Bacurius was not a magister militum, but only rose as high as the position of comes domesticorum, and therefore should not be counted.248 Furthermore, there is no certainty from the primary material on the identity of Merobaudes or Butherichus. Teitler included these two into the ‘Certain’ category because the authors of the PLRE single them out as particularly likely to be Germans based on their names, but this is hardly evidence beyond doubt.249 Therefore these men belong in the category of ‘Probable Barbarians’, which Teitler populated with ten other men based on their names. He does not provide a list for this group, so it is difficult to comment

244 Nicasie 1998, 102–3 also briefly summarized Teitler’s analysis in 1998 and added to it some of his own calculations, but this was not a significant enough development in the study of barbarian magistri militum to warrant a full discussion here. 245 Teitler 1989, 53. The Prosopography of the Later Roman Empire will be referred to as the PLRE from here onwards. 246 PLRE I, 1112–14. 247 Teitler 1989, 54. 248 See note 119n325. Bacurius is also not counted as a magister militum in the PLRE’s fasti, so we can conclude that he is at least one of the differences between Tetiler’s total and the total in the fasti. 249 A greater exploration of onomastic evidence is placed below.

190

The Prosopography of the Magistri Militum

on the plausibility of his deductions.250 Overall, he calculated that roughly 40 % of the fourth-century magistri had barbarian origins, and he guessed that the real number was probably lower because of the uncertainties in the onomastic method of populating the ‘Probable Barbarian’ category. However, the opposite may be more likely. Rather than some Romans being hidden behind German names, Germans might be obscured by names with Latin or Greek etymology. For example, the Alamanni general Serapio, whose name is Graeco-Egyptian in origin and therefore would make a purely onomastic analysis conclude he was Roman, was formerly known as Agenarichus, a Germanic name.251 This adoption of Greco-Roman names can also be seen in the papyrological record. A group of Germans can be found in Egypt, most likely some bucellarii or foederati recruited from north of the limes and settled in Egypt with their wives and children.252 The Germans clearly intermarried with Romans, or adopted Roman names, as one Onnakea, a Germanic name, was married to an Anastasios, a common Greco-Roman name. Moreover, the sons of this group seem more likely to adopt Greco-Roman names as the group assimilated into the “onomastic landscape of the region”.253 This potential for re-branding thus suggests that the barbarians might be under- rather than over-represented in onomastic data.254 Elton’s work was the first to approach the ethnic identities of all the magistri militum comprehensively to assess whether it is possible the Roman army was weakened by increased barbarization in this period. Elton’s process included three misleading methodologies that are not significant problems on their own, but each slightly skewed the statistics in the same direction, which when aggregated, resulted in an over-representation of Roman magistri. Firstly, Elton decided to emphasize geographical provenance, which meant that any individual born within the limits of the Empire, even if they had barbarian heritage, was considered to be Roman.255 This is not an inherently incorrect choice, and was a method used previously to identify Romans.256 But status was not determined solely by place of birth. As stated above, the sources identify both Stilicho and Silvanus as barbarians despite their place of birth. Thus, this method of classification is an arbitrary use of geography not based on the sources, and placed more magistri into the Roman categories. The second methodological problem stems from the decision to create an additional category, labelled ‘Others’. Elton populated this group with the non-European barbarians, such as those from Africa or Persia. He offers no explanation for why he chose to do this, or why he did not split the Romans along similar lines, such as east250 251 252 253 254 255 256

Teitler 1989, 54. Teitler 1989, 53; AM 16.12.25. Mallobaudes, see: PLRE I, 539. Carrié 2003, 85–86. Carrié 2003, 86. Liebeschuetz 1990, 8; Elton 1996, 146. Elton 1996, 146. Teall 1965.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

191

west, or European and non-European. For the fourth century there were only two in the ‘Others’ category, 3.33 % of Elton’s total. While this is a small percentage, when the total volume of verifiable data is so low, even a few percentage points become important, and splitting the barbarian categories in this manner resulted in the appearance of lower numbers of total barbarians. Furthermore, it reveals an incorrect equivalence of ‘barbarian’ with ‘Germanic’. The third and greatest problem with Elton’s methodology stems from his process of populating the ‘Probable Roman’ group. Elton accepts a total of sixty magistri for the period of 350 to 399.257 He provides no list of whom he included or excluded. Once he had filled the ‘Certain Barbarian’, ‘Probable Barbarian’, ‘Others’, and ‘Certain Roman’ categories, he allocated all the remaining men into the ‘Probable Roman’ group. This included no critical assessment of the evidence, nor any justification for why they belonged in this category. For example, the sources contain no indication about the origins or ethnic identity of the magister Addaeus, nor is it possible to make a judgement based on information like contemporaries who share his name. Nonetheless, Elton includes Addaeus in the ‘Probable Roman’ category. This causes this group to be significantly larger than any other. Elton acknowledges that some of the individuals in this category are likely to have been barbarians, but dismisses this problem as insignificant.258 The research conducted during the course of the present study, however, will show that this in fact had a significant effect on Elton’s results. To correct the issues identified in these previous studies, the procedure employed by this project has been to divide the magistri into similar categories as Elton: ‘Definite Barbarians’, ‘Definite Romans’, ‘Probable Barbarians’, ‘Probable Romans’. A fifth group has also been added, the ‘Unattested’. A full categorization of each individual assessed in this project can be found in Table Three below. Of course, identity was not truly so starkly separated; identity was multi-layered and could change over time.259 While these groupings are imperfect, individuals must be assigned into more or less practical categories if a prosopographical study is to be conducted, and utilizing these particular divisions will enable us to continue the investigation into the barbarization of the late Roman army. The two ‘Certain’ categories of this table are quite straightforward: if the sources specify an individual was a Roman or a barbarian, they are placed into these groups. The two ‘Probable’ categories will always be more controversial and require an explanation of the methods used to generate them. One method of categorization that has not been used to study the barbarization of the magistri militum before is the utilization of the apparent taboo on foreigners becoming emperors to place men into the ‘Probable Romans’ category. If someone was considered to be a viable emperor, such 257 Elton 1996, 148. 258 Elton 1996, 146. 259 Bjornlie 2018, 82.

192

The Prosopography of the Magistri Militum

Table Three The Perception of Identities of the Fourth-Century Magistri Militum in the Sources. The gens of the barbarians, if known, is recorded in parentheses. If the evidence for an individual’s identity is not clearly stated in his PLRE entry and has instead been developed from inference or complex synthesis, a symbol will dictate how this conclusion has been reached, as follows: a Was considered to be a viable imperial successor and was therefore probably Roman b Became emperor and was therefore probably Roman c Son became emperor and was therefore probably Roman d Name is only used by other known Romans e Indicated to be a Roman from descriptions f Germanic name g Iranian name Certain Romans

Probable Romans

Certain Barbarians

Probable Barbarians

Unattested

Marcellus

Hermogenes d

Silvanus (Frank)

Salia f

Arbitio

Equitius

Bonosus

Agilo (Alaman)

Ursicinus

Timasius

Eusebius d

Nevitta

Gaiso f

Abundantius

Sallustius d

Malarichus (Frank)

Gomoarius f

Jovinus

Vetranio

Julius

d

Barbatio

f

Lucillianus260

Victor (Sarmatian)

Dagalaifus

Romulus d

Modares (Goth)

Arinthaeus f

Severus 8

Bauto (Frank)

Merobaudes

Sabinianus e

Richomeres (Frank)

Sapores g

Andragathius262

Lupicinus d

Gildo (Mauri)

Ellebichus f

Promotus

Severus 10 a

Arbogast (Frank)

Rumoridus f

Flavius Theodosius 3 c

Stilicho (Vandal)

Butherichus

b

d

f

Traianus f

Majorianus261

Quintinus f

Addaeus

Saturninus d Sebastianus a Theodosius 4 b Nannienus e

260 Lucillianus was the father-in-law of Jovian, but because Stilicho was the father-in-law of Honorius, this does not preclude Lucillianus from being a barbarian. 261 Majorianus was the maternal grandfather of emperor Majorian, but because the maternal grandfather of emperor Theodosius II was the barbarian Bauto, this does not preclude Majorianus from being a barbarian. 262 Zos. 4.35.6 says that he came from the Black Sea, but does not make it clear whether this was the Roman or barbarian parts.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

193

as Equitius or Severus, or actually became a recognized and legitimate emperor, like Theodosius, the taboo makes it likely that they were Romans. We can also be confident that their immediate relatives, such as Flavius Theodosius père, were Romans. A difficulty arises because the barbarian stigma could disappear from the family line in successive generations. As stated above, Ursicinus and Stilicho may have been illegible for the imperial office themselves, but they were both suspected of seeking the imperial throne for their children rather than themselves.263 Another problematic case is that of Aelia Eudoxia, daughter of Bauto, wife of emperor Arcadius, and mother to future emperor Theodosius II. We remain unaware of her mother’s identity; if she was a full barbarian, that would make Eudoxia also a barbarian, and thus emperor Theodosius II would be a semibarbarus. There is, however, the likelihood that Bauto, like so many magistri, married a Roman woman and thus Eudoxia herself was a semibarbara and Theodosius II’s identity was acceptably Roman. Similar can be assumed for Majorianus, maternal grandfather of emperor Majorian. We can thus tentatively deduce that after enough generations, ‘barbarism’ was considered sufficiently diluted, although without more ancestral records it is difficult to ascertain whether this was universal, or whether something like Stilicho’s high degree of Romanization and closeness to the royal family quickened this process. We can only make the best attempt possible at identification, and we must remember that the conclusions drawn here are only for a specific snapshot of people and time and will not necessarily hold true outside of the fourth-century magistri militum. Despite these difficulties, using this method to determine likely Roman identities remains the best possible conjecture based on the extant evidence. Onomastic analyses, as employed by Teitler, was also used to gauge the origins of the magistri. For example, we can see that the name ‘Merobaudes’ has the same termination as ‘Mallobaudes’, who was a known German.264 Comparable relationships can be seen between the magister ‘Ellebichus’ and the German ‘Allobichus’, or ‘Dagalaifus’ and ‘Gadalaifus’, among others.265 A similar methodology can be applied to create a Probable Roman category. For example, there are 42 entries in the PLRE for individuals with the name ‘Eusebius’, with Eusebius 39 being our magister militum.266 The sources do not provide his place of origin, but we do know that Eusebius 5, 10, 12, 13, 14, 16, 17, 18, 24, and 40 were all native Romans. Likewise, when looking at the entries for ‘Severus’ in the PLRE, we find 31 individuals with this name.267 Severus 8 and 10 were both magistri militum. We know that Severus 10, as well as 30, were Romans, and there

263 Zos. 5.32.1–2; Olymp. fr. 5.2; Philost. HE 12.2; Oros. 7.38; Marc. com. s. a. 408; Soz. 9.4. 264 AM 31.10.6. 265 These relationships can be identified by Schönfeld 1911, which has been an invaluable source for this project. 266 PLRE I, 301–8. 267 PLRE I, 831–38.

194

The Prosopography of the Magistri Militum

are no known barbarians with the name Severus.268 On the basis of these analyses, we can conclude that the names ‘Eusebius’ and ‘Severus’, along with some other names used by the magistri, were typically only used by Romans, and thus they fall into the ‘Probable Roman’ category. It is apparent that this is not a foolproof method; as mentioned, Serapio’s original name was the more Germanic Agenarichus, and in Egypt it is clear that some non-Romans, and especially their children, adopted Greco-Roman names.269 This shows that barbarians did not always have barbarian names, and a Roman name did not make one a Roman. Furthermore, a barbarian trying to appear as a Roman might very well take on a typical ‘Roman’ name like Severus or Eusebius. One of the more difficult cases to make a conclusion on is that of the magister militum Ursicinus. His name is probably related to the Latin word ursus, meaning bear, and the related word ursinus. The name appears for Marcus Aurelius Ursicinus, a praetorian guardsman in the third century; Valerius Ursicinus, brother of Aurelia Ursicina; and on the Iberian Peninsula, there was an Ursicinus, son of Siluius Paternus and P[ublius?] Ursicinus.270 The name Aurelius Ursicinus also appears frequently in the Hoxne Hoard, the largest collection of silver and gold artefacts ever discovered in Britain.271 This hoard was probably buried in the early fifth century, but many of the items themselves date earlier. This name was also shared by high-ranking officials in the civilian bureaucracy in the fourth century.272 None of these individuals, however, have any clear features that identify them as Romans or barbarians, although the setting up of funerary inscriptions in Iberia, a significant distance from barbaricum, would be more likely done by Romans than not. Positions in the civil bureaucracy rarely went to barbarians, and thus these might have also been Romans, although we can be even less certain about the praetorian guardsman. There is also evidence for a barbarian identity for Ursicinus. As previously stated, he was suspected of seeking the throne for his sons and not himself, even though he was a popular and competent general, at least according to Ammianus’ testimony, and thus in potentially a prime position to gain the throne.273 Thus he may have fallen on the wrong side of the social taboo stopping barbarians becoming emperor. The most solid evidence for a barbarian identity for this magister militum is that one of the Alamannic chieftains who fought against Julian in 357 was named Ursicinus.274 Thus this name could clearly be used by definite barbarians – but it could be a case like that of Serapio where a Roman name was taken by a foreigner. Onomastics like this are thus at most an effective guide to the cultural aspirations of parents at the time of a person’s 268 Aur. Vict. 40.1. 269 See page 190. 270 CIL II/2, 14, 1632, III 5449, 11036; AE 1997, 877. 271 Johns 2010, 166–73, 168, 206, 263–64. 272 Ie. Ursicinus 6 and Ursicinus 7, see: PLRE I, 987. 273 AM 14.9.1, 11.2–3. 274 AM 16.12.1, 18.2.18–19.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

195

birth, or at the time they changed their name. We can understand that Mederichus, the brother of the Alamanni rex by the name of Chonodomarius, had been kept as a hostage in Gaul and inducted into the traditional religious cults, or ‘mysteries’, of the Greek-speaking eastern Mediterranean, most probably that of the Egyptian deity Serapis, and thus changed his son’s name from the Germanic Agenarichus to the Graeco-Egyptian Serapio.275 We have no such knowledge on the Alamannic chieftain Ursicinus’ parents’ cultural aspirations. Ultimately, we must consider which weighs heavier, a small amount of much more certain evidence (the Alamannic chieftain Ursicinus, and the rumor of seeking the throne for the magister militum Ursicinus’ sons, à la Stilicho) or a large amount of weak evidence (archaeological references to Ursicini that might be Romans). Scholars could come down on either side of such a question. I think that we should consider the temporal nature of the evidence; the epigraphic evidence dates to the third century, and it has been noted that the name Ursicinus seems to be “old-fashioned” in Roman circles by the fourth century, but was clearly still in use by Alamanni.276 Thus Ursicinus has been placed in the Probable Barbarian category. In summation of this methodology, it has been acknowledged by modern scholars to be flawed, but it is nevertheless agreed that onomastic analysis still represents an important source of information, because is it often the only instrument we have.277 The evidence is less certain about many other magistri, but we can look for context clues in the sources. For example, when describing Gratian’s campaign in 378 against the Lentienses, Ammianus writes that Nannienus and Mallobaudes were given command, and he distinguished Nannienus as virtuous and Mallobaudes as a Frank.278 While this is not an explicit statement that Nannienus was a Roman, the juxtaposition is a strong implication that he was, and thus he is placed in the ‘Probable’ category. The final group in the current survey is labelled as ‘Unattested’ and comprises those individuals where no information can be deduced about their identity. This group includes some quite important and influential individuals such as Arbitio, and others that only feature briefly in the sources, such as Quintinus or Majorianus. All these individuals were of course identified as something, and if it could be discovered that this group was mostly Roman or mostly barbarian it would greatly change the results, but this will likely always be a mystery. The argument could be made that the ancient sources were more likely to record the barbarian rather than Roman identity of an individual, and this appears somewhat true. As can been seen in Table Three, eleven barbarians are recorded specifically as such in the sources, while only four Romans are recorded as clearly. This could lead one to believe many of the ‘Unattested’ were 275 276 277 278

AM 16.12.25. Johns 2010, 168. Pohl 1998, 10; Parnell 2017, 51. AM 31.10.6.

196

The Prosopography of the Magistri Militum

Romans and ancient authors simply did not feel this needed to be specified. However, we must wonder if this is actually correct, seeing as there were also eleven magistri who were probably barbarians but the sources do not explicitly say so. Thus, we cannot be certain the sources were more thorough in recording barbarians, and therefore cannot be too overreaching in making assumptions about the ‘Unattested’ category. Using the detailed breakdown of the results in Table Four, we can see that, overall, the ethnic identities of the magistri is a quite well-documented topic. Of the fifty-two individuals considered to be magistri, only eleven individuals, i. e. twenty-one percent, have no information that we can use to deduce their identity. Table Four also shows that the data is split fairly evenly across all most categories, the exception being the ‘Definite Romans’. Table Four A comparison of Teitler 1989, Elton 1996, and the present study’s results on the distribution of Magistri Militum by identity (barbarian and Roman). Teitler (1989)

Elton (1996)

Bendle

Definite Romans

23 %

8 %

Probable Romans

43 %

29 %

Definite Barbarians

23 %

17 %279

21 %

Probable Barbarians

17 %

17 %

21 %

60 (342–392)

60 (350–399)

52 (342–392)

Unattested Total

21 %

In Figure Four below, which displays the data without the ‘Unattested’ category, we can see that there are slightly more in both ‘Barbarian’ categories combined compared to the ‘Roman’ ones, with the lowest category being ‘Definite Romans’ by a large margin. This therefore suggests that a magister militum in the fourth century was approximately as likely to be a barbarian as he was to be a Roman, and therefore the highest rank of the Roman military was thoroughly barbarized.280 If we only look at the more certain data, the results are quite different: there are nearly three times as many ‘Definite Barbarians’ than ‘Definite Romans’, however I would caution against anyone taking this conclusion as the reality, as this is surely a result of the inconsistent sources. The combined data that implies a similar number of Roman and non-Roman magistri militum in the fourth century, with maybe only slightly more barbarians, is the most enlightening.

279 This number combines Elton’s ‘Barbarian’ and ‘Other’ categories. 280 Waas 1971, 6, came to this same conclusion, although without explaining his methodology or providing a list of names, only stating that there were twenty barbarians and forty-four total magistri militum in the same time period.

