‘THE FITTING FACE OF EMPIRE’ PALACES AND POWER IN THE EARLY MIDDLE AGES A Dissertation Presented to the Faculty of the Graduate School of Cornell University in Partial Fulfilment of the Requirements for the Degree of Doctor of Philosophy by Samuel James Barber May 2021 © 2021 Samuel James Barber ‘THE FITTING FACE OF EMPIRE’: PALACES AND POWER IN THE EARLY MIDDLE AGES Samuel James Barber Cornell University, 2021 This dissertation argues that early medieval palaces were not passive expressions of royal power, but agents in the constitution of new forms of political authority in post-Roman Europe. Focusing on palaces in the territory of the Roman Empire’s western provinces between the fourth and ninth cen- turies CE, this study adopts an interdisciplinary methodology to reinterpret how palaces were concep- tualised by contemporaries, their role in early medieval societies, and how these characteristics changed over time. Drawing on the approaches of architectural history, archaeology, and social theory, I rein- terpret palaces as simultaneously architecture designed to assert the ruler’s power and social spaces where the foundations and limits of political authority were negotiated. My investigation is divided into two co-ordinating parts with complementary analytical per- spectives. In Part I, ‘Architecture and Space’, I show that a coherent ‘grammar’ of Late Roman palaces precipitously broke down following the disintegration of the Western Roman Empire. In its wake coalesced a new spatial configuration which responded to a changed lexicon of material display and to new ideals of legitimate authority. The architecture of the palace was not just a stage for ‘royal ideology’, however, but a medium of symbolic communication between the ruler and the elite that was given expression in ritual. In Part II, ‘Infrastructure and Landscape’, I draw on archaeological theories of landscape to examine the broader political structures in which palaces were embedded. In two case studies, I explore how royal authority was ‘spatialised’ through networks of palaces in the Ostrogothic kingdom of Italy and in the Frankish kingdoms of Gaul and Germany. By charting the fluctuations of these geographies over time, I demonstrate the instrumental role played by palaces in the reordering of political space in the post-Roman West. BIOGRAPHICAL SKETCH Sam’s research focuses on the visual and material cultures of late ancient and early medieval Europe, with a special interest in the intersections of the History of Art with social and anthropological theory. Before joining Cornell’s Medieval Studies program, Sam studied History at Durham University and earned a Master’s degree in Medieval Studies at the University of York in the United Kingdom. In 2017 Sam earned a further MA in Medieval Studies in the course of his doctoral studies at Cornell, with a major concentration in the History of Medieval Art. From 2019 to 2020, he was a visiting doctoral student at the Freie Universität Berlin in the Friedrich-Meinecke-Institut (Department of History and Cultural Studies). His research has been supported by the Cornell University Graduate School, the International Center for Medieval Art, the Lemmermann Foundation, and the Deutscher Akademischer Austauschdienst. iii ACKNOWLEDGMENTS This dissertation was finished in difficult circumstances. It has been a welcome antidote to the isola- tion of the past twelve months, however, to reflect on all those who have left a mark on this project and without whose encouragement it would not have been brought to completion. First and foremost, I would like to thank my advisor, Ben Anderson, who first suggested a study of palaces and whose influence can be felt throughout the following pages. I owe much to his example. I am no less grateful to Éric Rebillard and Cynthia Robinson, whose guidance and advice over the past seven years have significantly shaped how I approach Late Antiquity and the Middle Ages. At Cornell I have had the privilege of being part of a unique community of scholars, and a great many people have shaped this project in ways both large and small. Special thanks go to Max McComb and Nick Bujalski for their critical eye, patient ear, and spirited comradeship. I must also thank the members of the Society for the Humanities dissertation writing group ‘Constructing Com- munity in the Early Middle Ages’, namely Abby Sprenkle, Patrick Naeve, Paul Vinhage, and Tyler Wolford, for their feedback on draft chapters as this project approached its latter stages. From the Medieval Studies Program I would be remiss not to also mention John Wyatt Greenlee, Anna Way- mack, Danielle Reid, and Professors Andrew Galloway, Andrew Hicks, and David Powers for their feedback and thoughts at various points over the years. In presenting various pieces that eventually made it into this dissertation at the monthly graduate student ‘roundtable’, I have also benefited from the questions and comments of other friends and colleagues, too numerous to name individually. Sincere thanks to you all. The patience of Jim Utz and, more recently, Marion Penning—though we still have not met in person!—in helping me out with the endless bureaucracy of international student life should also not go unrecognised. Beyond Cornell I have also been fortunate to kick ideas back and forth with James Harland and Mateusz Fafinski. However much my friends and colleagues have iv influenced this project, all errors to be found in the following pages are, it must be said, wholly my own. The undoubted highlights of writing this dissertation were the opportunities to travel and study abroad – and those opportunities were only possible with the generosity of various funding institutions. In 2017 the Fondazione Lemmermann subsidised a semester-long stay in Rome, in which I dug into the history and archaeology of the Palatine, the Villa of Maxentius, and the Sessorian Palace at the American Academy and the British School. In 2018 the International Center for Medieval Art (ICMA) generously funded a research trip to Spain—a simultaneously exhilarating and exhausting month-long itinerary from Asturias to Andalucía—that has profoundly informed my thinking throughout this project. Most recently, in 2019–2020 I was able to spend several months studying at the Freie Universität Berlin in the department of Prof. Dr. Stefan Esders with the support of a One- Year Research Grant from the Deutscher Akademischer Austauschdienst (DAAD). Though my planned one-year stay was ultimately cut short by the onset of the COVID-19 pandemic, I extend my sincere thanks to Prof. Dr. Esders and to the community of scholars at the Friedrich-Meinecke-Insti- tut for their hospitality. Various other bodies have funded research trips and professional training, including the Cornell Institute of Archaeology and Material Studies (CIAMS), the Cornell University Graduate School, and the Einaudi Center for International Studies. All of the above have my sincere gratitude for supporting my scholarship. Parts of what would become Chapter 3 were presented at the 2018 International Congress on Medieval Studies in Kalamazoo, and I was glad to have the opportunity to discuss Ravenna and The- oderic with Jonathan Arnold and Deborah Deliyannis. Another section of the same chapter was pre- sented virtually at the 2020 Haskins Society Conference, and I am grateful for the audience’s questions and feedback. A draft of Chapter 2 was presented to the Forschungskolloquium zur Geschichte der Spätantike und des Frühmittelalters at the FU, and the comments of my colleagues there were helpful v for thinking more carefully about ‘public’ and ‘private’ power in the Middle Ages. Methodological considerations from the Introduction were presented at the 2019 International Medieval Congress in Leeds, and I would like to thank James Harland and Andrew Welton for including me in their series of panels on ‘Material Narratives of Late Antiquity’. I would like to close these Acknowledgments by thanking the three people who have, in fact, shaped this project the most. My parents, Debbie and Tony, have been unwaveringly supportive throughout the last seven years—always ready with a pep talk or a photo of their ever-excitable dachs- hunds when required—and it is no exaggeration to say that this process would not have been possible at all without them. I also would never have reached these final stages without the love and encour- agement of my partner, Kristen, who has had my back throughout. Despite all the challenges of this past year, I am beyond delighted that I get to celebrate this achievement with you. vi TABLE OF CONTENTS Biographical Sketch iii Acknowledgments iv Table of Contents vii List of Figures x List of Abbreviations xviii Introduction I. ‘The Fitting Face of Empire’ 1 II. The Historiography of Early Medieval Palaces 3 The problem of ‘palatial’ architecture: disciplinary compartmentalisation 4 Rupture and degradation: temporal compartmentalisation 7 Politics in place: the sedes regiae and the Hauptstadtproblem 10 III. What is a Palace, Anyway? 15 Towards a socio-spatial approach 17 IV. Looking Ahead 19 Part I. Architecture and Space Chapter 1. Palace Architecture, Between Imperium and Regna I. Introduction 23 Public and private in palace architecture 24 Overview 27 II. ‘Rome is where the Emperor is’: The Palace in the Later Roman Empire 27 New seats for new emperors: Milan and Trier in the fourth century 28 Milan and Trier in context 32 From the periphery to the Eternal City: the building projects of Maxentius 35 III. Reimagining Rome: Palaces and the Iconography of Roman Authority 41 Emulation and distinction: Carthage, Ravenna, and Reccopolis 42 The iconography of Roman authority in the Carolingian world 54 IV. Architectures of Authority in an Age of Transformation 60 Onward and upward: transformations in secular architecture 61 The ‘great hall’ in early-medieval Britain: an alternative model? 69 V. The Palace Chapel: Its Genesis and Architectural Significance 73 Private oratories and public basilicas in the Later Roman Empire 75 Palace chapels and the ‘liturgification’ of authority in the Early Middle Ages 78 VI. Conclusion 83 Chapter 2. The Ritual Shaping of the Early Medieval Palace I. Introduction 85 Ritual matters: what is ‘ritual’ and what does it do? 86 Overview 89 II. The City 90 The emperor’s changing image 91 The streets, the circus, the church: the spaces of imperial ritual 94 The afterlife of imperial ritual 101 vii III. The Audience Hall 103 The spatial formation of the audience hall 104 In praesentia regis: the experience of the royal audience 106 IV. The Treasury 110 The treasury as social space 112 Rituals of gift exchange 114 V. The Forest 118 Hunting and aristocratic masculinity 119 Hunting grounds: parks and forests 121 Rituals of the hunt in the Carolingian period 124 VI. The Palace Church 127 The ‘liturgification’ of royal ritual in Francia 128 Space, liturgy, and royal vision in Asturias 130 Martyrial saints and public ritual in Benevento 132 VII. Conclusion 137 Part II. Infrastructure and Landscape Chapter 3. Palaces, Militarisation, and the Political Landscape of Ostrogothic Italy I. Introduction 140 From histories of the State to archaeologies of landscape 142 Overview 146 II. Northern Italy as a Royal Landscape 147 Mapping Ostrogothic palaces 148 How mobile was the Ostrogothic court? 154 III. ‘The Fear of Neighbouring Peoples’: Militarisation and the Spatial Architectonics 158 of the Italian Kingdom A ‘militarised’ landscape? The late-Roman background 158 Verona, Pavia, and the defence of Italy in the sixth century 165 IV. Feeding the People, Forming the Polity: Food Supply and the Fiscal Administration 176 in Ostrogothic Italy Subsistence crisis as political discourse in Ostrogothic Italy 176 Palaces and resource networks in Italia annonaria 181 V. Conclusion 187 Chapter 4. Mobility, Itinerancy, and Political Integration in the Regnum Francorum I. Introduction 189 The Frankish ‘palatial system’ and the problem of itinerancy 190 Overview 195 II. Authority on the Frontier: Austrasia 196 Military infrastructure and political space: northern Gaul from the third to fifth 197 centuries CE Palaces and the Austrasian political landscape 200 ‘The whole of the royal retinue was afloat on the waves’: mobility on the 204 Austrasian frontier III. From civitas to pagus: Neustria 207 viii Paris as symbolic centre 208 A place in the country: the palaces of the pagus Parisiacus 211 Regional politics and political integration 215 IV. Forging a New Empire: Carolingian Royal Landscapes 220 Carolingian royal mobility 222 Austrasia as a Carolingian royal landscape 235 Life on the limes: palaces beyond the Rhine 239 V. Conclusion 244 Conclusion 246 Appendix I: Royal Mobility under the Carolingians (751–840 CE) 249 Figures 256 Bibliography 359 ix LIST OF FIGURES Fig. 1 View of the so-called ‘Peristyle’ of the Palace of Diocletian, Split (Croatia). Late- third/early-fourth century CE. Photo: Wikimedia Commons. Fig. 2 Axonometric plan of the Villa Romana del Casale outside Piazza Armerina, Sicily. Fourth century CE. Wulf-Rheidt, ‘Die Bedeutung der severischen Paläste für spätere Residenzbauten’, fig. 5. Fig. 3 Axonometric reconstruction of the excavated villa at Cercadilla (Córdoba, Spain). Fourth century CE. Hidalgo Prieto, Espacio publico y espacio privado, fig. 5. Fig. 4 Plan of the fourth-century palace in the western quadrant of late ancient Milan. Soci- età Lombarda di Archeologia, http://milanoarcheologia.beniculturali.it. Fig. 5 The so-called ‘Torre di Ansperto’, San Maurizio al Monastero Maggiore. Photo: Wi- kimedia Commons. Fig. 6 Tower of San Maurizio al Monastero Maggiore. Photo: Wikimedia Commons. Fig. 7 Plan of the structures on the Via Brisa, Milan. Mirabella Roberti, Milano romana, fig. 79. Fig. 8 The so-called ‘Temple of Minerva Medica’, Rome. Fourth century CE. Photo: author. Fig. 9 Plan of the excavations at the site at 2–4 Via Gorani, Milan. Società Lombarda di Archeologia, http://milanoarcheologia.beniculturali.it. Fig. 10 Plan of late-ancient Trier. Trier: Kaiserresidenz und Bischofssitz, 351. Fig. 11 Exterior of the Constantinian aula, Trier. Early fourth century CE. Photo: Wikimedia Commons. Fig. 12 Detail of the late-Roman ceiling discovered under Trier Cathedral. Early fourth-cen- tury CE. Museum am Dom, Trier, Germany. Photo: ArtStor. Fig. 13 Plan of the palace-quarter in Thessaoliniki. Dey, Afterlife, fig. 2.4. Fig. 14 Plan of the Great Palace at Constantinople in the sixth century. Dey, Afterlife, fig. 3.4. Fig. 15 Plan of the Sessorian Palace. Fourth century CE. Colli, ‘Il palazzo sessoriano’, fig. 23. Fig. 16 Remains of the carceres of the Circus Varianus by Santa Croce in Gerusalemme, Rome. Photo: author. Fig. 17 The terminal wall (exterior) of the ‘Civil Basilica’ of the Sessorian Palace, Rome. Fourth century CE. Photo: author. Fig. 18 View of the Palatine from the southwest across the Circus Maximus. Photo: author. Fig. 19 Plan of the Palatine residences in the fourth century (upper level). Wulf-Rheidt, ‘The Palace of the Roman Emperors on the Palatine in Rome’, fig. 3. x Fig. 20 Fragments 8f and 8g of the Severan Marble Plan. Photo: Stanford Digital Forma Urbis Romae Project. https://formaurbis.stanford.edu/fragment.php?record=31. Fig. 21 Remains of the Severan baths on the southwestern platform of the Palatine (Domus Severiana). Late-second/third century CE. Photo: author. Fig. 22 Plan of the ‘Villa of Maxentius’ on the Via Appia. Ioppolo and Pisani Sartorio, eds., La villa di Massenzio sulla Via Appia, vol. 2, fig. 1. Fig. 23 The carceres of the Circus of Maxentius, Rome. 306–312 CE. Photo: author. Fig. 24 Elevation of the remains of the pulvinar at the Villa of Maxentius on the Via Appia. Pisani Sartorio and Calza, La villa di Massenzio sulla Via Appia, vol. 1, fig. 132. Fig. 25 Apse of the aula in the villa of Maxentius, Rome. 306–312 CE. Photo: author. Fig. 26 Plan of the centre of Rome in the fourth century. Ziemssen, ‘Die Kaiserresidenz Rom in der Zeit der Tetrarchie’, fig. 4. Fig. 27 Reconstruction of the southwestern Palatine in its Maxentian phase. Wulf-Rheidt, ‘Die schwierige Frage der Nutzung des römischen Kaiserpalastes auf dem Pa- latin in Rom in der Spätantike’, fig. 6. Fig. 28 View of the Basilica of Maxentius from the Palatine, Rome. 306–312 CE. Photo: au- thor. Fig. 29 Plan of the excavated structures on the Byrsa Hill, late-fifth or sixth century CE. Lé- zine, ‘Le “Palais” de Byrsa’, in Carthage-Utique, fig. 6, after Ferron and Pirard, ‘Les fouilles’, pl. 2. Fig. 30 Schematic plan of the representational structures on the Byrsa Hill: Lézine, ‘Le “Pal- ais” de Byrsa’, in Carthage-Utique, fig. 7. Fig. 31 Map of Ravenna, ca. 530 CE. Deliyannis, Ravenna, fig. 28. Fig. 32 Plan of the Palace of Theoderic. Early sixth century CE (Augenti Phase IV). Andrea Augenti, ‘Archeologia e topografia a Ravenna’, fig. 8. Fig. 33 Central field of the mosaic pavement of the triconchos in the Palace of Theoderic, Ra- venna. Early sixth-century CE(?). Berti, Mosaici antichi in Italia, cat. n. 60. Fig. 34 Palatium mosaic on the south nave wall of Sant’Apollinare Nuovo, Ravenna. Late- fifth/early-sixth century CE. Photo: author. Fig. 35 Detail of the Palatium mosaic, showing the severed hand of an excised figure. Sant’Apollinare Nuovo, Ravenna. Photo: author. Fig. 36 General plan of Reccopolis. Mid-sixth century CE. Henning et al., ‘Reccopolis Re- vealed’, fig. 3. Fig. 37 Plan of the ‘palace compound’ of Reccopolis. Mid-sixth century CE. Henning et al., ‘Reccopolis Revealed’, fig. 4. xi Fig. 38 View towards the basilica at Reccopolis. Mid-sixth century CE, with later additions. Photo: author. Fig. 39 The central piers of the lower storey of the residence at Reccopolis, looking towards the east. Mid-sixth century CE. Photo: author. Fig. 40 General plan of Caričin Grad (Justiniana Prima), with detail of the acropolis. Robert Ousterhout, Eastern Medieval Architecture, fig. 7.26. Fig. 41 Archaeological plan of the palace at Aachen. Müller et al, ‘Pfalz und vicus Aachen’, fig. 25. Fig. 42 Elevation of the palace chapel of Aachen. Late-eighth/early-ninth century CE. Kreusch, ‘Kirche, Atrium und Portikus der Aachener Pfalz’, figs. 4 and 5. Fig. 43 Interior view of San Vitale, Ravenna. 526–547 CE. Photo: author. Fig. 44 Interior of the palace chapel of Aachen, looking upwards into the dome. Architecture of the late-eighth/early-ninth century, with later additions and restoration. Photo: Lessing/Art Resource, NY (ArtStor). Fig. 45 Plan of the standing and excavated structures at Ingelheim. Lobbedey, ‘Carolingian Royal Palaces’, fig. 19b. Fig. 46 Isometric representation of the standing and excavated structures at Ingelheim. Chris- tian Rauch/Forschungstelle Kaiserpfalz Ingelheim. www.heritage- route.eu/ingelheim Fig. 47 Plan of the villa at Montmaurin (Haute-Garonne, France). Fourth century CE. Sfa- meni, Ville residenziali, fig. 12. Fig. 48 Detail of the palatium mosaic, Sant’Apollinare Nuovo, Ravenna. Photo: author. Fig. 49 Reconstruction of the villa at San Giovanni di Ruoti (Basilicata) in its late-fifth-century phase (Phase 3B). Small and Buck, Excavations, fig. 133. Fig. 50 Plan and elevation of the audience chamber in the episcopium of Poreč in its sixth- century phase (under bishop Euphrasius). After Matejčić and Chevalier, ‘The Episcopium of Poreč’, fig. 3. Fig. 51 Plan of the Merovingian ‘nobleman’s house’ at Camp de Larina, Lyon. Samson, ‘The Merovingian nobleman’s home’, fig. 1. Fig. 52 Axonometric reconstruction of the Merovingian ‘nobleman’s house’, Camp de Larina, Lyon. Samson, ‘The Merovingian nobleman’s home’, fig. 2. Fig. 53 An early medieval domus on the Forum Transitorium, Rome. Ninth century CE. Photo: author. Fig. 54 Plan of the palace of Paderborn ca. 799, overlain on the modern urban topography (shaded). Sveva Gai, ‘Die Pfalz Karls des Großen’, fig. 6. Fig. 55 Plan of the excavated palace structures in the region of Frankfurt Cathedral. Stamm, ‘Zur karolingischen Königspfalz in Frankfurt am Main’, fig. 3. xii Fig. 56 Reconstructed plan of the palace of Arechis in Salerno. Peduto et al., eds., Salerno. Una sede ducale della langobardia meridionale, plate 3, fig. 1. Fig. 57 Arcade of the palace incorporated into later structures on the Via Dogana Vecchia, Salerno. Photo: author. Fig. 58 Detail of the Corinthian capitals of the Arechian portico, Salerno. Photo: author. Fig. 59 Detail of the Corinthian capitals of the Arechian portico, Salerno. Photo: author. Fig. 60 Exterior of San Pietro a Corte, Salerno. Late eighth-century CE, with later additions. Photo: author. Fig. 61 Longitudinal section of San Pietro a Corte, Salerno. Paolo Peduto, 'Paolo Diacono e la Cappella Palatina di Salerno', fig. 5. Fig. 62 View of the eastern façade of Santa María del Naranco, Oviedo. 842–850 CE. Photo: author. Fig. 63 Isometric plan of Santa María del Naranco, Oviedo. Arbeiter and Noack-Haley, Christ- liche Denkmäler, fig. 101. Fig. 64 Detail of a roundel located in the spandrel of the interior arcade of Santa María del Naranco. 842–850 CE. Photo: author. Fig. 65 Yeavering, in Hope-Taylor’s phase IIIc. Mid-seventh century CE. Hope-Taylor, Yeavering, fig. 77. Fig. 66 Digital reconstruction of the so-called cuneus at Yeavering. Past Perfect. pastper- fect.org.uk. Fig. 67 Reconstructed axonometric plan of Santa Croce in Gerusalemme in the fourth cen- tury. Krautheimer, Corpus basilicarum, fig. 117. Fig. 68 Christ enthroned, flanked by angels, at the east end of the nave of Sant’Apollinare Nuovo, Ravenna. Late-fifth/early-sixth century CE. Photo: author. Fig. 69 Reconstruction of the palace group of Alfonso II, 791–842 CE. Arbeiter and Noack- Haley, Christliche Denkmäler, fig. 6. Fig. 70 The ‘Torre Vieja de San Miguel’, 791–842 CE, on the south side of the Cathedral of San Salvador, Oviedo. Photo: author. Fig. 71 The exterior of the Cámara Santa, 791–842 CE. Cathedral of San Salvador, Oviedo. Photo: author. Fig. 72 Silver missorium with image of Constantius II, found near Kerch. 350–361 CE. Her- mitage Museum, St. Petersburg. Photo: hermitagemuseum.org. Fig. 73 Silver missorium with image of Constantius II, found near Kerch. 350–361 CE. Her- mitage Museum, St. Petersburg. Photo: hermitagemuseum.org. Fig. 74 The Colonne di San Lorenzo, Milan. 376–378 CE. Photo: Scala/ArtStor. xiii Fig. 75 The submission of barbarians in the Hippodrome. Obelisk Base of Theodosius (north-western face), Istanbul. 390 CE. Photo: ArtStor. Fig. 76 Fourth-century mosaic of Christ with the Apostles, handing the text of the law to Peter (traditio legis), located in the chapel of S. Aquilino. San Lorenzo, Milan. Photo: Giovanni Dall’Orto/Wikimedia Commons. Fig. 77 Three-dimensional reconstruction of the Septizodium, Rome. Franck Deve- djian/Wikimedia Commons. Fig. 78 Nunburnholme Cross, Nunburnholme, East Yorkshire. Tenth century CE. Photo: Corpus of Anglo-Saxon Stone Sculpture, vol. 3, pl. 709. Fig. 79 Consular portrait of Constantius II in the Chronography of 354, f. 13r. 1620 CE. Vatican City, Biblioteca Apostolica Vaticana, Barb. lat. 2154. Photo: digi.vatlib.it Fig. 80 Missorium of Theodosius, 388 CE. Real Academia de la Historia Photo: Real Academia de la Historia, Madrid. Fig. 81 Christ and Barabbas before Pilate. Codex Purpureus Rossanensis, f. 16v. Sixth century CE. Museo Diocesano, Rossano. Photo: Archive for Research on Archetypal Symbolism/ArtStor. Fig. 82 Charles the Bald receiving his Bible from the monks of Tours. The First Bible of Charles the Bald, f. 423r. 845 CE. Paris, Bibliothèque nationale de France, MS. lat. 1. Photo: BnF Gallica. Fig. 83 Commemorative gold triple-solidus of Theoderic (The ‘Senigallia Medallion’), c. 500 CE(?). Museo Nazionale Romano, Palazzo Massimo alle Terme. Photo: fo- tosar.it Fig. 84 The Repton Stone. Seventh century CE? Derby Museum, Derby, UK. Photo: Wiki- media Commons. Fig. 85 The ‘Hornhausen Rider’. Seventh century CE? Photo: Landesamt für Denkmalpflege und Archäologie-Sachsen Anhalt, Halle, https://lda.sachsen-anhalt.de. Fig. 86 The Worcester Hunt Mosaic. Worcester Art Museum, Worcester, MA. Photo: Kon- doleon, ed., Antioch: The Lost Ancient City, fig. 2. Fig. 87 The Little Hunt mosaic from the Villa del Casale Romana, Piazza Armerina, Sicily. Early fourth century CE. Photo: Andrew Malone/Wikimedia Commons. Fig. 88 The so-called ‘Throne of Charlemagne’. Early-ninth century CE(?). Aachen Cathedral. Photo: Lessing/Art Resource, NY/ArtStor. Fig. 89 View upwards into the cupola of the Palatine Chapel, Aachen. Photo: Wikimedia Commons. Fig. 90 Engraving representing the cupola mosaic of the Palatine Chapel, Aachen, from Gio- vanni Giusto Ciampini’s Vetera monimenta in quibus praecipua (1690). Photo: Art- Stor. xiv Fig. 91 Exterior of the standing fabric of San Miguel de Lillo, Oviedo. 848 CE. Photo: author. Fig. 92 Isometric reconstruction of San Miguel de Lillo, 848 CE. Arbeiter and Noack-Haley, Christliche Denkmäler, fig. 83. Fig. 93 Plan of San Julián de los Prados, Oviedo. Arbeiter and Noack-Haley, Christliche Denk- mäler, fig. 58. Fig. 94 Interior view towards the eastern apses in San Julián de los Prados, Oviedo. Photo: Alamy. Fig. 95 Detail of the ‘basilica’ motif of the frescoes of San Julián de los Prados, Arbeiter and Noack-Haley, Christliche Denkmäler, pl. 17b. Fig. 96 Detail of the ‘basilica’ motif of the frescoes of San Julián de los Prados, Arbeiter and Noack-Haley, Christliche Denkmäler, pl. 17c. Fig. 97 Detail of the ‘basilica’ motif of the frescoes of San Julián de los Prados, Arbeiter and Noack-Haley, Christliche Denkmäler, pl. 17d. Fig. 98 Monumental cross in the upper register of San Julián de los Prados, located over the apse-arch. Photo: Alamy. Fig. 99 Plan of Santa Sofia, Benevento. Belting, Studien zur beneventanischen Malerei, fig. 12. Fig. 100 Southeastern exterior view of Santa Sofia, Benevento. Mid-eighth century. Photo: au- thor. Fig. 101 View across the nave of Santa Sofia, Benevento. Photo: author. Fig. 102 Arechis enthroned and crowned, ordering the construction of Santa Sofia in the Chronicon Sanctae Sophiae. Early-twelfth century CE. Bibliotheca Apostolica Vaticana, Vat. Lat. 4939, f. 26r. Photo: Digital Vatican Library. Fig. 103 Mosaic of the CIVI(TAS) CLASSIS. Sant’Apollinare Nuovo, Ravenna. Late- fifth/early-sixth century CE. Photo: author. Fig. 104 Detail of the ships in Classe’s harbour. Sant’Apollinare Nuovo, Ravenna. Photo: au- thor. Fig. 105 Palaces of the Ostrogothic rulers in Italy (493–553). Map produced with QGIS. Fig. 106 Insignia of the (western) magister officiorum in the Munich manuscript of the Notitia Dignitatum (Not. Dig. Occ. IX), 1542/1550–1551. Bayerische Staatsbibliothek 10291, fol. 207v. Photo: MDZ Digitale Bibliothek. Fig. 107 Plan of the fourth-century fort of Ad pirum (Hrušica): Ulbert, ed., AD PIRVM (Hrušica), appendix 1. Fig. 108 Insignia of the comes Italiae in the Munich manuscript of the Notitia Dignitatum (Not. Dig. Occ. XXIV). Bayerische Staatsbibliothek 10291, fol. 214r. Photo: MDZ Digitale Bibliothek. Fig. 109 Plan of late ancient Susa. Christie, From Constantine to Charlemagne, fig. 61. xv Fig. 110 Plan of Brescia, mid-fourth century to sixth century. Brogiolo, Brescia altomedievale, fig. 22. Fig. 111 Plan of the archaeological area of Castelseprio. Brogiolo and Gelichi, Nuove ricerche, fig. 38. Fig. 112 Plan of Pavia in the seventh century CE. Progetto Monasteri imperiali di Pavia. www.monasteriimperialipavia.it Fig. 113 The façade of San Giovanni in Borgo (Pavia) before its destruction in 1818. Ferdinand de Dartein. Musei Civici di Pavia. Fig. 114 The Iconografia Rateriana. Eighteenth century copy of a tenth century manuscript (based on a sixth century exemplar?). Verona, Bibliotheca capitolare MS CXIV. Photo: Wikimedia Commons. Fig. 115 Jerusalem on the Madaba Map Mosaic (Agios Georgios, Madaba, Jordan). Sixth cen- tury CE. Photo: Lessing/Art Resource NY. Fig. 116 Plan of the Sirmione peninsula in Late Antiquity. Christie, From Constantine to Charle- magne, fig. 65. Fig. 117 Plan of the site of the main building at Monte Barro (sector B). Brogiolo and Cas- telletti, Archeologia a Monte Barro, vol. 1, fig. 9. Fig. 118 Reconstruction of the grande edificio at Monte Barro (sector B). Brogiolo and Castelletti, Archeologia a Monte Barro, vol. 1, fig. 27. Fig. 119 Bronze hanging crown from Monte Barro (Sector B) and a reconstruction. Brogiolo and Castelletti, Archeologia a Monte Barro, vol. 1, figs. 52 and 54. Fig. 120 Map of public granaries in the Ostrogothic Kingdom of Italy. Map produced with QGIS. Fig. 121 Reconstruction of the late-ancient topography of Ravenna and Classe, superimposed on a modern satellite map. Augenti and Cirelli, ‘From Suburb to Port’, fig. 10.1. Fig. 122 Plan of the excavations at Podere Chiavichetta, Ravenna. Augenti and Cirelli, ‘From Suburb to Port’, fig. 10.2. Fig. 123 Buildings 1 and 2 at Podere Chiavichetta, Ravenna, viewed from the south west. Photo: author. Fig. 124 Section I of the Tabula Peutingeriana. Österreichische Nationalbibliothek, MS Cod. 324, first half of the thirteenth century, Colmar/Reichenau? Photo: ONB Digital. Fig. 125 Section II of the Tabula Peutingeriana. Österreichische Nationalbibliothek, MS Cod. 324, first half of the thirteenth century, Colmar/Reichenau. Photo: ONB Dig- ital. Fig. 126 Gaul on the Tabula Peutingeriana (Sections I and II), from Konrad Miller’s 1888 repro- duction. Wikimedia Commons. xvi Fig. 127 Rome and Ostia on the Tabula Peutingeriana (Section IV, here from Miller’s 1888 re- production). Wikimedia Commons. Fig. 128 ‘Francia’ on the Tabula Peutingeriana (Section II, from Miller’s 1888 reproduction). Wikimedia Commons. Fig. 129 Plan of the Roman fort at Divitia (Deutz). Caroll-Spilleke, ‘Das Militärlager Divitia in Köln-Deutz’, fig. 1. Fig. 130 Reconstruction of Cologne in the fourth century. Römisch-Germanisches Museum, Cologne. Wikimedia Commons. Fig. 131 Facsimile of the signet ring of Childeric I. Bibliothèque nationale de France, dépar- tement Monnaies, médailles et antiques, inv.56.460. Photo: BnF Gallica. Fig. 132 The palaces of Austrasia in the sixth- and early-seventh century. Map: author. Pro- duced with QGIS. Icons by Freepik. Fig. 133 Paris in Late Antiquity and the Early Middle Ages. Dierkens and Périn, ‘Les sedes regiae mérovingiennes’, fig. 3. Fig. 134 Palaces of the Pagus Parisiacus in the seventh century. Map: author. Produced with QGIS. Fig. 135 Palaces of the Carolingian Empire mentioned in text. Map produced with QGIS. xvii LIST OF ABBREVIATIONS Agnellus, LPRa Agnellus of Ravenna, Liber pontificalis ecclesiae Ravennatis, ed. De- borah Mauskopf Deliyannis, Corpus Christianorum. Continuatio Mediaevalis 199 (Turnhout: Brepols, 2006). Amm. Marc., Res Gestae Ammianus Marcellinus, Res Gestae = Ammiani Marcellini Rerum ge- starum libri qui supersunt, ed. Wolfgang Seyfarth, 2 vols., BSGRT (Leipzig: Teubner, 1978). Anon. Val. (II) Anonymus Valesianus, pars posterior, in Excerpta Valesiana, ed. Jacques Moreau, rev. Velizar Velkov, BSGRT (Leipzig: Teubner, 1968), 10–27. ARF Annales regni Francorum inde a. 741 usque ad 829, qui dicuntur Annales Laurissenses maiores et Einhardi, ed. Georg Heinrich Pertz, rev. Friedrich Kurze, MGH. SS rer. Germ. 6 (Hannover: Hahn, 1895). BSGRT Bibliotheca Scriptorum Graecorum et Romanorum Teubneriana Cass., Var. Cassiodorus, Variae, ed. Theodor Mommsen, MGH AA 12 (Ber- lin: Weidmann, 1894). CIL Corpus inscriptionum latinarum CISAM Centro italiano di studi sull’Alto Medioevo CSMLT Cambridge Studies in Medieval Life and Thought Deutsche Königspfalzen Deutsche Königspfalzen. Beiträge zu ihrer historischen und ar- chäologischen Erforschung (Veröffentlichungen des Max-Planck-Instituts für Geschichte 11/1–8). DOP Dumbarton Oaks Papers EME Early Medieval Europe Fred., Chron. Fredegar, Chronicon, in Fredegarii et aliorum chronica. Vitae sanctorum, ed. Bruno Krusch, MGH. SS rer. Merov. 2 (Hannover: Hahn, 1888), 1–168. Greg. Tur., Hist. Gregory of Tours, Historiae = Gregorii episcopi Turonensis libri Histo- riarum X, ed. Bruno Krusch and Wilhelm Levison, 2nd ed., MGH. SS rer. Merov. 1/1 (Hannover: Hahn, 1951). JRA Journal of Roman Archaeology JRS Journal of Roman Studies JSAH Journal of the Society of Architectural Historians MEFRA Mélanges de l’École française à Rome. Antiquité xviii MGH Monumenta Germaniae Historica AA Auctores antiquissimi Capit. Capitularia regum Francorum DD Arnulf Diplomata maiorum domus regiae e stirpe Arnulforum DD Merov. Diplomata regum Francorum e stirpe Merovingica DD Karol. 1 Diplomatum Karolinorum I Epp. Epistolae Fontes iuris Fontes iuris Germanici antiqui in usum scholarum separatim editi LL Leges LL nat. Germ. Leges nationum Germanicarum Poetae Poetae Latini medii aevi SS Scriptores SS rer. Germ. Scriptores rerum Germanicarum in usum scholarum se- paratim editi SS rer. Germ. N. S. Scriptores rerum Germanicarum. Nova series SS rer. Merov. Scriptores rerum Merovingicarum SS rer. Lang. Scriptores rerum Langobardicarum et Italicarum Paul. Diac., Hist. Lang. Paul the Deacon (Paulus Diaconus), Historia Langobardorum, ed. Ludwig Bethmann and Georg Waitz, in Scroptores rerum Langobardicarum et Italicarum, saec. VI–IX, MGH. SS rer. Lang. 1 (Hannover: Hahn, 1878), 12–187. PLRE Prosopography of the Later Roman Empire, ed. A. H. M. Jones et al., 3 vols. (Cambridge: Cambridge University Press, 1971– 1992). Proc., Wars Procopius, Wars = Procopii Caesariensis opera omnia, ed. Jakob Haury, vols. 1–2, BSGRT (Leipzig: Teubner, 1905). Ven. Fort., Carm. Venantius Fortunatus, Poèmes, ed. and French trans. Marc Reydel- let, 3 vols., Collection des universités de France (Paris: Les Belles Lettres, 1994–2004). xix INTRODUCTION Since our hall is recognised to have been constructed by skilled architects, the assidu- ous prudence of learned men ought to reside in it, as that wondrous beauty, if it is not promptly renewed, is marred by creeping age. These [halls] are the delights of our power, the fitting face of empire, the praiseworthy testimony of kingdoms: they are shown to legates caught in admiration and at first sight the master is believed to be as great as his dwelling. — Cassiodorus, Variae, VII.51 I. ‘The Fitting Face of Empire’ In the seventh book of his collected papers, the administrator and aristocrat Cassiodorus included a model for the formal letter appointing the curator palatii, the ‘guardian’ or ‘superintendent’ of the palace, a high-ranking functionary.2 His message to the curator-to-be is revealing on multiple levels, but none more so than its characterisation of the official’s charge, the royal palatium.3 Cassiodorus de- scribes the palace in terms that would not seem out of place today: the appearance of the residence as a manifestation of the owner’s status and prestige could just as easily apply to the Versailles of Louis XIV, or even the North American suburban home, as the court of the ‘barbarian’ king Theoderic (493–526 CE). In Cassiodorus’ prose, the palace was therefore the evidence of authority, the ‘testi- mony’ of kingly power, but at the same time a potent tool: a means of impressing, persuading, even dominating the viewer. Palaces and politics were thus inextricable. Two aspects of Cassiodorus’ letter should give us pause, however. First of all, little did Cassi- odorus realise that he wrote on the cusp of a revolution in not only the lexicon of architectural display, 1 Cass., Var., VII.5, 204: Aula nostra sicut agnoscitur peritis dispositoribus instituta, ita doctorum in ea diligens debet esse cautela, quoniam pulchritudo illa mirabilis, si subinde non reficitur, senectute obrepente vitiatur. haec nostrae sunt oblectamenta potentiae, imperii decora facies, testimonium praeconiale regnorum: haec legatis sub ammiratione monstrantur et prima fronte talis dominus esse creditur, quale eius habitaculum comprobatur. 2 For Cassiodorus’ career: James O’Donnell, Cassiodorus (Berkeley, CA: University of California Press, 1979), 13–32; Valérie Fauvinet-Ranson, Decor civitatis, decor Italiae, Monuments, travaux publics et spectacles au VIe siècle d’après les Variae de Cassiodore, Munera 23 (Bari: Edipuglia, 2006), 11–46. 3 The formulae of the sixth and seventh books of the collection were written by Cassiodorus self-consciously as models for future court officials and were, therefore, probably never officially sent: Cass., Var., preface, 14, 5. On the formulae: M. Shane Bjornlie, Politics and Tradition between Rome, Ravenna, and Constantinople: A Study of Cassiodorus and the ‘Variae’, 527–554, CSMLT, 4th series 89 (Cambridge: Cambridge University Press, 2013), 230–34. Although ‘abstracted’ from contemporary politics in this way (ibid., 234), they nonetheless offer a portrait of how legitimate, ethical governance was theorised in contemporary discourse. 1 but also the conceptualisation of the palace’s function within the political system. He certainly saw the palace as a means of connecting the Italian royal court at Ravenna to discourses surrounding ethical governance and legitimate authority inherited from the Roman imperial past.4 He exhorts the curator that he might ‘both preserve ancient things in pristine splendour and produce new things similar to the ancient: since, just as the body of ornaments combines to be clothed in one colour, in the same way the splendour of the palace ought to be diffused throughout every limb equally’.5 Yet with the collapse of Roman hegemony over its western provinces in the fifth century, the coalescence of new paradigms of authority, status, and prestige introduced equally transformative ideals of material dis- play. Cassiodorus lists the craftsmen and their luxury wares that would be under the curator’s charge— including wall masons, marble sculptors, brass casters, vault engineers, stucco carvers, and mosaic workers—yet within mere decades of his writing secular architecture in Europe, and particularly in its northern and north-western regions, would share few of these features until the turn of ninth century.6 In practical terms, the Western Empire’s fragmentation into smaller-scale, regional units, the collapse of the imperial fiscal system, and the shrinkage of the Mediterranean economy meant that in most areas the financial resources for architectural patronage on a ‘Roman’ scale was no longer possible. That the image of the palace we find in Cassiodorus’ letter seems so natural and familiar dis- guises the true nature and extent of its role in the political system. Surprising though it may seem, palaces have not formed an important component of the narratives surrounding the so-called ‘Trans- formation of the Roman World’. Palaces are, essentially, taken for granted: surely impressive stages 4 Another idea with arrestingly modern resonances, as witnessed in an executive order promulgated by U.S. President Donald Trump in December 2020, which states that ‘[the Founding Fathers] sought to use classical architecture to visually connect our contemporary Republic with the antecedents of democracy in classical antiquity, reminding citizens not only of their rights but also their responsibilities in maintaining and perpetuating its institutions’: Donald Trump, ‘Executive Order 13967: Promoting Beautiful Federal Civic Architecture (18 December 2020)’, Federal Register 85, no. 247 (December 23 2020): 83739. 5 Cass., Var., VII.5.3, 204: ut et antiqua in nitorem pristinum contineas et nova simili antiquitate producas: quia, sicut decorum corpus uno convenit colore vestiri, ita nitor palatii similis debet per universa membra diffundi. 6 Ibid., VII.5.5, 204–05: instructor parietum aut sculptor marmorum aut aeris fusor aut camerum rotator aut gypsoplastes aut musivarius. 2 for royal display, but separate to the power of the ruler who appears before it. This is where the majority of scholarship on post-Roman palaces has resided. Historians have described in great detail the ideologies of late ancient and early medieval regimes on the basis of textual sources, yet in most cases we are left with little sense of the role the palace played in the construction of authority in early medieval societies beyond thinly theorised ideas of symbolism or representation. More generally, the connection of the palace to broader social and economic systems through which the power of the ruler was actioned seldom enters into the discussion. We are thus left with a fuzzy, fragmented understanding of palaces, their socio-political func- tion, and how that function changed during the centuries following the Western Roman Empire’s putative ‘Fall’. This study seeks to unite these divergent—and all-too-often neglected—concerns to show that palaces were not passive expressions of royal power, but tools in the construction of new forms of rulership in the post-Roman West. By understanding them simultaneously as an architectural and a socio-political phenomenon, we will see that palaces did not just ‘advertise’ or ‘symbolise’ the authority of a ruler, but played a central role in its negotiation, legitimisation, and action. We will also see that these two sides of the palace, often reified as the ‘palace’ and the ‘court’, were neither separate nor strictly stable and constant; rather, palaces’ architectural form responded to new concepts of au- thority that were themselves in flux. Finally, we will question the centrality of ‘ideology’ for under- standing palaces in early medieval societies and, instead, show that they were essential instruments in the reordering of political space in the post-Roman West. II. The Historiography of Early Medieval Palaces Writing in 1993, the late Slobodan Ćurčić expressed optimism about the future of the study of late- ancient palaces. In his eyes, the field had reached an ‘important scholarly watershed’, with the ever- growing corpus of archaeological data offering an opportunity to ‘break the present major 3 historiographical stalemate’ that had plagued it for decades.7 This stalemate, as he saw it, was the result of polarisation between scholars of divergent methodological assumptions and objectives.8 Significant advances have indeed been made since the 1990s, but it must be said that Ćurčić’s prediction did not ultimately come to pass. The fact is—contrary to most people’s expectations—that palaces are not a usual feature in general histories either of Late Antiquity or the Early Middle Ages. The poor state of archaeological preservation and the paucity of structures that can be confidently identified in excava- tion only exacerbates matters. While emperors, kings, and their courts have traditionally been at the centre of historical imaginations of this period, a product of the authorship and interests of much of the surviving source material, their palaces have receded into the background. Tellingly, at the time of writing there is no general architectural or archaeological survey of palaces from the Late Roman period onwards.9 Before turning to this study’s approach, it is necessary first to sketch out how palaces have been treated in previous scholarship, especially in terms of the origins and intellectual stakes of Ćurčić’s ‘stalemate’. The problem of ‘palatial’ architecture: disciplinary compartmentalisation As Ćurčić noted, ‘palace studies’ has been stalled by the intersection of competing art-historical, his- torical, and archaeological methodologies. The dominant methodological strain in the twentieth cen- tury was art historical and, though not strictly a ‘school’, its proponents shared a common set of priorities. Karl Maria Swoboda was the first to make a systematic attempt at studying palace 7 Slobodan Ćurčić, ‘Late-Antique Palaces: The Meaning of Urban Context’, Ars Orientalis 23 (1993): 67. 8 Ibid. 9 Rollason’s recent The Power of Place, for example, though dealing with ‘royal sites’ generally and not only palaces, usefully highlights the role of architecture in the constructing an image of kingship (and thereby royal legitimacy), but his work is interested less in diachronic architectural or cultural change as much as illustrating what palaces have to tell us about the ‘unchanging’ aspects of power defined with reference to a Weberian tripartite schema: David Rollason, The Power of Place: Rulers and their Palaces, Landscapes, Cities, and Holy Places (Princeton, NJ: Princeton University Press, 2016). For ‘unchanging’: ibid., 4. Other studies are regionally or temporally bounded, about which see below. 4 architecture from Antiquity through to the close of the Middle Ages.10 He contended that there was a distinct typology of ‘palatial’ form in the Middle Ages characterised by an arcaded façade flanked by projecting towers (the Portikusvilla mit Eckrisaliten), which originated in Roman villa architecture and diffused widely in the medieval West and beyond.11 Influential too was the work of two Scandinavians: Ejnar Dyggve and Hans Peter L’Orange. Dyggve, a prolific archaeologist involved in the excavations of both the fourth-century palace of Galerius at Thessaloniki and the villa of Diocletian at Split, sought to demonstrate that a shared feature of late ancient palaces was the basilica ipetrale or discoperta, a portico- lined courtyard terminating in an elevated porch which was used for ceremonial appearances, analo- gous to the so-called ‘Peristyle’ at Split (Fig. 1).12 L’Orange, on the other hand, argued that a new, distinctly ‘late ancient’ aesthetic regime based on hierarchy, geometric order, and axial arrangement caused a revolution in the visual staging of imperial power – a cultural volte face that was vividly played out in the architecture of palaces.13 Common to all three is the implicit theorisation that a ‘palace’ could be defined as such on the basis of a select corpus of architectural forms that were not shared by other classes of building, and by extension that palaces of different periods shared a common set of features, transmitted iconographically over time and space. Their descriptions of such palatial forms were inevitably teleological, however. In L’Orange and Dyggve’s cases, for example, the Christian basilica is the ultimate end to which the ‘sacralised’ imperial palace led – an interpretation that is essentially circular in its logic, as the ascription of ‘sacral’ meaning to formal features such as axial 10 Karl Swoboda, Römische und romanische Paläste: eine architekturgeschichtliche Untersuchung (Vienna: Anton Schroll, 1919). See also: idem, ‘Palazzi antichi e medievali’, Bollettino del Centro di studi per la storia dell’architettura 11 (1957): 3–32; idem, ‘The Problem of the Iconography of Late Antique and Early Mediaeval Palaces’, JSAH 20, no. 2 (1961): 78–89. 11 Swoboda, Römische und romanische Paläste, 132–84. 12 Ejnar Dyggve, Ravennatum palatium sacrum: La basilica ipetrale per ceremonie. Studii sull’architettura dei palazzi della tarda antichità, Arkæologisk-kunsthistoriske Meddelelser 3, no. 2 (Copenhagen: E. Munksgaard, 1941) Cf. Géza de Francovich, Il Palatium di Teodorico a Ravenna e la cosiddetta architettura di potenza: Problemi d’interpretazione di raffigurazioni architettoniche nell’arte tardoantica e altomedioevale, Quaderni di commentari 1 (Rome: De Luca, 1970). 13 Hans Peter L’Orange, Art Forms and Civic Life in the Later Roman Empire (Princeton, NJ: Princeton University Press, 1965), esp. 69–85. See also: Irving Lavin, ‘The House of the Lord: Aspects of the Role of Palace Triclinia in the Architecture of Late Antiquity and the Early Middle Ages’, The Art Bulletin 44, no. 1 (1962): 1–27. 5 symmetry was made in the first place by analogy with the Christian basilica, as Kim Bowes has pointed out.14 This scholarship, and particularly the work of Dyggve, was the target of incisive criticism by the French archaeologist Noël Duval. In a series of articles, Duval advanced a restrictive view on what could be considered a ‘palace’ and, in so doing, thoroughly dismantled many of the prevailing inter- pretations advanced by Dyggve, L’Orange, and others.15 The idea of a recognisably ‘palatial’ architec- ture, he argued, placed too much emphasis on superficial similarity and overlooked fundamental differences between sites in both form and function.16 The strength of Duval’s arguments can be shown with the example of the so-called Villa del Casale Romana outside Piazza Armerina (Sicily) (Fig. 2). Piazza Armerina captured the scholarly imagination after its large-scale excavation in the mid- twentieth century and was immediately attributed to the emperor Maximian (286–305).17 Subsequent scholarship, however, has mostly placed the villa in a senatorial context, as the central mansion of the latifundia known as Philosophiana.18 A similar reflex has been evident in scholarly assessments of the fourth-century complex at Cercadilla, outside the Roman walls of Córdoba, which has also been at- tributed to Maximian’s patronage (Fig. 3).19 That there could be so much uncertainty about the 14 Kim Bowes, Houses and Society in the Later Roman Empire, Duckworth Debates in Archaeology (London: Duckworth, 2012), 22–23. 15 Particularly: Noël Duval, ‘Que savons-nous du palais de Théodoric à Ravenne ?’, MEFRA 72 (1960): 337–371 ; idem, ‘La place de Split dans l’architecture aulique du Bas-Empire’, Urbs 4 (1961): 67–95 ; idem, ‘La représentation du palais dans l’art du Bas-Empire et du Haut Moyen Age d’après le Psautier d’Utrecht’, Cahiers Archéologiques 15 (1965): 207–54 ; idem, ‘Comment reconnaitre un palais imperial ou royal : Ravenna et Piazza Armerina’, Felix Ravenna 115 (1978): 27–62. 16 Duval, ‘La place de Split’, 69–70. 17 Hans Peter L’Orange and Ejnar Dyggve, ‘È un palazzo di M. Erculeo che gli scavi da Piazza Armerina portano alla luce?’, Symbolae Osloenses 29 (1952): 114–28. Gino Vinicio Gentili, La villa romana di Piazza Armerina. Palazzo Erculio, 3 vols. (Osimo: Fondazione Don Carlo, 1999). Kähler maintained the ‘Herculean’ connection, but preferred an attribution to Maxentius: Heinz Kähler, Die Villa des Maxentius bei Piazza Armerina, Monumenta Artis Romanae 12 (Berlin: Mann, 1973), 32–35. For a review of the extensive scholarship, see: Carla Sfameni, Ville residenziali nell’Italia tardoantica, Munera 25 (Bari: Edipuglia, 2006), 29–46. 18 Andrea Carandini et al., Filosofiana. La villa di Piazza Armerina: immagine di un aristocratico romano al tempo di Costantino (Palermo: Flaccovio, 1982); Roger J. A. Wilson, Piazza Armerina, Archaeological Sites (London: Granada, 1983). 19 Rafael Hidalgo Prieto, Espacio publico y espacio privado en conjunto palatino de Cercadilla (Córdoba): El aula central y las termas, Colleción Arqueología: Cercadilla 1 (Seville: Junta de Andalucía, Consejería de Cultura, 1996), 149–56. 6 ownership of these complexes emphasises that, contra Swoboda and Dyggve, there was not a discrete category of architectural forms or features which were unique to imperial palaces per se. Duval’s arguments have been widely accepted, to such an extent that Late Roman palaces have by and large been abandoned as a field of study in favour of other forms of domestic architecture.20 Nonetheless, as Kim Bowes notes, the wider implications of the palatial architecture thesis, namely ‘the notion of a rigid, symmetrical architecture reflecting a rigid, ritualized society’ were left largely undisturbed.21 While archaeological investigation of late-imperial palaces continues to produce ex- traordinary discoveries, there has been little impetus to place these insights in a broader comparative framework for the most part. Rupture and degradation: temporal compartmentalisation The debate that raged over the palaces of the Late Roman period took place essentially separately to historical studies of palaces in the Early Middle Ages. If one were to take the scholarly landscape at face value, it would rather seem that palaces ceased to exist as a phenomenon after the fifth century, only reappearing in the age of the Carolingians in the eighth and ninth centuries.22 The palaces that, as written sources make abundantly clear, continued to form the basis of elite social and political life in the intervening centuries have not been the subject of systematic study to date. As a result, there is little sense in scholarship of how palaces changed over time formally, materially, or conceptually dur- ing this period. 20 Though see recently e.g.: Lynda Mulvin and Nigel Westbrook, eds., Late Antique Palatine Architecture. Palaces and Palace Culture: Patterns of Transculturation, Architectural Crossroads. Studies in the History of Architecture 5 (Turnhout: Brepols, 2019). 21 Bowes, Houses and Society, 23. 22 C.f.: Michael Featherstone et al, eds., The Emperor’s House: Palaces from Augustus to the Age of Absolutism, Urban Spaces 4 (Berlin: De Gruyter, 2015). This volume is especially welcome for its temporal and geographic scope and its comparative framing, but the lack of introductory or conclusory essays drawing out common themes between the individual contribu- tions means that there is little sense of comparison or diachronic evolution. 7 The palaces constructed on either side of the Western Roman Empire’s so-called ‘Fall’ have been treated as essentially separate phenomena. This is, to be sure, in large part a result of differing material evidence between these two periods. Late-imperial remains have both survived better and are better documented archaeologically (relatively speaking) than those of two centuries later, for example, and a general material-cultural shift in construction techniques from the sixth century onwards made secular architectures of that period so much less likely to survive intact and so much harder to detect in archaeology.23 This transformation is a fact of the material record—one explicable from multiple possible directions, from changing social mores to economic involution—yet it has frequently been taken to connote wholesale cultural change, even the tumult of the so-called ‘Migration Age’. Devia- tions from (an essentialised vision of) Roman monumentality in secular architectures are thus seen to indicate cultural degradation, at times under the influence of a separate ‘Germanic’ cultural tradition.24 The architecture of the early medieval palace has thus been assumed to be either archaeologically irrecoverable or a discrete phenomenon to the more complex Roman palaces of preceding centuries. This same dualism arises even in more sympathetic studies. Consider for a moment the so- called ‘Germanic hall’. Timber-hewn halls are overwhelmingly taken to belong to a very different ar- chitectural tradition to that of (stone-built) Roman architecture. For Charles McClendon, for example, the striking aspect of Charlemagne’s palace at Paderborn was ‘its apparent fusion of Germanic and Roman building traits. The design of the audience hall, with its long and narrow proportions, aisleless interior, and lateral entrance, belongs to the centuries-old tradition of a Germanic long hall.’25 It could equally be argued, however, that the aula at Paderborn—as at other sites in the Frankish world and beyond—was less a habit of a separate cultural tradition, but a recreation of a widespread and 23 On this material shift, see below, Chapter 1. 