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

27%

Definite Romans

10%

Probable Romans

36% 27%

197

Definite Barbarians Probable Barbarians

Fig. Four Pie Graph of the Identities of the Magistri Militum. Own illustration.

4.3.3 Discussion The calculations carried out and displayed above were reached by assigning an equal value to all the magistri. This means that the magistri with the least power and influence, such as Malarichus or Rumoridus, receive the same value as the most important and powerful, such as Arbogast or Stilicho. The magistri Sebastianus and Theodosius, who were only in the office for a few weeks or months, receive the same value as Victor or Stilicho, who were both magistri for eighteen and seventeen years, respectively. This method, while illuminating, will be further improved by adding the context that is necessary to fully understand the nature of the magistri.281 Therefore we will now explore some of the traits this study has identified that were shared by the magistri who had the most profound effect upon the changing nature of the magisterium. Figure Five is a chronological timeline diagraming the identities of the magistri from the 340s to 395. As can be seen, the magistri were fairly evenly divided between Romans and barbarians before 375. There were seven barbarian magistri in the west and four more in the east, for a total of eleven identifiable barbarians.282 There were slightly more Romans during the same period and they were more evenly divided between the two halves of the Empire, with six in the east and seven in the west.283 Thus, both Roman and barbarian magistri were common in the early decades that have been reviewed here. While there were slightly more Romans, only four consulships were given to them while the barbarians received six, perhaps indicating greater importance for the non-Romans.284 However, Arbitio was the longest serving and most influential magister of this period as a result of his close association with emperor Constantius,

281 Ie. Demandt 1970, 785. 282 These were Salia, Ursicinus, Gaiso, Silvanus, Gomoarius, Agilo, Nevitta, Malarichus, Dagalaifus, Victor, and Arinthaeus. 283 These were Hermogenes, Bonosus, Sallustius, Eusebius, Vetranio, Romulus, Marcellus, Severus 8, Sabinianus, Lupicinus, Equitius, Severus 10, and Flavius Theodosius 3. 284 Only two were given to magistri in the ‘Unattested’ category.

Fig. Five The Ethnic Identities of the Magistri Militum. Own Illustration. Bright green indicates ‘Definite Romans’, dark green for ‘Probable Romans’. Bright red is used for ‘Definite Barbarians’, and dark red for ‘Probable Barbarians’, and finally white shows ‘Unattested’. Those above the white center line served primarily in the east, while those below it served in the west. As with Figure Three, this image is not intended to be an absolutely accurate reference for the dates and places the fourth-century magistri were in office, but only to help illustrate the conclusions of this discussion.

198 The Prosopography of the Magistri Militum

The Identity of the Magistri Militum: Barbarian, Roman, or Something in Between?

199

but we remain unaware of his identity, making it difficult to conclude which group may have been more influential. This changes during the later decades of the fourth century. The period from 375 to 395 saw an increase of only a single extra barbarian magister, but a significant decrease in the number of Roman magistri, only four in the east and three in the west. At seven, the number of barbarians in the east surpasses that of only five in the west. This is attributable to a higher total number of eastern magistri during this period, rather than to differing preferences for barbarian employment across the two halves of the Empire. Although the remaining seven in the Unattested category adds an element of uncertainty, there are a significant number of both barbarians and Romans in the Definite categories. This trend therefore holds true even if we only look at the identities we are most confident in. Thus, the evidence suggests that the ratio of barbarian to Roman magistri militum changed slightly in favor of barbarians over the period examined. Furthermore, the most powerful barbarian magistri existed towards the end of the period. While most of the magistri were influential, some of them, like Promotus or Butherichus, did not have the power and influence to protect themselves from the attacks of other individuals or groups.285 Others, such as Rumoridus, were unable to gain enough personal power to operate with any discernible agency. Even some of the more powerful magistri, such as Arbitio and Bauto, only had limited successes in furthering their personal agendas. Merobaudes, Arbogast, and Stilicho, comparatively, stand out as having obtained for themselves a massive degree of personal power and agency. These magistri were at various times able to take initiatives to further their own plans and goals. As discussed in the previous prosopographical investigation, one feature that may have contributed in a self-sustaining manner to the power of these magistri was their long careers. We can now add that their barbarian identities may have contributed to this. The careers of the magistri Timasius and Abundantius made it clear that long tenures were not exclusive to the barbarians, but there may have been reasons that the barbarian magistri were predisposed to such long careers. If there was a taboo on barbarian magistri becoming long term legitimate emperors, then the magisterium was the peak of their career path. Conversely, Romans could aspire to progress into the imperial office. Such was the case with Theodosius, who became a magister in 378. If he had been a barbarian and was forced to remain in the magisterial position under Gratian, he could very well have presented a strong challenge to Merobaudes’ dominance of the west. However, he advanced to become the emperor of the east and this left Merobaudes alone in the powerful position he would retain for another eight years. Similarly, if Stilicho was also prohibited from becoming emperor, then all he could do to advance his own position was continue to gather power as magister militum, and

285 Soz 7.25.3; Ruf. HE 2.18; Zos. 4.51.3; Claud. de cons. Stil. 1.94–6, in Ruf. 1.316–17.

200

The Prosopography of the Magistri Militum

there is some indication he did so by reforming the western military to concentrate more of the armies under his official control.286 Thus, while having a barbarian identity did not give one a preternatural tendency to have a long career, it may have removed the path to a higher position, making barbarian magistri more likely to stay in the office, and make themselves more powerful and influential. If so, increased barbarization of the highest ranks of the Roman military did have significant destabilizing effects on the Roman world. In summary, contrasting with the conclusions drawn by previous studies, the prosopographical research conducted here indicates a significant trend towards barbarization in the highest rank of the Roman military, especially in the west and in the last decades of the fourth century.

286 Mommsen 1889, 265.

Conclusion This book has been an attempt to show that the instrumentality of individuals played a critical role in the changing tides of power in the fourth-century Roman Empire. Interpersonal relationships, factional alliances, and unique approaches to crisis and opportunity guided the nature of the magistri militum in this period. Previous works on the fourth-century generals have assessed the magistri as an institution, and have focused on questions about the creation of the office, the development and implementation of new positions and titles, and its internal hierarchy. These institutional perspectives have revealed a great deal, but they are only part of the picture. By synthesizing narrative archontology, network analysis, and prosopography, this monograph has gone beyond previous views and argued that there were certain decisive moments in the fourth century that were taken advantage of by particular magistri to vastly increase the powers they wielded and their role in the Roman military-political landscape. Looking at the first magistri that appeared in the historical record, Chapter One argued that the strong, powerful emperors like Constantius and Julian were able to maintain control over their generals. The magistri were kept in their intended role of commanding Roman armies and carrying out the orders of the emperors. This included defending the Empire from Germanic, Eurasian, and Persian attacks, overthrowing usurpers, and carrying out the emperors’ religious policies. Two magistri attempted to usurp imperial power, a strategy that had become common for high-ranking military men in the third century. However, both Vetranio and Silvanus were somewhat forced into usurping because of circumstance, and neither had significant success before being defeated. It has been proposed that, as Constantine probably intended, the magistri had difficulty creating the necessary partnerships with civil officials and lower ranked officers.1 Vetranio’s officials betrayed the usurper’s defenses to Constantius, and through network analysis it was explored how even though Silvanus had a large alliance with some extremely loyal members, the smaller and more cohesive network of Ursicinus was able to infiltrate and break it up, ultimately killing the rebel and collaps-

1

Lee 2015, 103 n. 17.

202

Conclusion

ing the alliance. Many other magistri, however, formed positive, working relationships with other military officers and officials in the civil sector, indicating a difficulty in forming alliances may not have been the critical factor that disinclined later magistri towards rebellion. Other usurpers continued to arise from other strata of the military hierarchy, in particular the palatine units that were close to the imperial courts. The decision to raise middling officers or incapable children as emperors during the later fourth century indicates that the powerful individuals and groups in the Empire did not want someone who could challenge their positions. These palatine officers were thus able to form the necessary connections for an imperial challenge, but were not so high-positioned and powerful that they would completely dominate the court after their accession and turn acceptance groups against them. Concurrent with Vetranio’s and Silvanus’ rebellion, other magistri were exploring new methods of increasing their power. Arbitio secured a powerful position as Constantius’ primary court magister militum, and from this position he was able to advise and influence the emperor, yet he also did not find significant success. Instead, Constantius and after him Julian remained in powerful, central positions, and they exerted solid control over their generals. It was noted, however, that following the death of emperor Julian in 363, the magistri attempted to utilize their position to dictate the appointment of a new emperor. Their need to find a candidate that suited their desires squandered the opportunity in the end, but this moment shows they were reconsidering what their role in shaping imperial power could be. Chapter Two showed that in the span of three decades, the magistri utilized key opportunities to greatly change the nature of their office, especially its relationship to the emperors. Only a few months after the wasted opportunity of Julian’s death, a similar prospect arose with the death of Jovian. The magistri and other officials were able to orchestrate the appointment of an officer of middling rank, the tribunus of the schola secunda Scutariorum, Valentinian. He was not too powerful nor was his position too secure, and thus he needed to favor those who had raised him and continued to support him. He increased the social rank of the magistri to make it equal to that of the civil administrators. His general lack of interest in internal affairs also gave the magistri and other officials great freedom to maneuver against each other, and they manipulated the information he received to further their own interests. We saw in Chapter Three that this period gave rise to powerful, competitive social networks that engaged in often deadly fighting with one another. By applying mathematical formula to a sociogram of relationships, we were able to see that the magister Merobaudes had a greater degree of connectivity than other contemporary magistri. He thus occupied an extremely influential and central position in the western administration, and he was able to take advantage of Valentinian’s unexpected death in 375 to take a dominant role in creating an emperor. Merobaudes’ possible barbarian heritage may have precluded him from becoming emperor, although the evidence for this social taboo is not certain. He may have decided that appointing a puppet and remaining as

Conclusion

203

sole magister militum in the west would afford him just as much power as ruling himself, and would remove the risk of losing his life in civil war like Silvanus. What is clear is that Merobaudes had learned from the mistakes of the previous succession crises, and he acted quickly to find a candidate that was acceptable to the rankand-file soldiery as well as the incumbent emperors, and one that would not overrule his own power. In stark contrast to the 364 succession, where the infant son of emperor Jovian was not considered a viable candidate, Merobaudes’ raised the four-year old son of Valentinian I as emperor. By raising Valentinian II, Merobaudes set the Empire down a new path, one where the emperor did not need to be the military leader that previous emperors had been, either in truth or symbolically. Instead, it became acceptable for the emperors to remain in highly defensible cities in the central provinces, attending to ceremonial and religious duties, while the magistri competed over control of the armies and took over the act of governing. Although it does not seem that Merobaudes was able to dictate all governmental policies, he was able to contravene Gratian’s orders to the military with impunity, and he therefore might represent the first magister militum to overshadow an emperor. After Merobaudes defected to Magnus Maximus in 383, Bauto became the primary magister militum of Valentinian II’s court. Whereas emperors Magnus Maximus and Theodosius, both militarily capable emperors, were able to keep their magistri constrained into obedience, Valentinian was not able to free himself from the controlling forces around him. The sources are not particularly detailed, but accusations of Bauto using Valentinian to create his own puppet kingdom illustrate the power he may have wielded, although ultimately it seems the magister’s power was inhibited by the weak position of the Italian army, and he was challenged by other officials in Italy. Bauto shares several characteristics with Merobaudes, including barbarian heritage, a longer than average time in office, and commanding armies in the western Empire, factors which may have prompted and contributed to his influential position. When Valentinian was later installed as the west’s sole ruler, the eastern emperor Theodosius placed him and the western government in the care of Arbogast. This general greatly outstripped his expected role, and he gained the loyalty of the bureaucrats, officials, and the army, totally sidelining the emperor from active governing. This control was so complete that Valentinian could not even publicly dismiss Arbogast in his own court. Arbogast also had a long career, was a barbarian, and was present in the west. With the three most powerful magistri sharing these traits, it becomes increasingly clear they may have contributed to their positions. Before 365, the magistri had quite a rapid turnover, with mostly one to three year tenures being common. This changed from the reigns of Valentinian and Valens, with Bauto’s seven years, Arbogast’s eight, and Merobaudes’ career of thirteen years as magister being much greater than the average of five. Such lengths of time, in addition to their preceding years as lower-ranked officers, would have afforded these magistri many opportunities to create alliances and partnerships, build networks of patronage and support, and gain the loyalty of impor-

204

Conclusion

tant groups in the government. Furthermore, specific regulations were made to grant preeminent authority to those who held magisterial power for long periods. The long careers were not only held by barbarians, but also by those with Roman identities. The social taboo preventing barbarians from becoming emperors, however, prohibited barbarian magistri from gaining power by taking the throne. Even when Arbogast was suspected of murdering emperor Valentinian II and a civil war was fought to overthrow him, powerful magistri had become such an entrenched feature of the western Roman Empire that Theodosius designated another magister militum, Stilicho, to be the guardian of his son, the young emperor Honorius. Maybe, if he had not unexpectedly died, Theodosius would have expanded the western magisterium and instituted a loyal and decentralized cadre of generals as he had done in the west, although it is unclear whether the power blocs of the west would have permitted this shake up. Once again, Stilicho shares the same traits as previous powerful magistri. Although much of his career is outside our purview, he had the second longest career of all magistri assessed here, at seventeen years. As a semibarbarus, Stilicho was affected by the social taboo on the position of emperor, and thus all he could do to advance his own position was continue to gather power in his office. Stilicho also had a number of additional advantages, especially his links to the royal family through marriage. Although his authority was challenged by other powerful officials in both the eastern and western halves of the Empire, he represents the solidification of the power of the magistri that developed over the fourth century, and generals more powerful than the emperors would remain a key feature of the last decades of Roman hegemony in western Europe. Other magistri did not fulfil all these same factors during their time in office, and they did not gain as much power. Others still, such as Bauto and Gildo, fulfilled the same three characteristics, but had their ability to exert power challenged in other ways: Bauto’s influence was limited by others at court and the relative weakness of Valentinian II’s Italian army, while Gildo’s ability to influence political matters was limited by his station in Africa, although in the fifth century this would not pose such a problem to other magistri.2 These shared characteristics thus did not grant power on their basis alone, but they seem to have prompted magistri to increase the position of the office and smoothed their path to greater powers. This theory is not intended to exclude other factors that were unique to individual magistri, such as the advantages of Stilicho’s marriage into the imperial family. It is also important to acknowledge that none of these magistri had ultimate, unassailable power. Merobaudes lost his preeminent position when a powerful new emperor emerged, Arbogast was defeated in civil war, and Stilicho was assassinated on the orders of other court officials. Nonetheless, these magistri were some of the most powerful people in the fourth century, and it

2

E. g. Bonifatius. See: Wijnendaele 2015.

Conclusion

205

would be impossible to reasonably dismiss the importance of these shared traits and their impact on the propensity and ability of certain magistri to take advantage of key moments during the fourth century to enhance the power and impact of the office of magister militum. It was recently asserted by Mischa Meier in his book on Late Antiquity and the Migration Period, that the rise of the magistri militum was a symptom, not a cause, of the disintegration of the western Empire.3 It has hopefully been shown in this book that the magistri actually had a synergetic, reciprocal relationship with the collapse of imperial centralization. Each step taken by the magistri that increased their power, took away power from the emperors. This naturally in turn allowed greater opportunity for future increases in magisterial power. I am not promoting a teleological view of the fourth century, where we see the finality of imperial dissolution and judge events on how they led to the deposition of Romulus Augustus in 476. But it seems beyond deniability that the magistri were some of the most critical actors pushing change in the character of imperial and military power, and not simply passive symptoms. In conclusion, the development of magisterial power in the fourth century was not an institutional, top-down change as previously theorized. By charting the role magistri played in the pivotal events of the fourth century’s military history, and adding the conclusions derived from network and prosopographical analyses, this work has hopefully shown that individual magistri with certain shared and advantageous characteristics took advantage of opportunities to impact the trajectory of imperial power. This resulted in a rapid change in the military and political landscape that set the tone for the western Empire’s final demise.

3

Meier 2020, 366–68.

Bibliography Abbreviations and Ancient Sources Acta synhodorum habitarum Romae AE AM

Acta synhodorum habitarum Romae, in Monumenta Germaniae Historica: Auctores Antiquissimi 12: 393–455. T. Mommsen (ed.). Berlin, 1894. L’Année épigraphique. Paris, 1888–. Ammianus Marcellinus. Rerum gestarum libri qui supersunt. C. U. Clark (ed.), 2 vols. 1910; reprint, Berlin, 1963; cf. Ammianus Marcellinus. J. C. Rolfe (trans.) 3 vols. Cambridge, Mass., 1935–39. Amb. De ob. Theod. Ambrose. De obitu Theodosii: Corpus Scriptorum Ecclesiasticorum Latinorum 73: 39–401. O. Faller (ed.). 1955. Amb. De ob. Val. Iun. Ambrose. De obitu Valentiniani: Corpus Scriptorum Ecclesiasticorum Latinorum 73: 327–67. O. Faller (ed.). 1955. Amb. Ep. Ambrose. Epistulae: Corpus Scriptorum Ecclesiasticorum Latinorum 82.1. O. Faller (ed.). 1968: books 1–6; Corpus Scriptorum Ecclesiasticorum Latinorum 82.2. M. Zelzer (ed.). 1990: books 7–9; Corpus Scriptorum Ecclesiasticorum Latinorum 82.3. M. Zelzer (ed.). 1982: book 10 and Epistulae extra collectionem. Amb. fid. Ambrose. De fide ad Gratianum Augustum: Corpus Scriptorum Ecclesiasticorum Latinorum 78. O. Faller (ed.). 1962. Arist. Pol. Aristotle. Politika. W. D. Ross (ed.). Oxford 1957. AP Passio Artemii, in Philostorgius Kirchengeschichte. J. Bidez and F. Winkelmann (ed.). Berlin, 1972 Athan. Apol. De fuga Athanasius. Apologia de fuga sua. Philip Schaff and Henry Wace (eds.). New York, 1892. Athan. Fest. Ep. Athanasius. Festal Letters (Syriac), in The Festal Epistles of S. Athanasius. W. Cureton (trans.), Oxford, 1854. Athan. Hist. Ar. Athanasius. Historia Arianorum ad Monachos. M. Atkinson (trans.), A. Robertson (rev.). New York, 1892. Aug. c. Lit. Petil. Augustine. Contra Litteras Petiliani Donatistae Cortensis, Episcopi. J. R. King (trans.). Buffalo, 1887. Aug. Civ. Augustine. De Civitate Dei. Marcus Dods (trans.). Edinburgh: T. & T. Clark. 1871. Aur. Vict. Aurelius Victor. Liber de caesaribus. F. Pichlmayr (ed.). Leipzig, 1961; cf. Aurelius Victor: De Caesaribus, H. W. Bird (trans.). Liverpool, 1994.

208 Aus. Praef.

Bibliography

Ausonius. Praefatiunculae, in Ausonius. H. G. Evelyn-White (ed. and trans.), 2 vols. Cambridge, Mass., 1919–21; cf. Decimi Magni Ausonii opera. R. P. H. Green (ed.). Oxford, 1991. Aus. Grat. Act. Ausonius. Gratiarum Actio ad Gratiarum, in Ausonius. H. G. Evelyn-White (ed. and trans.), 2 vols. Cambridge, Mass., 1919–21; cf. Decimi Magni Ausonii opera. R. P. H. Green (ed.). Oxford, 1991. Bas. Ep. Basil. Lettres. Y. Courtonne (ed. and trans.), 3 vols. Paris, 1957–66; cf. St. Basil: The Letters. R. J. Deferrari (trans.), 4 vols. Cambridge, Mass., 1926–34. BGU Aegyptische Urkunden aus den staatlichen Museen zu Berlin, Griechische Urkunden. Berlin, 1892–2005. Bull. Comm. Bullettino della Commissione archeologica comunale di Roma. Rome, 1872–. Caes. B. Gal. Julius Caesar. De bello Gallico. F. P. Long (trans.). Oxford, 1911. Cass. Cassiodori Senatoris chronica ad. a DXIX. Chronica Minora 2: 109–61. T Mommsen (ed.), 3 vols. Berlin, 1892–98. Cass. Dio Dio Cassius. Historia Romana. L. Dindorf (ed.), 5 vols. Leipzig, 1863– 65. Chron. Edess. Chronicon Edessenum: Corpus scriptorum Christianorum orientalium SS 3.4.1. I. Guidi (ed.). Paris, 1903. Chron. Min. Chronica minora, in Monumenta Germaniae Historica: AA IX, XI, XIII. T. Mommsen (ed.). Berlin: 1892–98. Chron. Pasch. Chronicon paschale, 284–628 a. d., M. Whitby and M. Whitby (trans.). Liverpool, 1989. CIG Corpus Inscriptionum Graecarum. Augustus Boeckhius (ed.). Berlin, 1828–. CIL Corpus Inscriptionum Latinarum. T. Mommsen et al. (eds.). Berlin, 1863–. CJ Codex Justinianus. P. Krüger (ed.) Corpus Juris Civilis. Berlin, 1892– 1895; repr. 1945–1963. Claud. de bello Gild. Claudius Claudianus. De Bello Gildonico, in Claudian. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76. Claud. de cons. Stil. Claudius Claudianus. De Consulatu Stilichonis, in Claudian. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76. Claud. de IV cons. Hon. Claudius Claudianus. Panegyricus de Quarto Consulatu Honorii Augusti, in Claudian. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76. Claud. de VI cons. Hon. Claudius Claudianus. Panegyricus de Sexto Consulatu Honorii Augusti. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76. Claud. III cons. Hon. Claudius Claudianus. Panegyricus de Tertio Consulatu Honorii Augusti, in Claudian. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76. Claud. in Ruf. Claudius Claudianus. In Rufinum, in Claudian. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76. Claud. Laus Serenae Claudius Claudianus. Laus Serenae, in Claudian. M. Plautnauer (ed. and trans.), 2 vols. Cambridge, Massa., 1972–76.

Abbreviations and Ancient Sources

Coll. Avell.

209

Collectio Avellana: Corpus Scriptorum Ecclesiasticorum Latinorum 35. O. Guenther (ed.). 2 vols. 1895–98. Cons. Const. Consularia Constantinopolitana, in The Chronicle of Hydatius and the Consularia Constantinopolitana: Two Contemporary Accounts of the Final Years of the Roman Empire. R. W. Burgess (ed. and trans.). Oxford, 1993. CTh Theodosiani libri xvi cum constitutionibus Sirmondianis. T. Mommsen, P. Meyer, et al. (eds.) Berlin, 1905; reprint, Hildesheim, 1990; cf. Codex Theodosianus: The Theodosian Code and Novels and the Sirmondian Constitutions (including the Minutes of the Senate: Gesta Senatus urbis Romae). C. Pharr (trans.) New York, 1952. Dion. Hal. Rom. Ant. Dionysius of Halicarnassus. The Roman antiquities of Dionysius of Halicarnassus. E. Cray (trans.). Cambridge, Massa., 1937. Epit. Epitome de caesaribus, in Sexti Aurelii Victoris Liber De Caesaribus. F. Pichlmayr (ed.), 131–76. Leipzig, 1966. Eun. Hist. Eunapius. Universal History, in The Fragmentary Classicising Historians of the Later Roman Empire: Eunapius, Olympiodorus, Priscus, and Malchus. R. C. Blockley (ed. and trans.), vol. 2, Liverpool 1983. Eun. Lives Eunapius. The Lives of the Sophists. William Cave Wright (trans.). London, 1968. Eutr. Eutropius. Breviarum ab urbe condita. C. Santini (ed.). Leipzig, 1897. Excerp. Val. Excerpta Valesiana. J. Moreau (ed.). Leipzig, 1968. Faust. Byz. Faustus of Byzantium. P’awstosi Buzandac’woy Patmut’iwn Hayoc’. K. Patkanean (ed.), R. Bedrosian (trans.). New York, 1985. Festus Festus. The Breviarium of Festus: A Critical Edition with Historical Commentary. J. W. Eadie (ed.). London, 1967. Gaius Gaius. The Institutes of Gaius. W. M. Gordon and O. F. Robinson (trans.). Ithaca: Cornell University Press. Greg. Naz. Ep. Gregory of Nazianzus. Lettres. P. Gallay (ed. and trans.), 2 vols. Paris, 1964–67. Greg. Naz. Or. Gregory of Nazianzus. Discours 20–23: Sources Chrétiennes 270. J. Mossay (ed. and trans.). 1980. Greg. Tur. HF Gregory of Tours. History of the Franks, in Gregorii Turronensis Opera, in Monumenta Germaniae Historica. W. Arndt and B. Krusch (eds.). Hanover, 1884; cf. Gregory of Tours, History of the Franks. L. Thorpe (trans.). London, 1974. Hdt. Herodotus. Herodoti Historiae. C. Hude (ed.), 2 Vols. Oxford 1927. Herod. Herodian. Herodian of Antioch’s History of the Roman Empire. E. Echols (trans.). Berkeley and Los Angeles, 1961. Hyd. Chron. Hydatius. The Chronicle of Hydatius. And the Consularia Constantinopolitana. R. W. Burgess (ed.). Oxford, 1993. ICUR Inscriptiones Christianae Urbis Romae. G. B. de Rossi (ed.), 2 vols. Rome, 1857–61. IGR Inscriptiones graecae ad res romanas pertinentes. R. Cagnat (ed.), 3 vols. Chicago, 1964. ILCV Inscriptiones Latinae Christianae Veteres. E. Diehl (ed.). Berlin, 1925.

210 ILS Inscr. Lat. Tun. IRT

Bibliography

Inscriptiones Latinae Selectae. H. Dessau (ed.). Berlin, 1892–1916. Inscriptions latines de la Tunisie. A. Merlin (ed.). Paris, 1944. Inscriptions of Roman Tripolitania. J. M. Reynolds, C. M. Roueché, G. Bodard, C. Barron and others. Rome: British School at Rome, 1952. j. Ber Tractate Berkahot, in The Jerusalem Talmud: First Order, Zeraïm Tractate Berakhot. H. Guggenheimer (ed., trans., and comm.). Berlin, 2000. j. Shebi Tractate Sevi’it, in The Jerusalem Talmud: First Order, Zeraïm Tractates Kilaim and Seviit. H. Guggenheimer (ed., trans., and comm.). Berlin, 2001. j. Yeb Tractate Yebamot, in The Jerusalem Talmud: Third Order: Nasim, Tractate Yebamot. H. Guggenheimer (ed., trans., and comm.). Berlin, 2004. Jer. Chron. Jerome. Chronicon, in Eusebius Werke 7. R. Helm and U. Treu (eds.), 3rd edn. Berlin, 1984. Jer. Ep. Jerome. Epistulae: Corpus Scriptorum Ecclesiasticorum Latinorum 54– 56. I. Hilberg (ed.). 1910–18. Joh. Ant. John of Antioch. Historiarum Fragmenta, in Fragmenta Historicorum Graecorum 4: 535–622. K. Müller (ed.), 5 vols. Paris, 1868; cf. Ioannis Antiocheni fragmenta quae supersunt. Corpus fontium historiae Byzantinae 47. Sergei Mariev (trans.). Berlin, 2008. Joh. Lyd. de mag. John Lydus. De magistratibus populi romani libri tres: On Powers, or, The Magistrates of the Roman State. A. Bandy (ed., trans., and comm.). Philadelphia, 1983. Jord. Get. Jordanes. Getica, in Monumenta Germaniae Historica: AA V: 53–138. T. Mommsen (ed.). 1882. Jord. Rom. Jordanes. Romana, in Monumenta Germaniae Historica: AA V: 1–52. T. Mommsen (ed.). 1882. Jul. Ep. ad Ath. Julian. Letter to the Senate and People of Athens, in The Works of the Emperor Julian. W. C. Wright (ed. and trans.), 3 vols. Cambridge, Mass., 1913. Jul. Or. Julian. Oeuvres completes. J. Bidez, C. Lacombrade, and G. Rochefort (ed. and trans.), 4 vols. Paris, 1932–64. Lib. Ep. Libanius. Opera, vols. 10 and 11. H. Foerster (ed.), 12 vols. Leipzig, 1963–85; cf. Autobiography and Selected Letters. A. F. Norman (ed. and trans.), 2 vols. Cambridge, Mass., 1992. Lib. Or. Libanius. Opera, vols. 1, 2, 3, and 4. H. Foerster (ed.), 12 vols. Leipzig, 1963; cf. Selected Orations. A. F. Norman (ed. and trans.), 2 vols. Cambridge, Mass., 1969–77. Livy Livy. History of Rome. B. O. Foster (ed. and trans.), 14 vols. Cambridge, Mass., 1919. Marc. com. Chronicon, in Monumenta Germaniae Historica: AA XI: 37–108. T. Mommsen (ed.). 1894; cf. Marcellinus Comes. The Chronicle of Marcellinus. B. Croke (ed., trans., and comm.). Sydney, 1995. Menander Menander Rhetor. Menander Rhetor. D. A. Russell and N. G. Wilson (ed. and trans.). Oxford, 1981. Nic. Call. HE Nicephorus Callistus Xanthopoulos. Ecclesiasticae Historiae. J-P. Migne (ed. and trans.). 1865.

Abbreviations and Ancient Sources

Not. Dig. Occ.

211

Notitia Dignitatum in partibus Occidentis, in Notitia Dignitatum. O. Seeck (ed.). Berlin, 1876. Olymp. Olympiodorus. Fragments, in The Fragmentary Classicising Historians of the Later Roman Empire: Eunapius, Olympiodorus, Priscus and Malchus. R. C. Blockley (ed. and trans.) Liverpool, 1983. Oros. Orosius. Historiarum adversum paganos libri VII: Corpus Scriptorum Ecclesiasticorum Latinorum 5. K. Zangemeister (ed.). 1882 ; cf. Paulus Orosius: the Seven Books of History Against the Pagans. R. J. Deferrari (trans.). Washington, DC, 1964. P. Abinn. The Abbinaeus Archive: Papers of a Roman Officer in the Reign of Constantius II. H. I. Bell et al. (eds.). Oxford, 1962. P. Col. Columbia Papyri VII: Fourth-century Documents from Karanis. R. S. Bagnall and N. Lewis (ed. and trans.). Missoula, 1979. P. Flor. Papiri greco-egizii. G. Vitelli and D. Comparetti (eds.). Milan, 1906–15. P. Gen. Les Papyrus de Genève. J. Nicole. Ville de Genève, 1896–1900. P. Lips. Griechische Urkunden der Papyrussamlung zu Leipzig. L. Mitteis (ed.). Leipzig, 1906. P. Oxy. The Oxyrhyncus Papyri. B. P. Grenfell et al. (eds.) London, 1898–. Pan. Lat. In Praise of Later Roman Emperors: the Panegyrici Latini. C. E. V. Nixon and B. S. Rodgers (ed. and trans.). Berkeley, 1994. Paul. V. Amb. Paulinus of Milan. Vita di S. Ambrogio. M. Pellegrino (ed.), Verba Seniorum, ns 1. Rome, 1961; cf. Vita Sancti Ambrosii Mediolanensis Episcopi, a Paulino Eius Notario ad Beatum Augustinum Conscripta. M. S. Kaniecka (trans.). Washington, DC, 1928. Petr. Patr. Peter the Patrician. The Lost History of Peter the Patrician: An Account of Rome’s Imperial Past from the Age of Justinian. T. M. Banchich (ed. and trans.). London and New York: 2015. Philost. HE. Philostorgius. Historia ecclesiastica: Kirchengeschichte. J. Bidez and F. Winkelmann (eds.). Berlin, 1972; cf. Philostorgios: Kirchengeschichte. B. Bleckmann and M. Stein (eds. and trans.), 2 vols. Paderborn, 2015. Polyb. Polybius. The Histories. W. R. Paton (ed. and trans.), F. W. Walbank and C. Habicht (rev.), 6 vols. Cambridge, Mass., 2010–12. Procop. BP Procopius. History of the Wars: The Persian War, vol. 1. H. B. Dewing (ed. and trans.), 7 vols. Cambridge, Mass., 1914. Procop. DV Procopius. History of the Wars: The Vandalic War, vol. 2. H. B. Dewing (ed. and trans.), 7 vols. Cambridge, Mass., 1916. Prosp. Tiro. Prosper Tironis. Epitoma Chronicon, in Chronica Minora 1: Monumenta Germaniae Historica: AA IX: 341–499. T. Mommsen (ed.), 3 vols. Berlin, 1892). Prud. C. Symm. Prudentius. Contra Symmachum, in Prudentius, vol. 1: 344–401 and 2: 2–97. H. J. Thomson (ed. and trans.), 2 vols. Cambridge, Mass., 1961. RIC Roman Imperial Coinage, Volume 9: Valentinian I–Theodosius I. J. W. E. Pearce. London, 1951. Ruf. HE. Rufinus. Historia ecclesiastica, in Eusebii historia ecclesiastica translate et continuata: 947–1040. T. Mommsen (ed.). Berlin, 1908; cf. The Church History of Rufinus of Aquileia: Books 10 and 11. P. Amidon (trans.). Oxford, 1997.

212 Sid. Ap. Carm.

Bibliography

Sidonius Apollinaris. Poems and Letters, vol. 1: 1–327. W. B. Anderson (ed. and trans.), 2 vols. Cambridge, Mass., 1963. Sid. Ap. Ep. Sidonius Apollinaris. Poems and Letters, vols. 1: 331–483, 2. W. B. Anderson (ed. and trans.), 2 vols. Cambridge, Mass., 1963. Soc. Socrates. Historia ecclesiastica: Kirchengeschichte. G. C. Hansen (ed.). Turnhout, 1995. Soz. Sozomen. Historia ecclesiastica: Kirchengeschichte (2nd edn.). J. Bidez and G. C. Hansen (eds.). Turnhout, 1995. Symm. Ep. Symmachus. Q. Aurelii Symmachi quae supersunt: Monumenta Germaniae Historica: AA 6.1: 1–278. O. Seeck (ed.). 1883. Symm. Or. Symmachus. Q. Aurelii Symmachi quae supersunt: Monumenta Germaniae Historica: AA 6.1: 318–39. O. Seeck (ed.). 1883. Symm. Rel. Symmachus. Q. Aurelii Symmachi quae supersunt: Monumenta Germaniae Historica: AA 6.1: 279–317. O. Seeck (ed.). 1883. Syn. de reg. Synesius. Opere di Sinesio di Cirene: Epistole, operette, inni. A. Garzya (ed. and trans.). Turin, 1989. Tac. Germ. Tacitus. P. Cornelii Taciti libri qui supersunt. E. Koestermann (ed.). Leipzig, 1965–1970. Them. Or. Themistius. Orationes quae supersunt. G. Downey and A. F. Norman (eds.), 3 vols. Leipzig, 1965–74; cf. Politics, Philosophy, and Empire in the Fourth-century: Select Orations of Themistius. P. Heather and D. Moncur (ed. and trans.). Liverpool, 2001. Theod. HE. Theodoret. Historia ecclesiastica: Kirchengeschichte (3rd edn.). L. Parmentier and G. C. Hansen (ed.). Berlin, 1998. Theoph. Theophanes. Chronographia. C. de Boor (ed.). 1882; rep., Hildesheim, 1963; cf. The Chronicle of Theophanes Confessor. C. Mango and R. Scott (trans.). Oxford, 1997. Veg. Vegetius Renatus. Epitoma rei militaris. C. Lang (ed.). Leipzig, 1885; cf. Vegetius: Epitome of Military Science. N. P. Milner (trans.). Liverpool, 1993; Vegetius: Epitoma Rei Militaris. Michael D. Reeve (ed.) Oxford, 2005. Vita Isaacii Acta sanctorum Maii VII: 246–60. Zon. Zonaras. Epitomae historiarum libri XVIII. T. Büttner-Wobst (ed.). Bonn, 1897; cf. The History of Zonaras: From Alexander Severus to the Death of Theodosius the Great. T. Banchich and E. Lane (trans.). London, 2011. Zos. Zosimus. Histoire nouvelle. F. Paschoud (ed. and trans.), 3 vols. Paris, 1971–89; cf. Zosimus: New History. R. T. Ridley (trans.). Sydney, 1982.