24 E.g. Michelangelo Cagiano de Azevedo, ‘I palazzi tardoantichi e altomedievali’, in Atti del XVI congresso di storia dell’archi- tettura, Athene, 29 settembre – 5 ottobre 1969 (Rome: Centro di Studi per la Storia dell’Architettura, 1977), 303–18. 25 Charles McClendon, The Origins of Medieval Architecture: Building in Europe, A.D 600–900 (New Haven: Yale University Press, 2005), 106. 8 adaptable formal arrangement that held significance both in the Roman world and in the societies located in its penumbra. Methodologically speaking, the idea of a universal ‘Germanic’ culture detect- ible in material evidence—whether ‘barbarian jewellery’ discovered in early medieval graves or, as here, architecture—strikes a discordant note in the context recent debates in early medieval archaeology, which has stressed the situational, discursive nature of ethnic identities and the epistemological diffi- culty of reconstructing those identities from material correlates.26 In any case, it is unclear how and why timber construction would necessarily connote ‘Germanic’ or ‘barbarian’ in the context of the ninth century, particularly when its plastered interior decoration and arrangement in a courtyard was clearly intended to recall late-Roman villa-architecture. Moreover, it is unsatisfactory to assume on the basis of (relative) architectural simplicity that the materialised ideologies of a ‘Germanic’ hall and a ‘Roman’ villa are not comparable. After all, as James Gerrard has argued, to construct a hall required control over substantial labour and resources, and simplicity in ground-plan does not necessarily equate to simplicity in the hierarchical relations performed within.27 The strict division between late ancient and early medieval palaces in most scholarship means that, while they are tacitly acknowledged to constitute a thread of cultural continuity from Antiquity to the Middle Ages and beyond, scholarship has little sense of the dynamics of that continuity: that is, how the palace changed physically or ideologically in the intervening centuries. At the heart of this study lies the belief that more can be said about early medieval palaces and their role in the social formations and elite cultures in which they were situated, and that charting their material and concep- tual evolution from the fourth century onwards offers new perspectives on rulership after Rome. 26 For an overview, see recently: Mattias Friedrich and James Harland, eds., Interrogating the ‘Germanic’: A Category and its Use in Late Antiquity and the Early Middle Ages, Ergänzungsbände zum Reallexicon der Germanischen Altertumskunde 123 (Berlin: De Gruyter, 2021). McClendon’s interpretation that the palace’s ‘formal dichotomy’ therefore reflected ‘Charle- magne’s dual role […] as Germanic chieftain and Christian ruler, as conqueror of the heathens and patrician of the Romans’ is unconvincing: McClendon, Origins, 106. 27 James Gerrard, The Ruin of Roman Britain: An Archaeological Perspective (Cambridge: Cambridge University Press, 2013), 192–94. 9 Politics in place: palaces, the sedes regiae, and the Hauptstadtproblem In most recent scholarship, palaces have only played a role insofar as they defined the sedes regiae (‘royal seats’), the Roman civitates adopted by the rulers of the early medieval successor kingdoms as the sites of their palaces. That is to say, any interest in palaces has been subsumed by and large into urban history, especially in terms of the debates surrounding the fates of cities in the post-Roman West, rather than as architecture as such.28 All too often this scholarship is often specialised along regional or national lines, each with distinct historiographies and methods. The most extensive example (though principally focused on a later period) is the ongoing Deutsche Königspfalzen series, which has brought together archaeological and textual testimonia for German royal palaces in a geographically organised repertorium paired with a co-ordinating series of interpretive essays.29 For no other region is there a comparably comprehensive endeavour, and in most cases the substantial (and often underestimated) archaeological evidence for palaces is contained in individual site reports or, at most, regional studies. Palace studies in Germany, or Pfalzenforschung, has an independent historiography, methods, and set of priorities that—while in many ways useful as a methodological toolbox, as Chapters 3 and 4 of this study will show—can be difficult to reconcile with evidence from other regions and time periods.30 Despite the particularity of its concerns, scholarship in this tradition on the Frankish (and, later, Ger- man) Empires of the Early Middle Ages have proven influential in how the sedes regiae have been conceptualised.31 28 See especially: Gisela Ripoll and Josep Gurt, eds., Sedes regiae (ann. 400–800) (Barcelona: Rreial Acadèmia de Bones Lletres, 2000). The contributions to this often-cited volume are valuable, yet the focus lies mostly on matters of urban topography and its change over time rather than the palaces located within them as such. 29 Max-Planck-Institut für Geschichte, ed., Die deutschen Königspfalzen. Repertorium der Pfalzen, Königshöfe und übrigen Aufenthalts- orte der Könige im deutschen Reich des Mittelalters, 5 vols. (Göttingen: Vandenhoeck & Ruprecht, 1986–2016). On the project, see: Caspar Ehlers, ‘Pfalzenforschung Heute. Eine Einführung in das Repertorium der deutschen Königspfalzen’, in Orte der Herrschaft. Mittelalterliche Königspfalzen, ed. Caspar Ehlers (Göttingen: Vandenhoeck & Ruprecht, 2002), 25–54. 30 Stuart Airlie, ‘The Palace Complex’, in Diverging Paths? The Shapes of Power and Institutions in Medieval Christendom and Islam, ed. John Hudson and Ana Rodriguez (Leiden: Brill, 2014), 260–61. 31 Here, I will attend to the general methodological and theoretical background. The specific consequences of this influence as they pertain to Italy and the Frankish kingdoms will be set out in greater detail in Chapters 3 and 4. 10 To illustrate this point, let us begin with two German historians of the past century whose work has proven highly influential in the study of the sedes regiae: Eugen Ewig and Carlrichard Brühl.32 Post-war German scholarship understood Europe’s regionalisation after the fifth century through the lens of the persistence, disappearance, and transformation of ‘capitals’, Hauptstädte. It was understood that ‘capitals’ as such were a feature essentially of European Antiquity, disappearing at the end of the Roman period and emerging once again only from the twelfth century onwards.33 In an influential article, Ewig challenged this broad-strokes narrative, arguing that the sedes regiae represented a signifi- cant institutional continuity with the Late Roman period in many respects.34 The new ‘residences’ in Roman administrative centres might ‘recall’ (rappellent) the capitals of the Late Empire, constitute ‘more-or-less-exact replicas’ of an ‘imperial model’, or even transform into capitals over time, yet it is clear from his analysis that this continuity was not universal and he acknowledges his model is strained especially by evidence from the Frankish kingdom.35 Brühl, on the other hand, argued that the fifth and sixth centuries saw a shift from centralisation in the ‘capital’ (when the ruler’s seat and the centre of administration were one and the same) to decentralisation in the ‘residence’ (when they were sepa- rately located).36 At most, the new residences might have a ‘capital-like’ character, but were nonetheless to be distinguished.37 The transition from the centralised Roman capital to a multitude of residences 32 To be clear, Ewig and Brühl’s treatment of these themes is not unique, yet their work not only expresses the overall conceptual framework most clearly but has also proven influential in subsequent scholarship. 33 E.g. Wilhelm Berges, ‘Das Reich ohne Hauptstadt’, in Das Hauptstadtproblem in der Geschichte. Festgabe zum 90. Geburtstag Friedrich Meineckes gewidmet vom Friedrich-Meinecke-Institut an der Freien Universität Berlin, ed. Mitarbeiter des Friedrich-Mei- necke-Instituts (Tübingen: Niemeyer, 1952), 1–29. 34 Eugen Ewig, ‘Résidence et capitale pendant le haut Moyen Age’, Revue historique 230, no. 1 (1963): 25–72. 35 Ewig, ‘Résidence et capitale’, 25, 72. Ewig implies that Toledo began as a ‘residence’ that evolved into a ‘capital’ patterned after Constantinople: ibid., 32–36. He also argues that Paris became a ‘capital’ of the Merovingian kingdom, (ibid., 47–53), but in closing he admits that the development of the civitas regia was more-clearly perceptible in Spain and Italy than in Gaul on account of the the ‘deeper roots’ of Mediterranean Roman traditions and the frequent divisions of the Frankish regnum: ibid., 71. 36 Carlrichard Brühl, ‘Remarques sur les notions de «capitale» et de «résidence» pendant le haut Moyen Âge’, Journal des savants, 1967, 193–215. For the importance of the administration for ‘capital’ status, see e.g.: Aloys Schulte, ‘Pavia und Regensburg’, Historisches Jahrbuch 52 (1932): 472; idem, ‘Anläufe zu einer festeren Residenz der deutschen Könige im Hoch- mittelalter’, Historisches Jahrbuch 55 (1935): 132. 37 Brühl, ‘Remarques’, 207; idem, ‘Zum Hauptstadtproblem im frühen Mittelalter’, in Festschrift für Harald Keller: zum sech- zigsten Geburtstag, dargebracht von seinen Schülern, ed. Hans Martin von Erffa and Elisabeth Herget (Darmstadt: E. Roether, 1963), 45–70. 11 was a symptom of the disappearance of broader structures of rulership: a stable administration is incompatible with an itinerant court, which they agreed became the principal mode of rule from the Carolingian period onwards.38 There were, however, important exceptions: Ravenna during the reign of Theoderic and his successors, Langobard Pavia, and Visigothic Toledo. These cities maintained or revived an ancient tradition, namely a single city forming the focus of ‘national’ life as the principal royal residence, the foremost ecclesiastical centre, and the seat of the administration.39 For both, these cities had their archetype not in the immediate Roman past as such, but in the eastern imperial seat at Constantinople – or, significantly, in Ravenna as its proxy, itself assumed to be modelled on Constan- tinople.40 Even if more recent scholarship has eschewed such a fixation on ‘capitals’ as such, the basic shape of the narrative expressed by Ewig and Brühl is still discernible in general treatments of the sedes regiae.41 It often goes unacknowledged, however, that this debate arose out of a specific set of concerns in German academe. On the one hand, the Hauptstadtproblem or Hauptstadtfrage, as it was known in scholarship, was a product of contemporary politics. The partition of Germany (and Berlin itself) in the aftermath of the Second World War, as well the subsequent debate surrounding where the new western Bundesrepublik’s capital should be located, form the immediate political background.42 But it also expressed longer-term historiographical trends traceable to nineteenth century and earlier.43 A 38 Brühl, ‘Remarques’, 199–200; Ewig, ‘Résidence et capitale’, 53–65. For critique of how Ewig and Brühl’s work has shaped attitudes on the Carolingian period: Rosamond McKitterick, Charlemagne: The Formation of a European Identity (Cam- bridge: Cambridge University Press, 2008), 175–78. 39 Brühl, ‘Remarques’, 202–06; Ewig, ‘Résidence et capitale’, e.g. 36, 40, 42–47. 40 Ewig, ‘Résidence et capitale’, 29, 36, 39, 70; Brühl, ‘Remarques’, 206. On the framework of interpreting Ravenna’s topography in light of Constantinople’s, see below: Chapter 1. 41 See e.g.: J. H. W. G. Liebeschuetz, ‘Ravenna to Aachen’, in Sedes regiae (ann. 400–800), 9–30, as well as many of the contributions to the same volume. 42 Brühl, ‘Remarques’, 193. See generally the contributions to: Das Hauptstadtproblem. The editors explicitly acknowledge the political context of their chosen theme, and stress its relevance given Meinecke’s role in Berliner politics (including serving as the first president of the Freie Universität): ‘Vorwort der Herausgeber’, in Das Hauptstadtproblem, ix. 43 Walter Schmidt, ‘Die deutsche Hauptstadt-Diskussion in der Revolution von 1848/49’, Sitzungsberichte der Leibniz-Sozietät 6 (1995): 5–30; Zef Segal, ‘The Two-Edged Sword: Capital Cities and the Limits to State Centralization in Mid Nineteenth- Century Germany’, Journal of Historical Geography 60 (2018): 52–63 (for the political contexts of the Hauptstadt debate in the nineteenth century: ibid., 53–54). 12 stable, centralised administration was understood to be a precondition of Staatlichkeit (‘statehood’) and thus Germany’s particular historical trajectory from the medieval ‘empire without capital’ to modern nation-state revolved in part around its lack of a definitive capital, in contrast to France or England.44 What was at stake in this debate when applied to the Early Middle Ages, therefore, was not the disappearance of a particular urban form, or even type of rulership as such, but the persistence of the ‘State’ following the Roman Empire’s disintegration. Needless to say, the existence of a medieval ‘State’ has been a contentious subject of debate – and it is not my intention to relitigate it here.45 It is significant, however, that the presence of the State as the guiding framework has (tacitly) prioritised certain sets of institutional relationships as diagnostic of the transition from Roman to post-Roman in the political sphere. ‘We have not yet weaned ourselves away from a conception of history where the State is central, the source of all meaning and goodness’, as Conrad Leyser has put it.46 Many of the trends that Ewig, Brühl, and other scholars of their time identified do not hold up well when tested against the evidence. Take, for example, the conceptual opposition of stability and itinerancy. Despite how frequently it is repeated in scholarship, that early medieval kings were ‘itinerant’ is largely assumed rather than demonstrated. Itinerarforschung (‘itinerary studies’) has evolved as a distinct methodology in German historiography, using evidence for the ruler’s movements 44 Berges enumerates six forces that prevented the Reich from becoming a ‘State’ in the Middle Ages, which is equated with the presence of a stable central place of royal/imperial government (e.g. the frequent movement of the centre of power hindered the development of central places; the geographic instability and personality of central institutions; the promi- nence of the Church and ecclesiastical magnates in the empire’s governance): Wilhelm Berges, ‘Das Reich ohne Hauptstadt’, 9–16. 45 For an overview: Susan Reynolds, ‘The Historiography of the Medieval State’, in Companion to Historiography, ed. Michael Bentley (London: Routledge, 1997), 117–38. For a positive view, see e.g. the contributions to Staat im frühen Mittelalter, ed. Stuart Airlie et al., Forschungen zur Geschichte des Mittelalters 11 (Vienna: Verlag der Österreichischen Akademie der Wissenschaften, 2006), and esp. Walter Pohl, ‘Staat und Herrschaft im Frühmittelalter: Überlegungen zum Forschungs- stand’, 10–38, for a historiographical appraisal. The classic text in English is: Joseph Strayer, On the Medieval Origins of the Modern State (Princeton: Princeton University Press, 1970). Strayer was in no doubt that there was no ‘State’ before around 1200: ibid., 12–16. Scepticism of the utility of the ‘State’ as an object of analysis is particularly salient among historians influenced by social anthropology, especially in Britain. See the methodological comments in: Conrad Leyser, ‘Introduc- tion: Making Early Medieval Societies’, in Making Early Medieval Societies: Conflict and Belonging in the Latin West, 300-1200, ed. Kate Cooper and Conrad Leyser (Cambridge: Cambridge University Press, 2016), 1–15. 46 Leyser, ‘Making Early Medieval Societies’, 2. 13 through the realm (principally charters) as evidence for social, economic, and ‘constitutional’ relations in the absence of central administrative apparatuses.47As we will see in Chapters 3 and 4, however, the focus on the royal iter was originally a means of analysing the near-constant peripatesis of the Ottonian and Salian emperors; its projection onto earlier periods presents methodological challenges. Likewise, that post-Roman regimes patterned their rule on the Byzantine Empire, adopting singular ‘quasi-cap- itals’ in which their power was centralised, is also largely assumed. Chapter 3 shows that the stability of Theoderic and his royal successors in a singular ‘capital’ at Ravenna falls apart upon closer inspec- tion of the evidence. Likewise, in Chapter 4, we will witness the essential continuities in the geogra- phies of power in the Frankish regnum between the fifth and ninth centuries; the ascent of the Carolingians represented less a revolution in the practices of rulership than an intensification of exist- ing ones. For this reason, the more neutral ‘mobility’ is preferred to ‘itinerancy’ and its cognates for denoting movements between palaces. At a more basic level, the binary opposition of ‘itinerancy’ and ‘stability’ (reified as the ‘residence’ and the ‘capital’) as mutually incompatible ideal types does not stand up well when applied to the rather more flexible politics of the Early Middle Ages. Where are palaces in all of this? To summarise, the dominant approaches towards palaces in the Early Middle Ages have treated them as an abstraction: indicators of the extent, economic basis, and intensity of royal power, rather than as architecture as such. Moreover, as we will see in greater detail below, the theoretical orientation of the prevailing frameworks towards institutional histories of the State risks decontextualising the production of authority from the places and spaces in which it occurred. While the focus on the sedes regiae has shed valuable light on the transformations of Europe’s urban geography, and the methodologies of Pfalzen- and Itinerarforschung have usefully directed our 47 For an overview of Itinerarforschung, see e.g.: Rudolf Schieffer, ‘Von Ort zu Ort. Aufgaben und Ergebnisse der Erfor- schung ambulanter Herrschaftspraxis’, in Orte der Herrschaft, 11–24. For a discussion of the methodological problems of ‘itinerancy’, see below: Chapter 4. 14 attention to the connection between royal authority and the economic and political structures that supported it, palaces offer an opportunity for a much-needed reassessment. III. What Is A Palace, Anyway? If our goal here is to move palaces from their current place on the periphery to the centre of our understandings of early medieval politics, we must first define the subject matter itself. What counts as a ‘palace’ is far from self-evident, and there have been strikingly few attempts to define it as a category of analysis. Rollason, to cite one recent example, adopts a deliberately broad definition: the palace as simply ‘the residence of a ruler’, which he acknowledges ‘embraces many types of residence which would not necessarily have been termed palaces by contemporaries. In other words, what char- acterizes them is simply their construction for, or use by, rulers.’48 Taking a ‘common sense’ approach like this has both advantages and disadvantages. Usefully, the breadth of such a definition does not impose expectations of architectural form or function, but it risks becoming so broad as to be un- workable. The difficulties here are well-demonstrated by Split. Constructed at the beginning of the fourth century to serve as the quiet retirement home of the emperor Diocletian, Split is undoubtedly important for discussions of secular architectural patronage, but by definition did not house the ruling emperor and is not termed palatium in any near-contemporary source.49 While certainly part of this story, to base our theorisation of a ‘palace’ on its basis—as did L’Orange and Dyggve as we saw above—would be misleading. Contemporary sources are, unfortunately, little help here. No contemporary author defines what, to them, constituted a palace as such, and the slipperiness of the term is reflected in the lexical variety we encounter in the sources: palatium, aula, villa, mansio, curtis, and so on. To compound matters, 48 Rollason, Power of Place, 9–10. 49 For analysis of the term villa and its use to describe Split: Ivan Basić, ‘Diocletian’s villa in Late Antique and Early Medieval Historiography: A Reconsideration’, Hortus Artium Medievalium 20, no. 1 (2014): 63–76. The best treatment of Split’s archi- tecture remains: John Joseph Wilkes, Diocletian’s Palace, Split: Residence for a Retired Roman Emperor (Sheffield: University of Sheffield, 1986), 15 these words could carry specific force to differentiate between sites with differing functions—some- thing we will encounter often with palatium and villa—but such distinctions are not universal nor con- sistently applied. Moreover, and as was implied already by Cassiodorus’ letter with which we began, ‘palace’ denoted not just the physical structure, but could also (as in modern English) stand metonym- ically for the royal entourage, the institutions of royal power, or government writ large.50 To cite but one example, in the Passio Leudegarii, the late-seventh century life of the bishop of Autun, Leudegar (bishop ca. 662–676), the narrator blurs these potential meanings together: ‘Then from the palace came a royal decree, the court’s command: that Leudegar must live no longer.’51 Contemporary terminology, though not useless, is a poor guide on its own. As we have seen, ‘objective’ criteria, such as a structure possessing a particular formal arrangement or architectural fea- tures are a poor indicator of ownership in the absence of other information, and additionally presents challenges in accounting for long term cultural change.52 It is better, therefore, to adopt a flexible and functional definition of a ‘palace’ rather than one rooted in the lexicon of any one period. A ‘palace’ is thus best described as 1. a residence in the sole possession of the ruler or dedicated solely for their use, that 2. formed the symbolic and institutional seat of a regime’s administration, either permanently or intermittently, and relatedly, 3. provided a spatial focus for the interaction of the ruler with repre- sentatives of the political community and/or of other rulers.53 Importantly, what constitutes a 50 Brühl, ‘Remarques’, 208; Josiane Barbier, ‘Le système palatial franc : genèse et fonctionnement dans le nord-ouest du regnum’, Bibliothèque de l’École des Chartes 148, no. 2 (1990): 249. 51 Passio Leudegarii episcopi et martyris Augustodensis, ed. Bruno Krusch, c. 34, in Passiones vitaeque sanctorum aevi Merovingici, ed. Bruno Krusch and Wilhelm Levison, vol. 3, MGH. SS rer. Merov. 5 (Hannover: Hahn, 1910), 315: Tunc a palatio sententia mandatur decreti, Leudegarium diutius vivere non debere. Trans.: Passio Leudegarii, in Late Merovingian France: History and Hagiography, 640–720, trans. Paul Fouracre and Richard Gerberding, Manchester Medieval Sources (Manchester: Manchester University Press, 1996), 249. 52 Needless to say, any ‘subjective’ assessment of size or grandeur as defining of a palace would be both overly broad and theoretically unsatisfactory. 53 Helpful here is the definition advanced in: Josiane Barbier, ‘Les lieux du pouvoir en Gaule franque: l’exemple des palais’, in Places of Power – Orte der Herrschaft – Lieux du Pouvoir, ed. Caspar Ehlers, Deutsche Königspfalzen 8 (Göttingen: Vanden- hoeck & Ruprecht, 2007), 228–30. 16 ‘residence’, or ‘administration’, must be defined contextually, rather than absolutely on the basis of external criteria.54 Towards a socio-spatial approach If we return momentarily to Cassiodorus’ formula on the curator palatii, we observe that the was an effective ‘testimony’ of royal power due to the effect it had on visitors, ‘legates’ (legati), plausibly both foreign embassies and delegations from local elites. That is to say, Cassiodorus recognises that the palace does not construct authority on a purely symbolic or ideological plane—what we might term ‘propaganda’—but a discursive one, in the context of the social relations in which it is implicated. As Emanuel Mayer has persuasively argued, while most research has been content to treat palaces and other ‘official’ architectures as either ‘propaganda’ or even part of a ‘subtle game of staged applause’, in his words, their political role is better explained as originating out of the discursive exchanges be- tween ruler and the social elite – in the shared values, norms, and expectations governing the action and distribution of political power.55 We can, therefore, read Cassiodorus’ description of the palace, the ‘fitting face of empire’, in a different way. Rather than a mere backdrop for royal power already understood to be per se legitimate, the palace was an agent in producing that very legitimacy. A brief example from the fourth-century poet Ausonius (d. 395) is useful for explicating this connection. When he writes that he ‘returned from the palace after many years as a most-honoured man, even a consul, to his roots, his little estate’, the reader clearly understands that he refers to the 54 ‘Administration’ is meant here in the vein of Weber and his concept of ‘administrative staff’ (Verwaltungsstab). By ‘staff’ Weber does not designate (only) a professional bureaucrat, rather the figures surrounding the ruler involved in organising and enacting political power. The ‘staff’ need not be a ‘professional’ in the modern sense, but could also include (for example) domestic office-holders—e.g. the chamberlain, the maior domus, the chancellor—who held aristocratic rank and were bonded to the ruler through oaths: see e.g. Max Weber, Economy and Society: A New Translation, trans. Keith Tribe (Cambridge, MA: Harvard University Press, 2019), 3.3.6, 354–74 (Weber’s definition of ‘Traditional Rule’, with frequent analogy to the Middle Ages). As is clear in Weber’s text, what constituted the ‘administration’ and the ‘staff’, however, must be defined in context rather than with analogy to modern expectation. 55 Emanuel Mayer, Rom ist dort, wo der Kaiser ist: Untersuchungen zu den Staatsdenkmälern des dezentralisierten Reiches von Diocletian bis zu Theodosius II. Römisch-Germanisches Zentralmuseum. Monographien 53 (Mainz: Verlag des Römisch-Germanischen Zentralmuseum, 2002), 4–27. 17 palatium not only as a physical building but simultaneously as an abstract, social space.56 The simplicity of Ausonius’ statement not only disguises a number of interrelated ideologies, but also how those ideologies were expressed spatially. For example, underlying the opposition between his herediolum (‘little patrimony’) and the palatium is an understanding that public service in the imperial household (negotium), at that time in Milan, was where one made one’s name, whereas the provincial estate was for quiet retirement (otium) – a duality which was characteristic of the ideal Roman aristocratic lifestyle. On a broader scale, his departure from court and return to Bordeaux reflects the spatial organization of the Roman Empire writ large: power was concentrated at the imperial centre—produced in the personal interactions of the ruler with courtiers, officials, and other members of the regime—but exercised in comparatively peripheral provinces.57 Ausonius’ characterisation of the palatium’s connection to the spatial ordering of the wider Roman world points us towards two essential points for this study. Firstly, the palatium was ontologi- cally dual. It is simultaneously a physical, architectural space, ‘space per se, space as a contextual given’, but also a locus of ‘socially-based spatiality, the created space of social organization and production’, the medium of social relations overlying physical space through which power is actualised.58 As Lefebvre perceptively argued, this abstract space of social relations is permeated by ideology – as perceived in the routes and places by which social life is physically organised; as conceptualised in representation; as experienced through signs and symbols.59 Its contingency is well expressed by 56 Ausonius, De herediolo, proemium, in Decimi Magni Ausonii Burgidalensis Opuscula, ed. Sextus Prete, BSGRT (Leipzig: Teu- bner, 1978), 89: Cum de palatio post multos annos honoratissimus, quippe iam consul, redisset ad patriam, villulam, quam pater reiquerat. Ausonius’ career: PLRE, 1:140–41. 57 ‘Centre’ does not (necessarily) refer to a geometrical/geographical middle point, but a ‘center of the order of symbols, of values and beliefs, which govern the society’, thus denoting where legitimacy is negotiated and authority is produced: Edward Shils, ‘Center and Periphery’, in Edward Shils, Center and Periphery: Essays in Macrosociology, Selected Papers of Ed- ward Shils 2 (Chicago: University of Chicago Press, 1975), 3–16 (Quote at 3). The presence of a ‘centre’ presupposes a ‘periphery’ – an opposition of those with access to power and those disbarred from it: Edward Soja, Postmodern Geographies: The Reassertion of Space in Critical Social Theory (London: Verso, 1989), 149–50. 58 Soja, Postmodern Geographies, 79, 120. 59 Henri Lefebvre, The Production of Space, trans. Donald Nicholson-Smith (Oxford: Blackwell, 1991), 38–39. 18 Lefebvre’s famous maxim ‘(social) space is a (social) product’.60 We will see in Chapter 2 how the architecture of the palace was inextricably connected to its conceptual evolution—its ‘shaping’ in re- sponse to evolving concepts of legitimate authority—through the lens of ritual. Ausonius also points us towards a second important point: the social space of the palace is not separate to the rest of a society’s organisation. ‘Societies are constituted of multiple overlapping and intersecting sociospatial networks of power’, as Mann described it.61 Viewing palaces in this way provides a vocabulary for investigating the link between the images, discourses, and rituals by which a ruler’s right to command obedience (legitimacy) was communicated and the action of those commands on bodies subjected to them, in for example the organisation of labour, the extraction of surplus, or physical violence.62 As we will see in Chapter 3, palaces were not separable from the economic and institutional apparatuses by which political power was enacted, and in Chapter 4 that their very ‘cen- trality’ made them an instrument of political integration and intensification. IV. Looking Ahead This study is organised into two co-ordinating parts, each with distinct (though related) foci and meth- odological orientations. Part I, ‘Architecture and Space’, considers ‘the palace’ in the Early Middle Ages as an architectural and conceptual category, in order to chart how it changed visually and sym- bolically between the fourth and ninth centuries CE. Through an architectural analysis of palace sites, Chapter 1 argues that palaces underwent a radical transformation after the Western Roman Empire’s fragmentation in the fifth century. Whereas the palaces of the Late Roman Empire were configured 60 Ibid., 26. 61 Michael Mann, The Sources of Social Power, vol. 1, 2nd ed. (Cambridge: Cambridge University Press, 2010), 1. Emphasis added. 62 Weber understood legitimacy to be the ‘belief’ that underlay a ‘minimum willingness to obey’ or ‘interest in obedience’ in the context of a relationship of rule: Weber, Economy and Society, 3.1.1, 338–41. It is thus related to Herrschaft, ‘rulership’, being the ‘Chance that a command of a particular kind will be obeyed by given persons’: ibid., 1.2.16, 134. Smith adds a useful gloss to this, defining legitimacy as ‘the ability of a regime to synchronize practices that perpetuate the existing political order within a discursive framework that generates the allegiance of subjects’: Adam T. Smith, The Political Land- scape: Constellatons of Authority in Early Complex Polities (Berkeley, CA: University of California Press, 2003), 108. 19 according to a common ‘grammar’, one attuned to the particular symbolic languages of imperial au- thority, this ideal broke down precipitously in the sixth century. In its place coalesced a new material and spatial configuration that responded to a new lexicon of social display. In Chapter 2 we turn from architectural to social forms, to examine the rituals that mediated communication between ruler and elite inside the palace. First, I show that ritual was not a passive expression of ‘royal ideology’ or similar, but a means of negotiating and defining ideals and expectations of authority and legitimacy. By considering these social and material exchanges from the perspective of the palace’s ‘social space’ I argue that it was through ritual that the palace was designated as a political locus. In Part II, ‘Infrastructure and Landscape’, our perspective pans outwards to consider the ma- terial and symbolic networks formed by and between palaces—what are here termed ‘royal land- scapes’—in two case studies: the Italian kingdom of Theoderic and his successors in the sixth century, and the Frankish regnum between the sixth and ninth centuries. In Chapter 3, I show that royal palaces were not only markers of romanitas, or continuity with Roman modes of authority, but tools for reor- dering the political landscape of Northern Italy. Their emplacement in the militarised landscape of the Po plain and Alpine foothills, and their integration into the fiscal infrastructure of northern Italy, were instrumental in the creation of new territorialisations of power in the aftermath of empire. Finally, in Chapter 4 we will again witness the importance of palaces in the renegotiating political space. As in northern Italy, we will see the importance of the Roman militarised frontier in shaping early medieval political formation, yet in the Frankish regnum we will trace these processes further, to highlight the coalescence of new conceptual geographies that were independent of earlier Roman strata. Whereas earlier scholarship has fixated on the question of Frankish ‘itinerancy’, I show instead that royal au- thority continued to be localised within narrowly defined royal landscapes in both the Merovingian and Carolingian periods. In Francia, however, we are able to discern the force palaces exerted in pro- cesses of political integration and centralisation. 20 It is worth stating at the outset that a history of palaces is emphatically not a narrow story of History’s ‘Great Men’, nor even of elite culture in isolation, but of how early medieval societies were ‘put together’: how social order and political power were conceptualised and the mechanisms that put those concepts into action in the lives of early medieval people. In the coming chapters this study will show that the palace was not just a ‘stage’—a passive expression of royal power—but an agent in reformulation of authority in the post-Roman West. In a broader sense, however, beginning from the spaces and places in which political power was negotiated, defined, and authorised, has the potential to offer a new perspective on early medieval political culture more generally, one attuned not just to the appearance of post-Roman ‘continuity’ but also its contradictions: how the rhetoric of continuity or revival—such as we saw in Cassiodorus’ letter to the curator palatii—could disguise, or exist in ten- sion with, a political world otherwise profoundly transformed. 21 PART I ARCHITECTURE AND SPACE CHAPTER 1 Palace Architecture, Between Imperium and Regna A court (aula) is a royal house, or a spacious dwelling enclosed by four colonnades. An atrium (atrium) is a large building, or a very roomy and spacious house, and it is called an atrium because three (tres, neuter tria) colonnades are added to it on the outside. Others say it is ‘atrium’ as if blackened (ater, neut. atrum) by fire and a lamp, for the blackening is caused by smoke. A palace (palatium) is named after Pallas, prince of the Arcadians, in whose honor the Arcadians built the town Pallanteum, and they called the royal palace that they founded in his name ‘Palatium.’ — Isidore of Seville, Etymo- logiae, XV.iii.63 I. Introduction Isidore illustrates the challenges of approaching early medieval palaces from an architectural perspec- tive. Throughout his lengthy examination of public and private buildings in the fifteenth book of the Etymologies, Isidore makes it clear that a structure could be recognisable on the basis of its exterior aspect – in this case, the aula or atrium’s external colonnades. When he comes to define a ‘palace’, however, he hesitates. Though implied to be essentially synonymous with an aula in his usage—since the latter was, he tells us, the domus regia, the ‘royal house’—Isidore does not elaborate on what is architecturally characteristic of a palatium, instead providing a derivation of the term drawn from Ro- man classical knowledge.64 From the perspective of architectural history, Isidore’s descriptions evi- dently leave much to be desired. Not only does he omit valuable details about actually existing structures, but it cannot even be guaranteed that his words necessarily reflected seventh-century real- ities. 63 Isidore of Seville, Etymologiae, XV.iii.3–5, in Isidori Hispalensis episcopi Etymologiarum siue Originum libri XX, ed. W. M. Lind- say, vol. 2 (Oxford: Clarendon Press, 1911): Aula domus est regia, sive spatiosum habitaculum portibibus quattuor conclusum. Atrium magna aedes est, sive amplior et spatiosa domus; et dicutum atrium [eo] quod addantur ei tres porticus extrinsecus. Alii atrium quasi ab igne et lychno atrum dixerunt ; atrum enim fit ex fumo. Palatium a Pallante principe Arcadum dictum, in cuius honore Arcades Pallanteium oppidum construxerunt, et regiam in ipsius nomine conditam Palatium vocaverunt. Trans.: idem, The Etymologies of Isidore of Seville, trans. Stephen A. Barney et al. (Cambridge: Cambridge University Press, 2006), 308. 64 The derivation of palatium is dependent on: Livy, Ab urbe condita, I.5, in T. Livi Ab urbe condita libri, ed. Wilhelm Weissen- born, rev. Hermann Johannes Müller, vol. 1 (Berlin: Weidmann, 1885), 95–98. 23 Isidore is, in a certain sense, in good company. Recent scholarship on the sedes regiae has gen- erally been more invested in questions of ‘ideology’ than material form, and undoubtedly the poor state of the archaeological evidence for palaces from this period is partly to blame here. The principal narrative framework in which the palaces of post-Roman rulers have been invoked is one of ‘emula- tion’ or the symbolic evocation of Roman models, a phenomenon known in scholarship as imitatio imperii.65 The sedes regiae are thus valued insofar as they reproduced features or topographical layouts of Roman palaces – and by the sixth century, specifically the surviving pole of imperial authority in the Mediterranean, the Great Palace of Constantinople. This chapter will show that, while the palaces of the post-Roman regna certainly did provide opportunities for performing ‘Roman-ness’ (romanitas), this is only part of the story. Rather, palaces were sites in which ideals of authority were negotiated and reinvented – and this includes the visual semantics of Roman imperial power. By reorienting our focus to the architectural transformation of the palatium in a diachronic perspective, we are able to discern not only the relations of palace architecture to norms of political behaviour, but also how the concept of ‘palace’ itself changed in response to changing ideals of political authority. ‘Public’ and ‘private’ in palace architecture Discouraging as it may be on first reading, Isidore’s cursory treatment of palatia also indicates possible routes forward. Implicit in his account is that palatium was not a distinct architectural typology in itself, but a relational category: a point on a spectrum including other forms of elite, domestic architecture. After all, Isidore categorises the palatium as a dwelling (habitaculum), alongside other structures that he states share common features: the atrium and aula are both recognisable by their colonnaded exterior and termed domus in his lexicon. The palatium, which he glosses only as referring to the ‘royal residence’ 65 On the origins of this term, see: Christian Scholl, ‘Imitatio imperii? Elements of Imperial Rule in the Barbarian Successor States of the Roman West’, in Transcultural Approaches to the Concept of Imperial Rule in the Middle Ages, ed. Christian Scholl et al. (Frankfurt am Main: Peter Lang, 2017), 19–20. 24 (regia), is not categorically distinct, but differentiated by its owner and its singularity. As we saw above, we should not expect that the palace of the ruler should necessarily differ from the houses of their greatest subjects.66 Considering palaces in the context of domestic architecture brings with it a number of fraught associations that must be addressed before we go any further. Foremost among these are notions of ‘public’ and ‘private’. Whatever the reality, in modern culture the interior of the house is commonly understood to be at a remove from society and the reach of government apparatuses. Premodern houses, however, had a dual status: as both a ‘private’ residence, but also a site of ‘public’ business. The interior of the home was where guests were entertained, business was undertaken, and patronage dispensed, among other functions the twenty-first-century reader might well consider ‘public’. Vitru- vius, for example, gave a neat summation of the distinction between ‘public’ and ‘private’ in the Roman home: Personal areas are those into which there is no possibility of entrance except by invi- tation, like cubicula, triclinia, baths, and the other rooms that have such functions. Public areas are those into which even uninvited members of the public may also come by right, that is, vestibules, cavaedia, peristyles, and any rooms that may perform this sort of function.67 Strict divides between ‘public’ and ‘private’ areas, however, are a construct of Vitruvius’ text and do not necessarily reflect contemporary practice.68 For one, privileged access to spaces in the Roman home (and palace) was variable according to factors such as the status of the viewer: a particularly honoured guest could be invited into rooms off-limits to the average visitor or client.69 This is not to 66 Above, Introduction, 4–7. 67 Vitruvius, De architectura, VI.5.1, in Vitruvius, De l’architecture, vol. 6, ed. and French trans., Louis Callebat (Paris: Les Belles Lettres, 2004), 20–21: Namque ex his quae propria sunt, in ea non est potestas omnibus introeundi nisi inuitatis, quemadmodum sunt cubicula, triclinia, balneae ceteraque quae easdem habent usus rationes. Communia autem sunt quibus etiam inuocati suo iure de populo possunt uenire, id est uestibula, caua aedium, peristylia quaeque eundem habere possunt usum. Trans.: Vitruvius, Ten Books on Architecture, trans. Ingrid Rowland (Cambridge: Cambridge University Press, 1999), 80. 68 Shelley Hales, The Roman House and Social Identity (Cambridge: Cambridge University Press, 2003), 5. 69 Katherine Dunbabin, ‘The Use of Private Space’, in La ciudad en el mondo romano. Atti del XIV congresso internazionale di archeologia classica (Tarragona, 1993), vol. 1 (Tarragona: Consejo Superior de Investigaciones Cientificas-Institut d’Etudis Catalans, 1994), 165–66. 25 say that space was not manipulated to express social hierarchies, however. Vitruvius advises that Ro- man aristocrats should include an audience chamber (basilica) in their home for conducting business – an architectural form which, as we will see, was designed to spatially manifest distinctions in social status.70 Likewise, the façade of the palace (or elite house) was designed to be viewed by an external audience barred from entry via social or legal (and possibly even physical) means, thereby establishing a social distinction between those excluded and those admitted within.71 As Vitruvius tells us, appear- ances were everything. ‘Vestibules should be constructed that are lofty and lordly, the atria and peri- styles at their most spacious, lush gardens and broad walkways refined as properly befits their dignity,’ to achieve the desired effect.72 To treat palaces as ‘houses’, therefore, is not to diminish them, but to recognise how they materialised contemporary conceptions of statues, prestige, and authority.73 In this light, Vitruvius indicates an important point, one that is equally applicable for seventh-century Visigothic Iberia as for first-century Rome: houses (and palaces) were not one-sided assertions of the inhabitant’s power and status, but a medium of social relations, ‘machines for competition’ as Kim Bowes has usefully put it.74 By treating palaces as an instantiation of a broader class of domestic architecture, therefore, our per- spective is reoriented from what was unique to them to how they manipulated the same visual language as other luxury residences in local discourses of authority – and how these discourses changed over time.75 70 Vitruvius, De architectura, VI.5.2, 6:21. See below, Chapter 2, 104–12. 71 Lefebvre, Production, 193. 72 Ibid.: faciunda sunt uestibula regalia, alta atria et peristyla amplissima, siluae ambulationesque laxiores ad decorem maiestatis perfectae. Trans. Rowland, 81. 73 For a modern perspective, see e.g.: John Archer, Architecture and Suburbia: from English Villa to American Dream House, 1690–2000 (Minneapolis, MN: University of Minnesota Press, 2005), esp. 1–16. 74 Bowes, Houses and Society, 23. 75 On houses being indexical of socio-cultural change, see e.g.: Federico Guidobaldi, ‘Le domus tardoantiche di Roma come “sensori” delle trasformazioni culturali e sociali’, in The Transformations of Vrbs Roma in Late Antiquity, ed. William V. Harris, Journal of Roman Archaeology Supplements 33 (Portsmouth, RI: Journal of Roman Archaeology, 1999), 53–68. 26 Overview Charting the architectural development of palaces offers a distinct perspective on political authority in the Early Middle Ages. We begin in the fourth century, with the palaces of the Later Roman Em- perors in Milan, Trier, and Rome. Late Roman palaces were configured according to a common spatial ‘grammar’ that reproduced the most salient topographical relationships of the archetypical palace, that of the Palatine in Rome. The spatial strategies that were characteristic of the provincial sedes were effectively reimported into Rome itself during the reign of Maxentius, transforming the relationship of palace and city in the metropolis. The fracturing of the Roman political economy in the fifth century triggered a revolution in the lexicon of architectural display. The Late Roman palace conceptually dissolved; those of the sixth century and later conformed to a new architectural typology. These trans- formations were simultaneously material and ideological. On the one hand, since the Roman house reflected—and, crucially, enabled—the reproduction of the Roman social order, the breakdown of Ro- man imperial hegemony necessarily entailed a re-evaluation of its distinctive architectural spaces. At the same time, we see a concomitant revolution in the materials of construction. Although a dichot- omy between ‘Roman’ brick and marble and ‘medieval’ timber and stone would be misleading, a stark transition in architectural materials is visible throughout northern and western Europe after the sixth century. Into this new conception of palace architecture was introduced a significant new element: the chapel. Chapels were not a natural feature of the palace, but one that originated in the specific context of the Christianisation, or ‘liturgification’, of political authority in the seventh century. Altogether, we will see that a continuous story of palaces is not only possible, but also provides a new perspective on the evolving symbolic languages of political authority in the early medieval world. II. ‘Rome is where the emperor is’: The Palace in the Later Roman Empire It is no exaggeration to say that the elevation of Maximian as Diocletian’s colleague in the West in 286, with the addition of Constantius Chlorus and Galerius as caesares in 293, revolutionised Roman 27 understandings of their political world. With the institution of the Tetrarchy, it became possible to conceptualise permanent imperial courts outside of Rome with little to no connection with the Eternal City. The new sedes brought imperial architectural patronage to the provinces on a scale hitherto un- seen, something which did not escape the notice of contemporaries.76 It also set in motion the trans- ference of political power away from the traditional centre of Roman power to new provincial elite audiences, a shift which brought about sweeping changes in modes of imperial representation in word and image.77 These new seats of power were legitimised not only by the emperor’s presence, but also through their architecture. The new palaces established in the West, at Milan and Trier, profoundly altered the topographies of their respective host cities, and the architectural strategies that the imperial architects adopted highlight the establishment of a flexible visual ‘grammar’ for palace architecture that responded to the emperor’s changing public image. New seats for new emperors: Milan and Trier in the fourth century Ausonius, tutor (and later praetorian prefect) of the emperor Gratian (367–383), surveys Milan’s sali- ent features in his enumeration of the ‘noble cities’ of the Roman world: At Milan also are all things wonderful, abundant wealth, countless stately houses, men able, eloquent, and cheerfully disposed; besides, there is the grandeur of the site en- larged by a double wall, the Circus, her people’s joy, the massive enclosed Theatre with wedge-like blocks of seats, the temples, the palatine citadels [palatinaeque arces], the wealthy Mint, and the quarter renowned under the title of the Baths of Herculeus; her colonnades all adorned with marble statuary, her walls piled like an earthen rampart round the city’s edge. All these, as it were, rivals excel in the magnitude of their work- manship, and the proximity of Rome does not diminish them.78 76 Emanuel Mayer, ‘The Architecture of the Tetrarchy’, in A Companion to Roman Architecture, ed. Roger B. Ulrich and Caroline K. Quenemoen (Oxford: Blackwell, 2014), 107–11. 77 Mayer, Rom ist dort, 4–27. 78 Ausonius, Ordo urbium nobilium VII, in Decimi Magni Ausonii Burdigalensis Opuscula, ed. Sesto Prete, BSGRT (Leipzig: Teubner, 1978), 195: Et Mediolani mira omnia, copia rerum innumerae cultae que domus, facunda uiuorum ingenia et mores laeti; tum duplice muro amplificaa loci species populi que uoluptas circus et inclusi moles cuneata theatri templa Palatinaeque arces opulens que moneta et regio Herculei celebris sub honore lauacri; cunctaque marmoreis ornata peristyla signis moeniaque in ualli formam circumadata limbo: omnia quae magnis operum uleut aemula formis excellunt: nec iuncta permit uicinia Romae. Trans.: Ausonius, The Order of Famous Cities, in Ausonius, trans. Hugh G. Evelyn White, Loeb Classical Library 96 (Cambridge, MA: Harvard University Press, 1919), 273 (amended). 28 Though little of the imperial palace’s architecture remains, nearly all the structures named by Ausonius are known archaeologically and the overall design of the complex is clear. The imperial palace was located on the city’s western periphery in a sector bounded by the intersection of the cardo and decu- manus on one side and the city walls on the other, themselves likely renewed by Maximian (Figs. 4– 5).79 The best attested structure in this sector is the substantial circus (c. 460m long by 67–68m wide), located in a projecting annex of the mural circuit.80 One of the towers for the starting gates, the carceres, was re-used as the belltower of San Maurizio al Monastero Maggiore (Fig. 6), and on the southwestern side were discovered in the 1930s the remains of a structure which abutted the circus, probably the entrance to either the judge’s tribunal or perhaps the emperor’s private box. 81 The palace lay alongside the circus on its eastern flank. Ongoing excavations have revealed a host of representational spaces of various kinds. Adjacent to the circus’ carceres, for example, stood a rotunda supported by an inner arcade of ten columns and preceded by a vestibule, likely dating to the mid-fourth century (Fig. 7). The intended function of this space is unclear, but by analogy with other contemporary structures, such as the erroneously named ‘Temple of Minerva Medica’ in Rome (Fig. 8), it is best explained as a suite of reception spaces with the centrally planned space serving as a junction between them; the elongated aula located axially with the entrance would also suggest a representational function.82 Recent excavations have brought to light other structures associated with the palace. Excavations on the Via Gorani, for example, have revealed that a peristyle domus occupied from the first century CE was 79 Mario Mirabella Roberti, Milano Romana (Milan: Rusconi, 1984), 29–31. 80 For an overview: John Humphrey, Roman Circuses: Arenas for Chariot Racing (Berkeley, CA: University of California Press, 1986), 613–20. The circus is incorporated into the wall-circuit at its southern end. For the excavation: Alberto de Capitani D’Arzago, Il circo romano, Istituto di Studi Romani – Sezione Lombarda. Ricerche per la Forma Urbis Mediolani (Milan: Ceschina, 1939). 81 Imperial box: Humphrey, Roman Circuses, 620. Judge’s tribunal: Mirabella Roberti, Milano romana, 66. 82 Theories range from a private bath complex to a large dining hall: Antonio Frova, ‘Trovamenti e scavi dal 1950 al 1953’, in Ritrovamenti e scavi per la “Forma Urbis Mediolani” IV, ed. Aristide Calderini and Carla Gerra, Quaderni di Studi Romani 5 (Milan: Ceschina, 1955), 11–12; Massimiliano David, ‘“Palatinaeque arces…” Temi di architettura palaziale a Milano tra III e X secolo’ in ‘Ubi palatio dicitur’. Residenze di re e imperatori in Lombardia, ed. Massimiliano David (Milan: Cinisello Balsamo, 1999), 26. 