Modern Scholarship

213

Modern Scholarship Alexander, Michael, and James Danowski. 1990. ‘Analysis of an Ancient Network’ Social Networks 12: 313–335. Alföldi, Andreas. 1952. A Conflict of Ideas in the Late Rome Empire: The Clash Between the Senate and Valentinian I. Harold Mattingly (trans.). Oxford: Clarendon Press. Alföldi, Andreas. 1967. ‘Die Hauptfaktoren der Geschichtsentwicklung zwischen 249–270 n. Chr.’ in Studien zur Geschichte der Weltkrise des 3. Jahrhunderts nach Christus, Andreas Alföldi (ed.). Darmstadt: Wissenschaftliche Buchgesellschaft. Alvarez-Jimenez, David. 2013. ‘Ammianus and Theodosius I concerning the Barbarica Conspiratio’ in The Theodosian Age (A. D. 379–455): Power, Place, Belief and Learning at the End of the Western Empire, Rosa Garcia-Gasco, Sergio Gonzalez Sanchez, and David Hernandez de la Fuente (eds.). Oxford: Gordon House. Amory, Patrick. 1997. People and identity in Ostrogothic Italy, 489–554. New York: Cambridge University Press. Arrizabalaga y Prado, Leonardo de. 2010. The Emperor Elagabalus: Fact or Fiction? New York: Cambridge University Press. Avi-Yonah, Michael. 1962. Geschichte der Juden im Zeitalter des Talmud in den Tagen von Rom und Byzanz. Berlin: Walter de Gruyter. Bagnall, Roger, Alan Cameron, Seth Schwartz, and Klass Worp. 1987. Consuls of the Later Roman Empire. Atlanta: Scholars Press. Balkanska, Anna. 1980. ‘Tirisis – Tirisa – Akra. Die thrakische und römisch-byzantinische Stadt am Kap Kaliakra (Scythia Minor)’ Klio 62: 27–46. Ball, Warwick. 2016. Rome in the East: The Transformation of an Empire. New York: Routledge. Banchich, T. and E. Lane (trans.). 2009. The History of Zonaras: From Alexander Severus to the Death of Theodosius the Great. London. Barlow, J. and Brennan, P. 2001. ‘Tribuni Scholarum Palatinarum c A. D. 353–64: Ammianus Marcellinus and the Notitia Dignitatum’ Classical Quarterly 51: 237–254. Barnes, Timothy. 1975. ‘Patricii under Valentinian III’ Phoenix 29.2: 155–170. Barnes, Timothy. 1980. ‘Imperial Chronology, A. D. 337–350’ Phoenix 34.2: 160–166. Barnes, Timothy. 1981. Constantine and Eusebius. Cambridge, Massachusetts: Harvard University Press. Barnes, Timothy. 1990. ‘Religion and Society in the Age of Theodosius’, in Grace, Politics and Desire: Essays on Augustine, H. A. Meynell (ed.). Calgary: University of Calgary Press. Barnes, Timothy. 1993. Athanasius and Constantius: Theology and Politics in the Constantinian Empire. Cambridge, Massachusetts: Harvard University Press. Barnes, Timothy. 1995. ‘Statistics and the Conversion of the Roman Aristocracy’ The Journal of Roman Studies 85: 135–147. Barnes, Timothy. 1996. ‘Emperors, Panegyrics, Prefects, Provinces and Palaces (284–317)’ Journal of Roman Archaeology 9: 532–552. Barnes, Timothy. 1998. Ammianus and the Representation of Historical Reality. Ithaca: Cornell University Press. Barnes, Timothy. 2014. Constantine: Dynasty, Religion and Power in the Later Roman Empire. England: Wiley Blackwell. Beauchamp, M. A. 1965. ‘An Improved Index of Centrality’ Behavioral Science 10: 161–163. Bennett, Julian. 2001. Trajan: Optimus Princeps. Bloomington: Indiana University Press.

214

Bibliography

Berchem, Denis van. 1952. L’armée de Dioclétien et la réforme constantinienne. Paris: Imprimerie Nationale et Paul Geuthner. Bidez, Joseph, and Friedhelm Winkelmann (eds.). 1972. Philostorgius: Kirchengeschichte. Mit dem Leben des Lucian von Antiochien und den Fragmenten eines arianischen Historiographen. Berlin: Akademie-Verlag. Bidez, Joseph. 1913. Philostorgius Kirchengeschichte: Mit dem Leben des Lucian von Antiochien und den Fragmenten eines arianischen Historiographen. Leipzig: J. C. Hinrichs’sche Buchhandlung. Bingham, Sandra. 2013. The Praetorian Guard: A History of Rome’s Elite Special Forces. Waco: Baylor University Press. Bird, H. W. (trans. and ed.). 1994. Aurelius Victor: De Caesaribus. Liverpool: Liverpool University Press. Bird, H. W. (trans. and ed.). 2011. Eutropius: Breviarium. Liverpool: Liverpool University Press. Birley, Anthony Richard. 1981. The Fasti of Roman Britain. New York: Oxford University Press. Birley, Anthony Richard. 1993. Marcus Aurelius: A Biography. London: Routledge. Bjornlie, M. Shane. 2018. ‘Romans, barbarians and provincials in the Res Gestae of Ammianus Marcellinus’ in Transformations of Romanness: Early Medieval Regions and Identities. Walter Pohl, Clemens Gantner, Cinzia Grifoni and Marianne Pollheimer-Mohaupt (eds.). Boston: Walter de Gruyter. Blanco-Perez, Aitor. 2013. ‘Themistius and the Accession of Theodosius I (Oratio XIV)’ in The Theodosian Age (A. D. 379–455): Power, Place, Belief and Learning at the End of the Western Empire, Rosa Garcia-Gasco, Sergio Gonzalez Sanchez, and David Hernandez de la Fuente (eds.). Oxford: Gordon House. Bleckmann, Bruno, and Markus Stein (eds.). 2015. Philostorgios: Kirchengeschichte. Volume Two. Paderborn: Ferdinand Schöningh. Bleckmann, Bruno. 1992. Die Reichskrise des III. Jahrhunderts in der spätantiken und byzantinischen Geschichtsschreibung. Untersuchungen zu den nachdionischen Quellen der Chronik des Johannes Zonaras. Munich: Tuduv. Bleckmann, Bruno. 1994. ‘Constantina, Vetranio und Gallus Caesar’ Chiron 24: 29–68. Bleckmann, Bruno. 2000. ‘Silvanus and his Followers in Italy: On the Interpretation of Two Campanian Inscriptions for the Usurper Silvanus [CIL X 6945 and 6946]’ Athenaeum 88: 477–483. Bleckmann, Bruno. 2003. ’Der Bürgerkrieg zwischen Constantin II. und Constans (340 n. Chr.)’ Historia: Zeitschrift für Alte Geschichte 52.2: 225–250. Bleckmann, Bruno. 2009. Die Germanen: Von Ariovist bis zu den Wikingern. Munich: C. H. Beck Verlag. Bleckmann, Bruno. 2012. ‘Sources for the History of Constantine’ in The Cambridge Companion to the Age of Constantine (Revised Edition), Noel Lenski (ed.). Cambridge: Cambridge University Press. Bloch, Herbert. 1945. ‘A new document of the last pagan revival of the west, 393–394AD’ The Harvard Theological Review 38.4: 200–244. Blockley, Roger. 1972. ‘Constantius Gallus and Julian as Caesars of Constantius II’ Latomus 31.2: 433–468. Blockley, Roger. 1980. ‘The Date of the ‘Barbarian Conspiracy’’ Britannia 11: 223–225. Blockley, Roger. 1981. The Fragmentary Classicising Historians of the Later Roman Empire: Eunapius, Olympiodorus, Priscus and Malchus. Vol. I. Liverpool: Francis Cairns. Blockley, Roger. 1983. The Fragmentary Classicising Historians of the Later Roman Empire: Eunapius, Olympiodorus, Priscus, and Malchus: Text, Translation, and Historiographical Notes. Vol. II. Liverpool: Francis Cairns.

Modern Scholarship

215

Blockley, Roger. 1987. ‘The Division of Armenia between the Romans and the Persians at the End of the Fourth-century AD’ Historia 36.2: 222–234. Blockley, Roger. 1992. East Roman Foreign Policy: Formation and Conduct from Diocletian to Anastasius. Leeds: Francis Cairns. Boak, Arthur. 1915. ‘The Roman Magistri in the Civil and Military Service of the Empire’ Harvard Studies in Classical Philology 26: 73–164. Boak, Arthur. 1955. Manpower Shortage and the Fall of the Roman Empire in the West. Ann Arbor: The University of Michigan Press. Boer, W. den. 1960. ‘The Emperor Silvanus and his Army’ Acta Classica 3:105–109. Böhme, Horst Wolfgang. 1974. Germanische Grabfunde des 4. bis 5. [i. e. fünften] Jahrhunderts zwischen unterer Elbe und Loire, Studien zur Chronologie und Bevölkerungsgeschichte. Munich: Beck. Borg, B. E. and C. Witschel. 2001. ‘Veränderungen im Repräsentationsverhalten der römischen Eliten während des 3. Jhs. n. Chr.’ in Inschriftliche Denkmäler als Medien der Selbstdarstellung in der römischen Welt, Geza Alföldy and Silvio Panciera (eds.). Stuttgart: Franz Steiner Verlag. Bowersock, G. W. 1978. Julian the Apostate. London: Duckworth. Bowersock, G. W. 1980. ‘Mavia, Queen of the Saracens’ in Studien zur antiken Sozialgeschichte: Festschrift Vittinghoff, Werner Eck, Hartmut Galsterer, and Hartmut Wolff (eds.). Cologne: Böhlau Verlag. Bowman, Alan, Averil Cameron, and Peter Garnsey (eds.). 2005. The Cambridge Ancient History, Volume 12: The Crisis of Empire, AD 193–337. Cambridge: Cambridge University Press. Brennan, P. 1996. ‘The Notitia Dignitatum’ in Les littératures techniques dans l’antiquité romain: statut, public et destination, tradition: sept exposés suivis de discussions. Genève: Fondation Hardt. Brennan, P. 1998. ‘The User’s Guide to the Notitia Dignitatum: The Case of the Dux Armeniae (ND Or. 38)’ Antichthon 32: 34–49. Brennan, P. 2007. ‘Zosimos 2.34.1 and ‘The Constantinian Reform’: Using Johannes Lydos to Expose an Insidious Fabrication’ in The Late Roman Army in the Near East from Diocletian to the Arab Conquest, Ariel Lewin and Pietrina Pellegrini (eds.). Oxford: Archaeopress. Broekaert, Wim, Elena Köstner, and Christian Rollinger. 2020. ‘Introducing the Ties that Bind’ Journal of Historical Network Research 4: i–xiii. Broekaert, Wim. 2013. ‘Financial Experts in a Spider Web. A Social Network Analysis of the Archives of Caecilius Lucundus and the Sulpicii’ KLIO 95: 471–510. Brown, Peter. 1971. The World of Late Antiquity: from Marcus Aurelius to Muhammad. London: Thames and Hudson. Browning, R. 1975. The Emperor Julian. London: Weidenfeld and Nicolson. Brunt, P. A. 1971. Italian Manpower, 225 B. C.-A. D. 14. London: Oxford University Press. Buckland, W. W., and Peter Stein. 1975. A Text-Book of Roman Law from Augustus to Justinian, Third Edition. Cambridge: Cambridge University Press. Burgess, R. W. 2008. ‘The Summer of Blood: The ‘Great Massacre’ of 337 and the Promotion of the Sons of Constantine’ Dumbarton Oaks Papers 62:5–51. Cameron, Alan. 1969. ‘Theodosius the Great and the regency of Stilicho’ Harvard Studies in Classical Philology 73: 247–280. Cameron, Alan. 2006. ‘Review: Umberto Roberto, Ioannis Antiocheni Fragmenta ex Historia chronica’ Bryn Mawr Classical Review 2006.07.37. Cameron, Alan. 2011. The Last Pagans of Rome. Oxford: Oxford University Press. Camps, Gabriel. 1984. ‘Rex gentium Maurorum et Romanorum’ Antiquités Africaines, 20: 183–218. Carrié, Jean-Michel. 2003. ‘The Contribution of Papyri to the Prosopography of the Ancient World: Evaluation and Prospects’ in Fifty Years of Prosopography: The Later Roman Empire, Byzantium and Beyond, Averil Cameron (ed.). Oxford/New York: Oxford University Press.

216

Bibliography

Cecconi, Giovanni Alberto. 1999. ‘I grandi magisteria tardoantichi. Ruolo istituzionale, attività e rapport con lestrutture amministrative territoriali (Italia, IV–VI secolo)’ in Magister: Aspetti Culturali e Istituzionali. Atti del Convegno Chieti, 12–13 Novembre 1997, Giulio Firpo and Giuseppe Zecchini (eds.). Alessandria: Edizioni dell’Orso. Chadwick, Henry. 1997. ‘Orthodoxy and Heresy from the Death of Constantine to the Eve of the First Council Of Ephesus’ in The Cambridge Ancient History, Volume 13: The Late Empire, AD 337–425, Averil Cameron and Peter Garnsey (eds.). Cambridge: Cambridge University Press. Chalmers, Walter R. 1960. ‘Eunapius, Ammianus Marcellinus, and Zosimus on Julian’s Persian Expedition’ The Classical Quarterly 10.2: 152–160. Chastagnol, André. 1962. Les Fastes de la prefecture de Rome au bas-empire. Paris: Nouvelles Editions Latines. Chastagnol, André. 1992. Le Sénat romain à l’époque impériale: recherches sur la composition de l’Assemblée et le statut de ses members. Paris: Les Belles Lettres. Cheesman, George Leonard. 1914. The Auxilia of the Roman Imperial Army. Oxford: Clarendon Press. Christie, Neil. 2013. ‘Wars Within the Frontiers: Archaeologies of Rebellion, Revolt and Civil War’ in War and Warfare in Late Antiquity, Alexander Sarantis and Neil Christie (eds.). Leiden: Brill. Cichocka, H. 1990. ‘Zosimus’ Account of Christianity’ Siculorum Gymnasium 43: 171–183. Cichorius, Conrad. 1986. Die Reliefs der Traianssäule. Erster Tafelband: “Die Reliefs des Ersten Dakischen Krieges”, Tafeln 1–57, Berlin: Verlag von Georg Reimer. Clarke, Graeme. 2005. ‘Christianity in the First Three Centuries’ in The Cambridge Ancient History, Volume 12: The Crisis of Empire, AD 193–337. Alan Bowman, Averil Cameron, and Peter Garnsey (eds.). Cambridge: Cambridge University Press. Cline, Diane, and Eric Cline. 2015. ‘Text Messages, Tablets, and Social Networks: The “Small World” of the Amarna Letters’ in There and Back Again – the Crossroads II. Proceedings of an International Conference Held in Prague, September 15–18, 2014, Jana Mynářová, Pavel Onderka, Peter Pavúk (eds.). Prague: Charles University in Prague. Cline, Diane. 2020. ‘Athens as a Small World’ Journal of Historical Network Research 4: 36–56. Coello, Terence. 1996. Unit Sizes in the Late Roman Army. Oxford: Tempvs Reparatvm. Conant, J. 2014. ‘Romanness in the Age of Attila’ in The Cambridge Companion to the Age of Attila, M. Maas (ed.). Cambridge: Cambridge University Press. Corcoran, Simon. 2000. The Empire of the Tetrarchs: Imperial Pronouncements and Government, AD 284–324, Second Edition. New York: Clarendon Press. Corcoran, Simon. 2008. ‘Diocletian’ in Lives of the Caesars, Anthony Barrett (ed.). Oxford: Blackwell. Corcoran, Simon. 2012. ‘Before Constantine’ in The Cambridge Companion to the Age of Constantine (Revised Edition), Noel Lenski (ed.). Cambridge: Cambridge University Press. Corcoran, Simon. 2014. ‘State Correspondence in the Roman Empire: Imperial Communication from Augustus to Justinian’, in State Correspondence in the Ancient World: From New Kingdom Egypt to the Roman Empire, Radner, Karen (ed.). Oxford: Oxford University Press. Coşkun, Altay. 2004. ‘Der comes Romanus, der Heermeister Theodosius und die drei letzten Akte der “Lepcis-Magna-Affaire” (a. 373–377)’ Antiquité Tardive 12: 293–308. Cosme, Pierre. 2007. ‘À propos de l’édit de Gallien’, in Crises and the Roman Empire: Proceedings of the Seventh Workshop of the International Network Impact of Empire, Nijmegen, June 20–24, 2006, Olivier Hekster, Gerda de Kleijn, and Daniëlle Slootjes (eds.). Boston: Brill. Croke, Brian. 1976. ‘Arbogast and the Death of Valentinian II’ Historia: Zeitschrift Für Alte Geschichte 25.2: 235–244.

Modern Scholarship

217

Croke, Brian. 2005. ‘Dynasty and Ethnicity: Emperor Leo I and the Eclipse of Aspar.’ Chiron 35: 147–203. Dal Santo, Gitte Lonstrop. 2015. ‘Concordia Apostolorum – Concordia Augustorum. Building a Corporate Image for the Theodosian Dynasty’, in East and West in the Roman Empire of the Fourth-century: An End to Unity?, Roald Dijkstra, Sanne van Poppel, and Daniëlle Slootjes (eds.) Boston: Brill. Dam, Raymond van. 1985. Leadership and Community in Late Antique Gaul. Berkeley: University of California Press. Davenport, Cailan. 2019. A History of the Roman Equestrian Order. Cambridge: Cambridge University Press. de Blois, Lukas. 1958. The Policy of the Emperor Gallienus. Leiden: Brill. de Boor, Carl. 1892. ‘Römische Kaisergeschichte in byzantinischer Fassung’ Byzantinische Zeitschrift 1:13–33. Dearn, Alan. 2003. ‘The Coinage of Vetranio: Imperial Representation and the Memory of Constantine the Great’ The Numismatic Chronicle 163: 169–191. Demandt, Alexander. 1969. ‘Der Tod des älteren Theodosius’ Historia: Zeitschrift für Alte Geschichte 18.5: 598–626. Demandt, Alexander. 1970. ‘Magister Militum’ Realencyclopädie der classischen Altertumswissenschaft Suppl. 12: 553–790. Demandt, Alexander. 1972. ‘Die Feldzüge der älteren Theodosius’ Hermes 100: 81–113. Demandt, Alexander. 1989. ‘The Osmosis of Late Roman and Germanic Aristocracies’ in Das Reich und die Barbaren, Evangelos K. Chrysos, Andreas Schawrcz (eds.). Vienna: Böhlau Verlag. Devijver, Hubert. 1999. ‘Les relations sociales des chevaliers romains’, in L’ordre équestre: histoire d’une aristocratie (IIe siècle av. J.-C.–IIIe siècle ap. J.-C.), Ségolène Demougin, Hubert Devijver, et Marie-Thérèse Raepsaet-Charlier (eds.), Rome: Ecole française de Rome. Dillon, John Noël. 2015. ‘The Inflation of Rank and Privilege: Regulating Precedence in the Fourth-century AD’ in Contested Monarchy: Integrating the Roman Empire in the Fourth-century AD, Johannes Wienand (ed.). New York: Oxford University Press. Dixon, Karen R., and Pat Southern. 1992. The Roman Cavalry: from the First to the Third-century AD. London: Batsford. Dixon, Karen R., and Pat Southern. 1996. The Late Roman Army. New York: Routledge. Dobson, Brian. 1974. ‘The Significance of the Centurion and Primipilaris in the Roman Army and Administration’, in Aufstieg und Niedergang der römischen Welt 2.1: 392–434. Dobson, Brian. 1978. Die Primipilares: Entwicklung und Bedeutung, Laufbahnen und Persönlichkeiten eines römischen Offiziersranges, Bonn: Rheinland-Verlag. Doležal, Stanislav. 2014. ‘Rethinking a Massacre: What Really Happened in Thessalonica and Milan in 390?’ Eirene. Studia Graeca et Latina 50.1–2: 89–107. Drake, Harold. 2005. ‘The Impact of Constantine on Christianity’ in The Cambridge Companion to the Age of Constantine, Noel Lenski (ed.). Cambridge: Cambridge University Press. Drake, Harold. 2009. ‘Solar Power in Late Antiquity’ in The Power of Religion in Late Antiquity, Andrew Cain and Noel Lenski (eds.). Farnham: Ashgate. Drew-Bear, Thomas. 1977. ‘A Fourth-Century Latin Soldier’s Epitaph at Nakolea’ Harvard Studies in Classical Philology 81: 257–74. Drinkwater, John. 1983. ‘The Pagan Underground, Constantius II’s “Secret Service”, and the Survival, and the Usurpation of Julian the Apostate’, in Studies in Latin Literature and Roman History III (Collection Latomus 180), Carl Deroux (ed.). Brussels: Latomus.

218

Bibliography

Drinkwater, John. 1994. ‘Silvanus, Ursicinus and Ammianus: fact or fiction?’, in Studies in Latin Literature and Roman History VII (Collection Latomus 227), Carl Deroux (ed.). Brussels: Latomus. Drinkwater, John. 2000. ‘The Revolt and Ethnic Origin of the Usurper Magnentius (350–53), and the Rebellion of Vetranio (350)’ Chiron 30: 131–45. Drinkwater, John. 2007. The Alamanni and Rome 213–496: Caracalla to Clovis. Oxford: Oxford University Press. Drogula, Fred. 2015. Commanders and Command in the Roman Republic and Early Empire. Chapel Hill: The University of North Carolina Press. Duncan, Mike. 2017. The Storm Before the Storm: The Beginning of the End of the Roman Republic, First Edition. New York: Public Affairs. Dušanić, Slobodan. 2004. ‘Julian’s strategy in AD 361’ Zbornik Radova Vizantološkog Instituta 41: 55–66. Elton, Hugh. 1996. Warfare in Roman Europe, AD 350–425. Oxford: Clarendon Press. Elton, Hugh. 2007. ‘Military Forces’ in The Cambridge History of Greek and Roman Warfare, Volume II: Rome from the late Republic to the Late Empire. Philip Sabin, Hans Van Wees, Michael Whitby (eds.). Cambridge: Cambridge University Press. Elton, Hugh. 2018. The Roman Empire in Late Antiquity: A Political and Military History. New York: Cambridge University Press Emerson, Richard. 1962. ‘Power-Dependence Relations’ American Sociological Review 27: 31–40. Enmann, Alexander. 1884. Eine verlorene Geschichte der römischen Kaiser und das buch De viris illustribus urbis Romae. Leipzig: Dieterich. Enßlin, Wilhelm. 1930. ‘Zum heermeisteramt des spätrömischen reiches: Die titulatur der magistri militum bis auf theodosius I’ Klio 23: 306–325. Enßlin, Wilhelm. 1931a. ‘Zum heermeisteramt des spätrömischen reiches: Die magistri militum des 4. Jahrhunderts’ Klio 24: 102–147. Enßlin, Wilhelm. 1931b. ‘Zum heermeisteramt des spätrömischen reiches: Der magister utriusque militiae et patricius des 5. Jahrhunderts’ Klio 24: 467–502. Enßlin, Wilhelm. 1934. ‘Theodosius’ Realencyclopädie der classischen Altertumswissenschaft 10A: 1937–1945. Eriksen, Thomas. 1993. Ethnicity and Nationalism: Anthropological Perspectives. London: Pluto Press. Errington, Malcolm. 1996a. ‘The accession of Theodosius I’ Klio 78.2: 438–53. Errington, Malcolm. 1996b. ‘Theodosius and the Goths’ Chiron 26: 1–27. Errington, Malcolm. 2006. Roman Imperial Policy from Julian to Theodosius. Chapel Hill: University of North Carolina Press. Faber, Eike. 2013. ‘The foedus of 382 or how the Goths did not become Integrated into the Roman Empire’ in in The Theodosian Age (A. D. 379–455): Power, Place, Belief and Learning at the End of the Western Empire, Rosa Garcia-Gasco, Sergio Gonzalez Sanchez, and David Hernandez de la Fuente (eds.). Oxford: Gordon House. Fanning, Steven. 2011. ‘Reguli in the Roman Empire, Late Antiquity, and the Early Medieval Germanic Kingdoms’ in Romans, Barbarians, and the Transformation of the Roman World: Cultural Interaction and the Creation of Identity in Late Antiquity, Ralph Mathisen and Danuta Shanzer (eds.). Burlington: Ashgate. Ferrill, Arther. 1986. The Fall of The Roman Empire: The Military Explanation. London: Thames and Hudon. Finley, M. 1958. ‘Review: Manpower Shortage and the Fall of the Roman Empire in the West by Arthur E. R. Boak’ The Journal of Roman Studies 48.1/2: 156–164.

Modern Scholarship

219

Frere, Sheppard. 1987. Britannia: A History of Roman Britain, Third Edition. London and New York: Routledge and Kegal Paul. Gardner, William Bell. 2005. Commanding Generals and Chiefs of Staff 1775–2005: Portraits & Biographical Sketches of the United States Army’s Senior Officer. Washington, D. C.: United States Army Center of Military History. Garsoian, Nina, 1999. L’Église arménienne et le grand schisme d’Orient. Lovanii: In Aedibus Peeters. Geary, Patrick. 1988. Before France and Germany: The Creation and Transformation of the Merovingian World. New York: Oxford University Press. Geiger, Joseph. 1979. ‘Ammianus Marcellinus and the Jewish Revolt under Gallus: A Note’ Liverpool Classical Monthly 4.4: 77. Geiger, Joseph. 1979–80. ‘The Last Jewish Revolt against Rome: A Reconsideration’ Scripta Classica Israelica 5: 250–257. Gibbon, Edward. 1781. The History of the Decline and Fall of the Roman Empire, Third Volume, originally published 1781, David Womersley (ed.). England: The Penguin Press, 1994. Gilliver, Kate. 2007. ‘The Augustan Reform and the Structure of the Imperial Army’ in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Girardet, Klaus M. 2004. ‘Die Erhebung Kaiser Valentinians II. Politische Umstände Und Folgen (375/76)’ Chiron 34: 109–144. Goffart, Walter. 1988. The Narrators of Barbarian History (A. D. 550–800): Jordanes, Gregory of Tours, Bede, and Paul the Deacon. Princeton: Princeton University Press. Goffart, Walter. 2006. Barbarian Tides. The Migration Age and the Later Roman Empire. Philadelphia: University of Pennsylvania Press. Goffart, Walter. 2009. ‘The Mirror of Jordanes: Concepts of the Barbarian, Then and Now’ in A Companion to Late Antiquity, Philip Rousseau (ed.). Malden: Wiley-Blackwell. Goldsworthy, Adrian. 2003. The Complete Roman Army. New York: Thames Hudson. Goldsworthy, Adrian. 2009. How Rome Fell: Death of a Superpower. New Haven: Yale University Press. Gracia, Javier Sánchez. 2014. ‘Amiano Marcelino y la usurpación de Silvano (15.5)’ Estudios Clásicos Anejo 2: 213–20. Grosse, Robert. 1920. Römische Militärgeschichte von Gallienus bis zum Beginn der Byzantinschen Themenverfassung. Berlin: Weidmannsche Buchhandlung. Gutmann, Bernhard. 1991. Studien zur römischen Aussenpolitik in der Spätantike: 364–395 n. Chr. Bonn: Rudolph Habelt. Haeling, Raban von. 1978. Die Religionszugehörigkeit der hohen Amtsträger des Römischen Reiches Seit Constantins I. Alleinherrschaft bis zum Ende der Theodosianischen Dynastie (324–450 BZW. 455 N. CHR.) (Antiquitas, Reihe III, XXIII). Bonn: Rudolf Habelt. Haeling, Raban von. 1988. ‘‘Timeo, ne per me consuetudo in regno nascatur’: Die Germanen und der romische Kaiserthron’ in Roma Renascens: Beiträge zur Spätantike und Rezeptiongeschichte: Ilona Opelt von ihren Freunden und Schülern zum 9.7.1988 in Verehrung gewidmet, Michael Wissemann (ed.). Frankfurt: Lang. Halsall, Guy. 2007. Barbarian Migrations and the Roman West, 376–568. New York: Cambridge University Press. Halsall, Guy. 2010. Cemeteries and Society in Merovingian Gaul: Selected Studies in History and Archaeology, 1992–2009. Leiden/Boston: Brill. Halsall, Guy. 2018. ‘Transformations of Romanness: The northern Gallic case’ in Transformations of Romanness: Early Medieval Regions and Identities. Walter Pohl, Clemens Gantner, Cinzia Grifoni and Marianne Pollheimer-Mohaupt (eds.). Boston: Walter de Gruyter.

220

Bibliography

Harmand, Jacques. 1967. L’armee et le Soldat a Rome de 107 a 50 avant notre ere. Paris: Editions A. et J. Picard et Cie. Harries, Jill. 2012. Imperial Rome AD 284 to 363: The New Empire. Edinburgh: Edinburgh University Press. Heather, Peter, and David Moncur. 2001. Politics, Philosophy, and Empire in the Fourth-century: Select Orations of Themistius. Liverpool: Liverpool University Press. Heather, Peter. 1991. Goths and Romans 332–489. Oxford: Clarendon Press. Heather, Peter. 1998. ‘Senatorial Careers’ in The Cambridge Ancient History, Volume 13: The Late Empire, AD 337–425, Averil Cameron and Peter Garnsey (eds.). Cambridge: Cambridge University Press. Heather, Peter. 1999. The Visigoths from the Migration Period to the Seventh-century: an Ethnographic Perspective. Rochester: Boydell Press. Heather, Peter. 2005. The Fall of the Roman Empire. London: Macmillan. Hebblewhite, Mark. 2017. The Emperor and the Army in the Later Roman Empire: AD 235–395. New York: Routledge. Heering, Walter. 1927. Kaiser Valentinian I. (364–375 n. Chr.). Ph. D dissertation, University of Jena. Heil, Wilhelm. 1966. Der konstantinische Patriziat. Basel: Helbing & Lichtenhahn. Hillner, Julia. 2019. ‘Imperial Women and Clerical Exile in Late Antiquity’ Studies in Late Antiquity 3.3:369–412. Hoepffner, André. 1936. ‘Les “Magistri Militum Praesentales” Au IV e Siècle’ Byzantion 11.2: 483– 498. Hoffmann, Dietrich. 1969–70. Das spätrömische Bewegungsheer und die Notitia Dignitatum. Two Volumes. Dusseldorf: Rheinland-Verlag. Hoffmann, Dietrich. 1974. ‘Der Oberbefehl des Spatromischen Heeres im 4. Jahundert n. Chr.’ in Actes Du IXeCongres International D’Etudes sur les Frontieres Romaines: Mamaia, 6–13 Septembre 1972, Dionisie Pippidi (ed.). Cologne: Böhlau Verlag. Honoré, Tony. 1998. Law in the Crisis of Empire 379–455 AD: The Theodosian Dynasty and its Quaestors. Clarendon: Oxford University Press. Howe, Laurence. 1942. The Pretorian Prefect from Commodus to Diocletian (AD 180–305). Chicago: University of Chicago Press. Hoyos, Dexter. 2007. ‘The Age of Overseas Expansion (264–146 BCE)’ in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Hughes, Ian. 2010. Stilicho: The Vandal Who Saved Rome. South Yorkshire: Pen & Sword Books Ltd. Hughes, Ian. 2012. Aetius: Attila’s Nemesis. South Yorkshire: Pen & Sword Books Ltd. Humphries, Mark. 2020. ‘The Memory of Mursa: Usurpation, Civil War, and Contested Legitimacy under the Sons of Constantine’ in The Sons of Constantine, AD 337–361: In the Shadows of Constantine and Julian, Nicholas Baker-Brian and Shaun Together (eds). Cham: Springer International Publishing. Hunt, David. 1997a. ‘Julian’ in The Cambridge Ancient History, Volume 13: The Late Empire, AD 337–425, Averil Cameron and Peter Garnsey (eds.). Cambridge: Cambridge University Press. Hunt, David. 1997b. ‘The Successors of Constantine’ in The Cambridge Ancient History, Volume 13: The Late Empire, AD 337–425, Averil Cameron and Peter Garnsey (eds.). Cambridge: Cambridge University Press. Hunt, David. 1999. ‘The Outsider Inside: Ammianus on the rebellion of Silvanus’ in The Late Roman World and Its Historian: Interpreting Ammianus Marcellinus, Jan Willem Drijvers and David Hunt (eds.). London and New York: Routledge.

Modern Scholarship

221

Isaac, B. 1988. ‘The meaning of the terms limes and limitanei’ Journal of Roman Studies 78: 125–47. Jacobs, Louis. 2008. Structure and Form in the Babylonian Talmud. New York: Cambridge University Press. Jans, N. A., and J. M. Frazer-Jans. 2001. ‘Job Rotation and Military Capability: Benefits, Certainly – but is Anyone Counting the Cost?’ Australian Defence Force Journal 146: 47–59. Johns, Catherine. 2010. The Hoxne Late Roman Treasure: Gold Jewellery and Silver Plate. London: British Museum Press. Jones, Arnold H. M. 1964. The Later Roman Empire, 284–602: A Social, Economic and Administrative Survey. Oxford: Blackwell. Kaldellis, Anthony. 2007. Hellenism in Byzantium: The Transformations of Greek Identity and the Reception of the Classical Tradition. New York: Cambridge University Press. Kelly, Christopher. 2006. Ruling the Later Roman Empire. Cambridge, Massachusetts: Harvard University Press. Kelly, Gavin. 2007. ‘The Sphragis and Closure of the Res Gestae’ in Ammianus after Julian: The Reign of Valentinian and Valens in Books 26–31 of the Res Gestae, J. den Boeft, J. W. Drijvers, D. den Hengst, and H. C. Teitler (eds.). Leiden and Boston: Brill. Kelly, Gavin. 2008. Ammianus Marcellinus: The Allusive Historian. Cambridge: Cambridge University Press. Kelly, Gavin. 2013. ‘The Political Crisis of AD 375–376’ Chiron 43: 357–410. Keyes, Clinton. 1915. The Rise of the Equites in the Third-century. Princeton: Princeton University Press. Kleeman, Jörg. 2005. ‘Die Trägen kamen zu spat – Zur ethnischen Interpretation ostgermanischen Fundstoffes’, in Bernd Päffgen, Ernst Pohl, and Michael Schmauder (eds.). Cum grano salis: Beiträge zur europäischen Vor- und Frühgeschichte; Festschrift für Volker Bierbrauer zum 65. Geburtstag. Firedberg: Memminger MedienCentrum Druckerei und Verlags-AG. Kovacs, Peter. 2017. ‘Constantius II und die Burgerkriege in Pannonien (350–351 n. Chr)’ Acta Archaeologica Academiae Scientiarum Hungaricae 68.2: 351–369. Krawczyk, Małgorzata. 2016. ‘Flawiusz Hermogenes – magister equitum Konstancjusza II czy praeses provinciae Scythiae?’ U schyłku starożytności – Studia źródłoznawcze 15: 31–40. Kulikowski, Michael. 2000. ‘The Notitia Dignitatum as a Historical Source’ Historia 49: 358–377. Kulikowski, Michael. 2002. ‘Nation versus Army: A Necessary Contrast?’ in On Barbarian Identity: Critical Approaches to Ethnicity in the Early Middle Ages, Andrew Gillet (ed.). Belgium: Brepolis. Kulikowski, Michael. 2007. Rome’s Gothic Wars: From the Third-century to Alaric. New York: Cambridge University Press. Kulikowski, Michael. 2009. Die Goten vor Rom. Stuttgart: Konrad Theiss Verlag. Kulikowski, Michael. 2011. ‘Barbarische Identität. Aktuelle Forschungen und neue Interpretationsansätze’ in Römische Legionslager in den Rhein- und Donauprovinzen: Nuclei spätantik-frühmittelalterlichen Lebens?, Michaela Konrad and Christian Witschel (eds.). Munich: C. H. Beck Verlag. Kulikowski, Michael. 2012. ‘Constantine and the Northern Barbarians’ in The Cambridge Companion to the Age of Constantine (Revised Edition), Noel Lenski (ed.). Cambridge: Cambridge University Press. Kulikowski, Michael. 2015. ‘Regional Dynasties and Imperial Court’ in Contested Monarchy: Integrating the Roman Empire in the Fourth-century AD, Johannes Wienand (ed.). New York: Oxford University Press. Kulikowski, Michael. 2016. The Triumph of Empire: The Roman world from Hadrian to Constantine. Cambridge, Massachusetts: Harvard University Press.