29 demolished at the end of the third to make way for a new structures (Fig. 9).83 This has been associated with the purchase and clearing of a residential district for the palace of Maximian, a trend attested for other architectural projects of the Tetrarchs.84 These structures seem to correspond to one range of a substantial peristyle courtyard, around which were arrayed structures including a large apsidal aula and a triconchos, which was accessible via a vestibule from the decumanus.85 The combination of curving forms, planimetric subdivision and complexity, tempered with a preference for axial foci reproduces on a vast scale the visual language of luxury residential structures from around the fourth- and fifth- century Mediterranean. The Constantinian palace at Trier presents a picture consistent with that sketched out for Milan. The anonymous orator of a panegyric delivered to Constantine at Trier in 310 also professed the grandeur of his emperor’s chosen seat: ‘I see a Circus Maximus to rival, in my opinion, that of Rome, I see basilicas and a forum, palatial buildings, and a seat of justice raised to such a height that they promise to be worthy of the stars and the sky, their neighbours.’86 The palace founded by Max- imian and elaborated under the Constantinians lay to the east of the cardo (Fig. 10), in a district periph- eral to the core of the ancient city around the forum. Its centrepiece was the colossal ‘basilica’, a reception hall associated with the palace complex (Fig. 11). In its dimensions the Trier aula is super- seded only by the near-contemporary Basilica of Maxentius in Rome in terms of secular architecture 83 Anna Ceresa Mori et al., ‘Via Gorani 2–4. Resti di domus di età romana e del Palatium imperiale’, Notiziario della Soprinten- denza Archeologica della Lombardia (2010–2011): 241–47; Anna Ceresa Mori, ‘Riflessioni sul palazzo imperiale di Milano alla luce delle recenti indagini’, in Milano e la chiesa di Milano prima di Ambrogio, ed. Raffaele Passarella, Studia Ambrosiana (Milan: Biblioteca Ambrosiana, 2018), 100–09. 84 Ceresa Mori, ‘Riflessioni’, 100. 85 The foundations of the aula continue beyond the limits of the excavation, but it was around 18m wide. The apse alone was 8m in diameter: Ceresa Mori, ‘Riflessioni’, 102. These are substantial dimensions, comparable with the aula at Piazza Armerina (30m by 16m: Sfameni, Ville residenziali, 88, 100) and that of the Villa of Maxentius on the Via Appia (see below). 86 XII panegyrici latini, ed. R. A. B. Mynors, Oxford Classical Texts (Oxford: Clarendon Press, 1964), VI.22.5, 202: Video circum maximum aemulum, credo, Romano, uideo basilicas et forum, opera regia, sedemque iustitiae in tantum altitudinem suscitari ut se sideribus et caelo digna et uicina promittant; Trans.: C. E. V. Nixon and Barbara Saylor Rodgers, trans., In Praise of Later Roman Emperors: The Panegyrici Latini, The Transformation of the Classical Heritage 21 (Berkeley, CA: University of California Press, 1994), 252. 30 in the Western provinces of this period.87 Dated securely on the basis of brick-stamps, the hall was built in the years around 310 over a first- or second-century domus, seemingly cleared (as at Milan) in the later third century.88 Built on a north-south orientation, the single nave terminates in a tall apse which rises nearly to the line of the trussed roof. The rhythmic character of the external pilasters we see today was interrupted by outer galleries on both sides, presumably for the maintenance of the windows and hypocaust flues innovatively integrated into the walls.89 On either side of the hall stood two peristyle courtyards; to the southern entrance was connected a apsidal vestibule which communi- cated with a cryptoporticus belonging to an earlier period, through which presumably were accessible other buildings of the palace complex.90 Given its size, it might not be too far a stretch to associate it with the towering ‘seat of justice’ remarked upon by Ausonius.91 The circus has been localised to a region between the palace and the city walls (Fig. 10, C).92 That the orator of 310 singled out the circus as a hallmark of the city under Constantine may suggest his patronage of its construction, or alternatively a substantial renovation of an earlier structure.93 To access the circus would require passing through the palace precincts, either in the vicinity of the aula palatina or, more likely, near the baths which probably indicated a ‘public’ point of entry. Set in line with the forum with respect to the street-grid, the imperial baths were built on the same north-south urban axis as the aula and were likewise a product of Constantine’s reign: during excavations, a coin 87 The Trier hall stands at 67m by 27.5m (Hans Eiden, ‘Ausgrabungen im spätantiken Trier’, in Neue Ausgrabungen in Deutschland, ed. Werner Krämer [Berlin: G. Mann Verlag, 1958], 547), or 1842.5m2 in area; the central nave alone of the Basilica of Maxentius comes in at 271ft by 83ft, or approximately 82.6m x 25.3m (Anthony Minoprio, ‘A Restoration of the Basilica of Constantine, Rome’, Papers of the British School at Rome 12 [1932]: 8), yielding an area of nearly 2100m2. 88 Wilhelm Reusch, ‘Die Aula Palatina in Trier. Ergebnisse der Untersuchungen in den Jahren 1951–1953‘, Germania 33 (1955): 194–96. 89 Ibid., 181–190; Wilhelm Reusch, ‘Die Außengalerien der sog. Basilika in Trier’, Trierer Zeitschrift (1949): 170–87; Fritz Kretzschmer, ‘Die Heizung der Aula Palatina in Trier: Ein Versuch ihrer Deutung und der Aufklärung ihrer Betriebsweise’, Germania 33 (1955): 200–10. 90 Eiden, ‘Ausgrabungen im spätantiken Trier’, 549. 91 Mayer, Rom ist dort, 34. 92 Wilhelm von Massow, ‘Der Circus des römischen Trier’, Trierer Zeitschrift 18 (1949): 161–67. 93 Humphrey, Roman Circuses, 602. Humphrey suggests that there was probably an earlier circus on the same site, perhaps only partially stone-built, dating to the second century: ibid., 603. 31 dateable between 314 and 317 was discovered in a foundation trench, indicating ongoing work at that time.94 In around 326 a wealthy residence in the northern sector of this eastern, ‘palatial’ area was razed to make way for the double-basilica which would come to serve as the city’s cathedral, despite only having been constructed itself (it seems) in around 315 and decorated with a beautiful, yet mys- terious figural painted ceiling (Fig. 12).95 Milan and Trier in context Visible at Milan and Trier are a core of architectural features which defined the palaces of the fourth- century Empire: a residential complex emplaced at the edge of the urban core, preceded by markers such as tetrapyla ‘separating’ it from the rest of the urban fabric, and associated with a monumental circus and mausoleum. No single site possesses the whole selection of features; we cannot talk in terms of a rigid, transferable model. Nonetheless, the frequency by which these elements recur points towards the existence of a more general ‘grammar’ of palace architecture that informed the design of all these sites.96 As Slobodan Ćurčić has usefully characterised it, these features were deployed ‘in standardized sequences, relating palaces to—and at the same time setting them apart from—their urban settings’.97 The elevation of provincial cities with the implantation of the imperial palace is a distinctive feature of the new sedes – and the same architectural strategies as those deployed at Milan and Trier are able to be discerned at other sites in the Mediterranean. The palace of Galerius at Thessaloniki has been excavated the most completely (Fig. 13), and its dense concentration of representational spaces, 94 Daniel Krencker et al., eds., Die Trierer Kaiserthermen, Trierer Grabungen und Forschungen 1 (Augsburg: Filser, 1929). 95 Theodor Kempf, ‘Trierer Domgrabungen, 1943–54’, in Neue Ausgrabungen in Deutschland, 370–74. On the ceiling: Theo- dor Kempf, ‘Konstantinische Deckenmalerei aus dem Trierer Dom’, Trierer Zeitschrift 19 (1950): 45–51. Attempts to asso- ciate these figures with women of the family of Constantine have been unconvincing. For summary and analysis, see: Irving Lavin, ‘The Ceiling Frescoes in Trier and Illusionism in Constantinian Painting’, DOP 21 (1967): 99–113. 96 For a summary: Verena Jaeschke, ‘Adapting to a New Concept of Sovereignty: Some Remarks on Tetrarchic Palace Architecture’, in Late Antique Palatine Architecture, 63–76; Hendrik Dey, The Afterlife of the Roman City: Architecture and Ceremony in the Early Middle Ages (Cambridge: Cambridge University Press, 2015), esp. 33–57, 68–84. 97 Ćurčić, ‘Late-Antique Palaces’, 72. 32 delimited from the rest of the urban area by the ‘junction’ of the Arch of Galerius, gives us a sense of how much of the framing armature that communicated the boundaries of the palace is missing from the Trierer and Milanese complexes.98 Likewise, Libanius, in an oration in praise of his home city of Antioch, described the palace occupied a full quarter of the Orontes island adjacent to the ancient core and was delimited by its own wall circuit.99 The separation of the palace from the rest of the urban area was marked by a tetrapylon at the intersection of the cardo and decumanus.100 The palace at Constantinople stands out for its unique longevity, but features such as its placement at the head of porticoed processional route and its architectonic conjunction to the adjacent Hippodrome place it firmly within the context of the late-ancient palatial idiom (Fig. 14).101 Given their intramural locations, all these projects—including those at Milan and Trier—must have required considerable resources and labour, and we should not underestimate the human cost of the expropriation of large areas of the urban area for their construction.102 This architectural arrangement was not the invention of a single moment in the fourth century, however. For example, the architectonic unification of palace and circus can be seen already in the Severan phase of the (then-suburban) Sessorian Palace in Rome (Fig. 15), located on the fundus Lau- rentum to the east of the Esquiline. Prior to the construction of the Aurelianic Walls, the fundus was dotted with estates once belonging to prominent Roman families and incorporated into the fisc by the Severans; sometime shortly after 312 it had become a personal possession of Constantine’s mother, 98 On Thessaloniki, see above all the analysis in: Mayer, Rom ist dort, 39–77. 99 Libanius, Oratio XI.203–07, in Libanii Opera, ed. Richard Förster, vol. 1, bk. 2, BSGRT (Leipzig: Teubner, 1903), 506– 07. 100 Ibid., XI.204, 507. On the use of urban ‘armature’ to spatially unite and separate: William MacDonald, The Architecture of the Roman Empire, vol. 2, Yale Publications in the History of Art 35 (New Haven, CT: Yale University Press, 1986), 5– 31. 101 Mayer, Rom ist dort, 93–97, 105–174; summary in: Dey, Afterlife, 77–84. 102 Lactantius commented that Diocletian’s ‘limitless passion for building’ (infinita quaedam cupiditas aedificandi) caused the removal of many from their homes in the construction of his new palace at Nicomedia: Lactantius, De mortibus persecutorum, VII.8–9, ed. and trans. J. L. Creed, Oxford Classical Texts (Oxford: Clarendon Press, 1984), 12–13. The palace of Galerius in Thessaloniki was constructed over existing buildings: Mayer, Rom ist dort, 30–31. The dedicatory inscription of the Baths of Diocletian in Rome records that the properties on the proposed site were bought up by the emperor: CIL VI, 1130. 33 Helena.103 In its Severan phase, the central aula of the Sessorian Palace was connected to the nearby Circus Varianus by a covered passageway, though the racetrack was subsequently severed by Aurelianic Walls (Fig 16).104 The fourth-century saw the addition of new structures, likely in connection with Helena’s patronage. The standing remains of a large, apsidal aula in the grounds of the nearby Museo della Fanteria belong to a common typology of fourth-century basilican audience chambers, witnessed not least in the Constantinian aula of Trier as well as, as we will see, at the Villa of Maxentius (Fig. 17).105 The buildings of the main complex, adjacent to the Circus Varianus and the Amphitheatrum Castrense, were joined by a frontal colonnade opening onto an open piazza accessed by a road ap- proaching from the north-west, in the vicinity of the Porta Maggiore.106 Crucially, however, the incor- poration of the hippodrome or aula was not itself limited to the imperial palace, but part of a broader architectural language of authority and status. With some rhetorical hyperbole, a fragment of the fifth- century historian Olympiodorus (transmitted via Photius’ ninth-century Bibliotheca) states that, ‘Each of the great houses of Rome contained within itself, as he [Olympiodorus] says, everything which a medium-sized city could hold, a hippodrome, fora, temples, fountains and different kinds of baths. At this the historian emotes: “One house is a town; the city hides ten thousand towns.”’107 103 The fundus Laurentum was a large fiscal property stretching between Via Praenestina and the Via Latina as far as Monte Cavo (Mons Gabus). See: Liber Pontificalis XXXIIII.27, in Le Liber pontificalis: texte, introduction et commentaire, ed. Louis Duch- esne, vol. 1, Bibliothèque des écoles franc ̧aises d'Athènes et de Rome, 2nd series, 3 (Paris: E. Thorin, 1886), 183; Hugo Brandenburg, Ancient Churches of Rome from the Fourth to the Seventh Century: The Dawn of Christian Architecture in the West, Bibliothèque de l’Antiquité Tardive 8 (Turnhout: Brepols, 2005), 103–04; Jan Willem Drijvers, ‘Helena Augusta and the City of Rome’, in Monuments & Memory. Christian Cult Buildings and Constructions of the Past: Essays in honour of Sible de Blaauw, ed. Mariëtte Verhoeven, Architectural Crossroads 3 (Turnhout: Brepols, 2016), 148. 104 Now Santa Croce in Gerusalemme, a dedication it received on its conversion into an oratory for the relics of the True Cross in the fourth century. On the Circus Varianus, see: Claudia Paterna, ‘Il circo Variano a Roma’, MEFRA 108, no. 2 (1996): 817–53. 105 Donato Colli, ‘Il palazzo Sessoriano nell’area archeologica di S. Croce in Gerusalemme : ultima sede imperiale a Roma?’, MEFRA 108, no. 2 (1996): 785–86. 106 Ibid., 795. For evidence for the road, and for further structures belonging to the palace (including a possible triclinium): Elisabetta Borgia et al., ‘Horti spei veteris e Palatium Sessorianum: nuove acquisizioni da interventi urbani 1996–2008’, in Roma. Scavi archeologici e scoperte degli ultimi 10 anni, ed. Helga Di Giuseppe and Elizabeth Fentress, Fold&Raccolte 1 (Rome: Scienze e Lettere, 2018), 118–21, 134–40. 107 Olympiodorus, frag. 41 (Photius, Bibliotheca, cod. 80), in Photius, Bibliothèque, ed. and French trans. René Henry (Paris: Les Belles Lettres, 1959), 185: Ὅτι ἕκαστος τῶν μεγάλων οἴκων τῆς Ῥώμης, ὥς φησιν, ἅπαντα εἶχεν ἐν ἑαυτῷ ὁπόσα πόλις σύμμετρος ἠδύνατο ἔχειν ἱππόδρομον καὶ φόρους καὶ ναοὺς καὶ πηγὰς καὶ λουτρὰ διὰφορα. διὸ καὶ ὁ συγγπραφεὺς ἀπεφθέγξατο· Εἷς δόμος ἄστυ πέλει· 34 From the periphery to the Eternal City: the building projects of Maxentius The Palatine was the centre of the Roman political imagination. As the original site of Romulus’ set- tlement it was a charged site of social memory; as the stable site of imperial power for centuries it defined Roman expectations of the architecture of authority. The predilection for plastic, curving forms and broken wall surface of its domestic and representational suites—a significant novelty in its first-century phase, as executed under Domitian—was influential in defining a persistent, but adapta- ble lexicon for flaunting status, prestige, and authority.108 The relationship forged between the new imperial sedes and the empire’s ideological axis was not just one of architectural style, however. The Trier orator’s boast that the city possessed its own ‘Circus Maximus’ that was the rival of that in Rome itself (for instance) is not just praise of Constantine’s architectural patronage, but an acknowledgment of the underlying motivation of these palatial projects: to reproduce an urban environment in which the new imperial structure could be presented to its publics as per se legitimate. The fourth-century sedes distilled and redeployed the most salient elements of the topography of the Palatine, ‘iconograph- ically’ legitimising the late-imperial sedes as permanent sites of political authority, in contrast to the temporary provincial seats adopted by emperors on campaign in previous centuries. The most vivid testimony of such a symbolic relationship forged between the imperial me- tropolis and ‘disembedded’ political centre is not to be found in the provinces, however, but paradox- ically in Rome itself. The buildings sponsored by Maxentius (306–312) constitute the best-preserved palatial complex of this period.109 Maxentius’ considerable programme of building was above all a πόλις ἄστεα μυρία κεύθει.. Translation: R. C. Blockley, The Fragmentary Classicising Historians of the Later Roman Empire: Eunapius, Olympiodorus, Priscus and Malchus, vol. 2, ARCA Classical and Medieval Texts, Papers and Monographs 10 (Liverpool: Fran- cis Cairns, 1983), 205. 108 Nero’s Domus Aurea was an important pathbreaker for the combination of the visual idioms of domestic, temple, and thermal architecture. Larry Ball, The Domus Aurea and the Roman Architectural Revolution (Cambridge: Cambridge University Press, 2003). 109 Hauke Ziemssen, ‘Die Kaiserresidenz Rom in der Zeit der Tetrarchie’, in Rom und Mailand in der Spätantike. Repräsentionen städtischer Räume in Literatur, Architektur und Kunst, ed. Therese Fuhrer, Topoi. Berlin Studies of the Ancient World 4 (Berlin: De Gruyter, 2011), 87–110. 35 reaction to the decentralisation of the Empire and the foundation of new imperial sedes.110 The Chron- ography of 354 provides a laconic report of his reign in the city: Maxentius ruled for six years. In his reign the Temple of Roma burned and was recon- structed. He built baths in the palace and a circus in the catacombs. There was great hunger. The Romans hung a Moesian soldier and six thousand men were killed by the soldiers. He taxed all the Romans in gold and they gave it up. He began work on a canal but he did not complete it. He was killed at the Milvian Bridge on the Tiber.111 The scale of the building programme undertaken under Maxentius was more impressive than the chronicler lets on, however. Maxentius’ interventions in Rome’s symbolic topography imported rep- resentational strategies developed in the Tetrarchic sedes to Rome itself, both within and without the city’s walls.112 Turning first to the conjunction of palace and circus on the Palatine, a defining element of the fourth-century sedes, we will then examine how Maxentius’ architects redeployed this formula in his extra-urban villa on the Via Appia – and even accentuated it on the Palatine itself. From a distance, the modern visitor to Rome is struck by the Palatine’s sheer scale, a visual impact that must have been even more arresting in Late Antiquity. Rising forty metres above the Roman Forum, the Palatine retains something of its ancient bearing thanks to the massive substruc- tures which supported the palaces above, constructed in three main phases between the third and fourth centuries CE. The amount by which the hill has been augmented and extended in this way is particularly evident on the southwestern side, overlooking the Circus Maximus (Fig. 18). Enlarged by this artificial platform, the palace complex stretches out over an area of around fourteen hectares. Imperial architects from the second century progressively augmented the Palatine’s façade on its southwestern edge to maximise its visual impact from the Circus Maximus (Fig. 19). Following a fire, 110 On Maxentius, see e.g.: Mats Cullhed, Conservator urbis suae: Studies in the Politics and Propaganda of the Emperor Maxentius, Skrifter utgivna av Svenska institutet i Rom, 8th series, 20 (Stockholm: P. Aströms Förlag, 1994); Simon Corcoran, ‘Ma- xentius: A Roman Emperor in Rome’, Antiquité Tardive 25 (2017): 59–74. 111 Chronography of 354, XVI.5.48, in Chronica Minora saec. IV, V, VI, VII, ed. Theodor Mommsen, vol. 1, MGH AA 9 (Berlin: Weidmann, 1882), 148: Maxentius imp. ann. VI. hoc imp. templum Romae arsit et fabricatum est. thermas in palatio fecit et circum in catecumbas. fames magna fuit. Romani traxerunt militem Moesiacum et occisi sunt Romani a militibus homines V ̄Ī. Romanis omnibus aureum indixit et dederunt. fossatum aperuit, sed non perfecit. occisus ad pontem Mulvium in Tiberim. 112 Ziemssen, ‘Die Kaiserresidenz Rom’, 104; Mayer, Rom ist dort, 185. 36 Trajan had the Circus reconstructed in brick-faced concrete and expanded. As part of this reconstruc- tion the pulvinar, the shrine which housed statues of the gods during the races, was transformed into a temple-like structure with a hexastyle façade built against the rear wall of the cavea.113 It is likely that, although the pulvinar was not primarily intended as such, the emperors from Trajan onwards used it as their private box.114 This visual connection was matched by a physical conjunction. Fortuitously, a surviving frag- ment of the Severan Marble Plan preserves the plan of the pulvinar (Fig. 20). Opposite its six-column façade, on the rear side facing the Palatine is indicated not a flat wall but another colonnaded opening with two small projections of the wall to either side. This could simply indicate that the entrance was open at the back and perhaps accessible via stairs. Ulrike Wulf-Rheidt has argued, however, that wall- remains on the slope below the domus Augustana are best explained as belonging to a bridge connecting directly to the box in the Circus.115 Further interventions on this side stressed the importance of the Palatine’s appearance from the Circus. Under the Antonines the façade of the Domus Augustana was also transformed into an arcaded exedra, opening itself to the external viewer while simultaneously screening the interior; under the Severans the platform of the Domus Augustana was extended yet fur- ther outwards towards the circus.116 The enlarged platform (the so-called Domus Severiana) housed a bath complex which was built up to multiple storeys thus offering new points from which the races could be observed in private (Figs. 19, D, and 21). The association of palace and circus was an essential component of the ‘grammar’ of late- ancient palaces, as we saw above. While the characteristic physical conjunction of the residential 113 There had been a pulvinar in the Circus Maximus as early as Augustus: Humphrey, Roman Circuses, 78. 114 Humphrey, Roman Circuses, 82–83. 115 Wulf-Rheidt attributes such a bridge to the time of Trajan: Ulrike Wulf-Rheidt, ‘Die Bedeutung der severischen Paläste für spätere Residenzbauten’, in Palast und Stadt im severischen Rom, ed. Natasha Sojc et al. (Stuttgart: Franz Steiner, 2013), 289–91; eadem, ‘Die schwierige Frage der Nutzung des römischen Kaiserpalastes auf dem Palatin in Rom in der Spätan- tike’, Antiquité Tardive 25 (2017): 131. 116 Jens Pflug, ‘Die Bauliche Entwicklung der Domus Augustana im Kontext des sudöstlichen Palatin bis in severische Zeit’, in Palast und Stadt im severischen Rom, 198. 37 quarters and imperial box cannot be demonstrated for every site, there are strong grounds to suppose that, as at Constantinople, it was a part of the basic design concept. The best evidence for this, how- ever, lies outside Rome’s walls. Between the second and third milestones of the Via Appia stands a colossal circus with associated villa and mausoleum (Figs. 22–23). Formerly considered a product of the reign of Caracalla, Maxentius’ ownership was confirmed in 1825 by Antonio Nibby after the dis- covery of two marble fragments in the circus with a dedication to Maxentius’ son, Romulus (d. 309).117 In formal terms, the grouping of a residence, circus, and mausoleum conforms to the general set of typological features that characterise the new imperial residences of this period.118 The circus, set be- low the villa on the adjacent ridge, is particularly grand, measuring around eighty percent that of the Circus Maximus in length and accommodating up to 10,000 spectators on twelve tiers of seating.119 Around halfway down the length of the cavea stood a tower with an internal staircase providing access between the villa and the pulvinar (Fig. 24), which was set into the banks of seating on a platform and from which were visible all the significant points of the circus and its course.120 This assemblage was joined to the cryptoporticus which ran along the outer terrace: it thus afforded the emperor the ability to enter the circus screened from view to appear suddenly before the assembled crowds.121 The villa itself was located on the low hill behind the circus. Its colonnaded façade, elevated over the nearby circus, would have recalled the exedra of the Domus Augustana overlooking the Circus Maximus. The ridge was dominated by a large aula in opus vittatum (33.1m by 19.45m), the apsidal wall 117 Antonio Nibby, Del circo volgarmente detto di Caracalla (Rome: Filippo e Nicola de Romanis, 1825), 9. CIL VI, 1138: DIVO ROMULO . N. M. V. / COS. OR[d. ii] .FILIO / D. N. MAXENT[ii] INVICT. / [ac perpet.] AVG NEPOTI / [di]VI [m]AXIM[i]ANI. SEN. [e]T. DIVI [maximiani iu]/ [ni]ORIS .AC[…]. ‘Maximianus iunior’ refers to Galerius. 118 Alfred Frazer, ‘The Iconography of the Emperor Maxentius’ Buildings in Via Appia’, Art Bulletin (1966): 385–88. 119 The external length of the Circus Maximus is c. 620m; the Circus of Maxentius 503m: Humphrey, Roman Circuses, 124, 586. Number of spectators: ibid., 592. For a comparison with other circuses: Giovanni Ioppolo, ‘La struttura architetto- nica’, in Giovanni Ioppolo and Giuseppina Pisani Sartorio, La villa di Massenzio sulla Via Appia, vol. 2, Monumenti Romani 9 (Rome: Editore Colombo, 1999), 115. 120 Giuseppina Pisani Sartorio and Raissa Calza, La villa di Massenzio sulla Via Appia, vol. 1, Monumenti Romani 6 (Rome: Istituto di studi romani, 1976), 107, 127; Ioppolo, ‘La struttura architettonica’, 166–67, 169. 121 Sartorio and Calza, La villa di Massenzio, 104. 38 of which is still standing, preceded in its original plan by a transverse vestibule (Fig. 25).122 The aula was the keystone of the villa’s plan, and it bears an immediate resemblance to the audience chambers of other imperial residences. As at Trier, the hall of Maxentius’ villa does not stand completely alone, being architecturally integrated into the complex through its vestibule and structures to either side. In plan, the apses are proportionally deep relative to their naves and thus fully encompass the image (or personage) within. The total arrangement of a vast single nave, entered through a vestibule and termi- nating in a deep apse, has definitively introduced a sense of directionality: the viewer is confronted by the unmediated impact of the space as they enter and observe the ‘façade’ at the opposite end. Despite the enormity of his project on the Via Appia, Maxentius did not neglect the centre of the city, and his highly visible interventions should alert us that the Palatine continued to be the centre of the Roman political imaginary (Fig. 26). The platform of the Domus Severiana and its bath complex were again expanded outwards towards the Circus Maximus on a promontory of massive brick sub- structures at the beginning of the fourth century, evidently the highly visible ‘baths in the palace’ attributed to Maxentius by the Chronography of 354 (Fig. 27). The new baths were built over multiple storeys with complex vaulting, the outermost structure taking the form of an octagonal ‘turret’ crowned with a vault which gave a ‘bird’s eye view’ of the events below.123 The distant spectacle of the baths was matched by a monumentalisation of the approaches to the Palatine. The enormous Basilica Nova (Fig. 28), the renewed Temple of Venus and Roma, and the vestibule added to the Temple of Peace (still erroneously known as the ‘Temple of Romulus’) were audacious new constructions which extended the political space of the palace outwards onto the Via Sacra. They display simultaneously a ‘conservative’ pietas to Rome’s protector goddesses and a markedly untraditional novelty, not only in 122 Pisani Sartorio and Calza, La villa di Massenzio, 91–97, 124. The vestibule: ibid., 97–99, 124.Conlin, Haeckl, and Ponti observe that it is unlikely that the excavated hypocausts were ever functional, perhaps indicating a planned but incomplete project: Diane A. Conlin et al., ‘The Villa of Maxentius on the Via Appia: Report on the 2005 Excavations’, Memoirs of the American Academy in Rome 51/52 (2007): 357–58. 123 Wulf-Rheidt, ‘Die schwierige Frage’, 131–32. 39 terms of their immense size but also in their daring use of forms.124 For example, the immense groin vaults used to span the wide space of the basilica is more typical of thermal architecture than a secular hall at this time. Its triple nave and lateral apse responded to new requirements for imperial ceremony at a time in which the emperor was presented ‘nakedly as a monarch’.125 The Basilica of Maxentius should thus be understood as an audience hall in the vein of the Trier aula, executed on a colossal scale and modified for the exigencies of the representational topography cultivated by Maxentius in Rome. Under Maxentius, Rome was transformed into, essentially, a ‘Tetrarchic’ seat, and this involved the orientation of the city’s space onto a new symbolic axis. There are indications that the Porta Appia and Porta Asinaria, the two gates through which one could reach the villa complex from the city’s core, were altered and expanded in this period, suggesting that conscious spatial connections were drawn between the symbolic core and the suburbium.126 Despite the brevity of his reign, Maxentius’ interventions in Rome’s built environment highlight the strength of a new architectural current for the design of a palatial seat that, though legitimised by their relation to the Palatine as a prototype were developed to facilitate an imperial image developed in the provinces. *** The seats of the Tetrarchs share an architectural grammar which symbolically connected them to an authorising prototype on the Palatine. Though not a strict ‘model’ which was translated wholesale 124 Mayer, Rom ist dort, 185. 125 Ziemssen, ‘Die Kaiserresidenz Rom’, 99. 126 Lorraine Kerr, ‘A Topography of Death: The Buildings of the Emperor Maxentius’, in TRAC 2001: Proceedings of the Eleventh Annual Theoretical Roman Archaeology Conference, Glasgow 2001, ed. Martin Carruthers et al. (Oxford: Oxbow Books, 2002), 27–28. On Maxentian additions to the walls: Ian Richmond, The City Wall of Imperial Rome: An Account of its Architec- tural Development from Aurelian to Narses (Oxford: Clarendon Press, 1930), 251–6, and for the Porta Appia: ibid., 121–42. Dey has cautioned that the additions attributed to Maxentius, though all built in opus vittatum, are heterogeneous in their execution and they likely reflect a lengthier process of repair than just his six-year reign: Hendrik Dey, The Aurelian Wall and the Refashioning of Imperial Rome, AD 271–855 (Cambridge: Cambridge University Press, 2011), 285–91. That accepted, Maxentius remains the most likely candidate for the additions to the gates (e.g. Corcoran, ‘Maxentius’, 69). Notably, the Porta Asinaria was also located adjacent to one of Maxentius’ nodes of power in the city – the camp of the equites singulares Augusti, which were subsequently destroyed under Constantine. 40 from site to site, the palaces at Milan and Trier selected from a common set of features and strategies for manifesting a new conception of the emperor and his status. They thus operated as symbolic claims to unity in the face of the empire’s decentralisation and served to anchor this new political situation in a visual rhetoric of historical continuity. Maxentius’ buildings ‘reimported’ these strategies which had been formulated in a period of imperial absence from the Eternal City into Rome itself. Even Maxentius’ villa on the Via Appia reproduced architectural and spatial arrangements found on the Palatine. As witnessed in the other palaces of the Late Empire, the residential area of Maxentius’ villa was structured around a large apsidal aula; an even more majestic iteration was added on the Via Sacra, and it is difficult to resist the temptation of reading a competitive architectural discourse into its bold and imposing appearance. III. Reimagining Rome: Palaces and the Iconography of Roman Authority The symbolic language of Roman political authority remained a powerful force after the Empire’s fragmentation, as has been amply demonstrated in recent scholarship.127 A striking feature of the po- litical cultures of the post-Roman regna that has received little sustained commentary, however, is the discursive connection drawn between the construction and renovation of palaces and the restora- tion—or, as we will see, rejection—of Roman authority itself. That is to say, in the contemporary imag- inary the palace was inextricably connected with concepts of romanitas, ‘Roman-ness’. In this section, we will consider two comparative contexts to scrutinise the architectural manifestation of romanitas and the strength of ideological continuity in palace architecture: the successor kingdoms of the fifth and sixth centuries, and the Carolingian Empire in ninth. Post-Roman palaces were not reproductions of existing archetypes (most notably the Great Palace of Constantinople), but rather sites where the 127 Especially: Michael McCormick, Eternal Victory: Triumphal Rulership in Late Antiquity, Byzantium, and the Early Medieval West, Past and Present Publications (Cambridge: Cambridge University Press, 1990). More recently, Hendrik Dey, building on the work of McCormick and others, has argued that the ‘urban paradigm’ of the Late Empire, in which the palace was the central element, provided a persistent model for advertising authority in Europe and the Mediterranean after the Ro- man Empire’s fragmentation: Dey, Afterlife, e.g. 137–40, and passim. 41 symbolic languages of political authority were reinvented and redeployed. In each case, the palace’s architecture engaged with an idea(l) of ‘Roman’ authority, but these regimes were as prone to adapt, reject, and innovate as to ‘imitate’. As Celia Chazelle has argued, engagements with ‘Roman’ images were inherently creative and performative – an act of imagination rather than a reproduction of inherited tradition.128 Viewing palaces in this light underlines the active role of early medieval regimes and their builders played in the reimagination of the visual language of Roman authority. Emulation and distinction: Carthage, Ravenna, and Reccopolis The idea of Roman authority, as manifested in the palace, was clearly attractive in the dynamic context of the late-fifth and early-sixth centuries. The symbolic opportunities offered to the new regimes in the Mediterranean by the iconographies of Roman authority are vividly shown by the palaces of the Vandals in and around Carthage, that of Theoderic in Ravenna, and in the probable palace of ‘Rec- copolis’ in Visigothic Iberia. Carthage. In 439, the Vandal King Geiseric (428–477) stood triumphant as the new master of Carthage, one of the great economic and political centres of the Roman Mediterranean. Since the early twentieth century, scholars have argued that the Vandal kings had adopted the proconsular prae- torium, located atop the Byrsa Hill at the centre of the ancient city, as their residence.129 Procopius, an eyewitness to events during the conquest of the Vandal kingdom, records that the general Belisarius ‘went up (ἀναβὰς) to the palace and seated himself on Gelimer’s throne’ following his entrance into Carthage, suggesting an elevated location.130 Excavations on the hill have brought to light the opus caementicium foundations of structures dating to the fifth or sixth centuries, tentatively attributed to the 128 Celia Chazelle, ‘“Romanness” in Early Medieval Culture’, in Paradigms and Methods in Early Medieval Studies, ed. Celia Chazelle and Felice Lifshitz, The New Middle Ages (New York: Palgrave Macmillan, 2007), 83–84. 129 Auguste Audollent, Carthage romaine, 146 avant Jésus-Christ–698 après Jésus-Christ (Paris: Thorin et Fils, 1901), 283. 130 Proc., Wars, III.20.21, 1:399: ἐς το Παλάτιον ἀναβὰς ἐν τῷ Γελίμερος θρόνῳ ἐκάθισεν. Trans.: Procopius, The Wars of Justinian, trans. H. B. Dewing, rev. Anthony Kaldellis (Indianopolis: Hackett, 2014), 185. 42 Vandal palace (Figs. 29–30).131 Lézine recognised in the foundations a small apsidal hall, divided into three naves and terminating in a semi-circular apse; behind the apse were further rooms, including a suspected triconch preceded by a vestibule.132 Most have placed these structures in a secular context, though the possibility that it relates to the ‘basilica of the palace’ dedicated to Mary, which existed already in the Vandal period and was restored by Justinian, is appealing.133 An assemblage of audience hall and triconchos would be typical for an elite residence of this period, yet the arrangement of these structures is unusual: the hall’s entrance faces towards the southern slope of the hill. It may be that we are looking at two separate assemblages within the same range of a larger complex.134 In this light, a small basilican church or chapel with two side-chambers either side of the central apse does not seem an implausible proposition. Unfortunately, all we can say with confidence about the structures on the Byrsa Hill is that they were constructed in the Vandal period with materials including spolia of earlier monuments that were otherwise left to decay (including a monumental basilica constructed in the Antonine period located nearby) and that they served some representational function.135 Nonetheless, this indicates a significant reimagination of public space in the ancient metropolis’ centre: rather than 131 Jean Ferron and Maurice Pirard, ‘Les fouilles de Byrsa, 1953–1954’, Cahiers de Byrsa 5 (1955) : 50–53. Aïcha Ben Abed and Noël Duval, ‘Carthage, la capital du royaume et les villes de Tunisie à l’époque vandale’, in Sedes regiae (ann. 400–800), 190. 132 Dimensions: 18m by 11m. The rectangular interior foundations were explained by him as belonging to a second phase of intervention in which a cistern was added to the site: Alexandre Lézine, Carthage-Utique. Études d’architecture et d’urbanisme (Paris : CNRS Editions, 1968), 177–80. Cf. Jean Deneauve, ‘Les structures romaines de Byrsa: Historique des recherches’, in Byrsa I: Rapports préliminaires des fouilles (1974–1976), ed. Serge Lancel, Collection de l’École française de Rome 41 (Rome: École française de Rome, 1979), 54, n. 63. 133 Anthologia Latina, no. 375, ed. David R. Shackleton-Bailey, vol 1, bk. 1, BGSRT (Stuttgart: Teubner, 1982), 291: Versus in basilica palatii sanctae Mariae; Procopius, De aedificiis (Περὶ κτισμάτων), VI.5.9, in Procopii Caesariensis Opera omnia, ed Jakob Haury, vol. 3, bk. 2, BSGRT (Leipzig: Teubner, 1913), 180: ἀνέθηκε δὲ καὶ ἱερὰ τεμενὴ, τῇ μὲν θεοτόκῳ, ὅπερ ἐν Παλατίῳ έστι, καὶ τούτου ἐκτὸς τῶν τινι έπιχωρίων [ἁγίων] ἁγίᾳ Πρίμῃ. (‘He also dedicated holy shrines, one to the Theotokos, which is in the palace, and another outside it to a certain native saint, Saint Prima’); Liliane Ennabli, Carthage: une métropole chrétienne du IVe à la fin du VIIe siècle (Paris: CNRS Editions, 1997), 38, 40, 84; Lézine, Carthage-Utique, 180. 134 Lézine noted that the corridors linking the space are conjectural: Lézine, Carthage-Utique, 179. 135 Bockmann, Capital Continuous, 50–52. Basilica: Pierre Gros, Byrsa III. Rapport sur les campagnes de fouilles de 1977 à 1980 : la basilique orientale et ses abords, Collection de l’École Française de Rome (Rome : École Française de Rome, 1985), 63. For the late ancient phases, see : ibid., 113–26. 43 simply appropriating and reusing the pre-existing architectures of authority, the Vandal regime made significant interventions in the urban fabric. In textual sources we see that extra-urban residences were also a focus of the regime’s ‘Ro- manising’ image. For example, a short anonymous poem in the Anthologia Latina gives an ekphrasis of a residence belonging to Hilderic (523–530) known as ‘Anclas’.136 The poet describes what are com- monly interpreted as wall paintings depicting his imperial forebears triumphing over barbarians.137 Elsewhere in the collection, Luxorius describes the ornate, possibly Proconnesian marble, flooring of the salutatorium, the reception hall: ‘Here the floor, fitted without fault, is believed to be sprinkled snow; when your feet stand upon it, you would think they could sink!’138 Other kings are likewise known to have possessed residences outside Carthage. Thrasamund (496–523) apparently also pos- sessed a residence at a place known as ‘Alianas’ or ‘Alianae’, another Carthaginian suburb. Five poems by Felix, a uir clarissimus, describe the baths at this estate, decorated with shining bronze statues and radiant marble.139 Anclas and Alianae have not been identified archaeologically, and thus it is impos- sible to verify these poetic images. Their specificity in singling out decorative elements, however, (the pavement at Anclas; the bronze statues and marble cladding at Alianae) would strike dissonant notes to their audience were they absent in reality. Within their residences, surrounded by expensive and craft-intensive materials, the Hasding kings must have seemed the image of a ‘Roman’ ruler. 136 Felix calls the (unnamed) addressee ‘powerful Vandal-king, heir to a twin diadem’ and ‘grandson of Valentinian’, appel- lations which can only apply to Hilderic, who was related to the imperial house through his mother, Eudocia. Anthologia Latina, no. 206, 154: Vandalrice potens, gemini diadematis heres, | […] Valentiniani […] nepotis. Anclas can probably be associated with the ‘Aklas’ (Ἄκλας) in which Belisarius stayed following the occupation of Carthage: Proc., Wars, IV.7.13, 1:450. The name may derive from a waterwheel (ancla, from antlia), the subject of another epigram in the Anthologia Latina (no. 278): Michel Chalon et al., ‘Memorabile factvm: Une célébration de l’evergétisme des rois vandals dans l’Anthologie Latine’, An- tiquités africaines 21 (1985): 243–4. 137 Chalon et al., ‘Memorabile factvm’, 242–47. 138 Anthologia Latina, no. 194, 145: hic sine labe solum nix [iuncta et] sparsa putatur; |dum steterint, credas mergere posse pedes. That marble floors could be insubstantial or even liquid was a common motif in Late Antiquity. See: Fabio Barry, ‘Walking on Water: Cosmic Floors in Antiquity and the Middle Ages’, The Art Bulletin 89, no. 4 (2007): 627–56. 139 Especially: Anthologia Latina, no. 201, 150: Hic ubi conspicuis radiant nunc signa metallis | et nitido clarum marmore fulget opus. Bockmann, Capital Continuous, 54–58; Chalon et al., ‘Memorabile factvm’, 231–41. 44 Strikingly, however, the Vandal regime deliberately rejected an ideology of explicit continuity with the imperial past. Rather, it appealed instead to a notion of ‘Felix Karthago’, ‘Fortunate Carthage’, even resurrecting Punic imagery as they did so, in the construction of an independent rhetoric of authority.140 The importance of the residences in and around Carthage is stressed by Florentinus in a panegyric delivered to Thrasamund. He pronounces that, ‘with you ruling, the citadels of Carthage shine forth uninterrupted; her daughter, Alianae, follows in her steps behind and she is not lesser in merit nor esteem,’ before closing with a rhythmically repetitive passage underscoring Carthage’s pre- eminence: For Carthage retains her glory through her summits, Carthage, for the king; the victor Carthage triumphs, Carthage, the mother of the Hasdings, Carthage glitters, Carthage, through all of Libya’s lands, Carthage is exalted.141 As Florentinus makes clear, palaces were not only a way to ‘look imperial’ (so to speak), but also a means of asserting independence from Roman hegemony. Ravenna. The reign of Theoderic, king of Italy from 493 to 526, stands out for the extent and consistency by which it was presented as a direct continuation of the Roman Empire.142 At the height of Theoderic’s success there are hints that he flirted with adopting the imperial title, and his reputation as an ‘imperial builder’ is vividly attested by the programme of palatial buildings the regime spon- sored.143 In a panegyric to Theoderic delivered probably in 507, Ennodius—Ligurian churchman, par- tisan of the Ravennate court, and later Bishop of Pavia—connected the restoration of palaces to the rejuvenation of the body politic: ‘I see that the unhoped-for beauty of cities has risen from the ashes 140 Clover, ‘Felix Karthago’, DOP 40 (1986): 1–16; Bockmann, Capital Continuous, 29–32. 141 Anthologia Latina no. 371, 287–88: Te regnante diu fulent Carthaginis arces, | filia quam squitur Alianas inpare gressu, | nec meriti(s) nec honore minor, […] Nam Carthago suman retinet per culmina laudem, | Carthago [in] regem; vi(c)trix Carthago triumphat, | Carthago Asdingis genetrix; Carthago coruscat, | Carthago, excellens Libyas per ora(s). Clover, ‘Felix Karthago’, 8–10; Chalon et al, ‘Memo- rabile factvm’, 249–54. 142 See especially: Jonathan Arnold, Theoderic and the Roman Imperial Restoration (Cambridge: Cambridge University Press, 2014). 143 For further examination of this programme of buildings and its use in the regime’s rhetoric, see below, Chapter 3. 45 and that under the generosity of your government everywhere palatine rooves glow.’144 The currency of this rhetoric is vividly attested by Theoderic’s adoption of the erstwhile imperial palace in Ravenna. At its core, the Ravennate palace reproduces the topographical arrangement familiar from other Late Roman sedes. Apparently known as the ‘Laurel Grove’ or ‘At the Laurel’, the palace was formed of a dispersed complex of buildings distributed through the eastern sector of city in conjunc- tion with a new north-south thoroughfare, the platea maior (Fig. 31).145 This street was apparently lined with porticoes that extended in front of the palace, thus tying together the whole district into an imposing ‘architectonic mise-en-scène’ for the apprehension of political authority.146 Ravenna’s circus cannot be localised with certainty, though later-medieval documentary sources plausibly suggest the presence of a circus across the platea maior from the palace.147 Excavations undertaken in this area between 1908 and 1914 brought to light the remains of a richly-decorated series of spaces arranged around a peristyle courtyard (Fig. 32), a discovery that was immediately hailed as belonging to Theod- eric’s palace.148 Theoderic’s ownership has been contested, notably by Noël Duval, on the grounds 144 Ennodius of Pavia, Panegyricus dictus Clementissimo Regi Theoderico, XI.56, in Magni Felicis Ennodi Opera, ed. Friedrich Vogel, MGH. AA 7 (Berlin: Weidmann, 1885), 210: video insperatum decorem urbium cineribus evenisse et sub civilitatis plenitudine palatina ubique tecta rutilare. 507 date: Arnold, Theoderic, 33. 145 According to Agnellus of Ravenna, the imperial palace Ad Laureta was built by Valentinian III, though Honorius must surely have possessed an earlier residence: Agnellus, LPRa, c. 40, 198–99. The palace is named in Lauretum by the Anonymus Valesianus: Anon. Val. (II), XI.55, 16. For summary of the debate over its location: Deborah Mauskopf Deliyannis, Ravenna in Late Antiquity (Cambridge: Cambridge University Press, 2010), 56–58. Ad Laureta (et sim.) could be explained as a calque on the name of the Daphne (Δάφνη) wing of the palace at Constantinople, traditionally interpreted as the oldest part of the palace complex: ibid., 56. 146 Dey, Afterlife, 115. See also: Mark Johnson, ‘Toward a History of Theoderic’s Building Program’, DOP 42 (1988): 82. The Anonymus Valesianus states that Theoderic ‘completed a portico around the palace’: Anon. Val. (II), XII.71, 20: portica circa palatium perfecit. A papyrus of 572, subscribed by a notary named Flavius Iohannis, describes his conducting his business ‘in the portico of the sacred palace, at the Gold Mint’: Jan Olof Tjäder, ed., Die nichtliterarischen lateinischen Papyri Italiens aus der Zeit 445–700, vol. 2, Skrifter utgivna av Svenska Instituet i Rom, 4th series, 19 (Lund: C. W. K. Gleerup, 1982), no. 35, 112: Fl(avius) Iohannis, for(ensis) huius splendedissimiae urbis Ravennatis, habens stationem ad Monitam auro in porticum Sacri Palati, scriptor huius instrumenti, complevi. 147 Johnson, ‘Theoderic’s Building Program’, 109, with n. 110; Enrico Cirelli, Ravenna: archeologia di una città, Contributi di Archeologia Medievale 2 (Florence: All’Insegna del Giglio, 2008), 90–91. For example, charters of 960 and 982 locate a circulus in the region of Sant’Agata Maggiore and the platea maior (by then known as the platea publica): Marco Fantuzzi, ed., Monumenti ravennati de’ secoli di mezzo per la maggior parte inediti, vol. 1 (Venice: Francesco Andreola, 1801), nos., 34, 52, pp. 151, 211. Agnellus at least, identifies what was probably a circus (the phrase he uses is stadium tabulae) outside the city’s walls: Agnellus, LPRa, cc. 22, 153, pp. 169, 330. 148 The excavations were not published beyond the Ghirardini’s initial report: Gherardo Ghirardini, ‘Gli scavi del Palazzo di Teoderico a Ravenna’, Monumenti antichi 24 (1917): coll. 737–888. Fuller chronological sequences have been proposed by Berti and Augenti: Fede Berti, Mosaici antichi in Italia. VIII: Aemilia: Ravenna (Roma: Istituto Poligrafico e Zecca dello 46 that there was no positive evidence for royal ownership.149 Duval’s caution is salutary, yet the topo- nymic evidence is unanimous that the palace was located in this region, and the peristyle’s remarkably precise alignment with the church of Sant’Apollinare Nuovo, founded by Theoderic himself, makes other attributions hard to credit.150 The best contemporary source, the sixth-century Anonymus Valesi- anus states only that Theoderic added the portico and ‘brought the palace, which he did not dedicate, to completion’.151 This phrase is ambiguous but the broader context of Theoderic’s building projects suggests that the regime renovated or added to the existing imperial palace at its northern extent, rather than building a palace ex novo.152 Typologically, in its sixth-century phase the complex conforms to the well-established con- ventions of elite Roman residences, with a deep apsidal hall (27 by 11m) placed axially on the main peristyle and, to the side, a triconch triclinium.153 The mosaic pavements that adorned each phase of the complex’s history give some idea of its original splendour.154 For example, the fragmentary pave- ment which adorned the central space of the triconch (Fig. 33) displayed personifications of the four seasons surrounding an image of Bellerophon burying his lance in the maw of the Chimera.155 A short hexametric inscription encourages the gathered diners to enjoy the fruits of the seasons: ‘take what Stato, 1976), 10–86; Andrea Augenti, ‘Archeologia e topografia a Ravenna: il Palazzo di Teoderico e la Moneta Aurea’, Archeologia Medievale 32 (2005): 7–23; idem, ‘The Palace of Theoderic at Ravenna: A New Analysis of the Complex’, in Housing in Late Antiquity: From Palaces to Shops, ed. Luke Lavan et al., Late Antique Archaeology 3/2 (Leiden: Brill, 2007), 425–53. 149 Duval, ‘Que savons-nous’, 369–71 ; idem, ‘Comment reconnaitre’, 32–39, 57–59. 150 Agnellus gives the text of the dedicatory inscription in the apse, now lost: Agnellus, LPRa, c. 86, 154: Theodoricus rex hanc ecclesiam a fundamentis in nomine Domini nostri Iesu Christi fecit (‘King Theoderic built this church from the foundations in the name of Our Lord Jesus Christ’). For toponymic evidence, see: Johnson, ‘Theoderic’s Building Program’, 81. Duval allows only that the structure belonged to an ‘urban villa discovered in the quarter attributed traditionally to the palace of Theoderic’: Duval, ‘Comment reconnaitre’, 32–33 (on the toponymic evidence), 59 (quotation). 151 Anon. Val. (II), XII.71, 20: palatium usque ad perfectum fecit, quem non dedicavit. 152 Deliyannis, Ravenna, 55–58; Cristina La Rocca, ‘Mores tuos fabricae loquuntur: Building Activity and the Rhetoric of Power in Ostrogothic Italy’, Haskins Society Journal 26 (2014): 1–29. 153 The main peristyle is rectangular, measuring 32.5m by 41m: Ghirardini, ‘Gli scavi’, col. 741. The aula seems to have been added sometime in the fifth century, prior to a second building phase when the triconch was added (Augenti’s phases III and IV): Augenti, ‘Archeologia e topografia’, 13–16. 154 Berti, Mosaici antichi, 30–86 (cat. nos. 1–65). 155 Ghirardini, ‘Gli scavi’, coll. 782–98. Berti has dated it to the first quarter of the sixth century: Berti, Mosaici antichi, 77– 81. 47 Autumn, Spring, Winter, and Summer supply in turn and which is engendered throughout the world’ – an unusual combination of motifs.156 Architecturally, the triconchos was a novel invention in late- Roman elite architecture. Its polylobed form responded to new norms for communal dining (materi- alised in the adoption of the semi-circular couch, the stibadium) and the value placed on the plasticity of curving forms in architecture as a sign of luxury.157 In both the aula and the triconchos, therefore, the Ravennate palace is on the cutting edge of aristocratic fashion of the fifth and sixth centuries. An idea of the palace’s external appearance is furnished by the mosaic on the south wall of the adjacent church, Sant’Apollinare Nuovo. An image of a pedimented façade flanked by porticoes and labelled PALATIVM in an inscription (Fig. 34) stands before the walled urban area of Ravenna.158 In the central space two curtains part to reveal a golden doorway; in the intercolumniations were origi- nally figures—whose hands, raised in excitable gestures of greeting or acclamation, remain visible— that were removed in the ‘reconciliation’ of the church to orthodoxy by the bishop Agnellus (577– 580) (Fig. 35).159 This image has provoked a great deal of debate about what precisely in represents and the conventions used by the artists to translate architecture into two dimensions.160 While it is clearly idealised and stereotyped, there is little reason to doubt its overall authenticity as a record of the palace’s external aspect.161 In any case, the palace’s façade—the ‘fitting face of empire’, as 156 Ghirardini, ‘Gli scavi’, coll. 788–89; Berti, Mosaici antichi, 79. SUME QUOD AUTUMNUS QUOD / VER QUOD BRUMA QUOD ESTAS / ALTERNIS REPARANT ET / TOTO CREANTUR IN ORBE. Simon Ellis, ‘Power, Ar- chitecture, and Decor: How the Late Roman Aristocrat Appeared to His Guests’, in Roman Art in the Private Sphere: New Perspectives on the Architecture and Decor of the Domus, Villa, and Insula, ed. Elaine Gazda, 2nd ed. (Ann Arbor: University of Michigan Press, 2010), 125–27. 157 Simon Ellis, Roman Housing (London: Duckworth, 2000), 67. Cf. Sfameni, Ville residenziali, 135. 