222

Bibliography

Kulikowski, Michael. 2019. The Tragedy of Empire: From Constantine to the Destruction of Roman Italy. Cambridge, Massachusetts: Harvard University Press. Ladner, Gerhart. 1976. ‘On Roman Attitudes Towards Barbarians in Late Antiquity’ Viator 7: 1–26. Landelle, Marc. 2014. ‘La Titulature des Magistri Militum au IVe Siècle’ Antiquité Tardive 22: 195–221. Landelle, Marc. 2016. ‘À Propos de la Création des Magistri Militum par Constantin Ier’ Revue des Etudes Anciennes 118.2: 493–509. Le Bohec, Yann. 2004. ‘Gallien et l’encadrement senatorial de l’armée romaine’ Revue des Etudes Militaires Anciennes 1: 123–32. Lee, A. D. 2007. War in Late Antiquity: A Social History. Melbourne: Blackwell Publishing. Lee, A. D. 2008. ‘The Army’ in The Cambridge Ancient History, Volume 13: The Late Empire, AD 337–425, Averil Cameron and Peter Garnsey (eds.). Cambridge: Cambridge University Press. Lee, A. D. 2015. ‘Emperors and Generals in the Fourth-century’ in Contested Monarchy: Integrating the Roman Empire in the Fourth-century AD, Johannes Wienand (ed.). New York: Oxford University Press. Lenski, Noel. 1999. ‘Assimilation and Revolt in the Territory of Isauria, from the 1st Century BC to the 6th-century AD’ Journal of the Economic and Social History of the Orient 42.4: 413–465. Lenski, Noel. 2000. ‘The Election of Jovian and the Role of the Late Imperial Guards’ Klio 82.2: 492–515. Lenski, Noel. 2002a. Failure of Empire: Valens and the Roman State in the Fourth-century A. D. Berkeley and Los Angeles: University of California Press. Lenski, Noel. 2002b. ‘Were Valentinian, Valens and Jovian Confessors before Julian the Apostate?’ Zeitschrift für antikes Christentum 6.2: 253–276. Lenski, Noel. 2004. ‘Valens and the Monks: Cudgeling and Conscription as a Means of Social Control’ Dumbarton Oaks Papers 58: 93–117. Lenski, Noel. 2007. ‘The Chronology of Valens’ Dealings with Persia and Armenia, 364–378 CE’ in Ammianus After Julian: The Reign of Valentinian and Valens in Books 26–31 of the Res Gestae, Jan den Boeft, Jan Willem Drijvers, and Hans Carel Teitler (eds.). Boston: Brill. Lenski, Noel. 2008. ‘Constantine’ in Lives of the Caesars, Anthony Barrett (ed.). Oxford: Blackwell. Lenski, Noel. 2017. ‘The Significance of the Edict of Milan’ in Constantine: Religious Faith and Imperial Policy, A. Edward Siecienski (ed.), Routledge: London and New York. Leppin, Hartmut. 1996. Von Constantin dem Großen zu Theodosius II: Das christliche Kaisertum bei den Kirchenhistorikern Socrates, Sozomenus und Theodoret. Gottingen: Vandenhoeck & Ruprecht. Lewis, Williams. 2020. ‘Constantine II and His Brothers: The Civil War of AD 340’ in The Sons of Constantine, AD 337–361: In the Shadows of Constantine and Julian, Nicholas Baker-Brian and Shaun Together (eds). Cham: Springer International Publishing. Lieberman, Saul. 1946. ‘Palestine in the Third and Fourth Centuries’ Jewish Quarterly Review 35: 329–370. Liebeschuetz, J. H. W. G. 1972. Antioch: City and Imperial Administration in the Later Roman Empire. Oxford: Clarendon Press. Liebeschuetz, J. H. W. G. 1990. Barbarians and Bishops: Army, Church, and State in the Age of Arcadius and Chrysostom. Oxford: Clarendon Press. Liebeschuetz, J. H. W. G. 1993. ‘The End of the Roman Army in the Western Empire’ in War and Society in the Roman World, John Rich and Graham Shipley (eds.). New York. Liebeschuetz, J. H. W. G. 1998. ‘Citizen Status and Law in the Roman Empire and the Visigothic Kingdom’ Strategies of Distinction: The Construction of Ethnic Communities, 300–800, Walter Pohl and Helmut Reimitz (eds.). Leiden: Brill.

Modern Scholarship

223

Liebeschuetz, J. H. W. G. 2007. ‘Was there a Crisis of the Third-century?’ in Crises and the Roman Empire: proceedings of the Seventh Workshop of the international network Impact of Empire, Nijmegen, June 20–24, 2006. Olivier Hekster, Gerda de Kleijn, Daniëlle Slootjes (eds.). Boston: Brill. Lunn-Rockliffe, Sophie. 2010. ‘Commemorating the Usurper Magnus Maximus: Ekphrasis, Poetry and History in Pacatus’ Panegyric of Theodosius’ Journal of Late Antiquity 3.2: 316–336. Luttwak, Edward. 2016. The Grand Strategy of the Roman Empire: From the First Century CE to the Third, Revised and Updated Edition. Baltimore: John Hopkins University Press. MacCormack, Sabine. 1981. Art and Ceremony in Late Antiquity. Berkeley: University of California Press. MacGeorge, Penny. 2002. Late Roman Warlords. Oxford: Oxford University Press. MacMullen, Ramsay. 1963. ‘Barbarian Enclaves in the Northern Roman Empire’ L’Antiquité Classique 32.2: 552–561. MacMullen, Ramsay. 1988. Corruption and the Decline of Rome. New Haven: Yale University Press. Mann, J. C. 1991. ‘The Notitia Dignitatum: Dating and Survival’ Britannia 22: 215–219. Martin, Bradly. 2004. Under the Loving Care of the Fatherly Leader: A History of North Korea and the Kim Dynasty. New York: Thomas Dunne. Masland, John and Laurence Radway. 1957. Soldiers and Scholars: Military Education and National Policy. Princeton: Princeton University Press. Mathisen, Ralph. 1997. ‘Les Barbares intellectuels dans l’Antiquité tardive’ Dialogues d’histoire ancienne Année 23.2: 139–148. Mathisen, Ralph. 2006. ‘Peregrini, Barbari, and Cives Romani: Concepts of Citizenship and the Legal Identity of Barbarians in the Later Roman Empire’ The American Historical Review 111.4: 1011–1040. Mathisen, Ralph. 2018. ‘‘Roman’ identity in Late Antiquity, with special attention to Gaul’ in in Transformations of Romanness: Early Medieval Regions and Identities. Walter Pohl, Clemens Gantner, Cinzia Grifoni and Marianne Pollheimer-Mohaupt (eds.). Boston: Walter de Gruyter. Mathisen, Ralph. 2019. ‘The End of the Western Roman Empire in the Fifth-century CE: Barbarian Auxiliaries, Independent Military Contractors, and Civil Wars’ in The Fifth-century: Age of Transformation: Proceedings of the 12th Biennial Shifting Frontiers in Late Antiquity Conference, Jan Willem Drijvers and Noel Lenski (eds.). Bari: Edipuglia. Matthews, John. 1971, ‘Gallic Supporters of Theodosius’ Latomus 30: 1073–99. Matthews, John. 1975. Western Aristocracies and Imperial Court, AD 364–425. Oxford: Clarendon Press. Matthews, John. 1989. The Roman Empire of Ammianus. London: Duckworth. Mayerson, Philip. 1980. ‘Mavia, Queen of the Saracens – A Cautionary Note’ Israel Exploration Journal 30.1/2: 123–131. McCall, Jeremiah. 2002. The Cavalry of the Roman Republic. Cavalry Combat and Elite Reputations in the Middle and Late Republic. London: Routledge. McEvoy, Meaghan. 2013. Child Emperor Rule in the Late Roman West, AD 367–455. Oxford: Oxford University Press. McLaughlin, Jonathan. 2014. ‘Bridging the Cultural Divide: Libanius, Ellebichus, and Letters to “Barbarian” Generals’ Journal of Late Antiquity 7.2: 253–279. Mclynn, Neil. 1994. Ambrose of Milan: Church and Court in a Christian Capital. Berkeley: University of California Press. Meier, Mischa. 2012. ‘Ostrom-Byzanz, Spätantike-Mittelalter. Überlegungen zum “Ende” der Antike im Osten des Römischen Reiches’ Millenium 9: 187–253.

224

Bibliography

Meier, Mischa. 2020. Geschichte der Völkerwanderung: Europa, Asien und Afrika vom 3. bis zum 8. Jahrhundert n. Chr. Second Edition. Munchen: C. H. Beck. Mendelsohn, Ludwig (ed.). 1887. Zosimus: Historia Nova. Leipzig. Milner, N. P. (trans). 1993. Vegetius: Epitome of Military Science. Liverpool: Liverpool University Press. Mitchell, Stephen and Geoffrey Greatrex (eds). 2000. Ethnicity and Culture in Late Antiquity. London: Gerald Duckworth & Co. Ltd. Mitchell, Stephen. 2015. A History of the Later Roman Empire, AD 284–641. Second Edition. Wiley Blackwell. Mommsen, Theodor. 1889. ‘Das Römische Militärwesen Seit Diocletian’ Hermes 24.2: 195–279. Mommsen, Theodor. 1901. ‘Aetius’ Hermes 36.4: 516–547. Mor, Menachem. 1989. ‘The Events of 351–352 in Palestine – The Last Revolt against Rome?’ in The Eastern Frontier of the Roman Empire. Proceedings of a Colloquium Held at Ankara in September 1988, David French and Chris Lightfoot (eds.). Oxford: B. A. R. Moreno, J. L. 1934. Who shall survive?: A New Approach to the Problem of Human Interrelations. Washington, D. C.: Nervous and Mental Disease Publishing Co. Näf, Beat. 1995. Senatorisches Standesbewusstsein in spätrömischer Zeit. Freiburg: Universitätsverlag. Nathanson, Barbara Geller. 1986. ‘Jews, Christians, and the Gallus Revolt in Fourth-Century Palestine’ The Biblical Archaeologist 49.1: 26–36. Nesselhauf, Herbert. 1938. Die spätrömische Verwaltung der gallisch-germanischen Länder. Berlin: Akademie der Wissenschaften. Nicasie, Martinus. 1998. Twilight of Empire: The Roman Army from the Reign of Diocletian until the Battle of Adrianople. Amsterdam: J. C. Gieben. Nischer, E. v. 1928. ‘Das Römische Heer Und Seine Generale Nach Ammianus Marcellinus’ Hermes 63.1: 430–456. Nitschke, Christian. 2020. ‘Network management in Ostrogothic Italy. Theoderic the Great and the Refusal of Sectarian Conflict’ Journal of Historical Research 4: 271–297. Nixon, C. E. V. 1991. ‘Aurelius Victor and Julian’ Classical Philology 86: 113–125. Nixon, C. E. V., and Barbara Rodgers. 1994. In Praise of Later Roman Emperors: the Panegyrici Latini. Berkeley: University of California. Noble, Thomas. 2006. ‘Romans, Barbarians, and the Transformation of the Roman Empire’ in From Roman Provinces to Medieval Kingdoms: Rewriting Histories, Thomas Noble (ed.). London: Routledge. Norman, Albert (trans.). 1969–77. Libanius: Selected Works. Two Volumes. Cambridge, Massachusetts: Harvard University Press. Nutt, D. C. 1973. ‘Silvanus and the Emperor Constantius II’ Antichthon 7: 80–9. O’Flynn, John. 1983. Generalissimos of the Western Roman Empire. Edmonton: University of Alberta Press. Omissi, Adrastos. 2018. Emperors and Usurpers in the Later Roman Empire: Civil War, Panegyric, and the Construction of Legitimacy. Oxford: Oxford University Press. Pabst, Angela (trans. and comm.). 1989. Quintus Aurelius Symmachus Reden. Darmstadt: WBG Academic. Pallottino, Massimo. 2014. A History of Earliest Italy. Martin Ryle and Kate Soper (trans.). Ann Arbor: University of Michigan Press. Parnell, David. 2012. ‘The Careers of Justinian’s Generals’ Journal of Medieval Military History 10: 1–16. Parnell, David. 2017. Justinian’s Men: Careers and Relationships of Byzantine Army Officers, 518–610. London: Macmillan.

Modern Scholarship

225

Paschoud, François. 1975. Cinq études sur Zosime. Paris: Les Belles Lettres. Paschoud, François. 1985. ‘Sur Eunape de Sardes’ Revue des Études Grecques 98: 395–398. Paschoud, François. 2006. ‘Chronique d’historiographie Tardive’ Antiquité Tardive 14: 325–344. Pellizzari, Andrea. 2011. ‘Le Armi e i Logoi: i Generali di Teodosio Nelle Lettere di Libanio’ Historia: Zeitschrift für Alte Geschichte 60.2: 191–218. Petit, Paul. 1955. Libanius et la vie municipale à Antioche au ive siècle. Paris: Geuthner. Petit, Paul. 1974. La Crise de l’Empire: des derniers Antonins a Diocletien (Histoire generale de l’Empire romain). Paris: Editions du Seuil. Pharr, Clyde, Theresa Sherrer Davidson, and Mary Brown Pharr. 1952. The Theodosian Code and Novels: and the Sirmondian Constitutions. Princeton: Princeton University Press. Piganiol, André. 1947 L’Empire chretien (325–395). Paris: Presses Universitaires de France. PLRE I, II. Arnold H. M. Jones, John R. Martindale, John Morris. 1971–80. The Prosopography of the Later Roman Empire: A. D. 260–395. Two Volumes. Cambridge: Cambridge University Press. Pohl, Walter. 1994. ‘Tradition, Ethnogenese und literarische Gestaltung: eine Zwischenbilanz’ in Ethnogenese und Überlieferung: angewandte Methoden der Frühmittelalterforschung. Karl Brunner and Brigitte Merta (eds.). Vienna: Oldenbourg. Pohl, Walter. 1998. ‘Introduction: Strategies of distinction’ in Strategies of Distinction: The Construction of Ethnic Communities, 300–800, Walter Pohl and Helmut Reimitz (eds.). Leiden: Brill. Pohl, Walter. 2018. ‘Introduction: Early Medieval Romanness – a multiple identity’ in Transformations of Romanness: Early Medieval Regions and Identities. Walter Pohl, Clemens Gantner, Cinzia Grifoni and Marianne Pollheimer-Mohaupt (eds.). Boston: Walter de Gruyter. Portmann, W. 1989. ‘Die 59. Rede des Libanios und das Datum der Schlacht von Singara’ Byzantinische Zeitschrift 82: 1–18. Potter, David. 2004. The Roman Empire at Bay AD 180–395. London: Routledge. Potter, David. 2015. ‘Measuring the Power of the Roman Empire’ in East and West in the Roman Empire of the Fourth-century: An End to Unity?, Roald Dijkstra, Sanne van Poppel, and Daniëlle Slootjes (eds.). Brill: Leiden. Rawlings, Louis. 2007. ‘Army and Battle during the Conquest of Italy (350–264 BCE)’ in in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Rebillard, E. 2013. ‘Review of The Last Pagans of Rome, by A. Cameron’ The Classical World 106.2: 297–298. Reeve, Michael D. (ed.). 2004. Oxford Classical Texts: Vegetius: Epitoma Rei Militaris. Oxford: Oxford University Press. Rich, John. 2007. ‘Warfare and the Army in Early Rome’ in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Ridley, Ronald (trans.). 2006. Zosimus: New History. Canberra: Australian Association for Byzantine Studies. Ridley, Ronald. 1970. ‘The Fourth and Fifth-century Civil and Military Hierarchy in Zosimus’ Byzantion 40: 91–104. Ridley, Ronald. 1972. ‘Zosimus the Historian’ Byzantinische Zeitschrift 65.2: 277–302. Roberts, C. H., and E. G. Turner (eds.). 1952. Catalogue of the Greek and Latin Papyri in the John Rylands Library, Manchester, Fourth Volume. Manchester: Manchester University Press. Rodgers, Barbara. 1981. ‘Merobaudes and Maximus in Gaul’ Historia: Zeitschrift für Alte Geschichte 30: 82–105. Rolfe, John C. 1939. Ammianus Marcellinus: Volume Three. Cambridge, Massachusetts: Harvard University Press.

226

Bibliography

Rollinger, Christian. 2020. ‘Prolegomena. Problems and Perspectives of Historical Network Research and Ancient History’ Journal of Historical Network Research 4: 1–35. Rosen, Klaus. 1970. Studien zur Darstellungskunst und Glaubwürdigkeit des Ammianus Marcellinus. Bonn: Rudolf Habelt. Ruffini, Giovanni Roberto. 2008. Social Networks in Byzantine Egypt. Cambridge University Press. Ruffini, Giovanni Roberto. 2020. ‘An Epilogue. Social Network Analysis and Greco-Roman Politics’ Journal of Historical Network Research 4: 325–339. Ruggini, Lellia Cracco. 1979. Il paganesimo romano tra religione e politica (384–394 d. C.). Rome: Accademia nazionale dei Lincei. Russell, Donald A. and Nigel G. Wilson (eds. and trans.). 1981. Menander Rhetor. Oxford: Clarendon Press. Sabidussi, Gert. 1966. ‘The Centrality Index of a Graph’ Psychometrika 31: 581–603. Salway, Benet. 2008. ‘Roman consuls, Imperial Politics, and Egyptian Papyri: The Consulates of 325 and 344 CE’ Journal of Late Antiquity 1.2: 278–310. Salzman, Michele Renee. 2002. The Making of a Christian Aristocracy: Social and Religious Change in the Western Roman Empire. Cambridge, Massachusetts: Harvard University Press. Salzman, Michele Renee. 2006. ‘Symmachus and the “Barbarian” Generals’ Historia: Zeitschrift für Alte Geschichte 55.3: 352–367. Sanchez, Fernando Lopez. 2003. ‘Left and Right in Roman Coins of the Fourth and Fifth Centuries A. D.’ in The Representation and Perception of Roman Imperial Power: Proceedings of the Third Workshop of the International Network Impact of Empire (Roman Empire, c. 200 B. C. – A. D. 476), Rome, March 20–23, 2002, Lukas de Blois, Paul Erdkamp, Olivier Hekster, Gerda de Kleijn, Stephan Mols (eds.). Brill. Šašel, Jaroslav, and Peter Petru. 1971. Claustra Alpium Iuliarum. Ljubljana: Narodni Muzej. Saxer, Robert. 1967. Untersuchungen zu den Vexillationen des römischen Kaiserheeres von Augustus bis Diokletian. Cologne: Böhlau Verlag. Schäfer, Peter. 1986. ‘Der Aufstand gegen Gallus Caesar’ in Tradition and Reinterpretation in Jewish and Early Christian Literature: Essays in Honour of Jürgen C. H. Lebram, J. W. van Henten, H. J. de Jonge, P. T. van Rooden, and J. W. Wesselius (eds.). Leiden: Brill. Scharf, Ralf. 1991. ‘Seniores-Iuniores Und Die Heeresteilung Des Jahres 364’ Zeitschrift Für Papyrologie Und Epigraphik 89: 265–72. Scharf, Ralf. 2005. Die Dux Mogontiacensis und die Notitia Dignitatum: Eine Studie zur Spätantiken Grenzverteidigung. Berlin: Walter de Gruyter. Schiffman, Lawrence. 1991. From Text to Tradition: A History of Second Temple and Rabbinic Judaism. Hoboken: KTAV Publishing House, Inc. Schmidt-Hofner, Sebastian. 2008. Reagieren und Gestalten: der Regierungsstil des spätrömischen Kaisers am Beispiel der Gesetzgebung Valentinians I. Munich: C. H. Beck. Schönfeld, Moritz. 1911. Wörterbuch der altgermanischen personen-und völkernamen nach der überlieferung des klassischen altertums bearbeitet. Heidelberg: Carl Winter’s Universitätsbuchhandlung. Schor, Adam. 2007. ‘Theodoret on the “School of Antioch”: A Network Approach’ Journal of Early Christian Studies 15.4: 517–562. Schor, Adam. 2011. Theodoret’s People: Social Networks and Religious Conflict in Late Roman Syria. Berkeley: University of California Press. Schuurmans, Constant. 1949. ‘Valentinien I et le senat romain’ L’Antiquité Classique 18: 25–38. Scott, John. 2000. Social Network Analysis: A Handbook, Second Edition. London: Sage Publications.

Modern Scholarship

227

Seeck, Otto. 1894. ‘Zur Echtheitsfrage des Scriptores historiae Augustae’ Rheinisches Museum für Philologie 49: 208–224. Seeck, Otto. 1906a. ‘Zur Chronologie und Quellenkritik des Ammianus Marcellinus’ Hermes 41.4: 481–539. Seeck, Otto. 1906b. Die Briefe des Libanius zeitlich geordnet. Leipzig: J. C. Hinrichs’sche Buchhandlung. Seeck, Otto. 1913. Geschichte des Untergangs der Antiken Welt. Volume Five. Berlin: Franz Siemenroth. Seeck, Otto. 1919. Regesten Der Kaiser und Päpste für die Jahre 311 bis 476 N. Chr.: Vorarbeit zu einer Prosopographie der Christlichen Kaiserzeit. Stuttgart: J. B. Metzlersche. Seeck, Otto. 1922. Geschichte des Untergangs der antiken Welt, Fourth Volume. Stuttgart: J. B. Metzlersche. Shahid, Irfan. 1984. Byzantium and the Arabs in the Fourth-century. Washington, D. C.: Dumbarton Oaks Research Library and Collection. Sivan, Hagith. 1993a. ‘Ammianus’ Terminus and the Accession of Theodosius I’ in Charistion C. P. T. Naude, Ursula Vogel-Weidemann, Jan Scholtermeijer (eds.). Pretoria: University of South Africa. Sivan, Hagith. 1993b. Ausonius of Bordeaux: Genesis of a Gallic Aristocracy. London: Routledge. Sivan, Hagith. 1996. ‘Was Theodosius I a Usurper?’ Klio 78: 198–211. Smith, William. 1875. A Dictionary of Greek and Roman Antiquities. London: John Murray. Speidel, Michael. 1978. Guards of the Roman Armies: An Essay on the Singulares of the Provinces. Bonn: Rudolph Habelt. Speidel, Michael. 1986. ‘The Early Protectores and Their Beneficiarius Lance’ Archäologisches Korrespondenzblatt 16: 451–4. Speidel, Michael. 1994. Riding for Caesar: The Roman Emperors’ Horse Guard. London: B T Batsford. St. John, Ronald Bruce. 2012. Libya: From Colony to Revolution. Second Edition. Oxford: Oneworld Publications. Stein, Ernst. 1928. Geschichte des spätrömischen Reiches. Vienna: L. W. Seidel. Steinacher, Roland. 2017. Rom und die Barbaren: Völker im Alpen- und Donauraum (300–600). Stuttgart: Verlag W. Kohlhammer Stickler, Timo. 2002. Aëtius: Gestaltungsspielräume eines Heermeisters im ausgehenden Weströmischen Reich. Munchen: Beck. Stickler, Timo. 2007. ‘The Foederati’ in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Stickler, Timo. 2011. ‘Die spätrömischen Heermeister bei Philostorg’ in Philostorge et l’historiographie de l’Antiquité tardive / Philostorg im Kontext der spätantiken Geschichtsschreibung, Doris Meyer (ed.). Stuttgart: Franz Steiner Verlag. Stoll, Oliver. 2007. ‘The Religions of the Armies’ in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Strobel, Karl. 2007. ‘Strategy and Army Structure between Septimius Severus and Constantine the Great’ in A Companion to the Roman Army, Paul Erdkamp (ed.). Oxford: Blackwell. Stroheker, K. F. 1961. ‘Alamannen im römischen Reichsdienst’ in Eranion; Festschrift für Hildebrecht Hommel, Dargebracht von seinen Tübinger Freunden und Kollegen, Jurgen Kroymann (ed.). Tubingen: M. Niemeyer. Syme, Ronald. 1958. Tacitus, Two Volumes. Oxford: Clarendon Press. Szidat, Joachim. 1979. ‘Die Usurpation Des Eugenius’ Historia: Zeitschrift Für Alte Geschichte 28.4: 487–508.

228

Bibliography

Szidat, Joachim. 1991. ‘Sabinianus: ein Heermeister senatorischer Abkunft im 4. Jh.’ Historia 40.4: 494–500. Szidat, Joachim. 1992. ‘Ammian und die historische Realität’ in Cognitio Gestorum: The Historiographic Art of Ammianus Marcellinus. J. den Boeft, D. den Hengst, and H. C. Teitler (eds.). Amsterdam: Royal Netherlands Academy of Arts and Sciences. Szidat, Joachim. 2010. Usurpator tanti nominis: Kaiser und Usurpator in der Spätantike (337–476 n. Chr.). Wiesbaden: Franz Steiner Verlag. Talbert, Richard. 1984. Barrington Atlas of the Greek and Roman World. Princeton: Princeton University Press. Talbert, Richard. 1996. ‘The Senate and Senatorial and Equestrian Posts’ in Cambridge Ancient History, Volume 10, Alan Bowman, Edward Champlin, and Andrew Lintott (eds.). Cambridge: Cambridge University Press. Tardieu, Michel. 1988. ‘Sebastianus étiqueté comme manichéen’ Klio 2: 494–500. Teall, John L. 1965. ‘The barbarians in Justinian’s armies’ Speculum 40.2: 294–322. Teitler, Hans. C. 1989. ‘Hariomundus en Haldagates: Enkele opmerkingen over barbaren in Romeinse krijgsdienst’ Lampas 22: 49–58. Testa, Rita Lizzi. 2007. ‘Christian Emperor, Vestal Virgins and Priestly Colleges: Reconsidering the End of Roman Paganism’ Antiquité Tardive 15: 251–162. Thelamon, Françoise. 1981. Païens et chrétiens au IVe siècle: l’apport de l’“Histoire ecclésiastique” de Rufin d’Aquilée. Paris: Études augustiniennes. Thompson, E. A. 1947. The Historical Work of Ammianus Marcellinus. Cambridge: Cambridge University Press. Tillemont, Louis-Sébastien Le Nain de. 1701–1704. Histoire des Empereurs et des autres Princes qui ont Régné Durant les Six Premiers Siècles de l’Église, Volumes Four and Five. Paris: chez Charles Robustel. Timpe, D. 1999. ‘Ethnologische Begriffsbildung in der Antike’ in Heinrich Beck, Germanenprobleme in heutiger Sicht. Berlin: De Gruyter. Tomlin, Roger. 1973. The Emperor Valentinian I. D. Phil dissertation, Oxford University. Tomlin, Roger. 1974. ‘The Date of the ‘Barbarian Conspiracy’’ Britannia 5: 303–309. Topchyan, Aram. 2006. The Problem of the Greek Sources of Movses Xorenac’I’s History of Armenia. Leuven: Peeters. Toynbee, Arnold J. 1965. Hannibal’s Legacy. The Hannibalic War’s Effects on Roman Life, Two Volumes. New York: Oxford University Press. Traina, Giusto. 2012. ‘La Production Rurale Dans La Phase Finale Du Royaume De La Grande Arménie: Le Témoignage De Moïse De Khorène’ Antiquité Tardive 20: 161–164. Treadgold, Warren. 1995. Byzantium and its Army, 284–1081. Stanford: Stanford University Press. Trombley, Frank. 1999. ‘Ammianus Marcellinus and Fourth-Century Warfare: A protector’s Approach to Historical Narrative’ in The Late Roman World and Its Historian: Interpreting Ammianus Marcellinus, Jan Willem Drijvers and David Hunt (eds.). London and New York: Routledge. Urbainczyk, Theresa. 1997. Socrates of Constantinople: Historian of Church and State. Ann Arbor: University of Michigan Press. Urbainczyk, Theresa. 2002. Theodoret of Cyrrhus: The Bishop and the Holy Man. Ann Arbor: University of Michigan Press. Vanderspoel, John. 1995. Themistius and the Imperial Court: Oratory, Civic Duty, and Paideia from Constantius to Theodosius. Ann Arbor: University of Michigan Press.

Modern Scholarship

229

Vanderspoel, John. 2020. ‘From the Tetrarchy to the Constantinian dynasty: A Narrative Introduction’ in The Sons of Constantine, AD 337–361: In the Shadows of Constantine and Julian, Nicholas Baker-Brian and Shaun Together (eds). Cham: Springer International Publishing. Verboven, Koenraad, Myriam Carlier, and Jan Dumolyn. 2007. ‘A Short Manual to the Art of Prosopography’ in Prosopography Approaches and Applications. A Handbook. Oxford: Unit for Prosopographical Research (Linacre College). Vervaet, Frederik Juliaan. 2002. ‘Domitius Corbulo and the Senatorial Opposition to the Reign of Nero’ Ancient Society 32: 135–193. Vervaet, Frederik Juliaan. 2007. ‘The Reappearance of the Supra-Provincial Commands in the Late Second and Early Third Centuries CE: Constitutional and Historical Considerations’ in Crises and the Roman Empire: proceedings of the Seventh Workshop of the international network Impact of Empire, Nijmegen, June 20–24, 2006. Olivier Hekster, Gerda de Kleijn, and Daniëlle Slootjes (eds.). Boston: Brill. Vetter, Emil. 1960. ‘Das Grab des Flavius Merobaudes in Trier’ Rheinisches Museum für Philologie 103.4: 366–372. Vrej, Nersessian. 1979. ‘Review of History of the Armenians’ Journal of Ecclesiastical History 30: 479–480. Waas, Manfred. 1971. Germanen im römischen Dienst (im 4. Jh. N. Chr.). Bonn: Rudolf Habelt. Wallraff, Martin. 1997. Der Kirchenhistoriker Sokrates: Untersuchungen zu Geschichtsdarstellung, Methode und Person. Gottingen: Vandenhoeck & Ruprecht. Wanke, Ulrich. 1990. Die Gotenkriege des Valens: Studien zu Topographie und Chronologie im unteren Donauraum von 366 bis 378 n. Chr. Frankfurt: Peter Lang. Ward, J. H. 1974. ‘The Notitia Dignitatum’ Latomus 33: 397–434. Warmington, B. H. 1956. ‘The Career of Romanus, Comes Africae’ Byzantinische Zeitschrift 49: 55–64. Wasserman, Stanley, and Katherine Faust. 1994. Social Network Analysis: Methods and Applications. New York: Cambridge University Press. Watson, Alaric. 1999. Aurelian and the Third-century. London: Routledge. Watts, Edward J. 2015. The Final Pagan Generation. Berkeley: University of California Press. Weisweiler, John. 2014. ‘Unreliable Witness: Failings of the Narrative in Ammianus Marcellinus’ in Literature and society in the Fourth Century AD: Performing Paideia, Contructing the Present, Presenting the Self, Lieve Van Hoof and Peter Van Nuffelen (eds.). Boston: Brill. Wenskus, Reinhard. 1961. Stammesbildung und Verfassung: das Werden der frühmittelalterlichen Gentes. Cologne: Böhlau Verlag. Werner, J. 1950. ‘Zur Entstehund der Reihengräberzivilisation’ Archaeologica Geographica 1: 23–32. Wijnendaele, Jeroen W. P. 2015. The Last of the Romans: Bonifatius – Warlord and Comes Africae. London: Bloomsbury Academic. Williams, Stephen and Gerard Friell. 1994. Theodosius: The Empire at Bay. London: B. T. Batsford Ltd. Williams, Stephen. 1985. Diocletian and the Roman Recovery. New York: Metheun. Wolfram, Herwig. 1988. History of the Goths, Thomas J. Dunlap (trans.). Berkley: University of California Press. Wolfram, Herwig. 1990. Das Reich und die Germanen: zwischen Antike und Mittelalter. Berlin: Siedler. Wolfram, Herwig. 1993. ‘Origio et religio. Ethnische Traditionen und Literatur in frühmittelalterlichen Quellen’ in Mittelalter. Annäherungen an eine fremde Zeit, Wilfried Hartmann (ed.). Regensburg: Universitatsverlag Regensburg.

230

Bibliography

Woods, David. 1995. ‘The Fate of the Magister Equitum Marcellus’ The Classical Quarterly 45: 266–268. Woods, David. 1997. ‘Ammianus and Some Tribuni Scholarum Palatinarum c. A. D. 353–64’ The Classical Quarterly 47: 269–291. Woods, David. 1998. ‘Valens, Valentinian I, and the Ioviani Cornuti’ in Studies in Latin Literature and Roman history IX, C. Deroux. Brussels: Peeters Publishers. Woods, David. 1999. ‘The Final Commission of Artemius the Former dux Aegypti’ Byzantine and Modern Greek Studies 23: 2–24. Woods, David. 2001. ‘Dating Basil of Caesarea’s Correspondence with Arinthaeus and his Widow’ Studia Patristica 37: 301–307. Woods, David. 2004. ‘Valens: Valiant but Flawed” Journal of Roman Archaeology 2004: 729–734. Woods, David. 2010. ‘Ammianus versus Libanius on Barbatio’s Alleged Bridge Across the Rhine’ Mnemosyne 63: 110–116. Woods, David. 2012. ‘Flavius Bonosus and the Consuls of AD 344’ Classical Quarterly 62.2: 895– 898. Woods, David. 2016. ‘Constantius, Julian, and the Fall of Sirmium’ in Beginning and End: from Ammianus Marcellinus to Eusebius of Caesarea, Álvaro Sánchez-Üstiz (ed.). Universidad de Huelva. Woolf, Greg. 2012. ‘Saving the Barbarian’ in Cultural Identity in the Ancient Mediterranean, Erich S. Gruen (ed.). Los Angeles: Getty Publications. Zuckerman, Constantin. 1998. ‘Comtes et Ducs en Égypte Autour de l’an 400 et la Date de la Notitia Dignitatum Orientis’ Antiquité Tardive 6: 137–147.