158 For visual comparanda for the palatium mosaic, see: Friedrich Deichmann, Ravenna, Hauptstadt des spätantiken Abendlandes, vol. 2, bk. 2 (Wiesbaden: Steiner, 1974), 143–5. 159 Agnellus, LPRa, c. 86, 253–54; Giuseppe Bovini, ‘Antichi rifacimenti nei mosaici di S. Apollinare Nuovo di Ravenna’, Corso di cultura sull’arte ravennate e bizantina 13 (1966): 51–81; Arthur Urbano, ‘Donation, Dedication, and Damnatio Memoriae: The Catholic Reconciliation of Ravenna and the Church of Sant’Apollinare Nuovo’, Journal of Early Christian Studies 13, no. 1 (2005): 71–110. 160 Deichmann, Ravenna, 2.2:141–43. Dyggve argued that it should be interpreted as a view into a colonnaded courtyard, a formal assemblage he termed the ‘uncovered basilica’ (basilica discoperta or ipetrale): Dyggve, Ravennatum palatium sacrum, 51– 53. This reconstruction was decisively critiqued by Duval and de Francovich: Duval, ‘Que savons-nous’, 345–55; de Fran- covich, Il Palatium di Teodorico a Ravenna. 161 At a basic level, the viewer would be immediately aware of significant divergences from reality given that the entrance to the palace was located nearby. The skyline of Ravenna behind the palatium seems to reference real structures, though 48 Cassiodorus’ letter to the curator palatii described it—was a powerful manifestation of the regime’s authority, both as physical architecture and as representational emblem.162 Theoderic’s program of buildings in Ravenna has usually been interpreted as imitatio imperii on a grand scale: an attempt to reproduce the symbolic topography of Constantinople in Ravenna. At root, this idea is founded on a passage in Agnellus of Ravenna’s ninth-century Book of Pontiffs of the Church of Ravenna. Agnellus describes a mosaic located in the apse of a triclinium called ‘By the Sea’ (Ad mare), which apparently depicted the king mounted on horseback and flanked by personifications of Rome and Ravenna. What follows is a (confusing) list of toponyms by which Agnellus triangulates this mosaic’s position: ‘above the gate and at the front of the main door that is called Ad Calchi, where the main gate of the palace was, in the place which is called Sicrestum, where the church of the Savior is seen to be’, at the ‘pinnacle’ (in pinnaculo) of which was located the mosaic.163 To be sure, these names seem to derive from the Byzantine Great Palace: the main gate Ad Calchi immediately evokes the Chalkē (Χαλκῆ) or ‘Bronze’ Gate, and Sicrestum is likely a degeneration of sekreton/sekretarion (σέκρετον/σεκρετάριον) or the Latin sacristia, referring to administrative departments.164 On this basis, it is frequently asserted that Theoderic intended for his royal seat to be, in effect, Constantinople-in-miniature.165 While persuasive on its face, and it is true that Theoderic spent a dec- ade in the imperial seat as a hostage, to read these toponyms as evidence for Theoderic’s intentions is a fraught exercise. Firstly, if Agnellus’ list of toponyms does indeed reflect sixth-century usage, it is unclear to whom these symbolic references would have held significance other than to the king himself again in a schematic, stereotyped fashion, thereby making their definitive identification impossible: Deichmann, Ravenna, 2.2:145; Deliyannis, Ravenna, 163; Johnson, ‘Theoderic’s Building Program’, 88. 162 Cass., Var., VII.5, 204 (as above, 1). 163 Agnellus, LPRa, c. 94, 258–59: Hic autem similis fuit in isto palatio quod ipse aedificavit in tribunale triclinii quod uocatur Ad mare, supra portam et in fronte regiae quae dicitur Ad Calchi istius ciuitatis, ubi prima porta palatii fuit, in loco qui uocatur Sicrestum, ubi ecclesia Saluatoris esse uidetur. In pinnaculo ipsius loci fuit Theodorici effigies. On the difficulties of this text: Duval, 'Que savons-nous’, 356–63; Deichmann, Ravenna. Hauptstadt des spätantiken Abdenlandes, vol. 2, bk. 3 (Stuttgart: Steiner, 1989), 53–54; Deli- yannis, Ravenna, 120–21. 164 Agnellus, LPRa, c. 94, 258. 165 See esp. e.g.: Johnson, ‘Toward a History’, 82–92; Dey, Afterlife, 119. 49 and to diplomats from the eastern court. More seriously, however, Agnellus himself states not only that he had not actually seen this mosaic, a fact underlined by his use of the imperfect tense and the explicit comparison with that in Pavia (which he had seen), but also that the gate Ad Calchi no longer existed at his time of writing.166 In any case, that Theoderic himself could serve as the link between the two representational topographies falls apart on the basis of chronology. The Justinianic Chalkē was only built following the Nika Riots in 532, and even the original gate commissioned by Anastasius seems to have been first built long after Theoderic had left for Italy, sometime after 498.167 Using such onomastic connections between Ravenna and Constantinople as a basis for reconstructing the ideo- logical intentions of Theoderic’s regime needs to be treated with caution; one suspects that they were, in fact, later accretions issuing from the presence of the Byzantine exarchs in the city until 751.168 The adoption and renovation of the imperial palace in Ravenna was undoubtedly a statement of Theode- ric’s legitimacy as ruler in Italy and of the continued currency of political representation before an urban public. Its ideological significance was not contingent, however, on its functioning as a material reproduction of an existing archetype so much as its activation in wider discourses of good govern- ance, urban renewal, and architectural spectacle – as indicated to us by Ennodius. Reccopolis. Despite the wealth of contemporary sources from the Visigothic kingdom of Iberia, the location and form of the palace in Toledo is lost.169 It is clear, however, that it was not the only royal seat. As one example, according to Julian of Toledo the king Reccesuinth died in 672 on a royal estate known as Gerticos in the vicinity of Salamanca, where also his successor Wamba (r. 672– 166 Deliyannis, Ravenna, 121. On the palace at Pavia, see below, Chapter 3. 167 Cyril Mango, The Brazen House: A Study of the Vestibule of the Imperial Palace at Constantinople, Arkælogisk-kunsthistoriske Meddelelser 4, no. 4 (Copenhagen: Munksgaard, 1959), 26–30. 168 Such a process can be witnessed in action by the fact that Agnellus’ own monastery, St Mary ad Blachernas—referencing the church of the Virgin at the Blachernae in Constantinople—seems not to have been founded earlier than the seventh century. Tellingly, it is the resting place of the exarch Theodore (678–687), who Agnellus also records as having founded a church dedicated to St Theodore the Deacon near to the place called ad Calchi: Agnellus, LPRa, c. 119, 290. 169 Isabel Velázquez and Gisela Ripoll, ‘Toletvm, la construcción de una vrbs regia’, in Sedes regiae (ann. 400–800), 521–78. 50 680) was elected by the aristocracy.170 Characteristically for the Iberian kingdom, Wamba’s acclamation and unction took place in Toledo; thereafter we do not hear again of Gerticos. The well-preserved Visigothic-period residence known as Reccopolis, however, is not only a remarkable record of secular architecture in this period, but also provides an insight into the strategies of romanitas in the later sixth century. The chronicler John of Biclaro, the most extensive narrative source for sixth-century Hispania, wrote that King Leovigild (568–586) founded a city in Celtiberia in 578 that he named ‘Recopolis’ after his son, Reccared – an act with obvious imperial precedents.171 The basic information of John’s account is recapitulated in other Latin narrative sources from the peninsula, stating only that it was located in ‘Celtiberia’ (roughly the region of Guadalajara). The tenth-century history of Ahmad ibn Muhammad al-Rāzī, associates Reccopolis with the fortress at Zorita and notes that the early medieval city’s stones were quarried for its construction.172 Investigations on a rise opposite the castillo at Zorita de los Canes, have uncovered structures which can be identified with Reccopolis with some confi- dence (Figs. 36–37).173 Reccopolis clearly engages with discourses of romanitas, but it betrays a concep- tual shift in its authorising archetype. 170 Julian of Toledo, Historiae Wambae regis, c. 3, in Passiones vitaeque sanctorum aevi merovingici III, ed. Bruno Krusch and Wilhelm Levison, MGH. SS rer. Merov. 5 (Hannover: Hahn, 1910), 502–03. 171 John of Biclaro, Chronicon, c. 50, in Victor Tunnunensis Chronicon cum reliquiis ex Consularibus Caesaraugustanis et Iohannis Biclarensis Chronicon, ed. Carmen Cardelle de Hartmann, CCSL 173A (Turnhout: Brepols, 2001), 70: ciuitatem in Celtiberia ex nomine filii condidit, que Recopolim nuncupatur, quam miro opera in minibus e suburbanis adornans, priuilegia populo noue urbis instituit. Examples of ‘eponymous’ foundations range from Adrianople (Hadrianoupolis; modern Edirne, Turkey), to the rededication of Byzantion as Constantinople. This had some currency in the post-Roman regna: Hadrumetum was renamed Hunericopolis during the reign of the Vandal king Huneric. 172 Al-Rāzī’s history is unfortunately only extant in a fifteenth-century Castilian version. Ahmad ibn Muhammad al-Rāzī, Chronicle, XIX–XX, in Cronica del Moro Rasis, ed. Diego Catalan and Maria Soledad de Andres, Fuentes cronísticas de la Historia de España (Madrid: Ediciones Gredos, 1975), 60–62: E la çibdat de Rracupel yaze entre Santa Bayra e Çerca. E poblola Lanbilote para su fijo, que avia nonbre Rracupel, e por eso puso a la çibdat el nonbre del fijo. […] E [Zorita] ay munchos buenos arboles que dan muchas espeçias e buenas; e es muy fuerte çibdat e muy alta e fizieronla de las piedras de Rracupel, que las ay muy buenas. 173 The identification of the site with Reccopolis is supported by the discovery of a hoard of bronze tremisses at Zorita with the legend RECCOPOLI FECIT: Lauro Olmo Enciso, ‘Fuentes escritas y primeras investigaciones sobre Recópolis’, in Recópolis y la ciudad en la época visigoda, ed. Lauro Olmo Enciso, Zona Arqueológica 9 (Alcalá de Henares: Museo Arqueoló- gico Regional, 2008), 33. The site seems to have been constructed ex novo in the sixth century but by the ninth centuries the stones were being quarried for use at Zorita de los Canes, a fact which lends credence to the picture derived from al- Rāzī: Lauro Olmo Enciso et al., ‘Recópolis y su justificación científica: la secuencia estratigráfica’, in Recópolis y la ciudad, 64–75. On the first excavation campaign, see: Klaus Raddatz, ‘Studien zu Reccopolis I. Die archäologischen Befunde’, Madrider Mitteilungen 5 (1964): 213–23. 51 Stretching out over the steep hillside overlooking the Tagus, Reccopolis’ plan can be broadly divided into two sectors.174 At the site’s crest, beyond a double-bayed gate separating the upper and lower zones, the site opens out onto a large plaza. To the east lies a cruciform basilica, likely a church based on the continuous transept ahead of the apse (Fig. 38).175 To the north is a long, ‘L’-shaped building commonly referred to as the palatium. This structure, a full 133m in length, is divided into two aisles by square piers and supported along its southern edge by six semi-circular buttresses (Fig. 39).176 This structure was evidently constructed over two storeys, with the principal representational spaces being located on the upper floor – a theme to which we will return below.177 The robust piers supported vaults for the upper floor which was paved in opus signinum.178 Reccopolis is an uncommonly complete assemblage of buildings, yet conversely its function is not totally clear. The sources are unan- imous in referring to it as a civitas (rather than a palatium) and the designation as the ‘city of Reccared’, if John of Biclaro is to be believed, does not guarantee that he, let alone his father Leovigild, officially resided there.179 Roger Collins has raised the possibility that the derivation of the name from Reccared, son of Leovigild, may have been a ‘post facto rationalization’ by John as it ‘makes no sense 174 Traces of additional structures have recently been discovered in the lower zone, as well as a possible extra-mural suburb, through geomagnetic survey: Joachim Henning et al., ‘Reccopolis Revealed: The First Geomagnetic Mapping of the Early Medieval Visigothic Royal Town’, Antiquity 93 (2019): 735–51. 175 Numerous sculptural fragments of a chancel barrier were also discovered. These have been reconstructed and are on display in the Museo Arqueólogico Nacional, Madrid (inv. 57872). See: Luis J. Balmaseda Muncharaz, ‘La escultura de Recópolis’, in Recópolis y la ciudad, 142–57. 176 Olmo Enciso, ‘Recópolis’, 47. 177 Ibid. Arce believed that, rather than a residential structure, this was a horreum: Javier Arce, ‘The so-called Visigothic “Palatium” of Recópolis (Spain): An Archaeological and Historical Analysis’, in The Emperor’s House, 69. It is likely that the vaulted substructures were used for storage among other ‘mundane’ functions, yet as we will see it was becoming common in this period for representational spaces to be located on the upper floors. 178 Olmo Enciso, ‘Recópolis’, 47. Additional, probably residential, structures relating to this complex were recently identi- fied: Henning et al., ‘Reccopolis Revealed’, 742–43. 179 According to John of Biclaro, Leovigild’s sons, Reccared and Hermenegild, shared in the kingship since 573: John of Biclaro, Chronicon, c. 27, 65. Hermenegild later rebelled against his father with the support of the Byzantines: ibid., c. 68, 74; Greg. Tur., Hist., V.38, 245. Note, however, that Leovigild granted Hermenegild and his Frankish wife, Ingund, ‘one of his cities, residing in which they would rule’ according to Gregory of Tours, opening the possibility that Reccopolis was in fact intended as Reccared’s residence during his co-rule with his father, though this is hypothetical: Greg. Tur., Hist., V.38, 244: Leuvichidus autem dedit eis unam de civitatibus, in qua resedentes regnarent. I thank Damián Fernández for drawing my attention to this passage. 52 linguistically’.180 Collins instead sensibly suggests that ‘Rec(c)opolis was intended to denote ‘City of the King’, an interpretation that would make some sense as John of Biclaro also mentions the foundation of another such city only two years later named Victoriacum to commemorate his victory over the Basques.181 Despite its contracted urban area and distinct architectural forms, at Reccopolis we see a final, faint echo of concepts we witnessed in the fourth-century sedes, above. The palatial compound, inte- grated into the urban area yet set apart from it by a wall and gate, echoes the spatial strategies that related the new imperial seats to their urban audiences. Hendrik Dey is correct to underline the sense of culmination as one proceeds along the main thoroughfare lined with houses and shops towards the upper ‘acropolis’, the only access to which was through the main gate.182 Reccopolis’ architecture pa- tently engages with an image of Romanising authority, but its architecture is only partially explained by comparison with earlier imperial sedes, or indeed with Constantinople.183 Rather, the best analogue for Recopolis’ architecture is Justinian’s monumental Justiniana Prima (Caričin Grad, Serbia), founded to honour his birthplace in 535 (Fig. 40).184 The two sites share a topographic orientation towards graduated wall enclosures and a gravitational focus around the uppermost compound, which at Jus- tiniana Prima housed the cathedral, baptistery, and the bishop’s palace.185 Comparison with Justiniana Prima suggests something of Reccopolis’ purpose. Whether con- stantly occupied or not, the residence at Reccopolis manifested the regime’s power in the Iberian landscape. That the complex possessed a mint indicates that in all events that it was a favoured site of 180 Roger Collins, Visigothic Spain, 409–711 (Oxford: Blackwell, 2004), 55–56. Similar eponymous foundations from Late Antiquity do not abbreviate the founder’s name so it is unclear why ‘Reccaredopolis’ was not the choice here. 181 Ibid.; John of Biclaro, Chronicon, c. 60, 72. 182 Dey, Afterlife, 147. 183 Olmo Enciso has attempted to relate features of Reccopolis’ topography to Constantinople, arguing that the gate (for example) was ‘inspired’ by the Chalkē: Olmo Enciso, ‘Recópolis’, 49–50, 53. 184 For the archaeology of the site: Noël Duval and Vladislav Popović, eds., Caričin Grad, 4 vols., Collection de l’École française de Rome 75 (Rome: École française de Rome, 1984–2019). 185 Dey, Afterlife, 147–48. 53 royal authority and played a role in the administration of the region.186 Between the façade of the two- storey residence looming over the Tagus and the encircling mural circuit punctuated by towers, the site’s foreboding aspect clearly marked its elevated status as a locus of the regime’s authority. The supply of the intramural area by an aqueduct—exceptional as an ex novo construction in this period— may have extended its reach further still, as well as evoking memories of characteristically ‘imperial’ patronage.187 Rather than showing the extent of Constantinople’s reach as a model for imitation, how- ever, Reccopolis highlights the flexibility of romanitas as discourse and the singular potency of palaces as sites to manifest them. Given the political context of conflict between the increasingly-assertive Iberian kingdom and the Byzantine province of Spania in the south of the peninsula, instituted less than two decades before Leovigild’s accession, we can in fact interpret the echoes of Justiniana Prima at Reccopolis as agonistic in their intent: an appropriation—and thus new imagining—of ‘Roman’ au- thority in the rejection of Byzantine hegemony.188 The iconography of Roman authority in the Carolingian world Writing in the late-ninth century, the monk of Saint-Gallen, Notker, reflected on Charlemagne’s (768– 814) foundation of his palace at Aachen. The king ‘took pleasure in building in his own land’, Notker writes, and he founded, following ‘his own plan a basilica finer by far than the works of the ancient Romans’.189 Charlemagne’s courtier, Einhard, similarly placed Aachen at the head of the construction 186 Lauro Olmo Enciso and Manuel Castro Priego, ‘Coins, Cities, and Archaeological Contexts in the Centre of the Iberian Peninsula Between the 6th and 8th Centuries AD: Reccopolis and Toledo’, in Numismatica e archeologia. Monete, stragrafie e contesti: Dati a confronto, ed. Giacomo Pardini et al. (Rome: Quasar, 2018), 557–572 (esp. 558–61). 187 Javier Martínez Jiménez, ‘A Preliminary Study of the Aqueduct of Reccopolis’, Oxford Journal of Archaeology 34, no. 3 (2015): 301–20. 188 For the establishment of the Byzantine province: Isidore of Seville, Historia Gothorum, Vandalorum et Sueuorum, c. 47, in Chronica minora saec. IV. V. VI. VII, ed. Theodor Mommsen, vol. 2, MGH AA 11 (Berlin: Weidmann, 1894), 286. On the province’s extent and conflict with the Visigothic kingdom: Jamie Wood, ‘Defending Byzantine Spain: Frontiers and Di- plomacy’, EME 18 (2010): 292–319. 189 Notker, Gesta Karoli, I.28, 38: adeo ut in genitali solo basilicam, antiquis Romanorum operibus praestantiorem, fabricare propria dispositione molitus in brevi compotem se voti sui gauderet. Trans.: Notker, ‘The Deeds of Emperor Charles’, in Charlemagne and Louis the Pious: Lives by Einhard, Notker, Ermoldus, Thegan, and the Astronomer, trans. Thomas F. X. Noble (University Park, PA: Pennsylvania State University Press, 2009), 83. 54 projects sponsored by the king in his ‘Life of Charles the Emperor’ completed sometime in the reign of Louis the Pious (emperor 813–840), a sign in his view of a revitalised Christian imperium.190 The ascent of the Carolingians has traditionally been seen as a watershed at which ‘Roman’ forms were reintroduced into European culture, and the building projects associated with Charlemagne have thus occupied a privileged place in the historiography of medieval architecture.191 In general, scholars have understood the architecture of the palace at Aachen in similar terms to Notker and Einhard: as a revival of Roman forms to assert a programmatic vision of imperial power, a new ‘capital’ reviving the symbolic topographies of the Roman sedes after a long dormancy.192 For Dey, Aachen is a final instan- tiation of an ‘unbroken sequence of “capitals” built in response to a common paradigm’ that originated in Late Antiquity.193 By placing Aachen in comparative context, however, a different picture emerges. While its architecture was certainly used to perform an image of romanitas, its referents were not to be found in any singular, authorising archetype. Just like the palaces of the earlier regna, those of the Carolingians creatively construct and perform new images of imperium in a Roman mould. For all its evident monumentality, Aachen does not seem to have been a significant location prior to Charlemagne’s palace. There had been a Roman settlement associated with the hot springs, and certainly a royal estate of some kind by the Carolingian period if not before, though not one of any conspicuous prominence.194 Dendrochronological analysis has shown that the palace was 190 Einhard, Vita Karoli imperatoris, c. 17, in Einhardi Vita Karoli magni, ed. Georg Pertz, rev. Georg Waitz, 6th ed., MGH SS rer. Germ. 25 (Hannover: Hahn, 1911), 20. 191 For Roman forms in architecture, see e.g.: Kenneth Conant, Carolingian and Romanesque Architecture, 800–1200, 2nd ed., The Pelican History of Art (Harmondsworth, UK: Penguin Books, 1979), 43–55. For a reappraisal of romanitas at the court of Charlemagne: Ildar Garipzanov, The Symbolic Language of Authority in the Carolingian World (ca. 751-877), Brill Studies on the Early Middle Ages 6 (Leiden: Brill, 2008). 192 E.g. Ewig, ‘Résidence et capitale’, 58–59; Liebeschuetz, ‘Ravenna to Aachen’, 28. 193 Dey, Afterlife, 241. 194 Andreas Schaub and Tanja Kohlberger-Schaub, ‘On the Origins of the Great Carolingian Place of Power: Recent Excavations at Aachen Cathedral’, in Churches and Social Power in Early Medieval Europe: Integrating Archaeological and Historical Approaches, ed. José C. Sánchez-Pardo and Michael G. Shapland, Studies in the Early Middle Ages 42 (Turnhout: Brepols, 2015), 473–79. See further, below: Chapter 4. 55 constructed in a fairly narrow window between 793 and 803.195 The general layout has been well- established in archaeology and the palace chapel (now the city’s cathedral) is substantially original (Fig. 41).196 The main representational spaces are located at opposing ends of the complex, linked by a two- storey gallery. Late-ancient analogues present themselves throughout. For example, the large basilican hall (17.2 by 44m) on the gallery’s northern evokes in its size and form the austere grandiosity of the Constantinian aula at Trier, and the insertion of a secondary apse into the northern wall invites a wider range of possible comparisons, including the Basilica of Maxentius in Rome. More famously, the oc- tagonal palace chapel (Fig. 42), at the southern end of the gallery, clearly derives in its architecture and ornamentation from late-ancient traditions; its centrally-planned, double-shell design, is unique in this period. Consequently, art historians have argued that the rounded form was a device to link the chapel to a range of authorising structures from the past, such as the Constantinian Rotunda of the Holy Sepulchre or even the Temple of Solomon.197 The structure to which it is most closely filiated, how- ever, is the sixth-century church of San Vitale in Ravenna (Fig. 43), founded by bishop Ecclesius of Ravenna (522–532) in 526 and completed under Maximian (546–557). Interpretation of the chapel’s architecture have followed a familiar pattern. The palace at Aa- chen was, the argument runs, the result of Charlemagne’s desire to create a new ‘imperial’ image for his authority by drawing on Roman and Byzantine models. While few would now agree that, for ex- ample, the palace chapel was directly based on structures of the imperial palace such as the 195 Schaub and Kohlberger-Schaub, ‘Origins’, 487–90; Harald Müller et al. ‘Pfalz und vicus Aachen in karolingischer Zeit’, in Aachen: von den Anfängen bis zur Gegenwart, ed. Thomas Kraus, vol. 2, Veröffentlichen des Stadtarchivs Aachen 14 (Aachen: Mayersche Buchhandlung, 2013), 145–57. 196 For a balanced treatment, see: Uwe Lobbedey, ‘Carolingian Royal Palaces: The State of Research from an Architectural Historian’s Viewpoint’, in Court Culture in the Early Middle Ages: The Proceedings of the First Alcuin Conference, ed. Catherine Cubitt, Studies in the Early Middle Ages 3 (Turnhout: Brepols, 2003), 129–54. 197 Generally, see: Werner Jacobsen, ‘Die Pfalzkonzeptionen Karls des Großen’, in Karl der Große als vielberufener Vorfahr: sein Bild in der Kunst der Fürsten, Kirchen und Städte, ed. Lieselotte Saurma-Jeltsch, Schriften des historischen Museums Frank- furt am Main 19 (Sigmaringen: Jan Thorbecke Verlag, 1994), 46–48; Judith Ley, ‘Aquis palatium: Spätantiker Palast oder frühmittelalterliche Pfalz? Architekturhistorische Überlegungen zur Ikonographie der Aachener Pfalz’, in The Emperor’s House, 127–46; Müller et al., ‘Pfalz und vicus Aachen’, 213–65. 56 Chrysotriklinos (Χρυσοτρίκλινος, the ‘Golden Dining Hall’), the more general argument that Aachen rep- resented a programmatic statement of imperial intent by reproducing ‘Byzantine’ forms remains largely intact.198 San Vitale, the palace chapel’s closest formal antecedent, is thus envisaged as providing access to a world of ‘Byzantine’ or otherwise imperial associations.199 The chapel’s architects did not follow their models slavishly, however. When compared to San Vitale, its closest formal antecedent, the chapel’s interior possesses a more emphatic verticality, drawing the eye upwards into the cupola, a quality given further accent by the double rows of columns that screen the upper gallery and provide the semblance of an uninterrupted interior plane (Fig. 44). The chapel thus dispenses with the billow- ing effect of the sixth-century church’s supporting exedrae, which so successfully disguise the struc- ture’s massing, in preference for a more immediately imposing conception of architectural space. Aachen’s forms were implicated in a rich field of possible significations and could reference multiple structures at once. Inevitably, therefore, it does not reveal a singular guise of Carolingian authority.200 Rather, Aachen was born from an evolving symbolic language that brought together het- erogeneous signs, images, and referents in a new creative ‘imagining’. A case in point is the court’s interest in Ravenna. Charlemagne requested marble spolia from Roma and Ravenna in exchanges with Pope Hadrian from 787, and in 801 had a statue of Theoderic, which had once stood before the gate of the royal palace in Ravenna, transferred to Aachen.201 Despite the fact that Theoderic was no 198 Heinrich Fichtenau, ‘Byzanz und die Pfalz zu Aachen‘, Mitteilungen des Instituts für Österreichische Geschichtsforschung 59, no. 1 (1951): 1–54. Cf. the objections of Donald Bullough, ‘Imagines regum and the Early Medieval West’, in Donald Bullough, Carolingian Renewal: Sources and Heritage (Manchester: Manchester University Press, 1991), 57–58, with n. 85 (on 89–90). The Chrysotriklinos was established in the sixth century and formed the heart of the Byzantine palace’s ceremonial life by the tenth century, recurring frequently in the Book of Ceremonies. Dey has also recently argued that Aachen reproduces a topo- graphical idea derived from the Byzantine Great Palace: Dey, Afterlife, 240–41. 199 E.g. Jacobsen, ‘Pfalzkonzeptionen’, 47; Ley, ‘Aquis palatium’, 136; McClendon, Origins, 123; Bandmann linked it to the Theotokos of the Blachernae and Theotokos of the Pharos, two palace chapels of the Byzantine emperor: Günter Band- mann, ‘Die Vorbilder der Aachener Pfalzkapelle’, in Karolingische Kunst, ed. Wolfgang Braunfels and Hermann Schnitzler, vol. 3 of Karl der Große. Lebenswerk und Nachleben, ed. Wolfgang Braunfels, (Düsseldorf: L. Schwann, 1965), 424–62. 200 Unlike Samson, I do not see this as rendering these images ineffective: Ross Samson, ‘Carolingian Palaces and the Poverty of Ideology’, in Meaningful Architecture: Social Interpretations of Buildings, ed. Martin Locock, Worldwide Archaeology Series 9 (Aldershot: Avebury, 1994), 99–131. 201 Codex Carolinus 81 is a reply from Pope Hadrian to Charlemagne, in which he refers to an earlier request from the king ‘in which it is stated that that we should grant to you mosaic and marbles and other images placed both on the floor and 57 emperor—and the kingdom he founded was ultimately toppled by Eastern Roman armies—it seems that his image became assimilated with an ideal of ‘Romanised’ authority. The spoliation of Ravenna was thus a means of imbuing with Carolingian imperial project with ‘an imperial look’, in Janet Nel- son’s words.202 Even then, Theoderic’s image was not unambiguous. In his poem De imagine Tetrici (‘On the Image of Tetricus’, i.e. Theoderic) of 829, Walahfrid Strabo utilises it as a leitmotif in a discourse on tyranny and pride: Theoderich, once ruler in Italian lands, Being miserly kept much of his great wealth for himself. But the wretched man now walks alone along the pitch-black Avemus; Scarcely anything in the world is left to him save a sparse reputation, Even if the rabble of the baths make a ford for him. Nor is this without cause, for he is cursed in every mouth, And the reproach of God himself and the judgement of the world Consign him to eternal flames and the great abyss. […] If you know that the proud are wont to be placed in chariots and on horses, you will scarcely be astonished that he is sitting on a horse.203 Having contrasted Theoderic’s hubris with Louis’ Christian virtue, Walahfrid imagines the statue, an- imated, leading a procession into the palace itself. He exclaims: ‘Let the great image of your colossus depart, O Rome; it is excessive. Should great Caesar [i.e. Louis] will, whatever the wretched world stirs up will migrate to the Frankish citadel.’204 From model of ‘Romanised’ authority to image of excess, Theoderic’s statue alerts us to the shifting boundaries of romanitas at the Aachen palace. on the walls of the palace in the city of Ravenna’, a request to which the Pope accedes. Codex Carolinus 81, in Epistolae Merowingici et Karolini aevi, ed. Ernst Dümmler, vol. 1, MGH. Epp. 3 (Berlin: Weidmann, 1892), 614: In quibus referebatur, quod palatii Ravennate civitatis mosivo atque marmores ceterisque exemplis tam in strato quamque in parietibus sitis vobis tribuissemus. 202 Janet Nelson, ‘Charlemagne and Ravenna’, in Ravenna: Its Role in Earlier Medieval Change and Exchange, ed. Judith Herrin and Janet Nelson (London: University of London School of Advanced Study, 2016), 248–49. For the statue and its removal by Charlemagne: Agnellus, LPRa, c. 94, 259–60. 203 Walahfrid Strabo, De imagine Tetrici, ll. 30–37, 42–45, in Michael Herren, ed. and trans., ‘The “De imagine Tetrici” of Walahfrid Strabo: Edition and Translation’, 123: TETRICVS, ITALICIS QVONDAM REGNATOR IN ORIS, | Multis ex opibus tantum sibi seruat auarus, | At secum infelix piceo spatiatur Auerno, | Cui nihil in mundo, nisi uix fama arida restat. | Quamquam thermarum uulgus uada praeperat olii; | Hoc sine nec causa, nam omni maledicitur ore, | Blasphemumque dei ipsius sententia mundi | Ignibus aeternis magnaeque addicit abysso. […] Curribus atque in equis noris si stare superbos, | Non quod sedit equo, tecum miraberis umquam. Trans.: Herren, 132. As Herren notes, the name ‘Tetricus’ plays on the name Theoderic, bringing with it the association of darkness and cruelty of the adjective t(a)etricus: Ibid., 120. 204 Walahfrid Strabo, De imagine Tetrici, ll. 215–17, 129: Cedant magna tui, super est, figmenta colossi | Roma: uelit Caesar magnus, migrabit ad arces | Francorum, quodqumque miser conflauerit orbis. Trans. Herren, 137–38. 58 The same is true of other Carolingian palaces known in archaeology. Ingelheim was executed according to radically different design to Aachen, but a similarly creative attitude towards what appear to us to be distinctively ‘late ancient’ forms (Figs. 45–46).205 Arriving at Ingelheim, the visitor was confronted by the convex side of a wide exedra, probably constructed over two storeys, punctuated by a gatehouse at its centre.206 The so-called Halbkreisbau immediately recalls the embracing hemicycle forecourt of the fourth-century villas of Montmaurin (Haute-Garonne) or the semi-circular cryptopor- ticus at Cercadilla around which the representational suites were organised (Figs. 3 and 47), but inverts it to create a closed, ‘D’-shaped interior space. Sidonius Apollinaris describes a similar ‘sigmatic’ court- yard in his ekphrasis of the villa of Pontius Leontius, which he playfully terms a burgus (literally, ‘for- tress’): Behind this rises up, and joins together the doubled house, a portico—itself doubled— unknown to the doubled Plough.207 On one of these, having been gently drawn away by the arc, the curved wings look back somewhat with their facing corners. The right side spies the dawn, its front sees the noon, the left side will see the sunset. It does not lose anything from these three quarters of the sky and preserves the whole sun throughout the lunate halls.208 Ingelheim’s Halbkreisbau thus serves as an architectural equivalent of the paintings which apparently adorned an interior space of the palace according to Ermoldus Nigellus, which depicted a familiar cast of ‘classical’ figures, including Romulus and Remus, Hannibal, Alexander, and the emperors 205 Generally on Ingelheim, see: Christian Rauch, Die Ausgrabungen in der Königspfalz Ingelheim 1909–1914, ed. Hans Jörg Jacobi, Studien zur Königspfalz Ingelheim 1 (Mainz: Verlag des Römisch-Germanischen Zentralmuseums, 1976); Walter Sage, ‘Die Ausgrabungen in der Pfalz zu Ingelheim am Rhein 1960–1970’, Francia 4 (1976): 141–60; Holger Grewe, ‘Die Königspfalz zu Ingelheim am Rhein’, in 799. Kunst und Kultur der Karolingerzeit. Karl der Große und Papst Leo III. in Paderborn, ed. Christoph Stiegemann and Matthias Wemhoff, vol. 3 (Mainz: Phillipp von Zabern, 1999), 142–51. 206 Sage, ‘Ausgrabungen’, 153–59. The gatehouse at the exedra’s centre dates from the central Middle Ages, but seems to incorporate Carolingian fabric: Lobbedey, ‘Carolingian Royal Palaces’, 141. 207 I.e. the constellations of Ursa Major and Minor. Sidonius means that these constellations cannot ‘see’ into the double (two-storey) portico as it faces away from them. 208 Sidonius Apollinaris, Carmina XXII.150–55, in Sidonius Apollinaris, Carm. 22: Bvrgvs Pontii Leontii, ed. Norbert Delhey, Untersuchungen zur antiken Literatur und Geschichte 40 (Berlin: de Gruyter, 1993), 41–42: Haec post assurgit duplicemque superuenit aedem | porticus ipsa duplex, duplici non cognita plaustro; | quarum unam molli subductam uertice curuae | obuersis paulum respectant cornibus alae. | ipsa diem natum cernit sinuamine dextro, | fronte uidens medium, laeuo uisura cadentem. | Non perdit quicquam trino de cardine caeli |et totum solem lunata per atria seruat. Delhey’s commentary: ibid., 142–50. 59 Constantine and Theodosius, culminating in the Carolingian family.209 At Ingelheim, as at Aachen, the architecture of the palace was a medium of (re)interpretation of the ancient past in the formation of a new image of royal authority. *** In all these examples—from North Africa to Francia, from the fifth century to the ninth—we have seen palaces function as a focus for images of ‘Roman’ authority. The notion of ‘Rome’ to which they appealed was not constant or universal, however, but contextual, interpretive, and performative. In the fifth and sixth centuries, essential continuities in the language of power legitimised new regimes in the eyes of domestic elite audiences, but simultaneously served as a framework for rejecting the au- thority of the Eastern Roman Empire. We saw this same architectural hermeneutic in evidence in the palaces of the Carolingians. While the architecture of Aachen and Ingelheim have been claimed to constitute a revival of architectural romanitas, the ideal of ‘Rome’ to which they referred has drastically changed, a materialisation of the shifting semantic boundaries of ‘Roman-ness’ in the Early Middle Ages.210 The palaces of the Carolingian period drew from a heterogeneous corpus of signs, images, and models (including Theoderic!) to create new ‘imaginings’ of an imperial ruler. What marks out Aachen and Ingelheim as unique, however, are not just the evocations of romanitas, but that they take place against the background of a drastically changed field of architectural practice – the theme to which we turn next. IV. Architectures of Authority in an Age of Transformation While palaces evidently provided opportunities for regimes to portray themselves in the guise of Ro- man authority, the acceleration of formal and material change after the sixth century is plain to see, 209 Described by Ermoldus Nigellus in his panegyric delivered to Louis the Pious: Ermoldus Nigellus, In honorem Hludowici, IV.243–74, in Poetae latini aevi Carolini, ed. Ernst Dümmler, vol. 2, MGH. Poetae 2 (Berlin: Weidmann, 1884), 65–66. 210 Julia Smith, Europe after Rome: A New Cultural History, 500–1000 (Oxford: Oxford University Press, 2005), 253–92. 60 even in the structures already surveyed. Between the mid-sixth and the late-eighth centuries we hear often of early medieval rulers ruling from palaces and even, at times, adding new buildings to them – but in most cases material evidence is essentially non-existent.211 By replacing palaces into the wider context of elite secular architecture, however, it is possible to discern at least the outlines of their architectural evolution. Viewed in this way, the palace of Theoderic stands out as the last extant palace in the Western Mediterranean to reproduce the spatial grammar of the late-imperial sedes. Thereafter the forms of architectural display were reinvented in the context of a new political environment. Onward and upward: transformations in secular architecture Ravenna is a good place to start. As we saw above, the palace at Ravenna is rooted in the late-Roman tradition for elite, domestic architecture, with a suite of representational spaces (namely the apsidal aula and the triconchos) arrayed around a peristyle courtyard and executed in forms familiar from nu- merous other sites. At the same time, however, the palace of Ravenna points us towards new trends. Firstly, the suggestion of a clerestory in the palatium mosaic implies the presence of an upper storey, and Agnellus’ description locates the triclinium ‘By the Sea’ in an elevated location (Fig. 48). Secondly, the close integration of the palace church, Sant’Apollinare Nuovo, into the palace’s external aspect on the platea maior represents a new departure in palace design – a theme to which we will return shortly. In contrast to Ravenna, later palaces, such as those constructed in the Carolingian period, have at their core a distinct layout. From the later-sixth century onwards, the central representational spaces are increasingly concentrated in a single ‘box-like’ structure, built over two storeys, without significant differentiation between interior spaces. In materials too, much has changed: rather than brick and 211 See especially the palaces of Pavia and Toledo. For an addition to the palace of Pavia: E.g. Paul. Diac., Hist. Lang., 156: His diebus rex Perctarit in civitate Ticinensi portam contiguam palatio, quae et Palatensis dicitur, opere mirifico construxit. (‘In those days, king Perctarit [661–62, 671–688] built in the city of Ticinum a gate adjoining the palace with marvellous workmanship, which is called the ‘Palatine Gate’). On Pavia see further, below, Chapter 3. 61 concrete, we see in the main timber, quoined rubble, and ashlar.212 Though separated by less than two centuries, these structures seem worlds apart. These architectural changes were not exogenous to the evolutionary currents of Roman archi- tecture. Roman houses traditionally had their reception areas arrayed around the courtyard at ground level, but from the fifth century there was a gradual shift of their most important spaces onto an upper storey.213 Where they can be detected in elite domus of earlier periods, upper floors most often had the character of service or domestic spaces; the representational suites were principally located at ground level. From the fifth century, however, the traditional layout of the Roman domus began to break down more fundamentally. The villa at San Giovanni di Ruoti (Basilicata) in its late-fifth century phase, for example, lacked a courtyard and was instead focused on a large, apsidal aula on the complex’s northern side (Fig. 49).214 The northern range, encompassing the aula and its subsidiary spaces was elevated on a second storey, with the lower level being cellars or service areas, a development we saw already in the palatial area at Reccopolis.215 At the north-eastern corner is a space out of alignment with the other walls, accessible by an adjacent staircase, that the excavators interpreted as a tower.216 The general transformations of elite architecture can also be seen clearly in bishops’ palaces, episcopia, a number of which are preserved from the fifth and sixth centuries.217 In the episcopium of Ravenna, for example, bishop Neon (ca. 450–473) added a luxurious dining room known as the ‘Five Couches’ (Quinque Agubitas), the five apses of which, judging from Agnellus’ description, were ornately 212 On the speed and consistency of the formal and material transition: Chris Wickham, Framing the Early Middle Ages: Europe and the Mediterranean, 400–800 (Oxford: Oxford University Press, 2005), 201–02. 213 Barbara Polci, ‘Some Aspects of the Transformation of the Roman Domus between Late Antiquity and the Early Middle Ages’, in Theory and Practice in Late Antique Archaeology, ed. Luke Lavan and William Bowden, Late Antique Archaeology 1 (Leiden: Brill, 2003), 89–105. 214 Alastair Small and Robert Buck, The Excavations of San Giovanni di Ruoti, vol. 1, Phoenix Supplementa 33 (Toronto: University of Toronto Press, 1994), 75–121. An earlier first-century CE villa was demolished and reconstructed ca. 400 CE (Phase 3A); this was followed by a second major construction phase in the third quarter of the fifth century after a landslip (Phase 3B) which significantly enlarged the villa’s area. For a summary: Sfameni, Ville residenziali, 215–20. 215 Small and Buck, Excavations, 92–93. 216 Ibid., 94–95. 217 Polci, ‘Aspects’, 90. For episcopia: Maureen Miller, The Bishop’s Palace: Architecture and Authority in Medieval Italy, Conjunc- tions of Religion and Power in the Medieval Past (Ithaca, NY: Cornell University Press, 2000). 62 decorated in figural mosaics, biblical verses, and hexametric inscriptions.218 Based on Agnellus’ remark that the room was located ‘over the stream’ (super amnem), it is likely that the triclinium was elevated to overlook the Fossa Amnis, one of the waterways that ran through the city.219 Later, Bishop Peter II (494–520) added a three-storey building to the Salustra Tower, originally part of a gate in the walls of the colonia, on the top floor of which is located the episcopal chapel.220 An elevated reception hall is also still extant at Poreč, across the Adriatic but with close artistic ties to Ravenna (Fig. 50).221 Con- temporaneously with Theoderic’s triclinium Ad mare, therefore, the bishops were also moving the rep- resentational spaces of their residences onto upper floors. While, of course, specific constraints no doubt influenced the construction of episcopia, not least limitations of space in the vicinity of urban cathedral churches, there is no reason to assume that they differed in kind from contemporary royal palaces – especially given that bishops were, in the fifth and sixth centuries, increasingly taking on roles as alternative power brokers in the late ancient city. Replacing palaces into a broader architectural field provides a framework for describing the general character of their transformation through the later-sixth and seventh centuries, especially north of the Alps where material evidence for palaces is especially thin. The palace at Malay (Masolaco) is attested textually both by Fredegar and in late-seventh century charters, and structures discovered in 1986 outside Malay-le-Grand (Yonne) remain the most likely candidates to be identified as an exca- vated Merovingian palace.222 Discovered in excavations undertaken in the 1980s were the foundations 218 Agnellus, LPRa, c. 29, 175–77. 219 Ibid., 176. In a later passage (Ibid., c. 163, 341), Agnellus describes Archbishop John VI taking a meal in a dining hall ‘behind the apse of the church above the uiuarium’ – a location that overlooks the Fossa: Clementina Rizzardi, ‘Note sull’an- tico episcopio di Ravenna: formazione e sviluppo’, in Actes du XIe congrès international d'archéologie chrétienne. Lyon, Vienne, Grenoble, Genève, Aoste (21-28 septembre 1986), ed. Noël Duval, vol. 2 (Rome: École Française de Rome, 1989), 719. See further: Miller, Bishop’s Palace, 23–27 220 Miller, The Bishop’s Palace, 27–31. 221 Ivan Matejčić and Pascale Chevalier, ‘The Episcopium of Poreč, in Des domus ecclesiae aux palais épiscopaux, Actes du colloque tenu à Autun du 26 au 28 novembre 2009, ed. Sylvie Balcon-Berry et al., Bibliothéque de l'antiquité tardive 23 (Turnhout: Brepols, 2012), 151–60. 222 Fred. Chron., cc. 44, 79, pp. 143, 161; DD Merov. 1, nos., 121, 122; Dider Perrugot, ‘Le palais mérovingien de Malay (Yonne): histoire et archéologie’, in Palais royaux et princiers au Moyen Âge : actes du colloque international tenu au Mans les 6–7 et 8 october 1994, ed. Annie Renoux (Le Mans : Publications de l'Université du Maine, 1996), 149–50. Weise’s excavations of 63 of a sturdy, box-like structure standing on the bank of the ancient course of the river Vanne which the excavators interpreted as the Merovingian palace.223 The width of the foundations indicate a stone construction which probably supported an upper storey, possibly preceded by a portico or gallery.224 The excavations were not extensive enough, however, to reveal the whole structure nor to illuminate its wider context. The features found at Malay do, however, accord well with the better understood site at Camp de Larina, to the east of Lyon (Figs. 51–52). This complex of structures, including a sizeable stone hall with further outbuildings (perhaps a stable and a small chapel), underwent numer- ous phases of expansion in the Merovingian period.225 The general impression made by these buildings is one similar to the two-storey residence uncovered on the Forum Transitorium in Rome, dating from the early-ninth century, which was built from blocks of peperino tufa and preceded by a portico or porch (Figs. 53).226 This (admittedly sparse) picture can be augmented with textual sources. While contemporary observers like Gregory of Tours do not provide nearly as much detail about royal palaces as we would like, they do record important details about other structures, especially episcopia. Gregory, for example, narrates a story centred on the expansion of the episcopium of Lyon.227 Priscus, the bishop of Lyon, ordered for the episcopium to be ‘raised’ (exaltari), a phrase that, though ambiguous, probably denotes that the structure was ‘elevated’ through the addition of an upper storey.228 Following its completion, a deacon climbed onto the roof and sought to prise off its tiles, cursing the name of Priscus’ saintly the palace of Quierzy and the Carolingian palace of Samoussy, executed in the wake of the German advance into France during the First World War, do not inspire confidence: Georg Weise, Zwei fränkische Königspfalzen: Bericht über die an den Pfalzen zu Quierzy und Samoussy vorgenommenen Grabungen (Tübingen : Fischer, 1923). 223 Perrugot, ‘Le palais mérovingien’, 152. 224 Ibid., 153. Post-holes to the north attest to the presence of subsidiary structures in wood. 225 Ross Samson, ‘The Merovingian Nobleman’s Home: Castle or Villa?’, Journal of Medieval History 13, no. 4 (1987): 301. 226 Riccardo Santangeli Valenzani, ‘Residential Building in Early Medieval Rome’, in Early Medieval Rome and the Christian West: Essays in Honour of Donald Bullough, ed. Julia Smith, The Medieval Mediterranean 28 (Leiden: Brill, 2000), 101–12. 227 For the architectural consequences of this anecdote: Ross Samson, ‘The Residences of the Potentiores in Gaul and Ger- mania in the Fifth to Mid-Ninth Century’ (PhD Dissertation, Glasgow, University of Glasgow, 1991), 148. 228 Greg. Tur., Hist., IV.36, 169: Iusserat enim in primordio episcopatus sui aedificium domus ecclesiasticae exaltari. 64 predecessor and rival as he did so. Possibly unsurprisingly, the roof then collapses and the deacon perishes.229 While innocuous in itself, this episode suggests that the episcopium was probably built of stone and covered by a pitched, tiled roof supported on a timber frame. Certainly, the episcopium of Geneva was built of multiple storeys possibly as early as the late-fourth century and certainly in the fifth.230 What little ecclesiastical architecture remains extant, for example Saint-Pierre in Vienne or the baptistery of Saint-Jean at Poitiers, highlights that Merovingian builders could work creatively with late ancient construction techniques.231 In a broader sense, the materials used in construction were evidently mixed, including wood, masonry, and spolia. Venantius Fortunatus, for example, praises a two-storey ‘palace’ completely out of wood: Go back to Paros, walls built with materials of stone; because of an artist’s skill I prefer wood to you. A many-storied palace strikes the heavens with its size; no crack appears, for handicraft has made it solid. Whatever roles stones, sand, lime, and clay perform, that whole building a single forest joyfully constructed. It is higher in the middle, sur- rounded by a square colonnade, and ornamented by a craftsman’s skill in carving.232 While there is a hint of humour in Fortunatus’ words, which should alert us to what he describes not being usual, it nonetheless betrays that timber constructions were not necessarily valued less than work in stone and, moreover, were used in the development of an independent architectural idiom. The broader architectural context suggests we are dealing with relatively simple structures built predominantly of stone (though conceivably with wood annexes or outbuildings) over two storeys. We can see this new arrangement clearly in the explosion of palace construction in the late-eighth and ninth centuries as a new political order came into focus. In the first phase of construction of 229 Ibid.: hic ascendens super tectum domus illius, cum detegere coepisset. 230 Charles Bonnet and Alain Peillex, Les fouilles de la cathédrale Saint-Pierre de Genève. Les édifices chrétiens et le groupe épiscopal, Mémoires et documents 65 (Geneva: Société d'histoire et d'archéologie de Genève, 2012), 55, 79. A new residence was constructed in the sixth century also over two-storeys, accessed by a double staircase: ibid., 114. 231 McClendon, Origins, 42–49. 232 Ven. Fort., Carm., IX.15, 3:36–37: Cede Parum, paries lapidoso structe metallo: | artificis merito praefero ligna tibi. | aethera mole sua tabulata palatia pulsant, | quo neque rima patet consolidante manu. | quidquid saxa, sablo, calces, argila tuentur, | singula silva favens aedificavit opus. | altior inmitior quadrataque porticus ambit | et sculpturata lusit in arte faber. Translation: Venantius Fortunatus, Poems, trans. Michael Roberts, Dumbarton Oaks Medieval Library 46 (Cambridge, MA: Harvard University Press, 2017), 599–601. 65 Charlemagne’s palace at Paderborn, for example, the hall was built over two storeys and may have had a gallery or balcony on its front, possibly for the king to appear before his followers (Fig. 54).233 At Frankfurt too, we can faintly discern the connection of a two-storey hall, supported on sturdy square pillars, connected to a church by a gallery or porticus, a topographical arrangement we saw already in Aachen (Fig. 55).234 Excavations led by Michäel Wyss at Saint-Denis have demonstrated the presence of a Carolingian residence alongside the abbey church, similarly built over two storeys and presenting a long, ‘gallery’ façade.235 The coalescence of this new formal arrangement was not limited to the Carolingian world. In Salerno, for example, southern Italian sources tell us that the princeps of Benevento, Arechis II, had built a new palace to which he retreated in the face of Charlemagne’s southward advance in 787.236 While the overall structure of Arechis’ palace is all-but-lost, traces incorporated into later structures provide a general sense of its dimensions and design (Fig. 56). A portico of Roman spolia supporting an upper-storey loggia can be discerned in the later Palazzo Pinto (Figs. 57–59), at the intersection of the Via Dogana Vecchia and the Via Arechi.237 The overall impression is of a compact complex of buildings over two or more storeys, built ex novo with the use of spolia, and straddling the walls at their 233 Manfred Balzer, ‘Paderborn als karolingischer Pfalzort’, in Deutsche Königspfalzen 3, 9–85 (esp. 37–66); Sveva Gai, ‘Die Pfalz Karls des Großen in Paderborn: Ihre Entwicklung von 777 bis zum Ende des 10. Jahrhunderts’, in 799. Kunst und Kultur, 3:183–96. 234 Otto Stamm, ‘Zur karolingischen Königspfalz in Frankfurt am Main’, Germania 33 (1955): 391–401; Andrea Hampel and Heinz Schomann, Der Kaiserdom zu Frankfurt am Main: Ausgrabungen 1991–93, Beiträge zum Denkmalschutz in Frank- furt am Main 8 (Nussloch: R. Angerer, 1994). 235 Michäel Wyss, ‘Saint-Denis’, in 799. Kunst und Kultur, 3:138–41; idem, ‘Die Klosterpfalz Saint-Denis im Licht der neuen Ausgrabungen’, in Splendor palatii. Neue Forschungen zu Paderborn und anderen Pfalzen der Karolingerzeit, ed. Lutz Fenske, Deut- sche Königspfalzen 5 (Göttingen: Vandenhoeck & Ruprecht, 2001), 175–92. 236 Erchempert, Ystoriola Langobardorum beneventum degentium, III, in Erchempert, Piccola storia dei Longobardi di Benevento, ed. and Italian trans. Luigi Andrea Berto, Nuovo Medioevo 94 (Naples: Liguori, 2013), 86; Chronicon Salernitanum, X, in Chro- nicon Salernitanum: A Critical Edition with Studies on Literary and Historical Sources and on Langauge, ed. Ulla Westerbergh, Studia Latina Stockholmiensia 3 (Stockholm: Almsqvist & Wiksell, 1956), 13. Cf. ARF, a. 787, 74. For the transformations of the late ancient topography of Salerno under Arechis, see: Paolo Peduto, ‘Arechi II e Salerno’, in Tra i Langobardi del Sud: Arechi II e il Ducato di Benevento, ed. Marcello Rotili (Padua: Il Poligrafo, 2017), 355–65. 237 Paolo Peduto, ‘Consuetudine ed evoluzione dell’antico nelle costruzioni di Arechi II’, in Salerno. Una sede ducale della Langobardia meridionale, ed. Paolo Peduto et al., Studi e Ricerche di Archeologia e Storia dell’Arte 16 (Spoleto: CISAM, 2013), 7. 66 western limit.238 Better preserved is the palace church, known as San Pietro a Corte (Figs. 60–61).239 The church is constructed over the frigidarium of a Severan-period Roman bathhouse, its arches provid- ing the substructures for the upper-storey chapel.240 The diminutive church—single-naved with a flat, barrel-vaulted apse—respects the outlines of the Roman structures below and was only originally ac- cessible from the palace’s interior on its upper storey. The exterior façade was originally the part of a unified wall along the palace’s length; behind the eighteenth-century staircase are the traces of a blocked loggia241 The palaces established by the kings of Asturias highlight the final maturation of this new spatial arrangement. The late ninth-century Chronicon Albeldense states that Ramiro I of Asturias (842– 850) ‘constructed a church and vaulted palaces with wonderful skill in the place Ligno [The Wood]’.242 On this site, located on the Monte Naranco overlooking the royal seat of Oviedo, survives a uniquely well-preserved example of secular architecture from this period: one of the palace’s representational spaces, now known as Santa María del Naranco (Figs. 62–63). Perched on the hillside, the shallow ‘T’- shaped building was constructed over two storeys in an irregular yellow ashlar, buttressed by narrow engaged piers.243 At each end, and originally also on the southern side, are open loggias that give spectacular views over the surrounding countryside – surely a guiding motive in its design. An inscrip- tion on an altar in the western loggia, originally belonging to the nearby palace church of San Miguel 238 Paolo Delogu, Mito di una città meridionale: Salerno, secoli VIII–XI, Nuovo Medioevo 2 (Naples: Liguori Editore, 1977), 44. 239 Chronicon Salernitanum, XVII, 22: hanc civitatem undique munivit atque in ea mire magnitudinis immo et pulcritudinis palaccium construxit, et ibidem in aquilonis parte ecclesiam in honorem beatorum Petri et Pauli instituit. (‘He fortified this city on all sides and in it he constructed a palace of wondrous size and beauty, and there in the northern part he founded a church dedicated to Saints Peter and Paul’) 240 Paolo Peduto et al., ‘Un accesso alla storia di Salerno: stratigrafie e materiali dell’area palaziale longobarda’, Rassegna Storica Salernitana (new series) 10 (1988): 9–63. 241 Delogu, Mito, 45. 242 Chronicon Albeldensis XV.10, in Cronicas asturianas, ed. Juan Gil Fernández, Spanish trans. Jose L. Moralejo, Publicaciones del Departamento de Historia Medieval, Universidad de Oviedo 11 (Oviedo: Universidad de Oviedo, 1985), 175: In locum Ligno ecclesiam et palatia arte fornicea mire construxit. 243 Generally, see: Achim Arbeiter and Sabine Noack-Haley, Christliche Denkmäler des frühen Mittelalters vom 8. bis ins 11. Jahrhundert, Hispania Antiqua (Mainz: Phillipp von Zabern, 1999), 158–65, with references to further literature. 67 de Lillo, provides 23 June 848 as its date of dedication; we should expect that the belvedere was a part of the same project.244 The palace evidently made an impression on contemporaries. The vaulting caught the attention of the Albelda chronicler, and the Chronicle of Alfonso III remarks that Ramiro ‘built many arched buildings of granite and marble, without using any wood’, emphasising the palace’s ma- terial wealth as well as the skill of Ramiro’s masons for not using wooden centring.245 A number of factors strongly suggest its originally secular function. The uninterrupted, transversal main space with annexed loggias would be highly unusual in ecclesiastical architecture, and the eclectic sculptural or- namentation, including animal motifs set in roundels and confronted horsemen in combat, would suit a secular context most comfortably (Fig. 64). Moreover, the interpenetration of interior and exterior space might suggest the influence of the villa-estates (munya, pl. munyāt) of the Ummayad emir and aristocracy in the vicinity of Córdoba.246 Multiple lines of inspiration converge in the design of the Naranco belvedere, but underpinning it is a new general conception of secular architecture. In contrast to the Roman peristyle domus, early medieval residences typically had their representational spaces located on an upper storey, a spatial arrangement akin to the piano nobile of the later Italian palazzo. The inflection point of this transition cannot be pinpointed exactly. We have seen that the roots of this trend lay already in the fifth century, but the later-sixth or seventh century seems likely as the watershed after which secular architecture, including palaces, took on a new aspect. Alongside this typological shift, there was a concurrent tran- sition in materials of construction. Already in the sixth century, secular architectures were being 244 The altar was probably transferred to the belvedere after the church fell into disrepair: Arbeiter and Noack-Haley, Christ- liche Denkmäler, 142. 245 Chronica Adefonsi tertii regis (Rotensis), c. 24, in Cronicas asturianas, 144: multa edificia ex murice et marmore sine lignis opere forniceo in latere montis Naurantii duo tantum miliaris procul ab Oueto edificauit. Trans.: The Chronicle of Alfonso III, in Conquerors and Chron- iclers of Early Medieval Spain, trans. Kenneth Baxter Wolf, 2nd ed., TTH 9 (Liverpool: Liverpool University Press, 1990), 141. 246 Rose Walker, Art in Spain and Portugal from the Romans to the Early Middle Ages: Routes and Myths (Kalamazoo, MI: Medieval Institute Publications, 2016), 163. On munya architecture, see especially: Glaire Anderson, The Islamic Villa in Early Medieval Iberia (Farnham: Ashgate, 2013), especially e.g. 50–69 (on the tenth-century site al-Rumāniyya’s staged views over the valley and its terraced gardens, and its relation to earlier Roman and medieval architecture). 68 constructed from, on the face of it, humbler materials: rubble and fill, spolia, and timber, rather than newly-produced brick and concrete. While this transition took place at different rates in different areas, by the seventh century royal palaces were not just formally, but also materially distinct from their Roman predecessors. The ‘great hall’ in early medieval Britain: an alternative model? The fifth century saw a sudden and dramatic collapse of the material markers of Roman provincial life in the British Isles. At the same time, the British provinces swiftly dissolved into a patchwork of fissiparous, small-scale units that only began to approach the intensity of their continental neighbours from the later-eighth century onwards.247 For much of the period under discussion, distinctively ‘royal’ activities in the landscape are hard to detect; the lines dividing a royal estate from an aristocratic set- tlement or a monastery are slim indeed.248 Identifying ‘palaces’ in a society with such an ephemeral material culture is a fraught exercise—John Blair recently disavowed it entirely as ‘anachronistic’—but the architectures of early medieval Britain nonetheless provide an important counterpoint to the over- all narrative advanced in this chapter so far.249 Given the distinct idiom of secular architecture in the British Isles in general, and the phe- nomenon of the ‘Great Hall’ in particular, the notion of ‘Germanic’ migrants bringing with them a distinct visual and material culture looms ever in the background.250 Certainly, the lord’s hall, most famously Hrothgar’s Heorot in Beowulf, forms the setting for ‘courtly’ secular display and elite compe- tition in the Anglo-Saxon poetic corpus—a theme examined in the next chapter—and it seems to have figured as a discrete conceptual unit in the Anglo-Saxon imaginary as a site of political action.251 247 Simon Esmonde Cleary, The Ending of Roman Britain (London: Batsford, 1989). Wickham, Framing, 306–39, 345–51; Gerrard, Ruin. 248 John Blair, ‘Palaces or Minsters? Northampton and Cheddar Reconsidered’, Anglo Saxon England 25 (1996): 97–121. 249 John Blair, Building Anglo-Saxon England (Princeton, NJ: Princeton University Press, 2018), 104. 250 E.g. McClendon, Origins, 59. 251 Barbara Raw, ‘Royal Power and Royal Symbols in Beowulf’, in The Age of Sutton Hoo: the Seventh Century in North-Western Europe, ed. Martin Carver (Woodbridge: Boydell, 1992), 167–74; Heidi Stoner, ‘Heaven and Hall: Space and Place in Anglo- 69 In this sense, therefore, we can examine the halls of the early English kingdoms in terms of the same politics of secular display as their continental counterparts. Their distinctive material characteristics speak more to the limited resources available to the rulers of such small-scale polities than any distinct cultural tradition. The seventh century stands out for, in John Blair’s words, a ‘short but spectacular episode of secular architectural display’ in the context of the coalescence of new political units.252 Eighteen sites located in the eastern part of Britain from Lothian to Kent evince a strikingly uniform set of architectural groups all focused around a large timber hall.253 Of these, the best known is Yeavering, located on the edge of the Cheviot Hills in Northumbria, northern England. Excavated by Brian Hope-Taylor between 1953 and 1962, scholarly consensus holds that Yeavering is the site of the ‘royal villa’ of the Northumbrian king Edwin (616–632/3) named Ad Gefrin in Bede’s Ecclesiastical History of the English People.254 The estate would not remain a permanent fixture in the landscape, how- ever, since (Bede continues), ‘this palace [uilla] was left deserted in the time of the kings who followed Edwin, and another was built instead in a place called Mælmin’, probably the nearby site of Milfield.255 On Yeavering Bell were discovered the traces of a succession of large complex of timber buildings adjacent to a palisaded mound enclosure and other earthwork features (Fig. 65). At the centre of the site stood a large timber hall with subsidiary structures, built and rebuilt over successive phases.256 The largest of the halls, some twenty-four metres in length, was constructed with rectangular timbers set in deep trenches and buttressed by a fan of posts set at an angle to them.257 Hope-Taylor emphasised Saxon England’, in Place and Space in the Medieval World, ed. Meg Boulton et al., Routledge Research in Art History (New York: Routledge, 2018), 159–77. 252 Blair, Building, 104. 253 Ibid., 114–25. 254 Brian Hope-Taylor, Yeavering: An Anglo-British Centre of Early Northumbria, Department of the Environment Archaeo- logical Reports 7 (London: HM Stationery Office, 1977), 28–45. Bede, Histora ecclesiastica gentis Anglorum, II.14, in Bede, The Ecclesiastical History of the English People, ed. and trans. Bertram Colgrave and R. A. B. Mynors, Oxford Medieval Texts (Oxford: Clarendon Press, 1969), 188–189: in uillam regiam quae uocatur Adgefrin. 255 Bede, Historia ecclesiastica, II.14, 188–89: Haec uilla tempore sequentium regum deserta, et alia pro illa est facta in loco qui uocatur Mælmin. 256 Summary of the excavated features in the central area: Hope-Taylor, Yeavering, 46–69. 257 Building A4 in the excavation report: ibid., 58–62. It measures 80ft in length. 70 that the depth of the trenches showed that the walls were intended to be load-bearing and to ‘induce both the structural properties and the appearance of stone buildings’ as a display of romanitas.258 John Blair, on the other hand, has emphasised the impermanence of elite sites generally in this period and notes instead that the combination of a high, pitched roof gabled over low walls and supporting by projecting buttress-posts may have more resembled ‘tents-in-timber’.259 References to an idealised ‘Rome’ are not absent, however. To the west, at the centre of the hill, stood a structure akin in form to a section of a Roman theatre with, at its base, a podium or platform for a speaker (Fig. 66).260 After an expansion it has been estimated that this structure could seat as many as 320 people.261 Mysteriously, a grave sited prominently at the eastern door of the large hall (and thus at the eastern end of this axis) contained the traces of what has been plausibly explained as a groma, a Roman surveying tool.262 Something of the halls’ excessive quality is conveyed in the dissonance between their high visibility in the landscape but their fundamental (and seemingly deliberate) impermanence. The meagre resources—capital, materiel, labour—available to seventh-century kings inevitably had a profound ef- fect on architectural forms, but this does not mean that the social hierarchies performed within were simplistic or necessarily of a different order to those on the continent. Early English builders evidently did work in stone, particularly when it was available from Roman sites. We need look no further than the robust forms of St Mary’s, Brixworth or St Martin’s, Canterbury (both seventh-century in date) to suspect that Yeavering’s impermanence was, in fact, the point. As Blair has characterised it, royal 258 Ibid., 129, 140. Hope-Taylor also briefly explores an analogy of their floor plan with Roman ‘basilican’-plan structures, concluding that ‘both the native [i.e. British] and intrusive [i.e. Anglo-Saxon] ruling classes knew and envied the strength and solidity of Roman masonry buildings, which had long been associated with the exercise of power in the land’: ibid., 233–34. 259 Blair, Building, 123. 260 Hope-Taylor, Yeavering, 119–22. 261 Ibid., 161. 262 Ibid., 202. 71 display in the seventh-century British landscape was ‘heavy in treasure and display, light in permanent built installations and capital investment, but flexible and adaptive in its spatial organisation’.263 *** Two principal transformations in palace architecture have been identified between the fourth and eighth centuries. In the most general terms, we have witnessed the transition from one architectural idiom to another, a cultural shift remarkable in both its geographic extent and the velocity of change after the sixth century. While the material record for palaces is interrupted at the crucial juncture of the late-sixth/seventh century, by replacing palaces into the broader context of secular architecture, it is possible to distantly perceive the coalescence of a new concept of architectural space. Whereas elite Roman residences were predominantly built with the most important representational spaces on the ground floor, by eighth century palaces were designed around a single, rectilinear main building, often built over two storeys. Materially, too, we have seen a drastic change in architectural practice. Produc- tion of the materials characteristic of Roman urbanism, such as bricks and concrete, rapidly ceased following the breakdown of the Roman Mediterranean economy. In their place, timber, ashlar, and rough rubble-fill construction techniques become more prominent. There is no reason to suppose, however, that early medieval people understood these techniques to be any less ‘grand’ than Roman architecture. At the extreme of this trend, the exclusively timber-wrought halls of Yeavering were both keyed to the spatial particularities of Northumbrian political culture and still offered opportunities for reimaginations of the Roman past. The political-economic upheavals of the late-fifth century onwards were in large part the cause of these changes. Simply put, from the later-sixth century onwards it was no longer possible to make grand architectural statements on the scale of the later Roman emperors.264 Construction materials 263 Blair, Building, 103. 264 See Wickham’s description of the transition from the tax-based Roman Empire to the ‘weak states’ of the post-Roman regna: Wickham, Framing, 56–124. 72 bring together a range of relationships of production; in a sense, they describe a regime’s ability to mobilise labour and resources. A case in point is the production of bricks. Building in brick is labour- intensive and requires a significant infrastructural network to facilitate on any kind of large scale. To take Rome as an example, Cassiodorus’ Variae suggest that brick production had at least faltered al- ready in first half of the sixth century.265 As Riccardo Santangeli Valenzani has shown, the ‘decline’ of construction techniques in Rome’s buildings between the fourth and eighth centuries indexes chang- ing relations of production and levels of social stratification.266 At the same time, it is important not to be overly deterministic. It was patently not the case that early medieval builders strove towards a static idea of ancient grandeur that was beyond their grasp. We must, therefore, also look to cultural explanations for the instantiation of this new formal grammar. The fifth century set off a wide-ranging reorientation of the norms of, and lexicon for, exhibiting social status and political power, one hardly limited to architecture. While the visual styles of, for example, ‘barbarian’ metalwork might ultimately derive from Roman prototypes, they signalled the wearer’s relative status in a transformed semantic field.267 In this context, we would be mistaken to assume that formal qualities such as elevation, a ‘sturdy’ compact plan, and juxtapositions of spolia and new fabric were not desirable aesthetic phe- nomena in themselves, rather than simple ‘degradations’ from a pristine Antiquity. V. The Palace Church: Its Genesis and Architectural Significance In 812, an embassy from the Byzantine emperor Michael I (811–813) arrived at Aachen in the context of ongoing negotiations between the two courts. Once admitted to Charlemagne’s presence, the en- voys ‘received from him [Charlemagne] in church the document of the treaty, acclaimed him according 265 Cass., Var., I.25, 28 (to Sabinanus, vir spectabilis, ordering that the riverfront brick kilns be returned to their former purpose in order to provide tiles for the city walls). 266 Riccardo Santangeli Valenzani, ‘Public and Private Building Activity in Late Antique Rome’, in Technology in Transition, A.D. 300–650, ed. Luke Lavan et al., Late Antique Archaeology 4 (Leiden: Brill, 2007), 435–49. 267 See e.g.: Guy Halsall, Barbarian Migrations and the Roman West, 376–568, Cambridge Medieval Textbooks (Cambridge: Cambridge University Press, 2007), 482–88, 494–98. 73 to their custom, that is, in Greek, and called him “Emperor” and “Basileus”’.268 The Annales regni Francorum’s record of this event clearly marks it as significant in the regime’s propaganda. It was, in the annalist’s eyes, a de facto recognition of Charlemagne’s imperial title by his Constantinopolitan counterpart.269 In addition to what it says about Carolingian political culture, this episode also betrays a significant transition in the spatial ideologies of the palace: the transformation of the palace church into an arena for political ritual. The palace chapel would, as representational space, become a defining element of royal and aristocratic residence through the European Middle Ages and beyond.270 Remarkably, however, the architectural and conceptual genesis of the palace chapel has elicited little comment in scholarship. Josef Fleckenstein, in his classic institutional history of the German Hofkapelle, argued that the institution of the capella regis was essentially an invention of the Carolingian period.271 By prioritising the chapel’s institutional formation, however, he also dismisses its spatial/ar- chitectural manifestation as a significant theme. He notes that, for example, the capella could take the form of a simple hut, or even a tent, if required and need not have been housed in a recognisably ‘ecclesiastical’ structure.272 Fleckenstein is of course correct to observe that chapels do not conform to a single architectural typology, but he overlooks the fact that the conjunction of residence and religious space is itself a phenomenon in need of explanation. 268 ARF, a. 812, 136: Nam Aquisgrani, ubi ad imperatorem venerunt, scriptum pacti ab eo in ecclesia suscipientes more suo, id est Greca lingua, laudes ei dixerunt, imperatorem eum et basileum appellantes. Trans.: Bernhard Scholz, trans., Carolingian Chronicles: Royal Frankish Annals and Nithard’s Histories (Ann Arbor, MI: University of Michigan Press, 1970), 94–95. 269 Einhard claims that the Byzantine emperors were suspicious (and envious) of Charlemagne’s power and thus sought an alliance with him, which Charlemagne prudently granted and observed: Einhard, Vita Karoli imperatoris, c. 16, 19–20. 270 There is not yet a general study of palace chapels as an architectural phenomenon, though there have been a number of excellent studies on individual chapels and their use in specific contexts. See e.g.: Martha Mel Stumberg Edmunds, Piety and Politics: Imaging Divine Kingship in Louis XIV’s Chapel at Versailles, Studies in Seventeenth- and Eighteenth-Century Art and Culture (Newark, DE: University of Delaware Press, 2002); Andrea Longhi, ‘Palaces and Palatine Chapels in 15th– Century Italian Dukedoms: Ideas and Experiences’, in A Renaissance Architecture of Power: Princely Palaces in the Italian Quattro- cento, ed. Silvia Beltramo et al., The Medieval Mediterranean 104 (Leiden: Brill, 2015), 82–104; Meredith Cohen, The Sainte- Chapelle and the Construction of Sacral Monarchy: Royal Architecture in Thirteenth-Century Paris (Cambridge: Cambridge University Press, 2015), esp. 113–145. 271 Josef Fleckenstein, Die Hofkapelle der deutschen Könige, vol. 1, MGH. Schriften 16 (Stuttgart: Hiersemann, 1959), e.g. 13– 14, 37–39, 43. 272 Ibid., 95. 74 Here, we will untangle two strands of the palace church’s evolution: evidence for Late Roman archetypes for the palace chapel, and the ‘liturgification’ of authority in the seventh century. Our focus here is solely on the development of the palace chapel as an architectural ‘idea’; its activation in the context of political ritual is treated in the next chapter. By approaching chapels in diachronic perspec- tive, we will see that the assimilation of ‘sacred’ space into the ruler’s residence was neither natural nor inevitable, as has usually been assumed, but triggered by an intensification of the framing of authority in terms of Christian doctrine that took place in the seventh century. Oratoria were introduced into palaces and elite residences already in the Late Empire, but their transformation into ritual space was a later development, one specific to the Early Middle Ages Private oratories and public basilicas in the Later Roman Empire Churches and oratories were associated with imperial palaces from an early date, and these structures are usually presented as the archetypes from which palace chapels of the Early Middle Ages and later typologically derive.273 As a result, palace chapels appear to have been a constant feature since the fourth century, thus making their presence in medieval palaces seem unremarkable. If we consider the late Roman evidence in its own terms, however, it is far more ambiguous. In fact, while ‘private’ oratories—in the sense here of structures that were not important foci of political representation but reserved for the sole use of the patron—were an available concept in late-ancient palaces, they do not seem to have formed an important component of its spatial composition until the mid-fifth and sixth centuries. The evidence for oratories in late Roman palaces is thinner than one might expect. Prior to the translation of Caesarius in the later sixth century, there does not seem to have been a chapel 273 For example: André Grabar, Martyrium : recherches sur le culte des reliques et l'art chrétien antique, vol. 1 (Paris: Collège de France, 1946), 559–79; Fleckenstein, Hofkapelle, 3. 75 incorporated into the imperial residences on the Palatine, for example.274 Structures at its periphery were converted into churches, but their relationship with the palace itself is tenuous at best. Sant’An- astasia al Palatino (the titulus Anastasiae) at the hill’s foot, for example, is of a significantly earlier date than San Cesario: a verse inscription once in the apse, recorded in the Einsiedeln Itinerary, states that a fresco programme executed at the time of Pope Hilarus (461–468) replaced one commissioned by Damasus (361–366).275 Yet factors such as its peripheral location, tucked behind the Circus Maximus, and its status as a titular church make Sant’Anastasia an unlikely candidate for a chapel destined for the exclusive use of the imperial court. A similar story, in fact, applies to the Great Palace in Constan- tinople. Though Constantine and his successors were evidently avid church-builders, oratories within the palace seem to have been a comparatively late development. That Constantine included a ‘church of the Lord’ within the Great Palace is reported only in the tenth-century Patria, for example.276 Seem- ingly more clear-cut is Theophanes’ ninth-century account of the empress Pulcheria depositing relics of the protomartyr Stephen ‘into the palace’ (εἰς το παλάτιον) in 421.277 Yet even this is ambiguous. While an oratory dedicated to Stephen was used in imperial coronations in the seventh century, including that of Heraclius and Eudocia in 610, significant doubts have been raised that this later chapel had been founded by Pulcheria.278 Even if we accept that Pulcheria’s oratory and the shrine (εὐκτηρίον) attested later are one and the same, there is little evidence that chapels were significant spaces within the imperial palace before that time. 274 On Caesarius’ cult, see: Maya Maskarinec, City of Saints: Rebuilding Rome in the Early Middle Ages (Philadelphia, PA: Uni- versity of Pennsylvania Press, 2018), 53–70. 275 Gerold Walser, ed., Die Einsiedler Inschriftensammlung und der Pilgerführer durch Rom: Codex Einsidlensis 326 (Wiesbaden: Steiner, 1987), no. 23, 82–83. Richard Krautheimer, Corpus Basilicarum Christianarum Romae: The Early Christian Basilicas of Rome (IV–IX Century), vol. 1, Monumenti di antichità cristiana, 2nd series, 2 (Vatican City: Ponticio istituto di archeologia cristiana, 1937), 43–63. 276 Patria (Πάτρια Κωνσταντινουπόλεως), I.60, in Scriptores originum Constantinopolitanarum, ed. Theodor Preger, vol. 2, BSGRT (Leipzig: Teubner, 1901), 144–45: τοῦ Κυρίου τὴν ἐκκλησίαν. 277 Theophanes, Chronographia, a. 5920, ed. Carl de Boor, vol. 1 (Leipzig: Teubner, 1883), 86–87. 278 Ibid., a. 6102, 1:299. Simon Malmberg, ‘Triumphal Arches and Gates of Piety at Constantinople, Ravenna, and Rome’, in Using Images in Late Antiquity, ed. Stine Birk et al., (Oxford: Oxbow, 2014), 167–68; John Wortley, ‘The Trier Ivory Reconsidered’, Greek, Roman, and Byzantine Studies 21, no. 4 (1980): 381–94. 76 An exception to prove this rule is provided by the Sessorian Palace. In the mid-fourth century a Severan-period hall was converted into an oratory dedicated to Christ’s crucifixion, now Santa Croce in Gerusalemme, in honour of the relics of the True Cross brought to Rome by the augusta Helena.279 While today Santa Croce conforms to a standard basilican plan, in the fourth century it possessed a distinctive architectural form. Not only was it accessible from the exterior along its north-western wall through five wide openings, but scars above the side aisles point to the existence of transversal arches across the nave (Fig. 67).280 These arches are an unusual feature and raise the possibility that the space was intended to be subdivided along its length, possibly according to social hierarchy.281 Either way, it is hard to assess to what extent Santa Croce was used as a representational space. It is likely that its conversion took place after Helena’s death in 330; thereafter emperors were not regularly resident in Rome again before the fifth century. In a broader sense, the conversion of the Roman civic elite in the fourth and fifth centuries also saw the addition of private cultic structures to the residences of the wealthy, though this was not a universal nor an instantaneous development.282 Overall, while private oratories were an available architectural concept in the fourth century, they did not form an important component of the imperial palace at this time. In the palaces of the fifth- and sixth-century regna, however, we see the palace church taking on a more assertive character. As we saw above, the palace at Carthage seems to have had an associated Marian shrine, though its precise function and location is unclear. More concretely, however, assuming a privileged space in the ensemble facing the platea maior in Ravenna is the palace church, Sant’Apollinare Nuovo. In contrast 279 Colli characterises the Severan structure as an ‘architectural junction’ joining the inhabited parts of the palace with the Circus Varianus: Colli, ‘Il palazzo Sessoriano’, 777. I follow here Krautheimer’s suggestion that Santa Croce was initially converted under Constantius II rather than Constantine, and underwent a series of alterations between the mid-fourth to mid-fifth centuries: Krautheimer, Corpus Basilicarum, 192. 280 Krautheimer, Corpus Basilicarum, 1:180; Brandenburg, Ancient Churches, 106. 281 Colli, ‘Il palazzo Sessoriano’, 779. It seems that these arches were an addition of the fourth century but possibly post- dating the space’s initial conversion 282 Sfameni, Ville residenziali, 243–84. On the difficulties of interpreting this evidence: Kim Bowes, ‘Christianization of Villas’, in The Roman Villa in the Mediterranean Basin: Late Republic to Late Antiquity, ed. Annalisa Marzano and Guy Métraux (Cambridge: Cambridge University Press, 2018), 453–64. 77 to, for example, Santa Croce, Sant’Apollinare Nuovo was both closely associated with the active resi- dence of the ruler and advertised its connection with royal authority to a broader, ‘public’ audience. Its presence in the monumental topography, sited in a liminal position between public thoroughfare and the interior of the palace complex, cannot be neatly explained with reference to extant structures. Any impulse to subsume it into the framework of imitatio imperii quickly breaks down; Deichmann’s attempt to relate Sant’Apollinare Nuovo to the supposedly-Constantinian ‘church of the Lord’ re- ported in the Patria, for example, is not convincing.283 The paired images of Ravenna’s palatium and the harbour of Classe are matched with pendant representations of Mary and Christ enthroned at the opposite end of the nave (Fig. 68), introducing a spatial differentiation between the nave’s western end and the sanctuary.284 One is tempted to suppose that Sant’Apollinare Nuovo possessed a dual function as both chapel and ‘secular’ representational space, a phenomenon demonstrable for later structures as we will see in the next chapter. In any case, Sant’Apollinare Nuovo points us towards a new departure in palace design: the increasing prominence of the palace church as representational space. The ‘liturgification’ of authority and the palace chapel Conjunctions of secular and ecclesiastical space were a widely dispersed phenomenon in the Early Middle Ages, even if material evidence is lacking for many of these structures. The increased promi- nence of the chapel in palaces of the Early Middle Ages is a product of the increased assimilation of the languages of political authority and Christian doctrine, a process Mischa Meier recently described 283 Friedrich Deichmann, Ravenna. Hauptstadt des spätantiken Abendlandes, vol. 2, bk. 1 (Wiesbaden: Steiner, 1974), 128; follo- wed by Johnson, Theoderic’s Building Program’, 85. 284 Speculating on what originally occupied the bottom register of the nave programme, subsequently obliterated by the processions of martyrs and virgins, is not productive. Theories have typically reconstructed a similar procession, but of Theoderic and the court towards Christ. Cf. Ian Wood, ‘Theoderic’s Monuments in Ravenna’, in The Ostrogoths from the Migration Period to the Sixth Century: An Ethnographic Perspective, ed. Sam Barnish and Federico Marazzi, Studies in Historical Archaeoethnology (Woodbridge: Boydell, 2007), 256–57. 78 with the evocative term Liturgisierung, or ‘liturgification’.285 While Late Roman archetypes provided the raw material, the chapel was a product of a broader social and cultural processes. The earliest signs of the liturgification of political authority can be found in Francia. At a council of bishops and laymen convoked in the basilica of Saint-Denis in 626 or 627, the assembled bishops acclaimed the king of the newly-unified kingdom, Chlothar II (king of Neustria 584–613; king of the Franks 613–629) by likening him to David: we rejoice greatly in the Lord that you make known to us not just your commands, but also those things which are announced to you by divine voices. Just like David, who governed his kingdom carefully with grace, you even anticipate what we are about to say and fulfil the prophetic ministry.286 Rhetorical connections between temporal and divine power had a long heritage stretching back to the age of Constantine, but for the most part direct comparisons between the ruler and Christ or other biblical prototypes, such as David, did not form a large component of the rhetoric of medieval king- ship before this time.287 Chlothar’s acclamation reflects, therefore, a transition in early medieval polit- ical culture in which the Bible (especially the historical books of the Old Testament) served as the epistemic lens for legitimate authority.288 Similarly, while the royal office was already closely connected with ecclesiastical sanction in the Visigothic kingdom by early the seventh century, a fully-fledged political theology of kingship which related royal and divine did not come about until the Carolingian period in the Frankish world.289 Jonas of Orléans’ De institutione regia, for example, a ‘mirror for princes’ 285 Meier argues that religiosity permeated all facets of social life in the Eastern Empire from the middle of the sixth century onwards: Mischa Meier, Geschichte der Völkerwanderung: Europa, Asien und Afrika vom 3. bis zum 8. Jahrhundert n. Chr., Histori- sche Bibliothek der Gerda Henkel Stiftung (Munich: C. H. Beck, 2019), 40–41, 953–73, 1035–69. See also: Eugen Ewig, ‘Zum christlichen Königsgedanken im Frühmittelalter’, in Das Königtum: seine geistigen und rechtlichen Grundlagen, Vorträge und Forschungen 3 (Lindau: Thorbecke, 1956), 7–73; Smith, Europe after Rome, 217–52. 286 Concilium Clippiacense (626/627), in Concilia aevi Merovingici, ed. Friedrich Maasen, MGH. Concilia 1 (Hannover: Hahn, 1893), 196: Unde non mediogriter gratulamur in Domino, quod ea, quae vobic divinis vocibus nuntiantur, non solum precepta profertis, quin etiam a nobis dicenda prevenitis hac velut illi David et regni imperium gratia provide gubernantes et ministrationem propheticam adimpletis. 287 Donald Bullough, ‘Imagines regum’, 54. 288 Ewig, ‘Königsgedanken’, 17–24; Yitzhak Hen, ‘The Uses of the Bible and the Perception of Kingship in Merovingian Gaul’, EME 7, no. 3 (1998): 277–90. 289 See esp: Rachel Stocking, Bishops, Councils, and Consensus in the Visigothic Kingdom, 589–633 (Ann Arbor, MI: University of Michigan Press, 2000). 79 for the young Pepin of Aquitaine, distinguished between temporal and spiritual authority, arguing that the authority of the former had been granted only by divine permission.290 While the ecclesia provided a range of symbols and signs which were progressively adopted in royal representation, it was only from the seventh century that Christianity began to constitute an embracing framework for compre- hending political reality. An important current in this development is the conjunction of ecclesiastical and royal space in the palace. In a narrow sense, the capella referred originally not to a building per se, but the space housing the relics belonging to the Frankish kings. Chief among these was the cappa (‘mantle’) of Saint Martin, from which the chapel took its name. The cappa is first attested as belonging to the royal treasury in 682; its mention also in a formula included in the Formulary of Marculf likely suggests that it had been in royal possession for some time.291 At this same time, connections were being drawn be- tween palaces and churches to mark them as specifically royal space. The venue chosen by Chlothar II for the council of 626/627 is significant in this context. In the conciliar acta, Saint-Denis is claimed be ‘joined’ to the royal palace at Clichy (Saint-Ouen sur Seine), at that time the pre-eminent royal residence.292 The favour shown to Clichy established a new axis for the staging of royal power, one inde- pendent of the logics of late Roman political culture. Under Chlothar and Dagobert, Dionysius was elevated to a symbol of Frankish kingship, receiving the title peculiaris patronus noster (‘our special 290 Jonas of Orléans, De institutione regia, VII, in Le métier de roi, ed. and French trans. Alain Dubreucq, Sources chrétiennes 407 (Paris: Éditions du Cerf, 1995), VII, 216–18. 291 DD Merov. 1, n. 126, 320: in oraturio nostro super cappella domni Martine. Marculfi Formulae, I.38, in Formulae Merowingici et Karolini aevi, ed. Karl Zeumer, MGH. Formulae (Hannover: Hahn, 1886), 68: tunc in palatio nostro, super capella domni Martini, ubi reliqua sacramenta percurrunt, debent coniurare. The date of the collection’s origin is unclear, but sometime in the second half of the seventh century seems probable. For discussion of the Formulary of Marculf’s date and provenance, see: Alice Rio, trans., The Formularies of Angers and Marculf: Two Merovingian Legal Handbooks, TTH 46 (Liverpool: Liverpool University Press, 2008), 104–23. For more on these texts, see below, Chapter 2. 292 Concilium Clippiacense (626/627), 196: Suggerente gloriosissimo atque piissimo domno Hlothario rege cum in suburbano Parisius in basilicam dominae Mariae matris Domini, quae in atrium sancti Dyonisii martyris sita est, iuxta predium, quod Clipiaco dicitur, venissimus. (‘At the urging of the most glorious and pious Lord King Chlothar, we have convened in in the suburbs of Paris in the basilica of the Lady Mary, the mother of the Lord, which is located in the atrium of [the basilica of] the holy martyr Dionysius, joined to the manor known as Clichy’) 80 patron’) in charters until the accession of the Carolingians.293 Saint-Denis was expanded and lavishly endowed under Dagobert, raised to the rank of senioris basilica, and tasked with the perpetual office.294 This elevation in status was accompanied by a lavish refurnishing of the interior. According to his Vita, the courtier Eligius, later the bishop of Noyon, made for Dagobert, both the tomb of the holy martyr Dionysius in the city of Paris and a canopy of marble over it fantastically worked with gold and gems. He also marvellously composed the crest and the images on the front. He likewise covered the columns arranged around the high altar in gold and placed golden apples on them, rounded and gemmed. He also covered both the lectern and the doors with silvered metal; he also enclosed the canopy of the high altar with columns of gold, as well as further elaborations to the area before the saint’s confessio.295 The institution of the perpetual office would be reconfirmed in 654 by Clovis II in a charter which also granted the community im- munity from the control of the bishop of Paris.296 Removing Saint-Denis from the authority of the Bishop of Paris ‘diluted’ his local authority, thereby making the church defiantly ‘royal space’ in which the king held precedence.297 From the seventh century onwards, the chapel would become an integral element of the pal- ace’s structure. We have already seen that the Marian chapel at Aachen was a structuring pole of the complex and understood by contemporaries to be a symbol of Carolingian governance writ large. The centrality of the chapel to the concept of the palace itself is even more noticeable in Oviedo, the royal 293 Carlrichard Brühl, Palatium und Civitas: Studien zur Profrantopographie spätantiker Civitates vom 3. bis zum 13. Jahrhundert, vol. 1 (Cologne: Böhlau, 1975)., 1:27–28. See e.g. DD Merov. 1, nos. 22, 28, 29 (as dominus et patronus noster), 41, 89, 93, 94, 122, 123, 138, 142. 294 DD Merov. 1, n. 85.10, 219; Fredegar, Chron., IV.79, 161; Vita Balthildis, c. 9, in Fredegarii et aliorum Chronica. Vitae sanctorum, ed. Bruno Krusch, MGH. SS rer. Merov. 2 (Hannover: Hahn, 1888), 493–94. For the archaeology of Dagobert’s church: Sumner McKnight Crosby, The Royal Abbey of Saint-Denis from its Beginnings to the Death of Sugar, 475–1151, ed. Pamela Blum, Yale Publications in the History of Art 37 (New Haven, CT: Yale University Press, 1987), 29–50. 295 Vita Eligii, I.32, in Passiones vitaeque sanctorum aevi Merovingici II, ed. Bruno Krusch, MGH. SS. Rer. Mer. 4 (Hannover: Hahnsche, 1902), 688–89: Eligius fabricavit et mauseoleum sancti martiris Dionisii Parisius civitate et tugurium super ipsum marmorem miro opere de auro et gemmis. Cristam quoque et species de fronte magnifice conposuit necnon et axes in circuitu throni altaris auro operuit et posuit in eis poma aurea, retundiles atque gemmatas. Operuit quoque et lecturium et ostia diligenter de metallo argenti ; sed et tectum throni altaribus axibus operuit argenteis. 296 DD Merov. 1 n. 85, 218. 297 Richard Gerberding, The Rise of the Carolingians and the Liber Historiae Francorum, Oxford Historical Monographs (Oxford: Clarendon Press, 1987), 68–69. 81 seat of the Asturian kings. The Chronicon Albeldensis describes the extensive palace complex Alfonso II (783, 791–842) founded in the region now dominated by the city’s Gothic cathedral, which included several churches ‘diligently ornamented with arches and columns, with gold and with silver, and at the same time he decorated them with varied pictures of royal palaces’ (Fig. 69).298 Of the palace itself, only the so-called ‘Torre Vieja de San Miguel’ on the south side of the cathedral (Fig. 70) is still extant and attests to the palace’s pronounced verticality. The structure known as the Cámara Santa (Fig. 71) a heavily-altered two-storey building attached to the southern side of the cathedral has been generally interpreted as a chapel connected with Alfonso’s palace, accessible from its interior on its upper floor.299 The Cámara Santa originally communicated with a Marian church also founded by Alfonso, the narthex of which served as the burial chapel of the kings of Asturias from Fruela I to Alfonso III (848–910) and two of the early kings of León. This church was demolished in the early-eighteenth century, but the original structure was seen and described in some detail by the antiquarian Ambrosio de Morales in 1572. According to Morales, the large church (around 100 feet in length), which termi- nated in a triple apse, had been assimilated into the thirteenth-century cathedral and connected to the transept.300 Taking the complex as a whole, the close intertwining of palace and various ecclesiastical structures is striking—especially given that Oviedo had not been a significant settlement prior to the eighth century—and effectively communicates an ideal of political legitimacy founded in the alliance of ecclesiastical and secular power. 298 Chronicon Albeldensis, XV.9, 174: cum arcis atque columnis marmoreis auro argentoque diligenter ornauit, simulque cum regiis palatiis picturis diuersis decorauit. 299 The Cámara Santa is first textually attested by name only in the early-twelfth century, but few doubt that it dates from Alfonso’s reign and was originally intended to function as a chapel. The present structure had to be mostly reconstructed after it was destroyed during the Asturian Revolution of 1934. Most of the masonry of the lower storey is original: Arbeiter and Noack-Haley, Christliche Denkmäler, 110. Rose Walker has argued it dates from the mid-ninth century by analogy with Ramiro’s foundations on the Monte Naranco: Walker, Art in Spain and Portugal, 164. This seems doubtful, given the close material and technical similarities between the Cámara Santa and other Alfonsine foundations, but in either case its close conjunction with the Oviedan palace is clear. 300 Ambrosio de Morales, Viaje de Ambrosio de Morales por orden del rey D. Felipe II, a los reinos de León y Galicia, y Principado de Asturias, para reconocer las reliquias de santos, sepulcros reales y libros manuscritos de las catedrales y monasterios, ed. Enrique Florez, Gran Biblioteca Historica Asturiana: Parte Eclesiástica 1 (Oviedo: Regareda y Comp., 1866), 37–42. 82 *** The palace chapel arose out of the combination of a range of influences and stimuli. While there was precedent for inserting private oratories and other religious structures into palaces in the Late Roman period, we have seen that there was not an unbroken tradition from the Late Roman period onwards in either the East or West. Rather, the seventh century stands out as the moment in which the prece- dents of earlier structures came together, possibly in parallel in the Eastern and Western Mediterra- nean, to produce the palace chapel in the course of the ‘liturgification’ of political action. While there was no consistent architectural typology, the introduction of the chapel into the ideal of a palace is nonetheless an architectural development, one reflecting changed ideals of legitimate authority. VI. Conclusion The preceding discussion has ranged widely over the evidence for palaces in the Early Middle Ages to make, at its core, three broad points. Firstly, palaces in the Late Roman Empire were designed accord- ing to a common spatial ‘grammar’, focused on the conjunction of palace and circus, as witnessed at Milan and Trier in the West, as well as Thessaloniki and Constantinople in the Eastern Mediterranean. This spatial grammar reproduced the salient topographical features of the Palatine, thereby legitimising them as sites of political authority. Despite the brevity of his reign, Maxentius successfully reimported the architectonic language of the provincial sedes to Rome itself, transforming the Eternal City into essentially, a ‘Tetrarchic’-style residence geared towards a new conception of imperial power. The idea of a palace expressed by the Late Roman sedes was short-lived after the Western Em- pire’s disintegration, however. We traced its dissolution in two respects. First, in ideological terms, we saw that palaces provided opportunities for post-Roman rulers to ‘look Roman’, to clothe themselves in the signs of Roman authority. At the same time, however, palaces were sites at which the languages of Roman authority were reinvented and redeployed to suit new contexts. Secondly, a gulf soon opened between rhetorical ideal and material reality. The breakdown of the Roman political economy 83 triggered wide-ranging formal and material changes in elite architecture. Although the palaces of, for example, the Merovingians and other rulers of the seventh-century regna are lost to us, it was argued here that the outlines of their palaces could be detected, however faintly, by replacing them into the broader context of architectural practices and technologies between the sixth and ninth centuries. During the nearly two centuries separating the accession of Theoderic and the accession of Charle- magne, the architectural ‘ideal’ of a palace as both ideological statement and material object was fun- damentally renegotiated. Finally, the Early Middle Ages saw the introduction of a new assemblage into the palace’s spatial composition: the palace chapel. Contrary to common perception, chapels were not significant sites of political representation in the Late Empire or in Byzantium prior to the sixth century. Conse- quently, the palace chapel (as representational space) was not inherent in medieval culture, but a de- velopment tied to social and cultural trends of a broader nature. Here it was argued that the specific conceptual merging of sacred and representational space was inextricably tied to what was termed here the liturgification of authority from the seventh century onwards. Overall, we have seen that palaces were not static as a building type. Rather, their forms and the conceptual syntheses that underlay them were keyed to local and contextual discourses concerning the nature of political authority and its proper action. After around the middle of the sixth century the form and function of the palace drastically changed, signalling the more-limited economic means of early medieval regimes, but also that concepts of political authority and legitimacy entering a state of flux. This was not limited to palace’s visual characteristics, however, but also in the activation of these spaces in political ritual, the subject to which we now turn. 84 CHAPTER 2 The Ritual Shaping of the Early Medieval Palace Consider also what you will obtain from such favour: that among the crowd of officials you may be seen to proceed first before the feet of the king, decorated with a golden rod, and by this sign of your proximity to us, we recognise that the palace has been entrusted to you. — Cassiodorus, Variae, VII.5301 Proof of this is still to be seen not only in the divine but also in the human basilica at Aachen, and in the houses for every man who held an office, which were so con- structed around the palace, on the plan of the most perceptive Charles, that he himself could see through the balustrade of his balcony what anyone who was entering or leaving was doing, as if he were right there with them. — Notker, Gesta Karoli, I.30302 I. Introduction The palace was always more than bricks-and-mortar architecture. In the rhetorical constructions of contemporary authors, the palace was both physical and social space, enlivened and governed by often- unspoken codes of behaviour which mediated social relations between ruler and elite. For Cassiodo- rus, the elevated position of the curator palatii is signalled by his right to walk before the king in pro- cession, marked out by the golden rod which denoted his office. Notker’s description of Charlemagne’s Aachen, on the other hand, melds architectural and social logic to imagine a proto- panoptic Charlemagne, surveillant over all from his balcony at the centre of a regime of regulation and discipline.303 We are reminded of Foucault’s reading of Bentham, not only in Notker’s all-seeing king, but in the figuration of the strategies of power in architecture, a ‘spatialization’ of authority.304 Notker was especially sensitive to the visual aspects of court life – the competitions of status communicated 301 Cass., Var., VII.5.6, 205: illud quoque considera, qua gratificatione tracteris, ut aurea virga decoratus inter obsequia numerosa ante pedes regios primus videaris incedere, ut ipso testimonio vicinitatis nostrae agnoscamur tibi palatia commisisse. 302 Notker, Gesta Karoli, I.30, 41: sicut adhuc probat non solum basilica divina, sed et humana apud Aquisgrani et mansiones omnium cuiusdam dignitatis hominum, quę ita circa palatium peritissimi Karoli eius dispositione constructę sunt, ut ipse per cancellos solaria sui cuncta posset videre, quęcumque ab intrantibus vel exeuntibus quasi latenter fierent. Trans. Noble, 85. 303 Mayke de Jong, ‘Charlemagne’s Balcony: The Solarium in Ninth-century Narratives’, in The Long Morning of Medieval Europe: New Directions in Early Medieval Studies, ed. Jennifer Davis and Michael McCormick (Aldershot: Ashgate, 2008), 277– 89. 304 Michel Foucault, Discipline and Punish: The Birth of the Prison, trans. Alan Sheridan, 2nd ed. (New York: Vintage, 1995), 195–228. See further the discussion in: Michel Foucault, ‘Questions on Geography’, in Power/Knowledge: Selected Interviews and Other Writings, 1972–1977, trans. Colin Gordon (New York: Pantheon, 1980), 63–78 (esp. 70–77). 85 in everything from speech and gesture, to ceremonial precedence, to clothing. ‘Before we have only seen men of clay, but now we have seen men of gold’, a caliphal envoy exclaims in one of his tales.305 This chapter seeks to advance two claims. The first is that the rituals of the early medieval palace were not unmitigated assertions of royal or imperial power in either the late- or post-Roman periods. Rather, it was a mode of symbolic communication between sovereign and elite through which political norms were negotiated, common identities were performed and, ultimately, legitimate author- ity was produced. Secondly, I will argue that the rituals of the palace were a function of the formation of the court as a social and spatial entity – both a network of relations between social actors and a place, the political centre. The sociogenesis of early medieval courts brought about a fundamental transformation of the spatial expressions of political relations. Whereas in the Late Empire, political authority was legitimised principally before an urban citizenry in architectonic settings designed and curated for this purpose, after the sixth century the most important components of this function were steadily transferred into the palace’s interior among a smaller clique of elites, with the palace church becoming a particularly charged site of political ritual. Altogether, the transition from Late Antiquity to the Early Middle Ages saw a revolution in the nature of the ‘public sphere’ in which authority was legitimised. Ritual matters: what is ‘ritual’ and what does it do? First, a word on the subject matter. ‘Ritual’ or ‘ritualised behaviour’ has been preferred here to describe the phenomena described by Cassiodorus and by Notker, a choice which requires some justification. Even a quick glance at the historiography reveals a variety of terms used nearly synonymously, with ‘ceremony’ and ‘ceremonial’ being especially common.306 These terms are useful for their evocation of 305 Notker, Gesta Karoli, II.8, 60: Prius terreos tantum homines vidimus, nunc autem aureum. Trans.: Noble, 97. 306 See, for example, the contributions to: Rituals of Royalty: Power and Ceremonial in Traditional Societies, ed. David Cannadine and Simon Price, Past and Present Publications (Cambridge: Cambridge University Press, 1987). For reflections on the utility of ‘ritual’, see: Walter Pohl, ‘Staat und Herrschaft’, 16–27. 