Register A Abundantius  113, 119, 157, 169, 192, 199 Addaeus  113, 119, 157, 169, 191–192 Adrianople 96 Adrianople, Battle of  90n102, 97–102, 104, 158 Ad Salices, Battle of  94–96n146, 166 Agenarichus see Serapio Agilo  66, 68–69, 89–90, 154n32–155, 158, 169, 189, 192, 197n282 Aginatius  135, 139 Alamanni  57, 59–62, 79, 81–82, 97, 112, 141, 160, 190, 194–195 Alani  96, 101, 106, 108 Alaric I 175, 180–181 Altar of Victory  108n247–109, 116n299 Ambianum 80–82 Ambrose  26, 98, 105, 107n233–109, 111, 115, 117, 137 Amida  13, 24, 58, 61, 64, 68, 158 Ammianus  24–25, 53–54, 69, 81n37, 93, 98, 126, 133, 137n99, 141, 161 – historians’ assessments of  22–23, 49–51, 53–54, 65, 68n297, 70n311, 71n321, 82, 86, 178 – on Arbitio  54, 56, 58–59, 113, 129, 154 – on Barbatio  63 – on Constantius II 48, 56–57, 114 – on Ursicinus  14, 22, 54–58, 61, 130 – on Julian  60–61, 63, 66–67, 114, 165 – on Nevitta  65, 69–70, 72 – on Lucilianus  72 – on Marcellus  61 – on Merobaudes  85–86, 139 – on Vetranio  46 – on Gomoarius  47 – on Silvanus  49–51, 56–57, 128–33

– – – – – – – – – – – –

on Sabinianus  54, 57–58 on Sebastianus  85 on Severus (8) 62 on Severus (10) 80–81 on Valentinian I 81 on Traianus  94 on Flavius Theodosius  81, 141 on Richomeres  97 on successions  76, 85–86, 105 on Gratian  105 on Valentinian II 105 relation to other sources, 44–45, 54, 69–70, 71n321, 72, 96, 100n183 Ampelius  134, 139, 141 Anatolius 44 Ancyra 77 Andragathius  106–107, 111, 169, 192 Antioch  33, 56, 58, 81n37, 89n87, 92, 98, 110 Apodemius 129–130 Aquileia  66, 111 Aquincum  84, 178 Arbitio  54, 58–59, 63, 65–66, 68–69, 74, 89–90, 113, 128–129, 132, 154–155, 161, 168– 170, 192, 195, 197, 199, 202 Arbogast  13–14, 16, 18, 102, 105, 109, 111–112, 114, 116–121, 146, 157, 165–167, 169–170, 185, 187, 189, 192, 197, 199, 203–204 Arcadius  112–113, 119–120, 166, 193 Arianism  26, 33–34, 43, 93, 96, 98, 127, 185 Arinthaeus  69–71, 73nn331–332, 76n2, 91, 96, 127, 154–158n50, 169, 192, 197n282 Armenia  91–92, 110 Artemius 68n297 Aspar 185–186 Athanaric  91, 100 Aurelius Victor  25, 29, 43, 50, 114, 186

232

Register

Ausonius, Decimius Magnus  138–139, 163 Ausonius, Iulius  138 B Bacurius  119n325, 189 Barbaricum  174, 194 Barbarization  17, 174–175, 188–191, 196, 200 Barbatio  52, 59, 61–63, 67, 74, 155, 160, 169, 192 Bauto  16, 99, 102–105, 107–109, 111, 117, 157, 164–167, 169–170, 182, 189, 192–193, 199, 203–204 Bithynia 43 Bleckmann  22, 41, 45, 47, 131 Bononia 65 Bonosus  20, 34–36, 161, 169, 192, 197n283 Brigetio 84–87 Britain  33, 36–37, 63–65, 74, 81–82, 106– 107n232, 137, 141, 159, 165, 194 Burgarii see limitanei Butherichus  110, 113, 169, 189, 192, 199 C Caecilius 138–139 Cameron, Alan  109n247, 149 Caracalla  173, 175, 179 Centrality  125, 133, 144–145 Cerealis  87, 136, 141 Chalcedon  68, 74 Claudian 167 Clique  83, 99, 125, 131, 133, 139, 142–145, 147, 165, 186 Codex Theodosianus  19, 24n3–25, 36, 64, 106, 111n262, 119, 135, 148 Comes  16, 55, 70n312, 80, 91, 107n232, 132 – Comes rei militaris  31–32, 37, 68n297, 82, 84, 153, 158–60, 164–65 – Comes domesticorum  65, 80–82, 95, 102, 119n325, 158–60, 164–65, 168, 189 – Comes et magister  49n146, 92, 110, 118–19 – Comes rei privatae 128 – Comes sacrarum largitionum 138 – Comes africae 135 – Comes sacri stabuli  154, 168 – Utriusque militiae  16, 94n128, 102, 118–19. Comitatenses  31, 160 Constans  33–38, 41, 50–52, 74, 159, 161, 168, 182

Constantine I 16, 19–20, 30–33, 38–40, 50n146, 54, 58, 105, 115, 131, 159, 173–174 Constantine II 33, 36–37 Constantinople  33, 46–47, 68, 79, 89, 91n111, 96, 101–103, 105n218, 113–114, 120, 167 Constantius II 13–14, 33–36, 38–52, 54, 56–70, 73–74, 82n40, 89–90, 108n244, 114, 118, 127–133, 158, 160–161, 163, 165, 168, 173, 185, 197, 201–202 Consul  13, 32, 34–35, 90–91, 102, 108– 109n246, 112–113, 138, 160–161, 163–164, 166–167, 184 Consulship  15, 32–36, 38, 58–59, 66, 68, 78–79, 85, 88, 91, 95–96, 102, 105, 108, 110, 116, 148, 152–153, 155–157, 160–164, 166–167, 184, 197 D Dagalaifus  65–66, 70–72, 76, 78–79, 82n40, 156, 159, 161, 169, 192, 197n282 Danube – region  16, 28, 46, 59, 65, 73, 83–86, 90, 92–94, 98, 100, 102–103, 121, 141, 145, 159–160, 173 – army  84, 86, 90, 145, 159–160 Demandt, Alexander  20–21, 81–82, 142 Diocletian  19, 28, 159 Doryphorianus  134–135, 137, 139 Dracontius 135 Drinkwater, John  22, 39, 41, 50–51, 60, 62, 186 Dux  16, 54, 58, 68n297, 70n312, 99, 156n36, 158 Dynamius  50–52, 128–133, 147 E Edessa 92 Egypt  33, 45, 68n297, 100n185, 124, 152, 190, 194–195 Elagabalus 115 Ellebichus  110, 113, 169, 182, 184, 192–193 Equitius  76, 80, 83–87, 92n116, 134n64, 136, 141–142, 144–145, 147, 156, 158n49, 163n79, 165–166, 169, 192–193, 197n283 Ethnicity  129, 131, 149, 176–177, 179–180, 186 Ethnography  176, 178–179 Eudoxia 193

Register

Eugenius  18, 109n247, 112, 116n299–121, 164, 166 Eunapius  25, 43–45, 69, 101, 109, 166 Eusebius (magister militum) 35–36, 155, 169, 192–193, 197 Eutropius  25, 43, 114 F Firmus  119, 136–137, 143n129, 145 Flavianus, Virius Nicomachus  109n247, 112, 117, 138, 164n89 Foederati  175, 177, 190 Frank  52, 62, 81, 87, 102, 106, 116, 118, 129–131, 166, 176, 178, 181, 183, 185–186, 192, 195 Frigeridus 95 Frigidus 119 Fritigern 101 G Gainas  103n209, 119n325, 175 Gaiso  37–38, 155, 161, 169, 182, 192, 197n282 Gallic armies  39, 64, 72, 82n40, 85, 95, 105–106, 116, 159 Gallic courtly faction  80, 142 Gallienus 29 Gallus  49, 55–56 Germans  87n73, 90, 150, 173–174, 177, 179, 183, 189–191, 193–195, 201 Gildo  118–119, 157, 169–170, 174n142, 183, 189, 192, 204 Gomoarius  47, 64, 66–67, 89–90, 127, 155, 158, 169, 192, 197n282 Goths  90–98, 100–104, 108n242, 119, 127, 173–174, 176–177, 180–181, 185–186, 192 Gratian  78–80, 82n40, 84–88, 90, 95–100, 102–103, 105–108, 110–111, 115–116, 125, 137–139, 141–142, 145–146, 158, 163–165, 195, 199, 203 Greco-Roman Polytheism  15, 60, 69, 71, 108–109, 116n299–117, 181 Gregory of Nazianzus  27, 101, 182 H Hadrianople see Adrianople Heraclea 47 Herculiani see Ioviani  37, 68, 75 Hermogenes  33–34, 169, 192, 197n283 Hesperius, Decimius Hilarianus  138

233

Hierapolis 92 Honorius  113, 115, 118, 120–121, 168, 184, 186, 204 Hormisdas  69–70, 154 Huns  101, 108, 179 I Illyricum  33, 37–40, 44, 48, 61, 66, 76, 80, 84, 92n116, 100–104, 112, 135, 165 Ioviani and Herculiani  37, 68, 75 J Jerome  42, 137, 180 Johannes (emperor) 185–187 John of Antioch  43, 117, 180n189 Jovian  16, 45, 67–68, 71–73, 75, 85, 89, 135, 158, 166, 168 Jovinus  65–66, 68–69, 72–73, 78–79, 81–82, 155, 169, 192 Julian  14, 16, 24, 41, 52, 57, 60–75, 78, 82n40, 84–85, 89–90, 108n144, 114, 118, 120, 146, 159, 161, 165, 168, 194, 201–202 – works of  36, 42, 44, 47, 50, 186 Julius (magister militum) 92, 94, 96, 99–100, 102–104, 156, 160, 169, 192 Justina  84n59, 87, 108n242, 115, 136, 139, 141 K Kelly, Gavin  84, 86–87 Kulikowski, Michael  50–51, 176, 178 L Laeti  174, 177 Lenski, Noel  71, 89 Lentienses  97, 195 Leo (notarius) 134, 136, 138–139, 144, 179 Libanius  27, 97–98, 102, 149, 182 Limitanei  30, 94, 158 Lucillianus  41, 63n265, 65, 72, 155, 159, 169, 192n260 Lupicinus  63–65, 67, 73–74, 78, 89–91, 155, 169, 192, 197n283 M Magister militum  13–15, 18–27, 30n15–32, 34, 37, 52–53, 55–57, 66, 73–75, 93, 104, 108, 113–114, 116, 119, 123, 151–154, 158–161, 165– 168, 170–171, 184, 188, 196–197, 199–200

234

Register

– equitum  15, 19–20, 22, 31, 33–36, 38, 54n175, 61, 64–66, 69–70, 72–73, 78, 83n47, 94n128, 96, 110–111n262, 159, 171 – equitum et peditum  15, 35, 49n146, 64, 72, 92 – Legal duties  56, 58–59, 68, 82–83, 97, 110, 184–188 – officiorum  21, 32, 38n68, 135, 158 – Military duties  1, 20–22, 31–32, 55–62, 65–66, 70, 80–82, 165–166 – peditum  15, 19–20, 22, 31, 34–36, 38n71, 49, 61, 63–64, 66, 69, 72, 78n17, 80, 84n54, 92, 105 – praesens see praesentalis – praesentalis  16, 19, 22, 59 – Religious duties  15, 26, 33–34, 43, 73, 99, 101, 117, 149, 182 Magnentius  37–42, 44–52, 56–58, 74–75, 158, 161, 186–187 Magnus Maximus  106–112, 114–115, 119, 164, 203 Majorianus  100, 103–104, 169, 192n261–193, 195 Malarichus  72, 129, 131–133, 147, 156, 158, 166, 169, 183n214, 189, 192, 197 Mallobaudes  129, 131, 193, 195 Marcellianus  134, 138, 141 Marcellus  13–14, 61–62, 67, 74, 169, 192, 197n283 Mavia 93–94 Maximinus  134–139, 141–145, 179 McEvoy, Meaghan  14, 23, 78 Meier, Mischa  23, 180, 205 Merobaudes  16, 79, 83–88, 95–96, 99, 103, 105–107, 111, 116, 120–121, 125, 136–137, 139, 141–145, 147, 156, 163–164, 167, 169–170, 184, 187, 189, 192–193, 199, 202–204 Milan  33, 61, 72, 80, 82n40, 105, 112, 130, 138–139 Militiae equestres  29, 151 Modares  100–101, 103–104, 169, 182–183, 189, 192 Mursa, Battle of  35, 38, 49, 57 N Naissus  38n72, 48, 79, 89 Nannienus  106, 156, 160, 169, 192, 195 Nebridius 55 Neo-Flavian dynasty  38–40, 81

Nepotianus  38n68, 41, 45, 51–52 Nevitta  65–66, 68–73, 85, 155, 158n50, 161, 169, 189, 192, 197n282 Nicaea 76–77 Notitia Dignitatum  16, 19–21, 25–26, 31, 111 O Onomastics  87, 190, 193–195 P Pagan see Greco-Roman Polytheism Palladius  135–136, 141, 145 Pannonia  39, 54, 84, 95, 98, 100, 134, 142–143, 165, 179 Pap 91 Persia  24, 28, 36, 39, 44, 47, 54, 57–58, 64–67, 69–70, 73–74, 84, 89–94, 96, 102, 110, 120, 154, 159, 165, 167, 179, 190, 201 Peter the Patrician  38n68, 45, 47 Philippopolis 46 Philostorgius  26, 38, 42, 46, 118, 185–186 Praetextatus, Vettius Agorius  108, 164 Probus, Petronius  84–85, 87, 110n256, 115, 134–136, 139, 141, 144–146 Procopius (usurper) 53, 79–80, 89–90, 118, 146, 158n45, 160, 165 Profuturus  94, 96n146 Promotus  97n162, 110–111, 113, 157, 169, 192, 199 Protectores  24, 30, 54, 69, 71, 75, 89, 100n188, 130, 154, 159, 174 Prusa 48–49 Q Quadi  83–85, 145, 160, 173 Quintinus  97n162, 106, 169, 192, 195 R Radagaisus 181 Religion  15, 24, 28, 33, 34, 40, 71, 92n111, 101, 108, 114, 115, 131, 137, 149, 182, 195 Remigius  135–136, 141–142, 144 Res Gestae see Ammianus Marcellinus Rheims  72, 81–82 Rhine  28, 36, 50, 60, 62–63, 73, 81–82, 97, 116, 129, 173, 177

Register

Richomeres  71n320, 95–97, 102–103, 110–111, 113, 118–119, 157, 159, 161, 169, 182, 189, 192 Ripenses see limitanei Romanization  178, 193 Romanus  135–136, 138–139, 141–142, 144–145 Rome  38n68, 41, 45, 51–52, 61, 109, 134, 138, 142, 151, 171–172, 185 Romulus  38, 97n162, 169, 192, 197n283, 205 Rufinus (Praetorian Prefect  392–395) 105n218, 112–114, 119–120 Rumoridus  109, 169, 192, 197, 199 S Sabinianus  54, 57–58, 60, 66, 90, 165, 169, 192, 197n283 Salia  33–34, 37, 99, 155, 161n75, 169, 192, 197n282 Sallustius  20, 34–35, 155, 169, 192, 197n283 Sapores  98–99, 102–103, 169, 192 Saracen  93–94, 179 Sarmatian  57, 59, 83, 90, 98, 100, 145, 173, 192 Sassanid Persia  28, 33 Saturninus  96–97, 101–104, 156, 158n49, 161, 163–164, 169, 192 Saturninus Salutius Secundus  71–72, 76–77, 79 Sebastianus  68n297, 84–87, 96–97, 136, 139, 145, 149, 156, 158n49, 160, 165, 169, 192, 197 Semibarbarus  180, 183–185, 193, 204 Serapio  190, 194–195 Serdica 46–48 Serena  119, 167 Severus (8) 61–63, 67, 74, 155, 169, 192–194, 197n283 Severus (10) 61–62, 78, 80–82, 156, 159, 169, 192–194, 197n283 Shapur  54, 57–58, 91–93, 159 Silvanus  16, 22, 37, 49–53, 56, 60, 63, 74, 87, 125, 128–133, 135, 147, 155, 158, 169, 174n141, 176, 183–187, 189–190, 192, 197, 201–203n282 Simplicius  134–135, 137, 139 Singara 54 Singidunum 46 Sirmium  33, 38n72, 41, 53, 65–66, 98, 101, 159 Stilicho  16, 103n210, 105n218, 113, 119–121, 157–158n49, 164n90, 167–170, 175, 180, 184, 189–190, 192–193, 195, 197, 199, 204

235

Strasbourg, Battle of  62–63, 65 Succession  16, 30, 36, 53, 67, 71–72, 75–78, 80, 84, 87–88, 117–118, 146, 185, 188, 203 Symmachus, Quintus Aurelius  26, 108–109, 112, 115, 137, 146, 163, 168, 182 T Teitler, Hans  21, 189, 193 Themistius  27, 42, 103, 186 Theodoret  26, 91n111, 96, 98–99n177, 127 Theodosian dynasty  115, 138 Theodosius, Flavius (the Elder) 53n168, 81–82, 99, 119, 136–138, 141–145, 147, 156n36, 165, 192–193, 197n283 Theodosius I 14, 16, 18–21, 24, 26, 37n63, 71n320, 81, 91n102, 99–105, 107, 110–114, 117–121, 137–138, 148, 156, 158, 163–165, 167– 169, 184, 192–193, 197, 199, 203–204 Theodosius II 25, 192–193 Thessalonica  101, 110–111 Thrace  33, 47, 90, 93, 95–96, 101–102, 104, 113 Timasius  110–111, 113, 119, 157, 169, 192, 199 Traianus  68n297, 94, 96–97, 127–128, 156, 158n49, 166, 169, 192 Trier  33, 51–52, 82, 105, 112, 114 U Ursicinus  13–14, 22, 24, 27, 50–52, 54–59, 61, 66, 70n312, 74, 90, 113, 130, 132–134, 147, 155, 168–169, 187, 192–195, 197n282, 201 Usurpation  14, 16, 19, 31–33, 36–37, 39, 41, 45, 48–53, 56, 64, 74–75, 79, 89, 91, 99n182, 103, 107, 125, 130–131, 164, 186, 188, 201–202 V Vacans  83, 119 Valens  43, 78–79, 86–98, 102, 104, 115, 120, 127–128, 137, 158, 160, 163, 165, 168, 173, 185, 203 Valentinian I 14, 16, 20, 75, 77–89, 91, 105, 107, 118, 125, 133–139, 141, 143, 147, 160, 163, 165, 168, 202–203 Valentinian II 13–14, 16, 23, 87–88, 95, 105, 107–118, 120–121, 136–137, 141–142, 145–146, 163–166, 185, 203–204 Valentinianus see Valentinian II Valentinus  141, 145 Vallio 107

236

Register

Vegetius  26, 31–32, 98, 153, 170, 174, 179 Vetranio  16, 22, 37–49, 51–53, 58, 64, 74, 80, 87, 127, 146, 155, 169, 187, 192, 197, 201–202 Victor  20, 40, 69–73n332, 78, 89–94, 96–99, 101–104, 127, 154, 156, 168–169, 189, 192, 197 Victorinus 134–135 Vienne  13, 117

W Woods, David  22, 35, 53, 65–66, 68n297, 89, 91–92, 96, 127n34 Z Zonaras  25, 45, 47, 50, 186 Zosimus  19, 25, 30, 44, 69–70, 72, 76, 85, 90, 100–101, 103, 111, 116–118, 152, 179, 186

This monograph presents a novel investi­ gation of the magistri militum, the highest­ ranking officers within the late Roman army. It posits that between 340 and 395 CE, specific magistri seized opportune moments, notably during the political voids following emperors’ deaths, to reshape the character of their office and expand its pivotal role in the military-­ political sphere. This transformation played a decisive role in the eventual dis­ solution of the Western Roman Empire. Furthermore, the study employs the prosopographical method to reevaluate previous scholarship regarding the pro­

ISBN 978-3-515-13614-3

9 783515 136143

portion of barbarian and Roman gene­ rals. Notably, the research posits that the balance between Roman and non-Roman officers was far more equitably distribu­ ted than hitherto conjectured. Additio­ nally, prosopography is used to recon­ struct the fourth-century cursus honorum. Finally, this work utilizes the analytical framework of social network analysis, predicated upon the application of mathe­matical equations and formulae to elucidate the intricate dynamics of positive and negative relationships. The findings of this study furnish valuable in­ sights and prospects for further research.

www.steiner-verlag.de Franz Steiner Verlag