86 the grand, symbol-laden occasions observed in political cultures from Antiquity to today, but their implication of a codified or ‘scripted’ order of proceedings strikes a dissonant note in the context of the early medieval West. As we will see, political ritual in the Early Middle Ages certainly could be highly choreographed, particularly in contexts of liturgy or in performances of universal consensus, but such terminology fails to capture smaller-scale use of gesture, emotion, or verbal formulae for contextual effect, as well as the unspoken ‘rules of the game’ which governed when such behaviours were legitimate or persuasive.307 In the present chapter, therefore, ‘ritual’ does not refer to a kind of ‘staged applause’, to use Emanuel Mayer’s useful term, but to a diffuse set behaviours and practices which mediated relations between social agents through which power was produced.308 In particular, court rituals defined the normative identities of members of the political community, established social ties, and represented hierarchies of status and prestige. The precise content of these behaviours is defined contextually and situationally, rather than (necessarily) in relation to a predetermined ‘script’.309 Attempting any such analysis is inevitably hampered by the dearth of detailed accounts of the rituals of the palace from contemporary sources. All too often we are left with allusions to protocols and behaviours that the authors of our sources take for granted will be understood by their audiences. There is also some truth to Buc’s admonition that we are not studying ‘ritual’ per se, but textual repre- sentations of epiphenomena that we are, epistemologically speaking, unable to recover in full.310 This 307 Gerd Althoff, Spielregeln der Politik im Mittelalter: Kommunikation in Frieden und Fehde (Darmstadt: Primus, 1997). In any case, we do not possess a ‘script’ for such occasions comparable to the tenth-century Book of Ceremonies of the Byzantine court, nor should we expect that one existed: Timothy Reuter, ‘Assembly Politics in Western Europe from the Eighth Century to the Twelfth’, in idem, Medieval Politics and Modern Mentalities, ed. Janet Nelson (Cambridge: Cambridge University Press, 2006), 203. 308 Catherine Bell, Ritual Theory, Ritual Practice (Oxford: Oxford University Press, 1992), 69–93 (esp. 88–93). For ‘staged applause’: Mayer, Rom ist dort, 1–27. 309 Bell writes of a ‘privileged distinction’ which separates ‘ritualised’ behaviour from everyday life: Bell, Ritual Theory, 90– 91. Viewed in the abstract, there might not be much which distinguishes (say) a banquet in honour of visiting guests from a regular meal, but in the former case the participants are expected to act in such a way as to establish it as distinct and symbolically ‘more important’. The ‘sense of ritual’ is thus a part of a society’s habitus: ibid., 80. 310 Philippe Buc, The Dangers of Ritual: Between Early Medieval Texts and Social Scientific Theory (Princeton, NJ: Princeton Uni- versity Press, 2001). For a robust rebuttal, see: Geoffrey Koziol, ‘The Dangers of Polemic: Is Ritual Still an Interesting Topic of Historical Study?’, EME 11, no. 4 (2002): 367–88. 87 need not be a cause for despair, however. As Geoffrey Koziol has argued, the point is not to isolate ‘ritual’ as a unique and discrete category of behaviour with specified symbolic referents, but to expli- cate the networks of signs—the ‘webs of significance’, in Geertz’s words—available to contemporary actors through which social values were manifested, manipulated, and mobilised in lived experience.311 Per Lefebvre, these ideologies are perceived in social space and embodied in codes of behaviour, thereby connecting the abstract to the lived.312 Elias was sensitive to such dynamics in the ‘court soci- ety’ of Louis XIV’s Versailles. Versailles, the physical place, was inextricable from the interdependent relations of the king and his aristocracy. Its architecture—a species of the aristocratic palais—was designed to facilitate particular contextual social interactions which were mediated through ritual.313 In this environment, elaborate codes of etiquette and behaviour created an ‘exhibition of court society to itself. Each participant, above all the king, has his prestige and his relative power confirmed by the others.’314 The consequences of these reflections are twofold. Firstly, ritual must be taken seriously as a process through which authority was defined, negotiated and legitimised rather than simply ‘adver- tised’. Secondly, it forces us to treat the palace as an agent in the production of that same authority. For political authority to be deemed legitimate, it must be perceived and experienced in environments (physical, social, institutional) created for that purpose. Adam T. Smith captures this in his reading of the courtroom. The judge possesses authority in the court, manifested in their ability to pass judge- ment on those before them, an authority that is not legitimately shared with others. The authority vested in the judge is thus dependent on its practice through that space: the words exchanged in a courtroom are only significant if the authority of the pro- ceedings has already been established, before the first word is uttered. […] The 311 Koziol, ‘Dangers of Polemic’, 385–87; Clifford Geertz, ‘Thick Description: Towards an Interpretive Theory of Culture’, in idem, The Interpretation of Cultures: Selected Essays (New York: Basic Books, 1973), 5. 312 Lefebvre, Production, 214–15; Yi-Fu Tuan, Space and Place: The Perspective of Experience (Minneapolis, MN: University of Minnesota Press, 1977). 313 Norbert Elias, The Court Society, trans. Edmund Jephcott (New York: Pantheon, 1983), 41–65; 78–116. 314 Ibid., 101. Emphasis added. 88 constitution of authority therefore hangs on the production of place as it organizes and conditions the actors who will practice through, rather than simply within, specific landscapes. Authority is constituted in the court through the production of its space providing an arena in which speech by designated actors can already be constituted as authoritative.315 This capability is not universal, however, but dependent on experience, in the sense of the knowledge of how to act in accordance with expected norms.316 The production of social space thus implies a community of knowledge, a group inducted into ceremonial protocols and skilled in their manipula- tion. How these practices created the ‘palace’ as the space in which politics legitimately took place is the subject of this chapter. The analysis here does not, therefore, attempt to distinguish a universal or defined corpus of ‘royal rituals’ and/or to trace their change over time in a formal sense, but to use textual, visual and architectural evidence to show that the ritual production of authority and the spatial production of the palace were inextricably entwined. Overview The present chapter is not organised according to a catalogue of named ceremonies, nor geographical or temporal case studies. Rather, it analyses in turn the social spaces which together constituted the early medieval palace. Behind this (possibly unusual) organisational principle lies the belief that, alt- hough the paths of the post-Roman kingdoms diverged in important ways throughout the period under discussion, within these polities the palace played a sufficiently similar role as to make them broadly comparable. First, however, we turn to Late Roman ‘public’ ritual. Political ritual plays an important role as a vector of cultural continuity in recent narratives charting the Roman world’s ‘trans- formation’. Here I will argue that, while the ritual forms of Roman public life had a long afterlife in the Early Middle Ages, the environments in which it took place were radically transformed – a sign that concepts of legitimacy and authority had entered into a state of flux by the mid-sixth century. 315 Smith, Political Landscape, 107. 316 E.g. Elias, Court Society, 103 on the exclusionary function of court etiquette. See also: Tuan, Space and Place, 68, on ‘spatial knowledge’. 89 Thereafter, investigation will move in turn through the spaces characteristic of the social life of the early medieval palace: the audience hall, the treasury, hunting grounds, and finally the palace church. The rituals that took place within, or through these spaces, not only mediated the interactions of the ruler with the political community but also negotiated the nature and limits of royal authority. II. The City Appearances of the emperor before the citizens of his chosen seat animated the public life of the Later Roman Empire. As we saw in the previous chapter, the sedes imperii of the fourth and fifth centuries were designed according to a common spatial ‘grammar’, which reproduced the most salient features of the topography of the Palatine in new urban centres. The sites chosen were dramatically altered to accommodate the vast palatial complexes inserted into their fabric. They were, in a very real sense, ‘palace-cities’. Wall circuits, gates, and porticoed streets in the imperial seat linked the most important loci of authority into a unified ‘armature’, at the centre of which was located the imperial palatium.317 Here we turn to the activation of this architecture in political ritual. Reports of imperial ritual from this period have exercised a powerful effect on the imagination of observers both ancient and modern as evidence for the final abandonment of any fig leaf of republicanism, leaving only a naked and decadent autocracy. The processions and laudatory acclamations of imperial power were more than ‘propaganda’ imposed on a pliant commons, however, and the ‘divinising’ guise of the later Roman emperor was not an unmitigated sign of his autocratic power. Rather, it was one side of a dialogue taking place with power holders in Roman society through which the imperial system was legitimised. At the same time, the symbolic languages of imperial ritual represented an important thread of continuity between the ancient and medieval worlds. Everything from ritual occasions (such as the adventus), to modes of address, to material objects such as items of regalia can be traced back from the 317 Dey, Afterlife, 21–77. For ‘armature’: MacDonald, Architecture, 2:5–31. 90 Early Middle Ages to Roman precedents. Hendrik Dey has recently argued that political ritual was, in fact, a persistent stimulus of the kind of urbanism witnessed in the imperial sedes over the longue durée in both East and West.318 This continuity was not absolute, however. Ritual forms inherited from the Later Roman Empire were creatively reimagined to suit new contexts; by the same token, the environ- ments in which political ritual took place were radically changed. We see this most clearly, in fact, in the relation of palace and city. By the seventh century, the city has ceased to be the foremost site of political ritual as in the Late Empire, reflecting new ideals of political authority and its distribution in society. The emperor’s changing image Constantius II’s (in)famous adventus into Rome in 357 encapsulates the complex relationship between public ritual and the production of imperial authority. Constantius’ entry into Rome was a major event of his reign, designed to celebrate his victories over the usurper Magnentius a few years earlier. Am- mianus Marcellinus gives a vivid account of the occasion that is worth briefly summarising. The em- peror arrived at Rome’s walls mounted in a golden chariot, flanked by heavy cavalrymen (catafracti) and infantry carrying standards.319 On entering the city and receiving the acclamation of the crowd, however, he stayed impassive: Though he was very short he stooped when he passed under a high gate; otherwise he was like a statue, gazing straight before him as if his head were in a vice and turning neither to right nor left. When a wheel jolted he did not nod, and at no point was he seen to spit or to wipe or rub his face or nose or to move his hand.320 318 Dey, Afterlife, passim. 319 Amm. Marc., Res Gestae, XVI.10.6–8, 83. 320 Ibid., XVI.10.10, 84: nam et corpus perhumile curuabat portas ingrediens celsas et velut collo munito, rectam aciem luminum tendens, nec dextra vultum nec laeva flectebat et (tamquam figmentum hominis) nec cum rota concuteret nutans, nec spuens, aut os aut nasum tergens vel fricans, manumve agitans visus est umquam. Trans.: Ammianus Marcellinus, The Later Roman Empire, A.D. 354–378, trans. Walter Hamilton, Penguin Classics (Harmondsworth: Penguin, 1986), 101. I have here amended the translation of figmentum hominis (literally, the ‘image of a man’) from Hamilton’s reading of ‘dummy’ to clarify the intended sense. 91 Riding alone in his chariot, Constantius merged with his own image. Maintaining a studied profile in relation to the amassed crowd, he thus simulated the eternal stillness of his own representations, as on his coinage or on luxury objects such as the silver missoria discovered near Kerch (Figs. 72–73).321 Once within the city, however, Constantius showed a different side. Awed by Rome’s architectural marvels—the Forum, the Capitoline, and especially the Forum of Trajan—he addressed the Senate in the Curia and the people from the Rostra, was received warmly in the palace itself, and finally presided over games in Circus Maximus.322 Sabine MacCormack has stressed that for every gesture which seems to confirm the ‘autocratic’ image of the Dominate, another emphasises the emperor’s civilian vir- tues.323 In Ammianus’ description, therefore, Constantius’ demeanour varied according to the spatial setting and, we might surmise, the composition of his audience. As MacCormack has argued, the rituals surrounding the emperor’s arrival ‘provided a vocabulary for the encounter of different types of persons, and for their convergence into one group’.324 A similar distinction is drawn by the anony- mous orator of the ‘birthday speech’ (Genethliacus) delivered before Maximian at Trier in 291. Describ- ing that emperor’s joint conference with Diocletian in Milan the preceding winter, the orator contrasts the meeting of the emperors and a select group of elites in the palace’s interior and the joyful crowds who clamoured to catch a sight of the Dyarchs on their procession through the city. On the one hand, the emperors were ‘gazed upon’ by the senators admitted into the palace, as if in a religious ceremony: What a spectacle your piety created, when those who were going to adore your sacred features were admitted to the palace in Milan you both were gazed upon and your twin deity suddenly confused the ceremony of a single veneration! No one observed the hierarchy according to the usual protocol; they all stopped still to spend more time in 321 Sabine MacCormack, Art and Ceremony in Late Antiquity, Transformation of the Classical Heritage 1 (Berkeley: University of California Press, 1981), 44. As Flower argues, however, Ammianus uses Constantius’ deportment as a vehicle of criti- cism: there was a wide gulf between Constantius’ self-image and how he was actually perceived. Richard Flower, ‘TAMQVAM FIGMENTVM HOMINIS: Ammianus, Constantius II and the Portrayal of Imperial Ritual’, The Classical Quarterly 65, no. 2 (2015): 822–35. 322 Amm. Marc., Res Gestae, XVI.10.13–15, 84–85. 323 MacCormack, Art and Ceremony, 42. 324 Ibid., 43. 92 adoration, stubborn in their duplicate pious duty. Yet this private veneration, as in in the inner shrine, stunned the minds only of those whose public rank gave them access to you.325 The emperors then process through the streets of Milan. The orator conjures the voices of the crowd—‘Do you see Diocletian? Do you see Maximian? Both are here! They are together! How closely they sit! How amicably they converse! How quickly they pass by!’—to draw attention to other ideal qualities of the ruling sovereigns: their collegiality, their friendly demeanour, their unity with the peo- ple.326 Back in Rome, conventions to which the distant emperor was accustomed are contrasted with what were, for Ammianus at least, ‘proper’ Roman values. For example, Constantius’ ‘statuesque’ de- meanour is dismissed as an affectation more suitable to a provincial (and more specifically, eastern) audience.327 Similarly, a parade of such military might was for Ammianus more appropriate for the limes: Constantius’ appearance was as if it had been ‘designed as a show of strength to overawe the Euphrates or the Rhine’.328 While Constantius’ ignorance of Rome’s traditions and amazement at her monuments is a rhetorical construction of Ammianus’ text, he points here towards a greater phenom- enon: differences in the imperial image at the centre and at the periphery communicated through differing ritual honours. Imperial representation took on a new mode from the late-third century onwards, a change noted by contemporaries.329 If we were to characterise this new imperial image merely in terms of 325 Pan. Lat. XI.11.1–3, 264–65: Quale pietas uestra spectaculum dedit, cum in Mediolanensi palatio admissis qui sacros uultus adoraturi erant conspecti estis ambo, et consuetudinem simplicis uenerationis geminato numine repente turbastis! Nemo ordinem numinum solita secutus est disciplina; omnes adorandi mora restiterunt duplicato pietatis officio contumaces. Atque haec quidem uelut interioribus sacrariis operta ueneratio eorum modo animos obstupefecerat quibus aditum uestri dabant ordines dignitatis. Trans. Nixon and Rodgers, 95–96. 326 Pan. Lat., XI.11.3–5, 265: ‘Vides Diocletianum? Maximianum vides? Ambo sunt! Pariter sunt! Quam iunctim sedent! Quam concorditer conloquuntur! Quam cito transeunt!’ Trans.: Nixon and Rodgers, 96. 327 Amm. Marc., Res Gestae, XVI.10.9, 84: talem se tamque immobilem, qualis in provinciis suis visebatur, ostendens. Trans.: Hamilton, 100–01: ‘[he] kept the same impassive air as he commonly wore before his subjects in the province.’ Constantius ruled from the East for most of his reign; the implication is that he had taken on the iconography of a Hellenistic monarch. 328 Amm. Marc., Res Gestae, XVI.10.6, 83: Et tamquam Euphraten armorum specie territurus aut Rhenum. Trans.: Hamilton, 100. 329 E.g Aurelius Victor, De caesaribus, c. 39, in Sexti Aurelii Victoris Liber de Caesaribus, ed. Franz Pichlmayr, BSGRT (Leipzig: Teubner, 1901), 117: Namque se primus omnium Caligulam post Domitanumque dominum palam dici passus et adorari se apellarique uti deum. Trans.: Aurelius Victor, Liber de Caesaribus of Sextus Aurelius Victor, trans. H. W. Bird, TTH 17 (Liverpool: Liverpool 93 ‘propaganda’, it would be underestimate its societal role, however. Following Emanuel Mayer, inno- vations in the emperor’s public image reflect more strongly a changed milieu of patrons and viewers than ‘staged applause’ for an autocratic regime.330 His ‘discursive’ model holds that differences in the imperial images manifested in Rome and the provincial sedes bespeak different shared values in each social space. We saw this in the case of Constantius II: the basis of Ammianus’ critique is that the representational strategies adopted in the (undoubtedly extensively choreographed) visit were inap- propriate in Rome. Underpinning these two examples is the contrast between the language of imperial praise in the social space of the imperial sedes and in Rome. The ‘veneration’ of the emperor described by the orator of 291 speaks to the performance of a ‘courtly’ identity among a small, privileged group. In doing so, the community represents itself to itself, distinguishing and confirming the status of each member relative to the others and asserting its cohesion as a group. The ‘divinised’ imperial image of the ‘Dominate’ can thus be said to have held significance in a specific social space and need not have been universally shared in society as such.331 The streets, the circus, the church: the spaces of imperial ritual The public experience of imperial ritual was understood to be an essential component of its legitima- tion by contemporaries. As in Milan in 291, Rome in 357, and many other examples described by late ancient authors, we frequently encounter the image of crowds of the common people surging to look upon the imperial countenance. The late ancient city was shaped by this ideology. That is to say, the particular forms of Roman urbanism expressed—and facilitated—the hierarchical relations of the Ro- man polity.332 The principal arenas of public life in the late Roman city were bound together by an University Press, 1994), 41: ‘[Diocletian] was the first after Caligula and Domitian to permit himself to be called “Lord” in public and to be worshipped and addressed as a god’. See also: Amm. Marc., Res Gestae, XV.5.18, 50. 330 Mayer, Rom ist dort, 19, 22–27. 331 This is also true of the earlier Empire: Emanuel Mayer, ‘Propaganda, Staged Applause, or Local Politics? Public Mon- uments from Augustus to Septimius Severus’, in The Emperor and Rome: Space, Representation, and Ritual, ed. Björn Ewald and Carlos Noreña, Yale Classical Studies 35 (Cambridge: Cambridge University Press, 2010), 111–34. 332 Lefebvre, Production, 40–45; Soja, Postmodern Geographies, 79–80. 94 architectonic ‘armature’, together forming a grammar of urban experience which was inextricable from the political structures to which it was connected. Monuments and their architectural framing were intended to be experienced in motion, both everyday movement through urban space and in ritual procession and were designed to frame semiotically-charged vistas.333 In the capitals of the late Roman Empire, authority was encountered in the streets. While such architectonic Inszenierung is most clearly attested in the Eastern Mediterranean, at Constantinople and Thessaloniki for example, traces are nonetheless perceptible in the Western sedes. In Milan, for example, Ausonius praised Milan’s marble clad colonnades as one of its noteworthy features alongside the circus, the theatre, and the palace itself, for example, yet within the walls they have proven elusive.334 Under Valentinian II the tract of the Via Romana approaching Milan’s walls was augmented with a double portico terminating in a triumphal arch, an assemblage monumentalising the approach to the city from Rome and embracing the Ambrosian basilica apostolorum (San Nazaro).335 It is tempting to suppose that the Via Ticinensis was similarly adorned with porticoes: the outer col- onnade and propylaeum of the portico which preceded the atrium of San Lorenzo (Fig. 74) along the street’s edge are still extant, yet unfortunately there is no evidence of their continuation in either di- rection.336 Likewise, as we saw in the previous chapter, the palatium mosaic of Sant’Apollinare Nuovo is a tantalising example of the continued valence, for some time at least, of this architectural frame in the former imperial seat of Ravenna. 333 E.g. Franz Alto Bauer, Stadt, Platz und Denkmal in der Spätantike: Untersuchungen zur Austattung des öffentlichen Raums in den spätantiken Städten Rom, Konstantinopel und Ephesos (Mainz: Phillipp von Zabern, 1996), esp. 307–388; Elizabeth Marlowe, ‘Framing the Sun: The Arch of Constantine and the Roman Cityscape’, Art Bulletin 88, no. 2 (2006): 223–42. 334 Ausonius, Ordo urbium nobilium, VII, 195: cunctaque marmoreis ornata peristyla signis. 335 Donatella Caporusso, ‘La zona di corso di Porta Romana in età romana e medioevale’, in Scavi MM3. Ricerche di archeologia urbana a Milano durante la costruzione della linea 3, ed. Donatella Caporusso, vol. 1 (Milan: Edizioni ET, 1991), 251–57. The martyrium was constructed in the early 380s and dedicated before 386: Neil McLynn, Ambrose of Milan: Church and Court in a Christian Capital, Transformation of the Classical Heritage 22 (Berkeley, CA: University of California Press, 1994), 230. 336 Jean-Charles Picard, ‘Le quadriportique de Saint-Laurent de Milan’, MEFRA 85, no. 2 (1973) : 699. The portico was constructed contemporaneously with the church: ibid., 708. For San Lorenzo’s dating and architectural history, see: Dale Kinney, ‘The Evidence for the Dating of S. Lorenzo in Milan’, Journal of the Society of Architectural Historians 31, no. 2 (1972): 92–97; Richard Krautheimer, Three Christian Capitals: Topography and Politics, Una’s Lectures 4 (Berkeley, CA: 1983), 69–92. 95 By the fourth century, the circus had taken on the role of the primary arena for interaction between the emperor and the urban citizenry as a group. Paul Veyne memorably described their meet- ing as a raucous back-and-forth, an exchange far removed from the usual image of solemn acclama- tions for an aloof and distant emperor: the circus crowd ‘honoured their Emperor, demanded that he should grant them pleasures, made known to him their political demands, and either hailed or attacked the Emperor under colour of applauding and booing the shows […]. The shows were a court cere- monial and a tête-à-tête between the Emperor and his court of citizens’.337 This was hinted at already in Constantius’ experience at Rome in 357. The emperor was ‘amused’ by the chants of the circus-going crowd who retained their ‘traditional freedom of speech without any loss of respect’ (unlike provincial audiences, we might surmise).338 The importance of the circus to Rome’s political culture was repro- duced and intensified in the provincial sedes. As shown in the previous chapter, the combination of palace and adjoining circus is consistently attested, thereby reproducing the most salient spatial axis of the Palatine. A strict architectural conjunction, as at the Circus of Maxentius and most famously at Constantinople, cannot be demonstrated in all cases but if the hypothesised bridge between the Domus Augustana and the Circus Maximus is taken seriously as part of the Trajanic arrangement there would be grounds to suspect its inclusion as part of the basic conceptual formulation. Games had been associated with such events as a part of the celebrations of important occa- sions, but certainly by the middle of the fifth century the circus had become a more general focus for imperial representation, a development shown in detail by Michael McCormick in relation to victory celebrations.339 The Hippodrome at Constantinople, for example, was associated with the submission of Rome’s enemies already in the fourth century. The Obelisk Base of Theodosius I (379–395 CE) 337 Paul Veyne, Bread and Circuses: Historical Sociology and Political Pluralism, trans. Brian Pearce (London: Penguin, 1990), 400– 01. 338 Amm. Marc., Res Gestae, XVI.10.13, 84: et saepe, cum equestres ederet ludos, dicacitate plebis oblectabatur nec superbae nec a libertate coalita desciscentis reuerenter modum ipse quoque debitum servans. Trans.: Hamilton, 101. 339 McCormick, Eternal Victory, 35–79. 96 depicts the imperial house, flanked by the Senate and by the imperial bodyguard, accepting the tribute of kneeling Persian and German barbarians from the kathisma (Fig. 75).340 The spina of the Circus Maximus likewise seems to have been adorned with statue groups of bound captives being trampled by victorious Romans.341 These images were animated by ritual spectacle. Philostorgius implies that Honorius performed the ritual trampling (calcatio colli) of the defeated usurper Priscus Attalus in Rome upon a dais before the mutilation of the usurper’s right hand, an action we should probably associate with the Circus Maximus.342 When news of the usurper Silvanus’ defeat reached Rome, the assembled crowd in the Circus Maximus shouted ‘Silvanus has been defeated’ (Siluanus deuinctus est), a phrase which possibly reflects an acclamatory formula.343 Procopius writes that Ioannes (423–425) was pa- raded in the circus of Aquileia, again with his hand mutilated, where he was publicly mocked before being executed.344 Finally, the other substantially new arena for imperial display was in church. There is an im- portant distinction to be made here: although, as we saw in the previous chapter, commemorative oratories were inserted into imperial palaces at a comparatively early date, they were not significant loci of imperial representation at this time. Rather, appearances of the emperor in church took place in the city, often in the bishop’s cathedral. Accommodating the emperor into the community of the 340 The Latin inscription records that the obelisk was erected to celebrate the defeat of ‘tyrants’, referring to Magnus Maximus and his son Victor, not barbarians per se: CIL III, 737: EXTINCTIS PALMAM PORTARE TYRANNIS. Thus, similarly to fourth-century reliefs on the Arch of Constantine, victory over internal enemies is elided with victory over the generic ‘barbarian’. On the obelisk, see: Linda Safran, ‘Points of View: The Theodosian Obelisk Base in Context’, Greek, Roman, and Byzantine Studies 34, no. 4 (1993): 409–35. 341 Cass., Var., III.51, 106: spina infelicium captivorum sortem designat, ubi duces Romanorum supra dorsa hostium ambulantes laborum suorum gaudia preceperunt. 342 Philostorgius, Historia ecclesiastica, XII.5, in Philostorgius, Kirchengeschichte, ed. Joseph Bidez, rev. Friedhelm Winkelmann, Die Griechischen Christlichen Schriftsteller (Berlin: Akademie Verlag, 1981), 144. Notably, Attalus is not executed but exiled. Olympiodorus confirms Attalus’ mutilation and exile but places it Ravenna: Olympiodorus, Historia, fg. 14, in Pho- tius, Bibliothèque, 1:170. McCormick, Eternal Victory, 97. 343 Amm. Marc., Res Gestae, XV.5.34, 53: Romae in Circo Maximo populus – incertum, relatione quadam percitus an praesagio – “Silvanus deuictus est” magnis uocibus exclamaret. 344 Proc., Wars, III.3.9, 2:320: ζῶντα δὲ Βαλεντινιανὸς Ἰωάννην λαβὼν ἔν τε τῷ Ἀκυληίας ἱπποδρομίῳ τὴν ἑτέραν ταῖν χεροῖν ἀποκοπέντα εἰσῆγεν ἐπόμπευσέ τε ὄνῳ ὀχούμενον, καὶ πολλὰ παρὰ τῶν ἀπὸ σκηνῆς ένταῦθα παθόντα τε καὶ ἀκούσαντα ἔκτεινεν. Cf. Philostorgius, Histora ecclesiastica, XII.12–13, 148–49. 97 faithful provoked tension, even competition, between the court and the bishops who presented, at times, a competing source of social power.345 The demonstration of Christian piety (and orthodoxy) became a standard virtue in the repertoire of imperial legitimation and thus quickly became incorpo- rated into imperial ritual.346 In the western sedes, we see efforts by imperial courts to ‘make a space’ for themselves in the Christian church through votive dedications which heralded close links with partic- ular churches. In Milan, for example have already seen that Ambrose’s basilica apostolorum was inte- grated into the ceremonial armature of the Via Romana. The basilica would receive further attention from the imperial court during the reign of Honorius on its rededication to Nazarius. Serena, Stilicho’s wife and Honorius’ cousin, donated Libyan marble for the ornamentation of the new shrine and prayed for the general’s safe return in an inscription.347 It is likely that another Milanese church, San Lorenzo, was also closely connected with the court. Its large size, the spoliation of the amphitheatre for its construction, and the likely original function of the Sant’Aquilino chapel as an imperial mauso- leum all point towards imperial patronage (Fig. 76).348 The so-called ‘basilica dispute’ in Milan in 386, in which Ambrose successfully resisted the appropriation of the Basilica Portiana (probably San Lo- renzo) for use by the imperial court at Easter, highlight the stakes of competition over ritual prece- dence inside the church.349 This same tension would bring Ambrose into conflict with Theodosius during his stay in Milan.350 Relationships between church and court were not universally hostile, 345 Neil McLynn, ‘The Transformation of Imperial Churchgoing in the Fourth Century’, in Approaching Late Antiquity: The Transformation from Early to Late Empire, ed. Simon Swain and Mark Edwards (Oxford: Oxford University Press, 2004), 235– 70. 346 Meaghan McEvoy, Child Emperor Rule in the Late Roman West, AD 367–455, Oxford Classical Monographs (Oxford: Oxford University Press, 2013), 23–47. 347 McLynn, Ambrose, 230–32; Dey, Afterlife, 75. Rededication: Paulinus, Vita Ambrosii, 32.2–33.4, ed. A. A. R. Basiaensen, Italian trans. Luca Canali, in Vite dei Santi, ed. Christine Mohrmann, vol. 3, Scrittori greci e latini (Fondazione Lorenzo Valla: A. Mondadori, 1975), 94–96. Inscription: Ernst Diehl, ed., Inscriptiones latinae christianae veteres, vol. 1, rev. ed. (Berlin: Weidmann, 1961), no. 1801, 352: exultat hunc tumuli esse locum, quem pius Ambrosius signauit imagine Christi, marmoribus Libycis fida Serena polit, coniugis ut reditu Stiliconis laeta fruatur, germanisque suis, pignoribus propriis. 348 Kinney, ‘Dating’; Krautheimer, Three Christian Capitals, 88. 349 See: Ambrose, Epistulae, 75–76, in Sancti Ambrosii Opera, ed. Michaela Zelzer, Corpus Scriptorum Ecclesiasticorum Latinorum 82/3 (Vienna: Verlag der Österreichischen Akademie der Wissenschaften, 1982), 74–125; Paulinus, Vita Am- brosii, c. 13, 68–70 For the identification of the Portiana with San Lorenzo: Kinney, ‘Dating’, 104–06. 350 McLynn, ‘Imperial Churchgoing’, 263. 98 however. Peter Chrysologus, for example, seems to have been a collaborator of the imperial court of Galla Placidia and Valentinian III in Ravenna and a number of his sermons imply the presence of the court among the congregation.351 What was the purpose of these rituals and their architectural framing? A common interpreta- tion is exemplified by Hendrik Dey’s recent work on the persistence of a late ancient ‘urban habit’ in the Early Middle Ages. The urban public—the ‘teeming and often very vocal masses’, in Dey’s words—might be a necessary audience, but their role in his analysis is limited to observance and ac- clamation.352 As he argues elsewhere, urban public ritual took place, in front of thronging onlookers whose role was typically limited to manifesting consent through chanted litanies and acclamations. The decision making quite literally occurred behind closed doors, among a tiny group of political actors, leaving to public proces- sional ceremony the role of proclaiming and enacting the authority of the decision makers and imbuing them with an aura of legitimacy enhanced by the spectacle of mass compliance.353 Of course, Dey is not incorrect to note the power differential between the emperor and the crowd – and his synthesis of architectural, archaeological, visual, and historical evidence makes a number of salutary contributions. But, as Dey would have it at least, the ritual itself is of little social consequence: it serves simply as an accoutrement of political power which exists irrespectively. The potentate pa- rades, the masses comply. 351 Vincenza Zangara, ‘Una predicazione alla presenza dei principi: la Chiesa di Ravenna nella prima metà del sec. V’, Antiquité Tardive 8 (2001): 265–304. Chrysologus was apparently represented prominently in the mosaics of San Giovanni Evangelista, founded by Galla Placidia in Ravenna. The unique program combined portraits of the Valentinianic-Theodo- sian dynasty with a depiction of the augusta’s rescue from a shipwreck by St John. On the mosaics: Agnellus, LPRa, cc. 27 and 42, pp. 174, 202; Friedrich Deichmann, Ravenna. Hauptstadt des spätantiken Abendlandes, vol. 2, bk. 1 (Wiesbaden: Steiner, 1974), 107–24. 352 Dey, Afterlife, 13. As Anderson notes, ‘[Dey’s] “crowds” and “throngs” regularly “cheer” and “teem” but seem never to work or build, mock or riot. […] There were other ways to inhabit a portico than “dutifully assembled”: proud in the product of one’s own labor, or in the wares of one’s shop; or alienated and afeared of state violence’: Benjamin Anderson, ‘In Search of Late Roman Porticoes’, Review of The Afterlife of the Roman City: Architecture and Ceremony in Late Antiquity and the Early Middle Ages, by Hendrik Dey, Journal of Roman Archaeology 29, no. 2 (2016): 971–74. 353 Hendrik Dey, ‘From “Street” to “Piazza”: Urban Politics, Public Ceremony, and the Redefinition of platea in Communal Italy and Beyond’, Speculum 91, no. 4 (2016): 938. 99 Part of the problem are the inherited perspectives of the source material. For observers with aristocratic sensibilities, or for an orator speaking at the court, it was axiomatic that imperial authority was per se legitimate and did not require the acclamation of the urban crowd to be perceived as such. We might suspect that unrest dismissed as mindless or trivial violence of the crowd by our sources were often expressions of real and substantive discontent.354 The same spaces in which imperial au- thority were legitimised could be transformed into sites of contestation and protest; imperial control of these spaces was not absolute. We can see this already in an episode of Ammianus Marcellinus’ Res Gestae, which describes tensions mounting in Rome in 355 against the administration of Leontius, the urban prefect.355 Ammianus dismisses the causes as trivial—the arrest of a popular charioteer, an al- leged scarcity of wine, and the general restiveness of the common people—but the results were grave, including a physical attack on the prefect.356 The mob assembled ‘at a much-frequented spot called the Septemzodium, where the emperor Marcus built a Nymphaeum in a pretentious style’.357 Despite the entreaties of his staff, Leontius confronted the crowd in person, identified the ring leader, a certain Peter Valvomeres (who was arrested and tortured), and dispersed the crowd.358 Two details are striking in Ammianus’ anecdote. Firstly, a shortage of wine, though sounding inconsequential to the modern ear, possessed a greater political valence in the context of fourth-century Rome. Wine had been guar- anteed as part of the annona, the supply guaranteed to Rome’s citizens, since the reign of Aurelian; even if Ammianus’ statement does not refer to a wider shortage of other foodstuffs, the failure to 354 E.g. Alan Cameron’s dismissal of the circus factions of Rome and Constantinople as analogous to traditional football hooligan firms: Alan Cameron, Circus Factions: Blues and Greens at Rome and Byzantium (Oxford: Clarendon Press, 1976), e.g. 271, 273, 277. Cf. Peter Bell, arguing that factional violence reflects the formation of group identities outside of official institutional structures: Peter Bell, Social Conflict in the Age of Justinian: Its Nature, Management, and Mediation (Oxford: Oxford University Press, 2013), 119–60. 355 PLRE 1:503. 356 Amm. Marc., Res Gestae, XV.7.2–3, 1:55: Prima igitur causa seditionis in eum concitandae vilissima fuit et levis. Philoromum enim aurigam rapi praeceptum, secuta plebs omnis, velut defensura proprium pignus, terribili impetu praefectum, incessebat ut timidum […] diebusque paucis secutis cum itidem plebs excita calore quo consuevit, vini causando inopiam, ad Septemzodium conuenisset 357 Ibid., ad Septemzodium conuenisset, celebrem locum, ubi operis ambitiosi Nymphaeum Marcus condidit. Trans: Hamilton, 79. 358 Ibid., XV.7.3–4, 1:55–56. 100 provide the annona was perceived as a violation of traditional rights.359 Secondly, the location chosen for their protest was carefully chosen. The Septizodium had been constructed by Septimius Severus as an outer façade for the palace on the Via Sacra (Fig. 77).360 While the architectonic conjunction of palace and city staged consent and legitimation when animated by public ritual, that same armature could be transformed into a site of dialogue, negotiation, resistance – and, as here, eventual violence. The afterlife of imperial ritual From the games thrown by Theoderic in the Circus Maximus in 500, fondly remembered by the Anon- ymus Valesianus, to the Lombard king Adaloald (616–628) being raised as co-ruler of the Lombards in the circus of Milan, the ritual armatures of the Roman emperors presented significant opportunities for post-Roman rulers.361 Ritual displays of romanitas are well documented in the post-Roman regna, and their forms are generally recognised to be an important thread of continuity binding together the Roman and post-Roman worlds. For example, the adventus, the ritualised processional entry, can be traced in various forms through the Early Middle Ages and beyond. Gregory of Tours provides a particularly clear example. Travelling from his favoured residence of Chalon-sur-Saône to Paris, Gun- tram (561–592) stopped at his nominal seat, Orléans.362 Arriving on the feast of Saint Martin (4 July) Gregory records his reception by the townsfolk: A huge crowd of people came out to meet him, carrying flags and banners and singing songs in his praise [laudes]. From here the language of the Syrians, here that of the Latins, here even that of the Jews resounded in different songs of praise, saying: ‘Long live the King! And may his rule continue over his peoples for innumerable years!’ The Jews, in fact, who were seen to have taken part in these acclamations, were shouting: ‘Let all peoples adore you and bow the knee to you and be subdued by you!’.363 359 Paul Erdkamp, ‘The Food Supply of the Capital’, in The Cambridge Companion to Ancient Rome, ed. Paul Erdkamp (Cam- bridge: Cambridge University Press, 2013), 267–69. 360 Septimius Severus’ patronage is confirmed by the dedicatory inscription, recorded in the inscription catalogue attached to the Einsiedeln Itinerary: Gerold Walser, ed., Die Einsiedler Inschriftensammlung und der Pilgerführer durch Rom: Codex Einsidlensis 326 (Wiesbaden: Steiner, 1987), no. 30, 33. 361 Anon. Val. (II), XII.67–69, 19–20; Paul. Diac., Hist. Lang., IV.30, 127. 362 On residences and villae in the Frankish regnum, see below, Chapter 4. 363 Greg. Tur., Hist., VIII.1, 370: Processitque in obviam eius inmensa populi turba cum signis adque vixillis, canentes laudes. Et hinc lingua Syrorum, hinc Latinorum, hinc etiam ipsorum Iudaeorum in diversis laudibus variae concrepabat, dicens: ‘Vivat rex, regnumque eius 101 What Gregory describes here is a classic adventus: the citizens meeting Guntram outside the gates and escorting him into the city while chanting formal acclamatory formulas, the ‘praises’ or laudes – a ritual ensemble that was activated by taking place within an architectural frame curated for this purpose.364 As Michael McCormick showed clearly in his seminal Eternal Victory, for example, the ritualised back- ground of Roman political culture provided the basis of the languages of early medieval kingship through the Carolingian period and beyond.365 Dey, building on the foundations lain by McCormick and others, has recently argued that such rituals reflected a continued ‘urban habit’ in the post-Roman West, in which the continued allure of ritual displays of romanitas spurred the maintenance of its archi- tectonic staging, even as other characteristics of Roman urbanism fell into abeyance.366 Even if the forms of the rituals themselves were persistent through the Early Middle Ages, their architectural framing was not. The settings and audiences of these rituals, as well as the ideals they embodied, were drastically transformed over the course of the sixth and seventh centuries: from being performed before an urban public, to being targeted towards smaller, elite audiences.367 From the seventh century, it was more usual for the rituals through which authority was legitimised to be performed away from the traditional centres of Roman power and among a smaller clique of aristo- cratic participants. As Halsall has recently argued, fundamental distinctions existed between Roman and early medieval ritual practice, despite the strength of continuity between them: the ‘liturgification’ of kingship, the decline of urban populations, and the militarisation of aristocratic culture were all factors in the transformation of royal ritual.368 That is to say, the late-ancient urban paradigm that in populis annis innumeris dilatetur’. Iudaei vero, qui in his laudibus videbantur esse participes, dicebant: ‘Omnes gentes te adorent tibique genu flectant adque tibi sint subditi’. Translation is here my own, adapting that of Thorpe to clarify the sense: Gregory of Tours, The History of the Franks, trans. Lewis Thorpe, Penguin Classics (London: Penguin, 1974), 433. 364 Dey, Afterlife, 166–67. On acclamations, see: Charlotte Roueché, ‘Acclamations in the Later Roman Empire: New Evi- dence from Aphrodisias’, JRS 74 (1984): 181–99. 365 McCormick, Eternal Victory. 366 Dey, Afterlife, esp. 127–220. 367 McCormick, Eternal Victory, 390–95. 368 Guy Halsall, ‘The Decline and Fall of the Ancient Triumph’, in Der römische Triumph in Prinzipat und Spätantike, ed. Fabian Goldbeck and Johannes Weinand (Berlin: De Gruyter, 2017), 562–65. 102 combined palace and city for ‘public’ ritual display was specific to the social structures of the Later Roman Empire.369 *** By the seventh century legitimate authority was no longer produced in the encounter of sovereign and crowd in an urban setting in the same way it had under the Roman emperors. Rather, the ‘public’ setting of political ritual—and thus the concept of legitimate authority which it expressed—had fun- damentally transformed: from before an urban crowd at the intersection of palace and city to the palace’s interior among a smaller clique of aristocratic participants. The early medieval palace, and the rituals which took place within it, defined new sets of social relations which, though not essentially separate to the Roman period, were nonetheless distinct. With this in mind, we can turn to the spaces that, together, constituted the early medieval palace. III. The Audience Hall A feature common to every palace we have encountered so far is the audience hall. Audience cham- bers of various kinds formed the architectural and social centrepiece of palaces in both the Roman period and after. The rituals associated with audiences were one side of a dialogue between ruler and elite in which the nature and extent of royal authority were (re)presented and negotiated. While early medieval sources only give us snapshots and vignettes of the rituals which accompanied audiences, it is nonetheless possible to glimpse at how their spatial experience was curated to distinguish ruler and subject. Analysis of the ‘audience hall’ as a distinct unit foregrounds the production of a space in which the speech-acts of the participants—both royal and elite—could be perceived as possessing authority. 369 Put in Lefebvre’s terms, as each social formation produces a set of spaces particular to it, the transformation of that society (in terms of the economic and ideological bases that supported it) necessarily entails the reorientation of its social space: Lefebvre, Production, 16–17 103 The spatial formation of the audience hall The audience hall represents a remarkably stable spatial ensemble in post-Roman Europe: longitudinal apsidal halls with thrones (often located on tribunals) are attested in all of the post-Roman kingdoms through to the Carolingian period and beyond. The strength of this continuity itself deserves com- ment. Continuities from the Late Roman Empire partly explain the prominence of the aula as space in which politics was conducted, but this Roman raw material was augmented and iterated over time with concepts drawn from other sources, including Christian doctrine. In addition to its archaeological and architectural consistency from the Late Roman period onwards, there is a similar continuity in its ritual use. Thus, for example, Sidonius Apollinaris provides an insight into the rituals of the Visigothic court at Toulouse under the king Theodoric II (r. 453– 466). Every morning, writes Sidonius, the king appeared before petitioners seated on a throne, flanked by guards arrayed a short distance to either side, ‘outside the curtains but inside the balustrade’.370 Isidore similarly tells us that the Visigothic king Leovigild appeared to his subjects in the palace of Toledo seated on a throne, ‘clothed in royal robes’. 371 The bishops of Clermont and Toledo were far from alone in presenting post-Roman rulers in this light. Procopius, for example, comments that the Vandal king Gelimer had ruled from a throne in the palace in Carthage.372 Venantius Fortunatus de- scribed a journey he took with the royal party from Metz to the palace of Andernach, where Brunhilda and her son Childebert II sat enthroned for a banquet in the royal hall.373 The enthronement of the ruler in the hall for the purpose of dispensing gifts and receiving oaths is present in the Old English poetic tradition and represented on later stone sculpture (Fig. 78).374 Jumping ahead to the Carolingian 370 Sidonius Apollinaris, Epistula, I.2.4, in Sidoine Apollinaire, ed. and French trans. André Loyen, vol. 2 (Paris: Les Belles Lettres, 1970), 5–6: Reliquum mane regni administrandi cura sibi deputat. Circumsistit sellam comes armiger; pellitorum turba satellitum ne absit, admittitur, ne obstrepat, eliminatur, sicque pro foribus immurmurat exclusa uelis, inclusa cancellis. 371 Isidore, Historia de regibus Gothorum, Vandalorum et Suevorum, in Chronica minora, 2:288: primusque inter suos regali veste opertus solio resedit: nam ante eum et habitus et consessus communis ut populo, ita et regibus erat. 372 Proc., Wars, III.20.21, 1:399: ἐς τὸ Παλάτιον ἀναβὰς ἐν τῷ Γελίμερος θρόνῳ ἐκάθισεν. 373 Ven. Fort., Carm., X.9, 3:86: Denique dum praesunt reges in sedibus aulae | ac mensae officio prandia festa colunt. 374 Stoner, ‘Heaven and Hall’. 104 period, Einhard’s satirical and polemical description of Childeric III (743–751) subverts expectations of the appearance of the king in the audience hall to present him as a figure of ridicule: In fact, nothing was left to the king except to be happy with the royal title and to sit on his throne with his flowing hair and long beard and to behave as if he had authority, hearing envoys who came from all over and dismissing them with answers that he had been taught, or even commanded, to give.375 That Carolingian kings greeted subjects and envoys enthroned in the hall is taken for granted by con- temporary observers.376 Needless to say, numerous other examples could also be adduced. Roman analogues for this ensemble are not difficult to locate. It is hard not to be reminded of the consular portrait of Constantius II, for example, as preserved in the Barberini manuscript of the Chronography of 354 (Fig. 79) or the missorium of Theodosius I (Fig. 80) for the enthronement of the emperor, elevated on a platform with a footstool, in an architectural frame which could reference the apse of the hall. In the former image we detect hints of the curtains (vela) mentioned by Sidonius, which served to divide space and to control lines of vision, and in the latter are witnessed the armed guards which regulated access to the ruler’s presence. It would be tempting to recognise in the staging of the audience a conscious emulation of the Roman emperor, along with its manifested significations, and certainly in the fifth and sixth centuries this must have been an important motivation. The sources and stimuli combined in the spatial production of the ‘audience hall’ were not strictly ‘imperial’, however. Ritual spaces like those described by Sidonius recall not only imperial au- diences, but a more general mode for the manifestation of authority in Late Antiquity. In the sixth- century Rossano Gospels, for example, Pilate is represented as a Roman magistrate, flanked by two at- tendants carrying standards bearing imperial portraits, and seated on an elevated throne (Fig. 81). The 375 Einhard, Vita Karoli, c. 1, 8: Neque regi aliud relinquebatur, quam ut regio tantum nomine contentus crine profuso, barba summissa, solio resideret ac speciem dominantis effingeret, legatos undecumque venientes audiret eisque abeuntibus responsa, quae erat edoctus vel etiam iussus, ex sua velut potestate redderet Trans.: Noble, 24. 376 E.g.: Notker, Gesta Karoli; II.6, 56–57; Ermold, In honorem Hludowici, II.5–6, 24; Thegan, Gesta Hludowici imperatoris, c. 9, in Thegan, Die Taten Kaiser Ludwigs. Astronomus, Das Leben Kaiser Ludwigs, ed. and German trans. Ernst Tremp, MGH. SS rer. Germ. 64 (Hannover: Hahnsche Buchhandlung), 190. 105 curvature of crowd surrounding the throne is continued in a thick grey line which arcs over Pilate’s head, an oblique reference to the common location of the tribunal in an apse. This arrangement also recalls the placement of the bishop’s cathedra in the apse of the Christian basilica, behind the altar and in front of the synthronon for the seated clergy. That apsidal reception halls were important structural elements of late ancient episcopia, including those of Ravenna, Poreč, Geneva, and possibly also Grado, show that this basic schema was not limited to agents of ‘secular’ authority per se.377 More generally, and as we saw above, apsidal aulae were a common feature of elite residences throughout the Medi- terranean, providing a spatial framework for the reception of clients – the subordination of one to the other underscored by the dominus being seated alone in the apse, visually separated from the rest of the hall by decorative motifs in the pavement or, more rarely, it being raised by a step (as at Piazza Armerina, for example).378 As we saw in previous chapter, the imperial palace existed on an architec- tural spectrum which included other forms of elite domestic architecture. The reproduction of this architectural feature need not, therefore, be seen as evidence for a ‘trickle-down effect’ from the im- perial palace into elite circles, but rather a diffuse language for asserting status, prestige, and hierar- chy.379 In praesentia regis: the experience of the royal audience These available meanings were mobilised through ritualised behaviours and practices in social space. Although we do not possess ‘scripts’ for the rituals of the audience hall, it is nonetheless possible to sketch out how such practices created meaning in the context of the social space of the palace.380 The 377 Generally, see: Miller, Bishop’s Palace, 16–53 (Late Antiquity), 54–85 (to the tenth century), and above, Chapter 1. For Geneva: Bonnet and Peillex, eds, Les fouilles. 378 Sfameni, Ville residenziali, 88. 379 Ibid., 141, and above, Chapter 1. 380 Useful in thinking through these issues has been: Josiane Barbier, ‘Un rituel politique à la cour mérovingienne : l’audi- ence royale’, in L’Audience. Rituels et cadres spatiaux dans l’Antiquité et le haut Moyen Âge, ed. Jean-Pierre Caillet and Michel Sot, Textes, images et monuments de l’Antiquité au Moyen Âge (Paris: Editions Picard, 2007), 241–63. 106 spatial experience of the audience hall acted out distance and difference between ruler and petitioner, thereby legitimising royal action as authoritative. First impressions were paramount. The petitioner, on being admitted into the royal presence witnessed a spectacle of royal authority. Notker captures this in an amusing anecdote in which legates from Constantinople were led through the palace of Aachen, abasing themselves in prokynesis errone- ously before, in turn, the count of the stable, the count of the palace, the master of the royal table, and the chamberlain, before they were finally led into Charlemagne’s presence.381 As shown above, Sidonius already gives us an idea of the distance imposed between the ruler and the petitioner by this spatial arrangement: on the one hand, the king seated in the apse of the hall, flanked by guards and senior officers of the court, the petitioner before them. While a stereotyped image, the large miniature of Charles the Bald placed at the end of his First Bible, recording its presentation in 846, also provides an insight into this spatial arrangement (Fig. 82). The king, coronate and seated on a high-backed throne, gestures in greeting or speech to the monks of Tours bearing the manuscript. He is flanked to either side by two nobles, clad in chlamydes, and at the outer edge, two soldiers.382 The miniature freezes in time two separate ritual vignettes: the presentation itself and, in the lower register, laudes sung by the monks, led by Vivian of Tours – in Dutton and Kessler’s words ‘a bold and precocious attempt to capture on parchment the passing of ceremonial and symbolic time’.383 These rituals were not solely gestural, but also had verbal components. Titulature in written sources reflected expected formulae of verbal address. In ‘official’ media, for example, the Ostrogothic king Theoderic adopted some aspects of Roman imperial titulature, notably the moniker ‘Flavius’— by this time serving more as an honorific than an actual nomen gentilicum—as well as other imperially- 381 Notker, Gesta Karoli, II.6, 56–57. 382 On this MS, see above all: Paul Dutton and Herbert Kessler, The Poetry and Paintings of the First Bible of Charles the Bald (Ann Arbor, MI: University of Michigan Press, 1997). Useful also is Garipzanov’s discussion of the corpus of royal images in Bibles associated with Charles the Bald: Garipzanov, Symbolic Language, 235–58 (On the First Bible: 244–48). 383 Dutton and Kessler, Poetry and Paintings, 78. 107 flavoured titles such as victor gentium, as on the Senigallia Medallion (Fig. 83), but he is otherwise con- sistently styled as rex or princeps, never augustus or imperator.384 Correspondence addressed to Theoderic use a variety of modes of address. Pope Gelasius refers to him, for example, as ‘Your Magnificence’ and ‘Most Excellent King’; the acts of the fourth synod convoked to adjudicate the Laurentian schism (23 October 501) refer to him as ‘the most glorious king Theoderic’.385 It is easy to forget that these texts were not encountered in the first instance as ‘texts’ as such, but as oral formulae which were read out to the recipient. We see a similar phenomenon in Gregory of Tours’ Histories. Gregory relates that a priest addressed the king Charibert in the palace at Paris with: ‘Greetings, glorious king! For the Apostolic Seat sends most abundant greetings to your eminence’.386 Later, Gregory reports that he himself used the same formula to address Guntram in the palace at Chalon-sur-Saône: ‘Your most glorious nephew Childebert sends most abundant greetings to you, o famous king.’387 When uttered in the ritual space of the audience hall, titles distanced and distinguished the ruler from the other participants, and by so doing they signalled a recognition of the ruler’s authority. Viewed in this light, we should not be so quick as to dismiss the inscriptions erected on the Via Appia near Terracina that refer to Theoderic as semper augustus as being somehow in error, but rather reflecting exchanges be- tween court and Italo-Roman elite concerning the nature of the king’s authority.388 384 The ambiguity is pointed to by Procopius, who notes that Theoderic never took the purple, only ever referring to himself as rex, but he behaved as an emperor: Proc., Wars, V.1.26–29, 2:8–9. The praeceptiones sent to the Roman synods of 501, for example, style him Fl. Theodericus rex. Acta synhodorum habitarum Romae, nos. 2, 4, in Cass., Var., 420, 424. For a reading of the Senigallia Medallion in terms of imperial exemplars, see: Jonathan Arnold, ‘Theoderic’s Invincible Mus- tache’, JLA 6, no. 1 (2013): 152–83. For ‘Flavius’ as a title: Alan Cameron, ‘Flavius: A Nicety of Protocol’, Latomus 47 (1988): 26–33. 385 Epistulae Theodericianae Variae, nos. 1, 3, 4, 6, in Cass. Var., 389–91; Acta synhodorum habitarum Romae, no. 6, 426: ex praecepto gloriosissimi regis Theoderici. On Theoderic’s titulature: Arnold, Theoderic, 88–91. 386 Greg. Tur. Hist., IV.26, 158: regis praesentiam adiit, haec affatus est : ‘Salvae, rex gloriosae. Sedis enim apostolica eminentiae tuae salute uberrimam mittit.’ Barbier, ‘Un rituel politique’, 258. 387 Greg. Tur., Hist., IX.20, 434: Salutem uberrimam mittit tibi gloriosissimus nepus tuus Childeberthus, o inclite rex. 388 CIL X, 6850–52. Though not ‘official’ statements per se, these inscriptions were erected by the former consul, urban prefect, and praetorian prefect Caecina Mavortius Basilius Decius: PLRE, 2:349. McCormick comments that these inscrip- tions reflect ‘senatorial wishful thinking’: McCormick, Eternal Victory, 278–79. As Arnold argues, that Theoderic and his sucessors ‘never openly declared themselves imperatores or Augusti should not suggest that they or their subjects understood their position to be otherwise’: Arnold, Theoderic, 77. 108 Gregory also provides an insight into an audience which did not go to plan, an episode which reveals much concerning contemporary expectations of behaviour in the audience hall. A large lega- tion was sent by Childebert to Guntram in around 584 following a conflict between the two kings.389 The party was met with a frosty reception. As each envoy approached the king in turn, Guntram berated them and refused their requests, accusing Guntram Boso, for example, of being a traitor for his support of the usurper Gundovald.390 When one of the legates replies mockingly to one of the king’s insults (that Guntram Boso was the son of a miller), they took their leave.391 Their parting shot reveals the heightened stakes: ‘we are mindful of the fact that the axe is still ready and waiting which split open the heads of your brothers. One day soon it will split open your head, too.’392 Gregory notes that the envoys then left ‘in disgrace’ (cum scandalum), implying that they had transgressed the norms of proper behaviour.393 In response to the envoys’ slight, the king had them pelted with manure, mouldy hay, and other foul substances as they departed.394 What Gregory here describes is that the disruption of normative behaviours brought danger to the legates. Guntram Boso, responding to the allegation of being a traitor, states the situation plainly: ‘You are a ruler […] and you sit as king upon a royal throne. That is why no one dares reply to what you say’.395 The usurpation of the authority of the king in his own space could not go unan- swered, but rather than physical violence (which could bring repercussions), Guntram orders instead 389 Gregory refers to it as a placitum. Among the legates were Bishop Egidius of Reims and the the duces Guntram Boso and Sigivald, as well as ‘many others’: Greg. Tur., Hist., VII.14, 334–35: Igitur, adveniente placito, directi sunt ad Childebertho rege Egidius episcopus, Gunthchramnus Boso, Sigyvaldus et alii multi ad Gunthchramno regem. 390 Ibid., 335: O inimici regionis regnique nostri, qui propterea ante hos annos Orientem adgressus es, ut Ballomerem quendam – sic enim vocabat rex Gundovaldum – super regnum nostrum adduceris, semper perfide et numquam custodiens quae promittis! Guntram Boso, no relation of the king Guntram, had travelled to Constantinople at some point between 578 and 582 and allegedly encouraged the exiled pretender, Gundovald, to assert his royal claim with Byzantine backing. See: PLRE, 3A:571–74. 391 Greg. Tur., Hist., VII.14, 336. 392 Ibid.: Scimus, salvam esse securem, quae fratrum tuorum capitibus est defixa. Celerius tuum libravit defixa cerebrum. Trans.: Thorpe, 398. 393 Ibid.: Et sic cum scandalum discesserunmt. Thorpe translates this as ‘Thereupon they departed, mad with anger’ (Thorpe, 398). However, this does not capture the sense of scandalum as a moral transgression, an insult, or a loss of reputation. 394 Greg. Tur., Hist., VII.14, 336: Tunc rex his verbis succensus, iussit super capita euntium proici aequorum stercoral, puterfactas astulas, paleas ac faenum putridine dissolutum ipsumque foetidum urbis lutum. 395 Ibid., 335: ‘Tu’, inquid, ‘dominus et rex regali in solio resedis, et nullus tibi ad ea quae loqueris ausus est respondere.’ 109 the public humiliation of the envoys – manifested in the besmirching of the outward signs of their status, their clothing. Yet in this episode Gregory’s words also betray that the unspoken rules of eti- quette applied to the king as well as the visitor. Gregory himself reserves criticism of Guntram, placing blame instead on the impertinence of the envoys, but the dux Guntram Boso’s recourse to offering a trial by combat in response to the allegation he supported Gundovald shows not only that words carried weight in this environment, but also that slights could not ignored – even when they issued from the throne.396 *** The rituals of the audience hall represented to its participants the distance between the king and the rest of the political community, not only through its mise-en-scène, but through the gestures and verbal responses which were expected within its bounds.397 The social roles imposed by behavioural norms were not limited to the petitioners, however. The king in the audience hall was expected to listen to advice, to behave morally, not to enrich himself at the expense of others, and to judge legal cases fairly, royal virtues made plain by Venantius Fortunatus in his panegyric to Sigibert I of Austrasia, performed before the court at Metz in 566.398Among the principal such expectations was generosity and gift giving. We turn next, therefore, to an architecturally less defined spatial unit: the treasury. IV. The Treasury In 811, with his health ailing, Charlemagne drew up his will.399 This document, quoted by Einhard in full, provides for the ‘description and division’ (descriptio et divisio) of his treasury: two thirds were to be reserved for the senior metropolitan sees of Francia and Italy, a further third was to be shared equally 396 Ibid., 336: Tunc, o rex piissime, ponens hoc in Dei iudicium, ut ille discernat, cum nos in unius campi planitiae viderit demicare. 397 Barbier, ‘Un rituel politique’, 262. 398 Ven. Fort., Carm., VI.1. 85–98, vol. 2, 47–49. 399 Einhard, Vita Karoli imperatoris, c. 33, 37–41. Einhard states that Charlemagne had intended to include his daughters and illegitimate sons in his inheritance in a new will which was unfinished on his passing: ibid., 37: Testamenta facere instituit, quibus filias et ex concubinis liberos ex aliqua parte sibi heredes faceret, sed tarde inchoate perfici non poterant. The document which Einhard quotes is concerned purely with the division of property and donations to the Church. 110 among various recipients.400 What is striking in reading the will is the treasury’s dual materiality and spatiality. The various shares were to be allocated from ‘all his goods and possessions, in gold and silver and gems and royal ornament, which could be found in his chamber on that day’.401 These objects were of great variety: in addition to objects fashioned of precious metals, there were also arms, clothing, furnishings, numerous textiles, leather goods, saddles, a library of books, as well as three ornate tables in gold and silver which were singled out for special attention.402 Despite lacking a definitive architectural form—it is not possible to archaeologically identify a palace treasury in the same way we could an audience hall—the treasury was an essential component of the palace and for the manifestation of authority within it. In Smith’s terms, the treasury is a space through which politics was practised, a spatial frame for material exchanges that mediated social rela- tions.403 The objects collected in Charlemagne’s treasury existed within networks of meaning of a much broader nature than his personal wealth. For one, the division of the goods was not a private matter, but a public one, enacted in the presence of his friends and the palace officials, and subscribed by fifteen counts and fifteen members of the clergy, some of whom were named as recipients in the division of the treasure.404 Treasure was not just for hoarding, however. It was effective only insofar as it was used, displayed, and mobilised in gift exchange; to hoard treasure was a sign of avarice.405 The treasury’s contents were thus constantly changing: the will only accounted for what was to be found in the treasury on that particular day.406 The spatial frame of the treasury thus created connections between the assembled objects such that it was conceptualised as a discrete discursive object in itself 400 Einhard Vita Karoli imperatoris, c. 33, 37–40. 401 Ibid., 38: omnem substantiam atque suppellectilem suam, quae in auro et argento gemmisque et ornatu regio in illa, ut dictum est, die in camera eius poterat inveniri. Translation here is my own. 402 Ibid., 39–40. 403 See above, 89–90. 404 Einhard, Vita Karoli imperatoris, c. 33, 37: coram amicis et ministris suis. 405 Ian Wood, The Merovingian Kingdoms, 450–751 (London: Longman, 1994), 215. See e.g.: Greg. Tur., Hist., VI.28, 295; Fred., Chron., IV.60, 151; Isidore, Historia Gothorum, c. 51, 288. 406 Einhard, Vita Karoli imperatoris, c. 33, 38: quae in illa die in camera eius inventa est. 111 which was expressive of social norms about legitimacy, authority, and proper social behaviour. The ritualised public exchange of these objects forged relations between political actors drawn from a narrow social circle and could thus be carefully staged.407 The performance of exchange represented hierarchies of power and status to those assembled, but they also imposed identities on both donor and recipient governed by communal norms and expectations of behaviour.408 The treasury as social space An episode in the early eighth-century Liber Historiae Francorum vividly demonstrates the entanglement of the treasury with contemporary ideas surrounding authority and legitimacy. Theuderic II of Bur- gundy (595–613), devastating the Rhineland after defeating his brother, Theudebert II of Austrasia (595–612) in battle at Zülpich in 612, received a delegation from the ‘people’ (populus)—probably referring, based on what follows, to members of the Austrasian aristocracy—begging for their lands to be spared.409 Theuderic agreed on the condition that they brought his brother to him, dead or alive. Meeting with the king inside Cologne, they reported Theuderic’s demands as the return of his portion of their father’s treasure. The message having been delivered, Theudebert ‘went into the palace treas- ury with his men. And once the chests of the treasury had been opened and while he searched for the royal ornaments, one of these men, his blade drawn, struck the king from behind and, having taken his head, they carried it around the walls of Cologne.’410 407 Florin Curta, ‘Merovingian and Carolingian Gift Giving’, Speculum 81, no. 3 (2006): 698–99; Janet Nelson, ‘The Setting of the Gift in the Reign of Charlemagne’, in The Languages of Gift in the Early Middle Ages, ed. Wendy Davies and Paul Fouracre (Cambridge: Cambridge University Press, 2010), 116–148. 408 E.g. Barry Schwartz, ‘The Social Psychology of the Gift’, American Journal of Sociology 73, no. 1 (1967): 1–11. 409 Liber Historiae Francorum, c. 38, in Fredegarii et aliorum Chronica. Vitae sanctorum, ed. Bruno Krusch, MGH. SS rer. Merov. 2 (Hannover: Hahn, 1888), 308: Theudericus itaque terra Riboariense succendens ac devastans, populus ille sub eius manibus se tradedit, dicens: 'Parce, domine rex, nos et terra nostra, – iam tui sumus, – noli amplius delere populum.' 410 Ibid.: ille cum eis in palacium thesauris sui intravit. Cumque apertas arcas thesaurorum, ornamenta requireret, unus ex eis, abstracto gladio, retro eum in cervice percussit, et accepto caput eius sustullerunt per murum civitatis Coloniae. Theudericus namque haec videns, ipsam civitatem adprehendens, thesauris magnis acceptis. Cf. ‘Fredegar’, who has Theduebert escape from Cologne before being captured by Theuderic’s chamberlain, Berthar and thence sent to Chalon-sur-Saône: Fred., Chron., IV.38, 139. 112 Underpinning the Liber’s presentation of Theuderic’s victory is an understanding that the treas- ury was both a physical and social object. Inside Theudebert’s palace in Cologne was the legitimate basis of his authority and its transfer into his brother’s possession is thus equated with the takeover of his kingdom; his men, having already transferred their loyalty to Theuderic, also handed over his treasure. The elision of the treasury with the loyalty of the court elite (and ultimately with royal au- thority itself) was not unique to the author of the Liber. Gregory of Tours, for example, pays close attention to the location of the royal treasury during his narrative and its transfer in moments of royal succession.411 That treasure, its possession, and its inheritance is a consistent leitmotif of his Histories signals its potency as a vehicle for ideologies of gender, power, and familial politics.412 Contention over treasure was not always violent. ‘Fredegar’ describes a ‘hearing’ (placitum) between Sigibert III of Aus- trasia and Clovis II of Neustria-Burgundy over the fair division of Dagobert I’s treasure on his death in 639.413 On the Austrasian party’s return to Metz, the treasure was inventoried, or ‘described’ (dis- cribetur), as in Charlemagne’s will.414 The conceptual elision between the royal treasury and royal authority is especially clear in mo- ments of defeat. For many authors, ‘the dispersal [of the treasury] was also a symbolic dispersal of the power, and indeed the existence, of the conquered enemy’.415 John of Biclaro, for example, equates the disappearance of the Gepid kingdom with the capture of the royal treasure by the Byzantines.416 Similarly, the murder of the Lombard king Alboin inside the palace by a faction led by his queen, 411 Greg. Tur., Hist., II.37, 88 (Clovis takes the Visigothic treasure from Toulouse); II.40–41, 91–92 (Clovis defeats Sigibert of Cologne and Chararic and takes their treasure); IV.20, 152 (Chlothar I inherits Childebert’s treasury); IV.22, 154 (Chilperic takes control of Chlothar’s treasury, held at Berny-Rivière); V.1, 194–95 (Sigibert dies and his treasure taken by his brother Childebert); VII.6, 329 (Guntram lays claim to the treasure of Charibert in its entirety). 412 Pauline Stafford, ‘Queens and Treasure in the Early Middle Ages’, in Treasure in the Early Medieval West, ed. Elizabeth Tyler (York: York Medieval Press, 2000), 61–82. 413 Fred. Chron., IV.85, 164. 414 Ibid.: Chunibertus et Pippinus hoc tinsaurum, quod pars fuit Sigyberti, Mettis facint perducere ; Sigyberto praesentatur et discribetur. 415 Timothy Reuter, ‘“You Can’t Take It With You”: Testaments, Hoards, and Moveable Wealth in Europe, 600–1100’, in Treasure in the Early Medieval West, 15. 416 John of Biclaro, Chronicon, c. 19, 63: Gepidarum regnum finem accepit qui a Longobardis prelio superati. Cuniemundus rex campo occubuit et thesauri eius per Trasaricum Arriane secte episcopum et Reptilanem Cuniemundi nepotem Iustino imperatori Constatninopolim ad integrum perducti sunt. 113 Rosamunda, and the subsequent capture of the royal treasure by the Byzantines, is marked as a disas- trous event which presaged both further conflict and political instability in the Lombard kingdom.417 Later, the Annales Mettenses Priores and other sources mark Charles Martel’s ascent to dominance in Francia with the capture of his father’s treasure from his stepmother, Plectrud.418 Taking these examples together, we see clearly that the material wealth of the treasury was tied up with notions of interpersonal loyalty and royal legitimacy. The treasury oscillated between a number of different conceptual poles in early medieval texts, creating a discursive space in which it could ultimately stand in for the health of the polity itself. With the dispersal of the treasury came the disso- lution of the material means by which an early medieval ruler could display their status and the eco- nomic incentives through which interpersonal power was exercised, themes to which we now turn. Rituals of gift exchange The importance of gift-exchange as a medium of symbolic communication has long been recognised in anthropological scholarship, but the line of inquiry pioneered by Marcel Mauss only began to exert a significant influence on studies of the Middle Ages from the 1970s and 1980s.419 Historical study of the medieval gift have produced numerous rich expositions of, for example, lay donations to religious institutions such as monasteries; by comparison, the operation of gifts within the political structures of the Early Middle Ages specifically has been thematised less strongly. Despite its importance in the establishment of social relations between rulers and aristocracy, detailed accounts of gift-giving are lacking in early medieval sources, probably a reflection of such exchanges being a usual and 417 Ibid., c. 23, 64. See also: Paul. Diac., Hist. Lang., II.29–30, 89–90; Agnellus, LPRa, c. 96, 260–64. 418Annales Mettenses Priores, a. 717, ed. Bernhard von Simson, MGH. SS rer. Germ. 10 (Hannover: Hahn, 1905), 25. Cf. Liber Historiae Francorum, c. 53, 327; Continuationes Fredegarii, c. 10, in Fredegarii et aliorum chronica, ed. Krusch, 174. 419 Marcel Mauss, The Gift: The Form and Reason for Exchange in Archaic Societies, trans. W. D. Halls (London: Routledge, 1990). For a treatment of the ample scholarship on medieval practices of gift exchange, see: Arnoud-Jan A. Bijsterveld, ‘The Medieval Gift as Agent of Social Bonding and Political Power: A Comparative Approach’, in Medieval Transformations: Texts, Power, and Gifts in Context, ed. Esther Cohen and Mayke de Jong, Cultures Beliefs and Traditions 11 (Leiden: Brill, 2001), 123–56; Curta, ‘Gift Giving’, 671–77. 114 commonly-understood part of political life.420 For this reason, many of the details concerning the modulations of speech and gesture in gift-giving rituals are beyond our reach. Nonetheless, the social dynamics that motivated gift exchange, and therefore the social contexts in which they had meaning, are clearly perceptible. To give a gift was not an economic transaction as such, but a means of mani- festing hierarchical relations and affirming ties of loyalty between the ruler and the elite, particularly in moments of political crisis. The objects of exchange were intended for displaying aristocratic status. As Timothy Reuter argued, ‘noble gifts’, as he termed them, were sourced largely, if not exclusively, through warfare and ‘ran parallel to and largely independent of’ usual economic logics.421 We saw this clearly already in Charlemagne’s will: in addition to luxury objects of high craftsmanship, such as textiles or the silver tables, included in the inventory were seemingly more ‘workaday’ items such as arms and saddles. As Régine Le Jan has argued, weaponry was a prerequisite for the performance aristocratic masculinity in elite spaces, and richly decorated swords were especially prized. The Lex Ribuaria, for example, held that a sword and sheath (presumably decorated or otherwise embellished) was worth seven solidi against a wergeld, rather than the usual three.422 Hunting equipment and companions also counted among ‘noble’ gifts. Boniface, writing to King Æthelbald of Mercia in 745/746, sent with his letter, ‘as a sign of sincere affection and devoted friendship’, a hawk, two falcons, two shields and two lances.423 Of signal importance in this respect were horses. Ownership of horses became symbols of elevated social status by the seventh century; from this point the ‘mounted warrior’ starts to be more 420 Nelson, ‘Setting of the Gift’, 140. 421 Timothy Reuter, ‘Plunder and Tribute in the Carolingian Empire’, Transactions of the Royal Historical Society, 35 (1985): 75–94. Reprinted in Medieval Politics and Modern Mentalities, 231–50 (quote at 240). 422 Régine Le Jan, ‘Frankish Giving of Arms and Rituals of Power: Continuity and Change in the Carolingian Period’, in Rituals of Power: From Late Antiquity to the Early Middle Ages, ed. Frans Theuws and Janet Nelson, The Transformation of the Roman World 8 (Leiden: Brill, 2000), 282–9; Lex Ribuaria, 40.11, ed. Franz Beyerle and Rudolf Buchner, MGH. LL nat. Germ. 3/ 2 (Hannover: Hahn, 1954), 94: Spatham cum scoligilo pro septem solid. tribuat. Spatam absque scoigilo pro tres solid. tribuat. 423 Boniface, Epistulae, no. 69, in Die Briefe des heiligen Bonifatius und Lullus, ed. Michael Tangl, MGH. Epp. sel. 1 (Berlin: Weidmann, 1916), 142: Interea pro signo veri amoris et devote amicitiae direximus tibi accipitrem unum et duos falcones, duo scuta et duas lances; Curta, ‘Gift Giving’, 681. On rituals of hunting, see below. 115 frequently attested as an iconographic type – something witnessed clearly on, for example, the Repton Stone (Fig. 84) and the seventh-century Hornhausen reliefs (Fig. 85).424 Many of these objects origi- nated either as spoils or tribute after the conclusion of a conflict. Pepin the Short, for example, ex- tracted three hundred horses each year from the Saxons as tribute in 758.425 The arrival of the Avar treasure, captured in 796 by forces commanded by Pepin of Italy and Eric of Friuli, in Aachen in 796 made a large impression upon contemporaries. Einhard claimed that ‘All the treasure they [the Avars] had accumulated over a long time was pillaged, and the mind of man could not recall any war in which the Franks were so endowed with booty and wealth’.426 To give generously was a sign of a king’s care for his people and, ultimately his success. But gifts also flowed in the other direction. In both cases, the rituals of gift giving not only affirmed social relationships between ruler and aristocrat, but also manifested social hierarchies. While Pepin receiving horses from the Saxons was a sign of their submission, for example, to receive a gift could also man- ifest subordination. For example, the opulent gifts bestowed upon Harald Klak by Louis the Pious on his baptism in 826, which included a purple, gem-studded cloak, a fine sword, gold-worked clothing, and a crown, came with the subtext that Harald’s fragile rule in Denmark was under Frankish hegem- ony.427 On the other hand, the bestowal of gifts was also a sign of a generosity and thus legitimacy. Hincmar of Reims presented the proliferation of gifts in the royal household as a sign of orderly governance. The responsibility for the ‘annual gifts’ given to the palace officers (milites) lay with the 424 The date (and original form) of the Repton Stone is unclear. It cannot be associated with any particular king, despite its usual identification with Æthelbald of Merica. See: Catherine Karkov, The Art of Anglo-Saxon England, Boydell Studies in Medieval Art and Architecture (Woodbridge: Boydell Press, 2011), 102–04. Generally: Martin Biddle and Birthe Kjølbye- Biddle, ‘The Repton Stone’, Anglo-Saxon England 14 (1985): 233–92. The Hornhausen reliefs: Kurt Böhner, ‘Die Relief- platten von Hornhausen’, Jahrbuch des Römisch-Germanischen Zentralmuseums 23-24 (1976–1977): 89–138. The reliefs probably originated in the first half of the seventh century in an ecclesiastical context and were later reused as grave markers. 425 ARF, a. 758, 17: ut promitterent se omnem voluntate illius esse facturos et annis singulis honoris causa ad generalem conventum equos CCC pro munere daturos. 426 Einhard, Vita Karoli imperatoris, c. 13, 16: Omnis pecunia et congesti ex longo tempore thesauri direpti sunt. Neque ullum bellum contra Francos exortum humana potest memoria recordari, quo illi magis ditati et opibus aucti sint.. Trans. Noble, 33. 427 Ermold, In honorem Hludowici, IV.373–84, 68. 116 queen and, under her, the chamberlain.428 The milites were, Hincmar claims, to be rewarded not just with food and clothing, but also gold, silver, and horses on special occasions and at regular intervals during the year, a sign ultimately of the ‘benevolence and care’ of the royal household.429 In the other direction, a core component of political life was the presentation of gifts to the king at the annual assembly, the so-called dona annua, a practice that approached the status of ‘internal tribute’.430 Such exchanges are particularly prominent in moments of political crisis and transition. The Annales Laureshamenses remembers that Charlemagne honoured the Franks assembled at the palace in Regensburg ‘abundantly with gold, silver, silk and many gifts’ in 793, a year after the rebellion of Pepin the Hunchback – a rising which, though short-lived, had enjoyed some aristocratic support and had posed a serious, if short-lived, threat to Charlemagne’s rule.431 Airlie hypothesised that Pepin had been pushed as a ‘front man those members of the aristocracy who felt that they were losing out under Charlemagne’.432 Looked at another way, one could say that this looks very much like the king buying loyalty. In a similar fashion, Louis the Pious regularly accepted gifts at each of the annual assemblies between 832 and 837, years marked by sustained conflict and challenges to his authority, a sign of ‘a royal power in need of continuous, if symbolic, manifestation of loyalty from supporters’ in Curta’s words.433 *** 428 Hincmar of Reims, De ordine palatii, c. 22, ed. and German trans. Thomas Gross and Rudolf Schieffer, MGH. Fontes iuris 3 (Hannover: Hahn, 1980), 72: de donis annuis militum […] ad reginam praecipue et sub ipsa ad camerarium pertinebat. 429 Ibid., c. 27, 80: Uno videlicet, ut absque misteriis expediti milites, anteposita dominorum benignitate et sollicitudine, qua nunc victu, nunc vestitu, nunc auro, nunc argento, modo equis vel ceteris ornamentis interdum specialiter, aliquando prout tempus 430 Ibid., c. 29, 82. Reuter, ‘Plunder and Tribute’, 85. Curta points out that this was in large part a regularisation of existing Merovingian practice: Curta, ‘Gift Giving’, 698 431 Annales Laureshamenses, a. 793, in Annales et chronica aevi Carolini, ed. Georg Heinrich Pertz, MGH. SS. 1 (Hannover: Hahn, 1826), 35: et cum cognovisset fideles suos, episcopos, abbates et comites, qui cum ipso ibi aderant, et reliquum populum fidelem, qui cum Pippino in ipso consilio pessimo non erant, eos multipliciter honoravit in auro et argento et sirico et donis plurimis. Jennifer Davis, Charlemagne’s Practice of Empire (Cambridge: Cambridge University Press, 2013), 149–52. 432 Stuart Airlie, ‘Charlemagne and the Aristocracy: Captains and Kings’, in Charlemagne: Empire and Society, 100. 433 Annales Bertiniani, aa. 832, 833, 834, 835, 836, 837, ed. Georg Waitz, MGH. SS. Rer. Germ. 5 (Hannover: Hahn, 1883), 5, 7, 9, 11, 12, 13. Curta, ‘Gift Giving’, 698. 117 While not easily delimited as physical space, as social space the treasury was essential to early medieval conceptualisations of the palace. The treasury situated the palace at the centre of the political system, creating opportunities for material advancement among the elite and a symbolic language for mani- festing hierarchies of status and prestige. So important was the treasury to contemporary conceptual- isations of political action that its dispersal was understood to herald the disintegration of the polity itself. V. The Forest In a poem to the Austrasian noble Gogo (d. 581), Venantius Fortunatus speculates where his friend might be. Perhaps he was on the banks of the Rhine, or among the vineyards of the Moselle, or maybe on the bustling Meuse.434 He ranges over the landscapes in which the Austrasian court was active, turning from the idyllic river valleys to the forests of the Ardennes and the Vosges.435 There, he ima- gines Gogo engaged in the hunt, pursuing with a net, spear, and bow a variety animals including goats, stags, elk, and aurochs.436 Taken in context, the image of Gogo’s exploits as a hunter (including the menagerie of animals who apparently fell prey to him) is intended as a praiseworthy one, a complement to his other qualities including his wisdom and popularity as a prominent member of the court.437 As Fortunatus’ praise of Gogo shows, hunting was understood to be a vital part of the social life of the Austrasian court. A consideration of hunting might seem a strange choice in the context of the present investi- gation. It has left no discernible architectural patterning at any of the sites surveyed so far and, by definition, many of the ritual aspects of the hunt took place outside of any such architectural context 434 Ven. Fort., Carm., VII.4.5–12, 2:89. 435 For more on this aspect of Fortunatus’ poetry, see below, Chapter 4. 436 Ven. Fort., Carm., VII.4.17–23, 89: Aut aestiua magis nemorum saltusque pererrans | cuspide, rete feras hinc ligat, inde necat? | Ardenna an Vosagus cerui, caprae, helicis, uri | caede sagittifera silua fragore tonat? | Seu ualidus bufalus ferit inter cornua campum | nec mortem differ ursus, onager, aper? 437 Ibid., ll. 25–26, 90: Siue palatina residet modo laetus in aula, cui scola congrediens plaudit amore sequax? Gogo served as mayor of the Austrasian palace under Sigibert and was the tutor of Childebert II: PLRE, 3A:541–42. 118 anyway. It is striking nonetheless that hunting is consistently associated with palace sites, and legal texts show an increased interest in demarcating the social and topographical boundaries of participa- tion. Hunting was not only an arena for early medieval kings to display their bravery and skill at arms, but also for the performance of a shared masculine, aristocratic identity between the king and mem- bers of the elite. Hunting and aristocratic masculinity Hunting was a widely diffused practice among early medieval aristocracies in all the post-Roman king- doms.438 Gregory of Tours references Frankish kings and nobles hunting in a number of places in his narrative, as well as noblemen keeping huntsmen in their retinue.439 The same was true elsewhere. Sidonius Apollinaris describes the Visigothic king Theodoric II’s skill with the bow while on hunting expeditions with guests (including, it seems, Sidonius himself).440 ‘Fredegar’, on the other hand, criti- cises Chlothar II for an ‘excessive’ love of hunting.441 A further indication of its social diffusion is provided by the post-Roman law codes. In the Lex Ribuaria, for example, an untrained hawk could be exchanged for three solidi against a wergeld, a falcon for hunting cranes was worth six, and a trained hawk twelve.442 In the Lex Alamannorum, hunting dogs of various kinds, if stolen, are worth from three 438 For the broader context of hunting as a representational practice from the Ancient world to Modernity: Thomas Allsen, The Royal Hunt in Eurasian History (Philadelphia: University of Pennsylvania Press, 2006). 439 Greg. Tur., Hist., IV.21, 154 (Lothar goes hunting at Cuise near Compiègne); V.14, 211 (Guntram Boso tries to lure Merovech into danger by proposing they go hunting); V.39, 246 (Chilperic hunting near Chelles); VI.46, 319 (Chilperic murdered returning from a hunt near Chelles); VIII.6, 374 (Guntram hunting); VIII.10, 377 (Guntram sets off to locate the grave of Clovis while pretending to go hunting); X.10, 494 (Guntram hunting in the Vosges). For hunters in aristocratic retinues, e.g.: ibid., V.12, 206–07 (Brachio, abbot of Ménat, had been a huntsman in the employ of the dux Sigivald). Also: Greg. Tur., Vita patrum, XII.2, in Gregorii episcopi Turonensis Miracula et opera minora, ed. Bruno Krusch, MGH SS rer. Merov. 1/2 (Hannover: Hahn, 1885), 262. 440 Sidonius, Epistulae, I.2.5, 3. 441 Fred., Chron., IV.42, 142: venacionem ferarum nimium assiduae utens. On Merovingian hunting: Eric Goldberg, In the Manner of the Franks: Hunting, Kingship, and Masculinity in Early Medieval Europe, The Middle Ages (Philadelphia: University of Penn- sylvania Press, 2020), 44–69. 442 Lex Ribuaria 40.11, 95: Aucceptorem indomito pro 3 solidi tribuat. Conmorsum gruarium pro sex solid. tribuat. Aucceptorem mutatum pro 12 solid. tribuat. Jörg Jarnut, ‘Die frühmittelalterliche Jagd unter rechts- und sozialgeschichtlichen Aspekten’, in L’uomo di fronte al mondo animale nell’alto medioevo, Settimane 31 (Spoleto: CISAM, 1985), 768–771. I offer thanks to Paul Vinhage for discussion of commorsum gruarium here. 119 to twelve solidi.443 Despite the patchiness of the extant sources, the clear picture is that hunting was a common elite practice among early medieval aristocracies throughout Europe. Hunting’s popularity among early medieval elites was in part an inheritance from Roman aris- tocratic culture. Mosaic pavements from luxury villas in, for example, Antioch, Italy, and North Africa vividly attest to the currency of hunting in an aristocratic cultural ethos (Figs. 86–87).444 This remained true despite fluctuations in the bases of aristocratic identity. Guy Halsall has suggested that hunting formed a common ground between diverging models of masculinity in the Late Empire: both the ‘Roman’ senatorial aristocracy and the new military elite of the fourth and fifth centuries, frequently coded as ‘barbarian’, could utilise hunting as a distinctive symbol of their class.445 Thus could both Symmachus opine that hunting was a fitting activity for a young nobleman and Claudian praise Stilicho for impressing with the bow on a hunt while serving as a legate to Persia.446 Hunting had a broad social valence as (following Halsall) a means of performing masculine identity as the bases for expressing, hierarchy, status and prestige were renegotiated in the course of the fourth and fifth centuries.447 This social function was only intensified thereafter. Whatever functional purposes the hunt possessed, such as subsistence, or practice with arms in preparation for warfare, among the elite it was principally a representational practice.448 Archaeology strongly suggests that game formed a small 443 Lex Alamannorum, LXXVIII.1–4, in Leges Alamannorum, ed. Karl Lehmann, 2nd ed., MGH. LL. 5/1 (Hannover: Hahn, 1966), 142–43. 444 Katherine Dunbabin, The Mosaics of Roman North Africa: Studies in Iconography and Patronage, Oxford Monographs on Classical Archaeology (Oxford: Clarendon Press, 1978), 46–64. 445 Guy Halsall, ‘Gender and the End of Empire’, Journal of Medieval and Early Modern Studies 34, no. 1 (2004): 23; idem, Barbarian Migrations, 96–99, 482–88. 446 Symmachus, Epistulae, V.68, in Q. Aurelii Symmachi quae superseunt, ed. Otto Seeck, MGH. AA 6/1 (Berlin: Weidmann, 1883), 144; Claudian, De consulatu Stilichonis, I.51–69, in Claudii Claudiani Carmina, ed. John Barrie Hall, BSGRT (Leipzig: Teubner, 1985), 192–93. 447 Halsall, ‘Gender’. 448 Lutz Fenske, ‘Jagd und Jäger im früheren Mittelalter. Aspekte ihres Verhältnisses’, in Jagd und höfische Kultur im Mittelalter, ed. Werner Rösener, Veröffentlichungen des Max-Planck-Instituts für Geschichte 135 (Göttingen: Vandenhoeck & Rup- recht, 1997), 44–47. 120 proportion of the early medieval diet even at elite sites.449 The point was display, not nutrition as such. Understood in these terms, hunting served as a medium of symbolic communication: a means of performing a common masculine, aristocratic identity and, in doing so, affirming social bonds.450 The symbolic aspects of hunting were not limited to the chase itself. The distribution of the trophies af- terwards represented and affirmed hierarchies of status to the participants, and the communal con- sumption of the meat fostered solidarity among them. Naomi Sykes, for example, has suggested that large knives deposited in male graves in England between the fifth and eighth centuries refer to the interred individual’s status in the community as possessing the authority to redistribute food to oth- ers.451 Hunting grounds: parks and forests The other side to the rituals of hunting is the demarcation of social space. In fact, hunting was a consciously spatial practice: a means of mastering landscapes by imposing physical and legal bounda- ries regulating access and use.452 It thus effected the palace’s separation from the ‘everyday’ lifeways of early medieval societies which shaped the natural landscape. We see this in the references to en- closed hunting parks in Frankish sources and, from the seventh century onwards, indications that land was being reserved for exclusive royal use. The Capitulare de villis stresses the conceptual connection between the palace and estate, forest and hunting. The capitulary required that royal ‘woods and forests’ (sylvae vel forestes) should be assidu- ously maintained, obliging the iudices (royal estate managers) to maintain and clear wooded areas where 449 E.g. Helen Hamerow, Early Medieval Settlements: the Archaeology of Rural Communities in North-West Europe, 400–900 (Ox- ford: Oxford University Press, 2003), 126–27, 148–52; Naomi Sykes, ‘Deer, Land, Knives and Halls: Social Change in Early Medieval England’, Antiquaries Journal 90 (2010): 175–93. 450 Janet Nelson, ‘The Lord’s Anointed and the People’s Choice: Carolingian Royal Ritual’, in Rituals of Royalty, 169; Paul Dutton, ‘Charlemagne, King of Beasts’, in idem, Charlemagne’s Mustache, and Other Cultural Clusters of a Dark Age, The New Middle Ages (London: Palgrave Macmillan, 2004), 43–68. 451 Sykes, ‘Deer, Land, Knives, and Halls’, 181–82. 452 Joseph Morsel, ‘Jagd und Raum. Überlegungen über den sozialen Sinn der Jagdpraxis am Beispiel des spätmittelalterli- chen Franken’ in Jagd und höfische Kultur im Mittelalter, 260, 277–85; Lefebvre, Production, 193. 121 necessary, to manage game stocks, and to keep hawks and falcons for the king’s use.453 Likewise, hunters and falconers are numbered as part of the palace’s standard staff.454 Though a projection of idealised, rational governance rather than a reflection of lived reality, the Capitulare de villis reflects the perceived importance of hunting to the court’s proper ordering. The same is true of Hincmar’s De ordine palatii, which includes four chief hunters and a head falconer among the court hierarchy.455 Palaces were associated with hunting grounds from an early date. Gregory of Tours, for ex- ample, writes that Guntram, hunting in the Vosges in the year 590, discovered an aurochs which had been killed. Upon discovery that the perpetrator was his own chamberlain, Chundo, the king has him arrested, taken to his palace at Chalon-sur-Saône where, following a trial by combat, he was subjected to a violent death by stoning.456 While Gregory does not specify that the killing of the aurochs was contrary to codified law, the fact that the chamberlain was sentenced according to a trial by combat implies that a legalistic judgment was made. This episode suggests that the Vosges, a large geographical area, was being claimed as explicitly royal space. Similarly, the foundation charter for Stavelot-Mal- médy, given by Sigibert III in 648, describes the double monastery’s location as ‘in our forest named the Ardennes’, a region notable for its ‘multitude of animals’.457 The term forestis quickly came to denote a legal category of land ownership and exclusive rights to the resources within.458 This process inten- sified in the Carolingian period. Aachen, for example, lies at the edge of the Ardennes and seems to 453 Capit 1, no. 32.36, 86: Ut sylvae vel forestes nostrae bene sint custoditae; et ubi locus fuerit ad stirpandum, stirpare faciant et campos de silva increscere non permittant; et ubi sylvae debent esse, non eas permittant nimis capulare atque damnare; et foramina nostra intra forestes bene custodian; similiter acceptors et spervarios ad nostrum profectum praevideant; et censa nostra exinde diligenter exacteant. 454 Capit 1, no. 32.47, 87: Ut venatores nostri et falconarii vel reliqui ministeriales, qui nobis in palatio adsidue deserviunt. 455 Hincmar, De ordine palatii, c. 17, 64: venatores principales quatuor, falconarium unum. Ibid., V.24, 76: quatuor venatores et quintus falconarius. 456 Greg. Tur., Hist., X.10, 494. 457 DD Merov. 1, no. 81, 206: in foreste nostra nuncupante Arduinna in locis vastę solitudinis consulere cupientes, in quibus caterva bestiarum germinat 458 Karl Bosl, ‘Pfalzen und Forsten’, in Deutsche Königspfalzen 1, ed. Adolf Gauert (Göttingen: Vandenhoeck & Ruprecht, 1963), 2. Its origins likely lay in the saltus (‘woods’) of the Roman imperial fisc, large tracts of uncultivated land used for pasturage claimed by the emperors: Heinrich Rubner, ‘Vom römischen Saltus zum fränkischen Forst’, Historisches Jahrbuch 83 (1964): 271–77. 122 have been the centre of a large estate including large tracts of forest judging from later sources.459 By the ninth century, Ermold could present the Vosges as exclusively royal space. In his poem in praise of the young Pepin of Aquitaine, the personified Vosges (Wasacus) boasts of its bountiful hunting: ‘Kings have been wont to roam in our woods to chase various beasts in the hunt. Here a hind flees into the springs now the arrow has been cast; there a foaming boar makes for the familiar streams. What can I say about fish? I am full of all kinds of fish, rich as I am in small rivers.’460 At the same time as the establishment of forest rights, royal estates were equipped with private hunting parks. The Capitulare de Villis commands that the king’s copses (lucos), also known as brogili, should be kept well-maintained and that the estate managers ‘should by no means wait until it neces- sary to rebuild them completely. Let them do similarly for every building’.461 Given that forests and woods were treated separately to other palace buildings, it is likely that what is meant are walled hunt- ing parks. Ermold certainly references such an enclosure, surrounded by an earthen rampart and stone wall, near to the palace at Aachen and another at Ingelheim.462 The anonymous poet of Karolus magnus et Leo papa similarly describes a lucus at Paderborn: ‘Not far from the peerless town are a wood and a pleasant lawn, holding in their midst a verdant glade, its meadows fresh from the streams, and encir- cled by many walls’.463 Given the variety of other palaces used by the Carolingians for hunting, 459 Bosl, ‘Pfalzen und Forsten’, 19; David Rollason, ‘Forests, Parks, Palaces, and the Power of Place in Early Medieval Kingship’, Early Medieval Europe 20, no. 4 (2012): 431, and especially: Dietmar Flach, Untersuchungen zur Verfassung und Verwaltung des Aachener Reichsgutes, Veröffentlichungen des Max-Planck-Instituts für Geschichte 46 (Göttingen: Vanden- hoeck & Ruprecht, 1976). 460 Ermoldus Nigellus, In laudem Pippini Regis I, in Poetae aevi carolini, 2:82–83: Saltibus in nostris soliti discurrere reges, | Venatu variae exagitare feras. Hic fugit ad fontes coniecta cerva sagitta, | Hac spumosus aper flumina nota petit. | Quid de pisce loquar? Vario sum pisce repletus, | Fluminibus parvis sum quia dives ego. 461 Capit 1, no. 32.46, 87: Ut lucos nostros, quos vulgus brogilos vocat bene custodire faciant et ad tempus semper emendent et nullatenus expectant, ut necesse sit a novo reaedificare. Similiter faciant et de omni aedificio. See further: Karl Hauck, ‘Tiergärten im Pfalzbereich’, in Deutsche Königspfalzen, 1:32–50. 462 Ermoldus Nigellus, In honorem Hludovici imperatoris, III.583–94, 57; IV.481–534, 72–73. 463 Karolus magnus et Leo papa, ll. 137–325, ed. and German trans. Franz Brunhölzl, in Helmut Beumann et al., Karolus magnus et Leo papa. Ein Paderborner Epos vom Jahre 799, Studien und Quellen zur Westfälischen Geschichte 8 (Paderborn: Verein für Geschichte und Altertumskunde Westfalens, 1966), 68: Non procul excelsa nemus est et amoena virecta | Lucus ab urbe virens et prata recentia rivis | Obtinet in medio, multis circumsita muris. Trans.: Peter Godman, trans. Poetry of the Carolingian Renaissance (London: Duckworth, 1985), 205. 123 including Nijmegen, Frankfurt, and Salz, as well as Paderborn, we should expect that such enclosures were not uncommon.464 Rituals of the hunt in the Carolingian period The ritualised aspects of hunting are best exemplified in Ermoldus’ description of a hunt near Ingel- heim following the baptism of Harald Klak in 826.465 While doubtless exaggerated for effect, Ermold describes a large hunting party, the park bustling with noble sons, huntsmen, and their dogs.466 This display had a ritualised aspect. The whole royal household set out to watch the chase unfold, and the trip seems to have included some kind of processional element. Louis’ wife, Judith, ‘clothed and coiffed stunningly’ was escorted by ‘the highest nobles and a troop of magnates […] in honor of the pious king’.467 Lothar, accompanied by the pueri regis, sons of noblemen educated at court, makes a heroic showing, apparently slaying several bears; even Louis’ younger son, Charles, at most four years old, gets his chance to snag a young deer.468 Notably, the action focuses on the actions of the Franks; Ermold implies that the Danes remained on the sidelines with the women.469 At its conclusion, the hunters retired to a pavilion made ready by Judith, where the spoils were shared out and the partici- pants relaxed with roasted meats and wine.470 Having returned to the palace, young men paraded their trophies—the antlers of stags, the pelts of bears, the skins of goats—trying to catch the king’s attention 464 Nijmegen: RFA, aa. 817, 825. Frankfurt: RFA, aa. 829, 836. Ingelheim: Ermoldus Nigellus, In honorem Hludowici, IV.481– 534, 72–73. Salz: RFA, a. 826. See below, Chapter 4, and Appendix I. 465 The analysis presented here was informed by Eric Goldberg’s analysis of hunting in the age of Louis the Pious: Eric Goldberg, ‘Louis the Pious and the Hunt’, Speculum 88, no. 3 (2013): 613–643. The recent publication of his monograph on hunting in the Frankish world has been most welcome: idem, In the Manner of the Franks, esp. 103–28 (on the reign of Louis), and 113–17 (on this episode). 466 Ermoldus Nigellus, In honorem Hludowici, IV.489–494, 71–72. 467 Ibid., ll. 497–500, 72: Iam pia scandit equum Iudith pulcherrima coniunx | Caesaris, ornata comptaque mirifice, | Quam proceres summi dominam, seu turba potentum | Praeterit et sequitur regis honore pii. Trans.: Noble, 180–81. 468 Ibid., ll.495–96, 72: Plurima turba fluit iuvenum, necnon puerorum, | Inter Hluttharius quos celer ibat equo.; ll. 509–10, 72: Hlutha- riusque celer florens, fretusque iuventa | Percutit ursorum corpora multa manu; ll. 515–34, 72–73. 469 Ibid., ll. 495–96, 72: Atque simul Deni, necnon Heroldus et hospes | Spectandi studio huc quoque laetus adest. 470 Ibid., ll. 535–56, 73, and in particular: ll. 537–40: Sed tamen in medio nemoris viridantia claustra | Iudith prudenter construit, atque tegit, | Vimine praeraso nec non et tonsile buxo,| Palleolis cingit linteolisque tegit. 124 and Louis ‘shared the bounty among all his attendants in his normal way, and the great part fell to the clergy.’471 In Ermold’s text, the rituals of the hunt represent and affirm expected gender roles, identities, and social hierarchies. Most obviously, a distinction is drawn between the men, who display their skills at arms and bravery, and the women and children, who watch. This boundary is broken with the young Charles’ dispatch of the deer. His participation in the blood sport is treated by Ermold as a kind of rite of passage – a precocious sign of his masculine virtus despite his age and a promising sign for his future.472 Judith’s ‘prudent’ preparations for the hunt’s end and the distribution of the fresh kill among the hunters reflects a contemporary expectation of pious, domestic femininity in contradistinction to the warlike masculinity of the rest of the king and his retinue.473 This understanding of the feminine role of the court mirrors the presentation of the queen’s responsibilities in the palace in later Carolin- gian texts. Hincmar’s De ordine palatii—possibly reflecting conditions in the palace under Louis in the 820’s—ascribes the ‘dignity’ (honestas) of the palace, the royal insigna (ornamento regali), and furthermore the annual gifts for the court officers, to the queen.474 ‘Dignity’ refers here not only to the palace’s physical beauty, but also its moral virtue, a discursive nexus that would be weaponised against Judith herself in 830.475 The third group are the young men, the pueri and iuvenes. For them, the hunt was an opportunity to make a name for themselves: their triumphant parade of the trophies and attempts to 471 Ibid., ll. 559–66. 74: Ecce manus iuvenum, venatus munera tollens. | Multa fluit, cupiens regis adesse oculis. | Milia cervorum praegran- dia cornua, nec non | Ursorum referunt tergora seu capita, | Plurima saetigerum revehunt et corpora aprorum, | Capreolos, dammas fert puerile decus. | Ille pius praedam famulos partitur in omnes | More suo, clero pars quoque magna cadit. Trans. Noble, 182. 472 Ibid., ll. 531–32, 73: Hunc puerile decus hinc inde frequentat et ambit, | Hunc patris virtus, nomen et ornat avi (‘It was a boyish achievement this time, but more would follow, and he would go forth marked by his father’s strength and adorned with his grandfather’s name’: trans. Noble, 181.) 473 Compare also the contrast drawn by Einhard on the education of Charlemagne’s sons and daughters: Einhard, Vita Karoli imperatoris, c. 19, 23. 474 Hincmar, De ordine palatii, c. 22, 72: De honestate vero palatii seu specialiter ornamento regali nec non et de donis annuis militum […] ad reginam praecipue et sub ipsa ad camerarium pertinebat. 475 Elizabeth Ward, ‘Caesar’s Wife: The Career of the Empress Judith’, in Charlemagne’s Heir: New Perspectives on the Reign of Louis the Pious (814–840) (Oxford: Clarendon Press, 1990), 226–27; Mayke de Jong, ‘Bride Shows Revisited: Praise, Slander, Exegesis in the Reign of the Empress Judith’, in Gender in the Early Medieval World: East and West, 300–900, ed. Leslie Brubaker and Julia Smith (Cambridge: Cambridge University Press, 2004), 267–72. 125 catch Louis’ eye are a testament to the prestige up for grabs. These young men were the next genera- tion of the aristocracy and, naturally, then, they formed the party of the future king, Lothar. The ties of interdependence between aristocracy and royal house were forged at an early age, in the palace. Finally, there is the king himself. In Louis’ reign, possibly due to the momentum of Frankish expansion having slowed, thus leading to fewer large-scale campaigns being waged, the hunt took on increased social significance. There are no less than twenty-seven attestations of Louis’ hunting, some- times multiple times in a single year, and the image of the ‘royal hunter’ was carefully cultivated by Louis’ literate courtiers (for example, Ermold, Walahfrid, and the Astronomer).476 As Goldberg has persuasively shown, hunting was a means for Louis to highlight his imperial legitimacy and martial prowess when opportunities for actual campaigning were more limited.477 Walahfrid Strabo makes this connection starkly in his De imagine Tetrici: Just as the bear, boar, timid hare and swift stags Antelope, wolf, and huge herd of wild cattle Fear your bow in the lovely glades, So the Bulgar and cur of Sarah, bad guest of the Spaniards, The brutish Britton, shrewd Dane and dreadful Moor Bow their necks in terror before your venerable hands.478 For authors in Louis’ reign and after, hunting provided a means of commenting on royal power and prestige. *** The rituals of the hunt cannot be boiled down to any single ‘functional’ meaning. Rather, hunting was an activity through which social bonds were forged between members of the ruling class through the performance of a common masculine identity. As the literate members of Louis’ court were keen to emphasise, the hunt also offered opportunities for royal display, to make arguments about the king’s 476 Goldberg usefully tabulates all references to Carolingian hunting to 987: Goldberg, ‘Louis the Pious and the Hunt’, 642–43. 477 Ibid., 623–24; Goldberg, In the Manner of the Franks, 109–12. 478 Walahfrid Strabo, De imagine Tetrici, ll. 250–55, 130 (Latin), 138–39 (trans.). 126 legitimacy in a kind of ‘proxy warfare’ through displays of skill with weapons and generosity with the spoils. The definition of the ‘forest’ established social, legal, and physical boundaries for participation: hunting was an activity which differentiated the palace from non-elite society. VI. The Palace Church As we saw in the previous chapter, the most significant innovation in the architecture of palaces in this period was the incorporation of spaces of religious ritual into its structure. By the Carolingian period, the chapel had become the most significant ritual space in the palace. More than just an archi- tectural wonder, it was a statement and symbol of the ideology of the Carolingian imperial project.479 While we cannot say how liturgies performed in the presence of the king looked in contrast to else- where—we lack specific ordines for palace churches, if they ever existed—it would be fair to begin from the hypothesis that space reserved for the ruler in the liturgy manifested religious sanction for their rulership to a specific, courtly audience. Certainly, ecclesiastical figures connected with the Car- olingian court saw the formulation and patronage of a ‘liturgy of authority’ as an integral component of the imperial project, a way of representing the fundamental unity of the polity as an ecclesia through- out the kingdom.480 The ritual uses of the chapel were not solely liturgical, however. Through the palace chapel we also get a glimpse at the ‘liturgification’, of previously secular (or at least not explicitly religious) rituals through their transposition into an ecclesiastical setting. An analysis of three contexts in the eighth and ninth centuries—the Frankish regnum under the Carolingians, the principality of Be- nevento, and the kingdom of Asturias—highlights similar strategies being pursued for manifesting royal authority in ecclesiastical space. 479 Einhard, Vita Karoli imperatoris, c. 26, 30–31. See also: Notker, Gesta Karoli, I.27–28 and 30, 38–41. 480 For royal liturgies in the Carolingian period: Yitzhak Hen, The Royal Patronage of Liturgy in Frankish Gaul to the Death of Charles the Bald (877), Henry Bradshaw Society. Subsidia 3 (London: Boydell Press, 2001). For ‘liturgy of authority’: Garip- zanov, Symbolic Language, 43–100. For the Carolingian Empire as an ecclesia: Mayke de Jong, ‘Ecclesia and the Early Medieval Polity’, in Staat im frühen Mittelalter, 113–32. 127 The liturgification of royal ritual in Francia The chapel at Aachen, being the most complete example of a palace church from the Early Middle Ages, is a good place to start. Einhard presents Charlemagne as regularly attending the Mass and participating in the monastic offices in his chapel, the interior of which he outfitted with beautifully crafted objects.481 The spatial organisation of the chapel’s interior visual manifested social and cosmo- logical hierarchies as conceptualised by intellectual circles connected with the court. In the western bay of the upper gallery stands a marble throne (Fig. 88), traditionally interpreted as belonging to Charlemagne – an association that has received strong support from recent analysis.482 Facing the throne in the cupola was a mosaic depicting Christ in Majesty (Figs. 89–90), himself enthroned and receiving the crowns offered by the Twenty Four Elders of the Apocalypse.483 The location of the throne, facing outwards into the empty vessel of the nave, would have provided a privileged view of the liturgical drama taking place below. Notker describes envoys from the caliphate circling this gallery and marvelling at the clergy and the milites arranged below them, giving a sense of the social segregation inside the space at least in the ninth century.484 From this location, the elevated host is perceived in vertical co-ordination with the enthroned Christ above, acting out their identity. The visual and spatial concordances were not limited to this, however. The palace chapel’s decorative programme, while drawing a comparison between earthly and heavenly rulership, consciously maintained an interstitial distance: the king’s throne is elevated above the ministering clergy, below, but there is no question of his equation with Christ above. 481 Einhard, Vita Karoli imperatoris, c. 26, 31. For Louis attending Mass in the chapel: Astronomer, Vita Hludowici, c. 28, 373–75. 482 Sven Schütte, ‘Der Aachener Thron’, in Krönungen: Könige in Aachen – Geschichte und Mythos, ed. Mario Kramp (Mainz: Phillipp von Zabern, 2000), 213–22. 483 Rev. 4.1–4. McClendon, Origins, 120. The mosaic that presently adorns the ceiling is mostly a reconstruction, executed by Antonio Salviati in 1880/1881, but it is generally accepted that it preserves the original design thanks to the testimony of earlier drawings underpainting discovered beneath the nineteenth-century tesserae. McClendon summarises the evidence: ibid., 113–14, with references to earlier literature. 484 Notker, Gesta Karoli, II.8, 59–60: Ascendentesque in solarium, quod ambit ędem basilicę, et inde despectantes clerum vel exercitum. 128 There are also hints that the palace chapel possessed a broader social function already in the Merovingian period. By the end of the seventh century it seems to have been customary for oaths made to the king or on the resolution of a legal dispute to be sworn in the chapel on relics kept in the royal treasury. As we saw in the previous chapter, the cappa of Saint Martin was already held in the royal treasury in the seventh century. These relics were activated in the context of oath swearing. For example, a placitum of Theuderic III from Luzarches in 682 records an oath sworn ‘over the mantle of the lord Martin in our chapel’.485 Likewise, a formula for a placitum in the Formulary of Marculf uses notably similar language, stating that the accused and witnesses ‘should swear at such a time in our palace, on the cape of the Lord Martin, where the other oaths (reliqua sacramenta) take place’.486 The implication of customary practice in the formula is marked, and it suggests that religious sanction was being sought to strengthen agreements already by the middle of the seventh century. That this was a more widespread practice than just oaths taken in the palace is implied by another formula, which requires that a count summon together all the ‘pagenses, Franks, Romans, and those of any other origin’ of his jurisdiction to take oaths in the presence of an officer of the court, ‘on the relics which we have sent there through the same person’.487 An otherwise uncorroborated reference in the Liber Historiae Francorum that Clovis II ‘cut off the arm of the blessed Denis the Martyr’, an impious act in the chron- icler’s eyes, could plausibly reflect the collection of a relic for a similar practice.488 The royal chapel was thus an instrument of establishing both intensive social bonds between the ruler and petitioners 485 DD Merov. 1, no. 126, 320: in oraturio nostro super cappella domni Martine. 486 Marculfi Formulae, I.38, 68: tunc in palatio nostro, super capella domni Martini, ubi reliqua sacramenta percurrunt, debent coniurare; Trans. Rio, 173. 487 Marculfi Formulae, I.40, 68: iubemus, ut omnes paginsis vestros, tam Francos, Romanos vel reliquia natione degentibus, bannire et locis congruis per civitates, vicos et castella congregare faciatis, quatenus presente misso nostro, inlustris vero illo, quem ex nostro latere illuc pro hoc direximus, fidelitatem precelso filio nostro vel nobis et leudesamio per loca sanctorum vel pignora, quas illuc per eodem direximus, dibeant promittere et coniurare. Trans. Rio, 176. 488 Liber Historiae Francorum, c. 44, 316: Eo tempore Chlodeus brachium beati Dionisii martyris abscidit, instigante diabulo. 129 in the social space of the palace itself, but also extensive ones, between centre and locality – a theme to which we will return in Chapter 4. Space, liturgy, and royal vision in Asturias In churches associated with the kings of Asturias, we see a similar attention to spatial and visual hier- archies in the palace church – distinctions that acted out social hierarchies between king and elite and articulated royal authority in the language of religious sanction. Conspicuously consistent in all of the churches associated with the royal house in Asturias is the presence of a gallery or tribunal raised above the floor of the nave that provided a privileged view of the liturgical drama.489 In San Juan Apóstol y Evangelista (Santianes de Pravia), founded by Silo (774–783) and his wide Adosinda to coincide with the establishment of his seat at Pravia, a raised platform in wood was constructed over the church’s entrance, providing a commanding line of sight down the wide central nave towards the sanctuary—a view possibly obscured to those in the side-aisles by the heavy arches on piers of the arcade—though it likely belongs to a second phase in the building’s history.490 More concretely, San Miguel de Lillo on the Monte Naranco (Figs. 91–92), consecrated in 848 by Ramiro I, likewise pos- sessed a tribune at its western end, supported by the first bay of the nave arcade and accessible by staircases in the westernmost compartments of the aisles.491 This platform provided a similarly com- manding view down the length of the nave, one unavailable to the rest of the congregation in the palace church, as well as a secondary architectural focus in tension with the usual insistence on the 489 In addition to the structures discussed here, see also e.g. Santa Cristina de Lena (Lena), and San Salvador de Valdedíos (Valdedíos). 490 For a detailed history of the church, see: José Menéndez Pidal, ‘La basílica de Santianes de Pravia, Oviedo’, in Actas del Simposio para el Estudio de los Códices del ‘Comentario al Apocalipsis’ de Beato de Liébana, vol. 2 (Madrid: Joyas Bibliográficas, 1980), 281–97. Silo’s patronage of the church is confirmed by a fragment of an acrostic dedicatory inscription reading SILO PRINCEPS FECIT: ibid, 282; Arbeiter and Noack-Haley, Christliche Denkmäler, 103–05. The ninth-century Chronicon Albeldensis records that Silo established the royal seat at Pravia: Chronicon Albeldensis, XV.6, 174: Iste, dum regnum accepit, in Prabia solium firmavit. For the tribunal: Menéndez Pidal, ‘La basilíca’, 290–292; Arbeiter and Noack-Haley, Christliche Denkmä- ler, 100–01. 491 Arbeiter and Noack-Haley, Christliche Denkmäler, 144–45. 130 apse. As at Aachen, the elevation of the royal party over the rest of the congregation gave visual and spatial form to the separation of ruler and elite congregation. A variation on this theme can be seen in San Julián de los Prados, a large basilica associated with a palace founded by Alfonso III (866–910) outside the walls of Oviedo. According to the Chronicle of Alfonso III, composed in the first decade of the tenth century, ‘some distance from the palace he [Alfonso] built a church in honour of Sts Julian and Baselissa with a pair of altars of wonderful work- manship and set up in a marvellous arrangement. He also directed and ordered to be built a royal palace, baths, and storehouses for all types of provisions’.492 San Julián’s spacious interior (c. 39m by 25m) is composed of what are, essentially, two distinct units.493 The nave is open and bright, but the longitudinal axis is startlingly truncated by the tall, continuous transept (Fig. 93). The sanctuary is emphatically separated from the nave by the heavy, semi-circular triumphal arch which all-but closes off the lateral aisles with their much-reduced openings (Fig. 94). A raised tribune accessed from a room communicating with the exterior is incorporated into the northern arm of the transept rather than the western end of the nave, thereby giving a commanding view down the length of the transept across the front of the triple apse. In this arrangement, the occupant of the tribune is granted a view which is unequivocally denied to the spectator in the nave. The superimposition of royal over ecclesiastical space is underscored by the uniquely rich fresco programme.494 Organised in three horizontal bands, they display geometric motifs, representa- tional emblems of buildings, and large, gemmed crosses (Figs. 95–98). Figural representation is wholly absent. Framed within illusionistic niches are painted ‘basilica’ motifs that stylistically, in terms of their 492 Chronicon Adefonsi tertii regis (Rotensis), c. 21, in Cronicas Asturianas, 140: necnon satis procul a palatium edificauit eclesiam in honorem sancti Iuliani et Baselisse cum uinis altaribus magno opere et mirauili conpositione locauit ; nam et regia palatia, balnea, promtuaria atque uniuersa stipendia formauit et instruere precepit. Trans. Baxter Wolf, 172. 493 Arbeiter and Noack-Haley, Christliche Denkmäler, 119. 494 The best single study of the paintings of San Julián remains: Helmut Schlunk and Magín Berenguer, La pintura mural asturiana de los siglos IX y X, trans. María de los Angeles Vázquez de Parga (Madrid: L. Sanchez Cuesta, 1957). See also: Helmut Schlunk, ‘Las pinturas de Santullano. Avance al estudio de la pintura mural asturiana de los siglos IX y X’, Archivo Español de Arqueología 25 (1952): 15–37. 131 general ‘flattened’ perspective as well as details such as the curtains hanging in their intercolumniations, immediately recall the art of Late Antiquity, particularly in mosaic.495 We should undoubtedly associate these images with the ‘varied pictures of royal palaces’ that, according to the Chronicon Albeldensis, Alfonso had painted in all the churches he founded.496 The rhythmic repetition of the ‘basilica’ motifs around the nave walls reduces them to an emblem, a ‘logo’, asserting royal ownership over the space to the spectator in the nave. That the crosses in the uppermost zone, the signs of Christ’s eternal rule, are pressed against so closely by further mirrored ‘basilicas’ is a graphic figuration of the alliance of religious and secular power performed in San Julián and Oviedo, attested architecturally in the con- junction of palace and church within the walled city. Martyrial saints and public ritual in Benevento We see a distinct conceptualisation of the palace church as ritual space in the foundations of the self- styled princeps of Benevento, Arechis II. His palace church in Benevento, Santa Sofia, was the palace’s public face and formed the central element of an urban, public-facing representational strategy. Topo- graphical references in primary sources indicate that Santa Sofia was closely associated with the ducal palace, situated inside the walled upper city at the terminus of its main thoroughfare, the platea maior publica (the modern Corso Garibaldi), close to the Porta Summa, the ‘Top Gate’.497 Though having suffered invasive alterations in the late-twelfth and the seventeenth centuries, a campaign of restora- tion begun in 1950 established Santa Sofia’s original form (Fig. 99).498 Revealed were carefully- 495 Schlunk and Berenguer, La pintura mural asturiana, 82–94. 496 Chronicon Albeldensis, XV.9, 174. 497 Marcello Rotili, ‘Arechi II e Benevento’, in Tra i Langobardi del Sud, 196–99. Charters in Santa Sophia’s cartulary chron- icle, the Chronicon Sanctae Sophiae, state that Santa Sofia was ‘joined to’ (iuxta) the palace: Liber preceptorum, I.43, in Chronicon Sanctae Sophiae (cod. Vat. Lat. 4939), ed. Jean-Marie Martin, vol. 1, Fonti per la storia dell’Italia medievale 3 (Rome: Istituto storico italiano per il Medio Evo, 2000) 414. For the platea maior, and its relation to Santa Sofia: ibid. I.44, 1:419, granting rights to the monks of Santa Sofia both along the ‘public main street that rises from the Top Gate’ and on the transversal street ‘that once was seen to pass by the apse of the aforementioned monastery’ (i.e. Santa Sofia): Subsequenter concedimus in eunde(m) venerabile(m) locum ut de quantum continunt ipse case ei(us)dem monasterii que hedificate sunt intus hanc Beneventum vetere(m) civitatem iuxta platea maiore illa puplica que ascendit de porta Su(m)ma et iuxta transonda puplica que olim p(er)gere videbantur erga epsida predic(ti) monast(erii). Dey, ‘From Street to Piazza’, 925. 498 Antonino Rusconi, ‘La chiesa di S. Sofia di Benevento’, Corso di cultura sull’arte ravennate e bizantina 14 (1967): 344. 132 constructed foundations and fabric of an unusual structure, in plan resembling a polygonal centrally- planned church with the lateral walls broken into ‘zig-zags’, 23.5m across at its widest point. 499 Inside, Santa Sofia is centrally-planned, with six columns arrayed in a ring at its centre, surrounded by eight square pilasters arranged non-radially around them (Figs. 100–101). An image in the Chronicon Sanctae Sophiae depicting Arechis’ foundation of the church (Fig. 102) represents its exterior (albeit schemati- cally) prior to the alterations of the twelfth century and likely indicates that the church originally pos- sessed an extradossal dome.500 Southern Italian sources from the ninth century onwards are near-unanimous that Arechis founded Santa Sofia in imitation of Hagia Sophia in Constantinople. Erchempert, a monk of Mon- tecassino writing in the final quarter of the ninth century remarks that ‘inside the walls of Benevento he founded a most opulent and handsome temple of the Lord, which he named with the Greek phrase Agian Sophian, that is “Holy Wisdom”’, and that Arechis handed control over it to the Benedictines of Montecassino.501 As far as we can discern, however, this connection was not drawn in any official documents of the court itself, and it seems Santa Sofia was designed with a distinct purpose in mind. Unlike its Constantinopolitan cousin, Santa Sofia was specifically intended as martyrial shrine and votive foundation. Records of donations made by Arechis to the monastery of Santa Sofia preserved in the twelfth-century Chronicon Sanctae Sophiae are prefaced by the formula, ‘I offer to the church of Santa Sofia, which I built from the foundations, for the redemption of my soul and for the salvation 499 There are strong arguments for the authenticity of the ‘zig-zag’ plan: firstly, the still-extant triangular corbels in the vertices for the springing of the vaults: Rusconi, ‘S. Sofia’, 352. Likewise, fragments of fresco – possibly early medieval, though they cannot be definitively dated – along the inner edges also prove that the zig-zags were actualised before the seventeenth-century rationalisation of the church into a circular plan: Hans Belting, ‘Studien zum Beneventischen Hof im 8. Jahrhundert’, DOP 16 (1962): 179. Belting has argued, however, that the church was originally a rotunda, and the ‘zig zag walls’ were perhaps added during or after construction for ‘non-artistic motives’ such as buttressing: ibid., 180. But it is hard to credit the ‘zig-zag’ walls with an important structural role (and even if this were so, it would still be a unique solution and resulting floor-plan). 500 Delogu, Mito, 30–31. 501 Erchempert, Ystoriola, 86–88: Infra Beneventi autem menia templum Domino opulentissimum ac decentissimum condidit, quod Greco vocabulo Agian Sophian, id est sanctam sapientiam, nominavit. 133 of our people and fatherland’.502 Further charters of Arechis reserved Santa Sofia from the control of the bishops of Benevento, asserting that it would answer ‘only to our sacred palace, just like our other monasteries’ with Arechis reserving the right to appoint ordain its priests.503 Santa Sofia was concep- tualised, therefore, as simultaneously ‘private’ and ‘public’ (if such a distinction is meaningful in this context): the liturgies that took place within were for the benefit of the princeps, but also, by extension, the whole Beneventan community. A remarkable collection of Beneventan texts offer vivid accounts of the public celebrations and rituals that accompanied the translation of the relics of twelve martyred brothers (known simply as the Twelve Brothers), collected from various locations in Southern Italy, and of Saint Mercurius, to Santa Sofia in 768 and 774 respectively. For example, the Historia corporis sancti Mercurii, one of a num- ber of texts collected in a thirteenth-century Beneventan lectionary, describes Arechis leading a festive procession into the city – an event dramatically marked by the saint’s refusal to let the wagon carrying his body cross the bridge into the city until Arechis, removing his regalia and donning a hair-shirt, prayed to him.504 Another, possibly earlier version also stresses Arechis’ ‘triumphal’ entry into Bene- vento with Mercurius, accompanied by townsfolk and clerics chanting hymns and carrying banners before the saint’s dramatic intervention.505 These martyr cults were to be the focus of the new church. 502 E.g. Liber preceptorum, I.18, 1:362: Do(m)nus Arichis piisimus atq(ue) excellentissimus princ(eps) gentis Langubardor(um), divino p(re)monitus nutu, offero in eccl(esi)a S(an)c(t)e Sophię qua(m) a fundam(en)tis edificavi, p(ro) rede(m)ptio(n)e a(nim)e meę seu p(ro) salva- tion(ne) gentis n(ost)re et patrię. This formula is consistent in all the post-774 charters of Arechis. 503 Ibid., III.23, 2:516: ut tantu(m)modo ad sacro n(ost)ro palatio audi-entiam habere debeas sicut et ceteros n(ost)ra monasteria, in n(ost)ra sit potestate(m) ordinandi ipse n(ost)ra ęccl(esi)a qualiter n(ost)ra exemietas placuerit, et nullus ep(iscopu)s huius s(an)c(t)e sedis ęccl(esi)ę n(ost)rę Bene-ventane vel ab ei(us) sacerdotibus nunqua(m) req(ui)ratur in suo iure d(omi)nio 504 Historia corporis sancti Mercurii, XI, in Acta passionis et translationes sanctorum martyrum Mercurii ac XII Fratrum, ed. Vittorio Giovardi (Rome: Typis Joannis Baptistae a Caporalibus, 1730), 60–61. The texts contained in this manuscript (Biblioteca Giovardiana, Veroli, MS Verolensis 1) belong to a number of different historical ‘strata’. Given their manuscript context, their date of composition can only be hypothesised. See: Ian Wood, ‘Giovardi, MS Verolensis 1, Arichis and Mercurius’, in Zwischen Niederschrift und Wiederschrift: frühmittelalterliche Hagiographie und Historiographie im Spannungsfeld von Kompendienüber- lieferung und Editionstechnik, ed. Richard Corradini et al., Forschungen zur Geschichte des Mittelalters 18 (Vienna: Verlag der Österreichischen Akademie der Wissenschaften, 2010), 197–210. 505 Translatio sancti Mercurii, c. 2, ed. Georg Waitz, in Scriptores rerum Langobardicarum, 577: Hinc melos et cantica, hinc ymni et iubili, clero populoque dulci modulamine decantantibus, celum faciunt resonare, sicque vexillis precedentibus, velut tropheo habito, martiris expeditio iam appropriat Benevento. Binon hypothesised that this was the earliest extant version, perhaps belonging to the eighth century: Stéphanie Binon, Essai sur le cycle de Saint Mercure: Martyr de Dèce et meurtrier de l’empereur Julien, Bibliothèque de l’École des Hautes Études. Sciences Religieuses 53 (Paris: Librarie Ernest Leroux, 1937), 44. 134 In the twentieth-century excavations the foundations of two altars were discovered: one in the centre of the main apse, the other in front of it in the aisle formed by the pilasters and columns. On the basis of these discoveries, we can tentatively trust the translatio narratives, which record that Arechis had the Twelve Brothers relocated in front of the main altar, which was given over to Mercurius.506 On the face of it, there is little which would recommend that the Beneventan Santa Sofia was directly based upon the Justinianic Great Church. Certain architectural features nonetheless signalled a link between the two structures, associations which were prompted to the observer by their shared dedication.507 Greek observers from Procopius to the tenth-century Diegesis placed consistent emphasis on Hagia Sophia’s dome as the crowning achievement of the church’s architecture and on its conse- quent ‘rounded’ interior experience; likewise, Adomnán of Iona’s On the Holy Places, a text which cir- culated widely in the eighth and ninth centuries, described Hagia Sophia as a ‘rotunda of astounding size’.508 Similarly, Santa Sofia’s architecture can be read as an analogous effort to combine the liturgical needs of a ‘basilican’ church with a ‘centrally-planned’ interior experience. Although the space-defining interior supports are organised in a ‘rounded’ fashion, clear directions of sight and movement are emphasised in the interior which distinguish it from a ‘round’ church with an ambulatory. The ar- rangement of the columns and pilasters create a definite central aisle from the door to the altar, wid- ening into a ‘nave’ in the centre under the dome, flanked by two narrower aisles on either side. The ‘zig-zag’ walls, while not fully blocking passage around the outer edge, de-emphasise it as a possible direction of movement, thus articulating a threefold division in space laterally (nave and aisles) and longitudinally (the entrance, widening into the nave, which is in turn divided from the area of the 506 Translatio Mercurii c. 4, 578; Historia corporis Sancti Mercurii XI, 61; Belting, ‘Studien’, 181. 507 Belting, ‘Studien’, 188–89. 508 Adomnán, De locis sanctis, ed. Denis Meehan, Scriptores Latini Hiberniae 3 (Dublin: Dublin Institute for Advanced Studies, 1958), 108: rotunda mirae magnitudinis lapidea ecclesia; Belting, ‘Studien’, 189. 135 sanctuary). That is to say, Santa Sofia is not a church which is meant to be circulated; rather, its interior disposition insists on its east-west axis, culminating in the altars to the saints. *** While the languages of Christian doctrine provided new symbolic opportunities in the Early Middle Ages, out of the liturgification of authority also arose new discourses of constraint and responsibility. Injunctions that kings should be wise, fair, and just are attested for the earlier period as we have seen, yet by the ninth century these had acquired the status of moral sanction. Legitimate authority, in Car- olingian political theology, was contingent on moral rectitude. As Hincmar puts it, The king ought to maintain within himself the dignity of his own name. For the name “king” intellectually signifies that he should fulfil the office of “corrector” for all his subjects. But how can he who does not correct his own morals be able to correct others when they are wicked? It is by the justice of the king that the throne is exalted, and by truth that the governments of people are strengthened.509 What Hincmar highlights is that discourses of Christian authority provided a further axis for local negotiations of authority and its limits at the political centre. In Aachen and Oviedo, the intended audience of any ‘royal liturgy’ must have been the court aristocracy and the ministering clergy. The ‘distancing’ of the king within ecclesiastical space, therefore, took place in the specific context of the dynamics of court society. In Benevento, on the other hand, we saw a different set of intended audi- ences and messages, keyed to local discourses of authority taking place between ruler, elite, and bishop – a balance that was undone in the decades following Arechis’ death.510 Despite the numerous archi- tectural differences, that Santa Sofia provided the focus for rituals manifesting Arechis’ authority in a 509 Hincmar, De ordine palatii, c. 6: Et rex in semetipso nominis sui dignitatem custodire debet. Nomen enim regis intellectualiter hoc retinet, ut subiectis omnibus rectoris officium procuret. Sed qualiter aloios corrigere ptoerit qui proprios mores ne iniqui sint non corrigit? Quoniam iustitia regis exaltabitur solium et veritate solidantur gubernacula populorum. Trans.: David Herlihy, trans., The History of Feudalism, The Documentary History of Western Civilization (London: Macmillan, 1970), 212. 510 Further ‘political’ appropriations of relics from regions of southern Italy under Byzantine authority took place under Sico (817–832) and Sicard (832–839). First, Januarius (San Gennaro), martyred bishop of Benevento though patron of Naples, was reinterred in the cathedral, refitted and rededicated for the purpose. In 838 the apostle Bartholomew was brought from Lipari by Sicard and placed in a new church in the vicinity of the cathedral, acts that mark the growing authority of the church in the city. Belting, ‘Studien’, 159–64. 136 local social field underlines its fundamental similarity of function, if not necessarily form, to Aachen and the churches of Oviedo. VII. Conclusion This chapter has focused on the manifold relationships between the architecture of the palace, its social space, and the production of legitimate authority in the encounter of ruler and elite. The rituals of the palace were not assertions of autocratic imperial power or untrammelled royal authority. Rather, as we saw first in the examples from Cassiodorus and Notker, the palace was a social space, participa- tion in which was governed by ritualised codes of behaviour. Acting in line with these expected norms—to be able to play by the ‘rules of the game’—defined the limits of the political community in the post-Roman world. In places, we have seen how these ritual codes could indeed be manipulated to elevate the position of the king over the rest of the community, as in the distance imposed between ruler and elite in the spatial configurations of the audience hall or the palace chapel. Yet these codes also constrained royal action: kings were expected to act in certain ways in the audience hall, for ex- ample, to preserve the equilibrium of the political system. Ritual was not always agonistic, however, nor was it a zero-sum game in which either ruler or elite ‘won’.511 As in the case of hunting, ritual communicated norms and identities shared between the political community writ large, including the king. As recent historical research has emphasised, medieval politics was based fundamentally on ne- gotiating consensus between the royal centre and regional aristocracies.512 The palace provided the symbolic arena in which these relationships were forged. On a wider level, I have also suggested that the fragmentation of the Western Roman Empire transformed the spatial expression of political ritual. As we saw in the previous chapter, the 511 It is essential to Elias’ analysis of Louis XIV’s Versailles that ritual was competitive and conflictual: the king’s authority depended on his curtailing the power of court aristocrats, on whose service he depended, through demotions in ritual hierarchies: Elias, Court Society, e.g. 120, 131, 181. 512 Stuart Airlie, ‘The Aristocracy in the Service of the State in the Carolingian Period’, in Staat im frühen Mittelalter, 93–111. 137 iconography of rulership was radically changed after the sixth century. Here, we saw that the audiences of political ritual, and its embodied ideologies, underwent a similar transformation at the same junc- ture. While Late Roman ‘public’ ritual had a significant afterlife in the post-Roman regna, by the seventh century, the spaces in which authority was legitimised had been fundamentally reimagined. Put simply, power had moved from the streets into the palace. 138 PART II INFRASTRUCTURE AND LANDSCAPE CHAPTER 3 Palaces, Militarisation, and the Political Landscape of Ostrogothic Italy I. Introduction Situated opposite the image of Ravenna’s palatium in Sant’Apollinare Nuovo is a similarly arresting representation of the CIVITAS CLASSIS, ‘the city of the fleet’ – otherwise known as Ravenna’s port, Classe (Fig. 103). Although much altered in the restoration campaigns of the nineteenth century (and, as with the palatium mosaic opposite, having already suffered the excision of figures in the sixth cen- tury), much of the Classe mosaic is original and the restored sections at least seem to reflect the original design.513 At left, adjacent to Classe’s walled urban area, three ships lie at rest within the golden walls of the harbour, two at rest and one under sail (Fig. 104). Their sleek profile and sharply upturned prows are recognisable as the fast warships known as dromones (δρόμωνες), ‘runners’, a mainstay of the Byzantine navy until at least the thirteenth century.514 Surprisingly, and in contrast to the palatium, the mosaic of Classe has elicited little comment in scholarship. Yet these ships raise important questions for a consideration of the integration of the Ravennate palace into wider systems through which po- litical authority was actioned. We must ask: what was the purpose of these ships, and why were they represented so prominently in Theoderic’s palace church? What relation do the goods or personnel they carried have to the court assembled below? A context for them may be offered by Cassiodorus in the Variae. A letter addressed to the Praetorian Prefect Abundantius from Theoderic in around 525 or 526 gives instructions that, 513 Bovini, ‘Antichi rifacimenti’, 77–81; Deichmann, Ravenna, 2.2:145–46, and above, Chapter 1. 514 See especially: John Pryor and Elizabeth Jeffreys, The Age of the ΔΡΟΜΩΝ: The Byzantine Navy, ca. 500–1204, The Medieval Mediterranean 62 (Leiden: Brill, 2006). Pryor and Jeffreys identify three features that came to distinguish the dromon from the Roman liburna galley: a smaller size but a full (rather than half) deck, the distinctive waterline ram being replaced by an above-water spur, and lateen, rather than square, sails. Iconographic evidence shows that this type had become current already in the late-fifth and early-sixth centuries: ibid., 127–62. Procopius writes that the force sent to Africa in 533 included dromons: Proc., Wars, III.11.15, 1:362. 140 since a worry frequently assaults our mind that Italy does not possess ships, where such a great abundance of timbers is produced that it is even exported to other provinces on request, with God’s inspiration we decree that you presently undertake the con- struction of a thousand galleys (dromones) which can both carry the public grain and, if it were necessary, stand against opposing ships.515 In the next letter in the collection, with work apparently underway, Cassiodorus congratulates Abun- dantius on his efforts and possibly reveals the motivation behind the command: ‘No longer will the Greek take advantage of us nor will the African insult us’.516 Although there is no evidence that these ships were ever actually built—and some scepticism towards the feasibility of such a large project would be justified—Cassiodorus’ letters to Abundantius are revealing. Written against the backdrop of increasing instability in the Mediterranean as the consensus of the previous three decades began to crumble, these letters speak to the growing challenges of securing sufficient foodstuffs to service the demands of the annona – the centrally-organised provisioning of the city of Rome, the army, and the bureaucracy. While we cannot know for sure that these are real-life counterparts to the ships repre- sented in Sant’Apollinare Nuovo, considering them together reminds us of a simple fact. The repre- sentation of royal authority (as we examined in the previous chapter) was not separable from the systems of surplus extraction and distribution through which it was exercised. In other words, the palace, as a localisation of authority, was not some neutral rhetorical stage for the powerful to parade themselves, separate to the rest of society, but integral to—and more, constitutive of—broader social and economic networks. The objective of this chapter and the next, therefore, is to shift our focus from the palatium as a spatial product to be interpreted singularly, to palatia, a plurality of localisations of political authority 515 Cass., Var., V.16.2, 152: Cum nostrum igitur animum frequens cura pulsaret naves Italiam non habere, ubi tanta lignorum copia suffragatur, ut aliis quoque provinciis expetita transmittat, deo nobis inspirante decrevimus mille interim dromones fabricandos assumere, qui et frumenta publica possint convehere et adversis navibus, si necesse fuerit, obviare. Nothing is known of Abundantius outside Cassiodo- rus’ Variae. He was still Praetorian Prefect in 527, when he receives a letter from Cassiodorus in the name of Athalaric (Cass., Var., IX.4). See: PLRE, 2:3–4. 516 Cass., Var., V.17.3, 153: Ornasti rem publicam tua institutione reparatam. non habet quod nobis Graecus imputet aut Afer insultet. Cassiodorus here refers to Byzantine and Vandal raids of Italian shipping lanes. 141 and their evolving role in the formation of early medieval polities. These chapters will show that pal- aces were not only the spaces in which power was produced, but also the tool of its action. In this way they embody the term ‘architecture’ in both its senses: as buildings, spaces designed and constructed for human habitation and use, but also as structures, as the supporting medium of systems and net- works.517 In Italy under the Ostrogoths, the subject of the present chapter, and in the Frankish regnum, that of the next, we will see how early medieval kings and courts drew on and reordered pre-existing Roman architectures to form new territories—new ‘landscapes’ through which their authority could be manifested—and, above all, that palaces were instrumental in the material and conceptual trans- formations of the Roman world. From histories of the State to archaeologies of landscape To take this interpretive step, let us ask what is, on the face of it, a simple question: what was the role of the palace in early medieval society? As we saw in Chapters 1 and 2, most scholarship has hewn to what we could term here an ‘ideological’ approach: that the palace’s impressive appearance compelled obedience from a surrounding populace, a tacit theorisation of the social function of architecture that can be detected in numerous contexts.518 To be sure, psychological domination through sheer visibility could be a possible design goal, but it was certainly not the only one.519 Eyal Weizman has shown that infrastructures and architectures facilitate Israeli domination of Palestinians in ways not limited to the visual and psychological: they disrupt existing networks of everyday life and form the basis of new 517 Recalling Easterling’s concept of infrastructure as an ‘operating system’ and an instrument in political domination: Keller Easterling, Extrastatecraft: The Power of Infrastructure Space (London: Verso, 2014), 11–23. 518 Take, for example, the summer capital of the American administration in the Philippines, Baguio, which was deliberately designed by its American architect, Daniel Burnham, to emphasize the visibility of the governmental buildings of the colonial administration and thereby elevate them as symbols of just, rational governance over its colonial subjects: Wolf- gang Sonne, Representing the State: Capital City Planning in the Early Twentieth Century (Munich: Prestel, 2004), 90–94. 519 As Paul Zanker has argued, ‘in any historical context the shape of a city presents a coherent system of visual commu- nication that may affect its inhabitants even at the subconscious level by its constant presence’: Paul Zanker, The Power of Images in the Age of Augustus, trans. Alan Shapiro (Ann Arbor, MI: University of Michigan Press, 1988), 18. 142 territorialisations enforced through economic, legal, and violent means.520 My point is not to draw equations between West Bank settlements and early medieval palaces, but rather to note that architec- tures of authority were not only visually imposing with calculated effect, but also instruments of control through less immediately ‘visible’ methods – for example, the control of movement, surveillance of infrastructure networks, or imposition of economic obligation.521 Establishing a connection between the palace and society, therefore, requires that we look beyond its walls – to its role in the spatial exten- sion of authority over the landscape of the early medieval polity. ‘Landscape’ is used here in a specific sense. Entering the purview of anthropology and archae- ology from human geography, landscape and its role in mediating social relationships has been reimag- ined over the last three decades. Whereas previously landscape was understood to form the mere backdrop against which social action took place, more recent scholarship has positioned it as an active participant in the formation of social relationships – and itself a social product.522 As a product (or, as Hirsch puts it, a dialectical process of social and cultural production), inherent to a landscape’s for- mation is its temporality, the build-up of ‘strata’ formed by the everyday lives of previous generations as much as large-scale interventions.523 The landscape thus accumulates over time and stands as a record, or ‘memory’, of the social relations implicated in its creation.524 By extension, however, the landscape is not politically neutral. As representation, landscapes (whether literary, pictorial, or physical) were 520 Eyal Weizman, Hollow Land: Israel’s Architecture of Occupation, 2nd ed. (London: Verso, 2017). For the terrifying possibilities of visual reciprocity between Israeli and Palestinian settlements: ibid., 111–37. 521 For the ‘unseen’ components of domination in this sense, see also: Easterling, Extrastatecraft. 522 There is not space here for a comprehensive history of this term and its varied uses by different fields. Specifically on theorizations of landscape in archaeology, see: A. Bernard Knapp and Wendy Ashmore, ‘Archaeological Landscapes: Con- structed, Conceptualized, Ideational’, in Archaeologies of Landscape, ed. Wendy Ashmore and A. Bernard Knapp (Malden, MA: Blackwell, 1999), 1–8; Smith, Political Landscape, 30–77. The critical re-evaluation of landscape mimics that of space set in train by Lefebvre (see above, Introduction). 523 Tim Ingold, ‘The Temporality of the Landscape’, World Archaeology 25, no. 1 (1993): 162; Eric Hirsch, ‘Introduction. Landscape: Between Place and Space’, in The Anthropology of Landscape, ed. Eric Hirsch and Michael O’Hanlon, Oxford Studies in Social and Cultural Anthropology (New York: Oxford University Press, 1995), 4, 22. 524 Ingold, ‘Temporality’, 154. Cf. Lefebvre, Production, 193. 143 and are intimately connected with colonial, imperial, and national projects.525 This re-evaluation of landscape and its role in history has inspired a number of ground-breaking studies of Late Antiquity and the Early Middle Ages, but in general its application to questions of social organisation have been muted outside of archaeology.526 Centring on palaces’ role in forming landscapes of political authority—what will be referred to here as ‘royal landscapes’, drawing on Adam T. Smith’s concept of the ‘political landscape’—pro- vides a theoretical vocabulary for bridging the gap in traditional accounts between the spaces in which authority was produced and its action in the polity.527 The authority of the new rulers of post-Roman Europe was not innate, but had to be established and extended. The ‘post-Roman kingdoms’ were not even the inevitable outcome of the Western Empire’s disintegration; the forms and structures of political authority itself were renegotiated in the decades and centuries that followed. As Smith writes, the central spatial problem in the formation of the polity is ‘the delineation of a bounded territory within which a sovereign regime rules a community of subjects integrated by a shared sense of identity that binds them together in place.’528 Surprisingly, a sense of the ‘emplacedness’ of palaces and early medieval politics is lacking in traditional accounts. As we saw previously, the sedes regiae are less arenas of social action so much as instantiations of broader trends of the post-Roman transition, such as trends of urban continuity or imitatio imperii. Early medieval politics are thus abstracted from the particularities of the spaces and 525 Essential here is: Edward Said, Orientalism (New York: Pantheon, 1978). Carter describes the reduction of the mytho- poetic and culturally significant places and spaces of Australia to a mere ‘historical stage’ in its colonisation, with the (imperial) historian as the spectator: Paul Carter, Road to Botany Bay: An Exploration of Landscape and History (London: Faber and Faber, 1987; repr. Minneapolis, MN: University of Minnesota Press, 2010). See also, e.g.: W. J. T. Mitchell, ‘Imperial Landscape’, in Landscape and Power, ed. W. J. T. Mitchell, 2nd ed.(Chicago: University of Chicago Press, 2002), 5–34; Simon Schama, Landscape and Memory (New York: Alfred A. Knopf, 1995). 526 To name but a few useful recent studies: Paolo Squatriti, Landscape and Change in Early Medieval Italy: Chestnuts, Economy, and Culture (Cambridge: Cambridge University Press, 2013); Susan Oosthuizen, The Anglo-Saxon Fenland (Oxford: Wind- gather, 2017); Darlene Brooks Hedstrom, The Monastic Landscape of Late Antique Egypt: An Archaeological Reconstruction (Cam- bridge: Cambridge University Press, 2017). 527 Smith, Political Landscape. 528 Ibid., 151. 144 places in which it took place. In this analysis, therefore, I seek to re-orient our attention towards the relationships palaces sustained between regimes, institutions, and elites.529 As we saw in the previous chapter, these ties had a spatial component. The presence of the court in a given place (whether per- manently or temporarily) overlaid onto the geography of the polity a differentiation between centre and periphery.530 Similarly, royal authority was not a kind of immanent force, permeating the polity without interruption to its bounds. Smith, building on the geographic metaphor of Solzhenitsyn’s The Gulag Archipelago, observes that the Soviet labour camp system was so successfully (and terrifyingly) coercive not only on the basis of the firsthand experience on the part of Soviet citizens, but also in the way that a highly ‘discontinuous series of spaces were articulated in the minds of those it ensnared into a coherent landscape of authority’.531 That is to say, a concentration on the political dimensions of the landscape centres our analysis on how the highly localised spaces where royal authority was apprehended were connected to form the boundaries of the polity itself. At stake here is emphatically not a deterministic account of how pre-existing Roman infra- structures laid the tracks that shaped subsequent social development, a perspective so ‘zoomed out’ that only extreme continuity is perceptible. Rather, a focus on royal landscapes highlights how fluctu- ating political relationships forged between regimes and subjects played out in the perception and experience of political authority within specified contexts. While Roman institutions, infrastructures, and mentalities most certainly did shape the potentialities of the post-Roman kingdoms (and, as we will see, in different ways on either side of the Alps), an attention to landscape recasts these threads of Roman continuity as ‘resources’ that were available to early medieval actors in the redefinition of 529 Smith usefully indicates four sets of relationships as pivotal: ties between polities (geopolitics), between subjects and regimes (constitutive of polities), between elites and ‘grassroots’ organizations (constitutive of regimes), between institu- tions within the governing apparatus: Smith, Political Landscape, 26. 530 See above, Introduction. 531 Smith, Political Landscape, 78–79 (quote at 79). 145 social and political relationships.532 Soja, reformulating Marx’s famous maxim from The Eighteenth Bru- maire of Louis Bonaparte, helpfully captures this understanding of the relationships between landscapes, regimes, and individuals: ‘We make our own history and geography, but not just as we please; we do not make them under circumstances chosen by ourselves, but under circumstances directly encoun- tered, given and transmitted from the historical geographies produced in the past.’533 Overview This chapter examines the royal landscapes produced in the post-Roman kingdom of Italy ruled by the Gothic king Theoderic and his successors. Whereas the majority of scholarly attention has been focused on Ravenna, the kingdom’s supposed ‘capital’, this chapter seeks to shift our focus to the broader network of royal centres owned by the Amal rulers, a theme that has been underexplored in scholarship to date. We will see that the palaces of this network were not secondary to Ravenna, but integral parts of the kingdom’s administration. By reframing our perspective in this way, the strategies by which it was ruled—and how they already differed from Roman practices—are shown in a new light. The palatial network shows that royal authority was principally encountered in the north, and its relation to defensive infrastructures in the Po Plain and in the Alps shows that palaces were located with military considerations in mind. These palaces, moreover, were not just sites of representation, but the supporting medium of a reoriented system of taxation and redistribution centred on the Po, with the maritime city of Ravenna at its head. 532 See Fafinski’s treatment of the infrastructures of early medieval Britain as a ‘governance resource’: Mateusz Fafisnki, Roman Infrastructure in Early Medieval Britain: The Adaptations of the Past in Text and Stone (Amsterdam: Amsterdam University Press, 2021), esp. 21–42. I thank Mateusz for our conversations on the relations of infrastructure and authority, and for sharing an advance copy of his book with me. See also: Clemens Gantner et al., eds., The Resources of the Past in Early Medieval Europe (Cambridge: Cambridge University Press, 2015). Though dealing principally with textual ‘resources’, the concept elaborated in this volume for the repertoire of ideas available for imagining new social groupings is a useful one for de- noting the agglomeration of meaning in significant places. 533 Soja, Postmodern Geographies, 129. 146 II. Northern Italy as a Royal Landscape Palaces were, for the likes of Cassiodorus and Ennodius of Pavia, symbols of the Italian polity’s health. As we saw previously, Ennodius praised the restoration of palaces as a sign of the beneficent govern- ment of Theoderic’s regime.534 Cassiodorus concurred, telling the urban prefect Agapitus that, ‘it is proper that the ruler should direct his attention to things that improve the commonwealth, and truly is it fitting that the king decorate his palaces with buildings’.535 Most treatments of the Italian kingdom have focused exclusively on the palace of Ravenna as the kingdom’s ‘capital’, with little to no consid- eration of the other sites mentioned in contemporary sources.536 The Italian kingdom thus becomes a mirror-image of the contemporary Byzantine Empire, in which Constantinople was the undisputed ceremonial and administrative centre of an immobile governing apparatus and emperors had by the sixth century ceased to range far from the imperial seat.537 As we saw already in Chapter 1, however, such similarity is illusory – a rhetorical device of sources produced at the Ravennate court that ‘masked a strategy for governing that was strikingly different from that of the emperor at Constantinople’, in Shane Bjornlie’s words.538 In a letter to the eastern emperor Anastasius only a year after Ennodius delivered his panegyric, Cassiodorus wrote that, ‘our kingdom is an imitation of yours, the model of good design, the exemplar of the singular imperium; however much we follow you, we surpass other peoples by the same amount’.539 We might say that his words have been taken too literally. 534 Ennodius, Panegyricus, XI.56, 210. See above, Chapter 1. 535 Cass., Var., I.6.1, 16: Decet principem cura quae ad rem publicam spectat augendam, et vere dignum est regem aedificiis palatia decorare. 536 E.g. Sauro Gelichi, ‘Ravenna, ascesa e declino di una capitale’, in Sedes regiae (ann. 400–800), 109–34; Johnson names Pavia and Verona ‘Theoderic’s secondary capitals’, but remarks that ‘the focus of Theoderic’s patronage […] was on his capital, Ravenna’: Johnson, ‘Toward a History’, 77–78. 537 E.g. Valérie Fauvinet-Ranson, ‘Capitales et residences du roi Théodoric à l’orée du VIe siècle’, in Le gouvernement en déplacement: Pouvoir et mobilité de l’Antiquité à nos jours, ed. Sylvan Destephen et al. (Rennes: Presses Universitaires de Rennes, 2019), 176. 538 M. Shane Bjornlie, ‘Governmental Administration’, in A Companion to Ostrogothic Italy, ed. Jonathan Arnold, et al., Brill’s Companions to European History (Leiden: Brill, 2016), 57. 539 Cass., Var., I.1.3, 10: Regnum nostrum imitatio vestra est, forma boni propositi, unici exemplar imperii; qui quantum vos sequimur, tantum gentes alias anteimus. 147 Mapping Ostrogothic Palaces Given the importance that palaces held in the rhetoric of the Ravennate court, it is striking that schol- arship has generally failed to consider them together, as a network. Ravenna has received by far the lion’s share of attention, and sites like Verona and Pavia have generally been treated individually (mostly in Italian) or cursorily at best. As a result, therefore, as much as recent scholarship has illumi- nated the reception of Theoderic’s quasi-imperial image, for example, we still possess a comparatively dim understanding of how it was put to use.540 Before turning, therefore, to the spatial practices that transformed a discontinuous network of royal sites into a landscape through which royal power was exercised, it is first necessary to catalogue as completely as is possible on the basis of the extent sources the palaces of the Ostrogothic kings (Fig. 105). Ravenna. Ravenna was clearly the most important residence of the Ostrogothic kings and contemporary sources are unanimous that it was the chief city of the kingdom. The preference for Ravenna over Rome has usually been interpreted as the Gothic king maintaining a respectful distance from the Roman Senate.541 While plausible, as we will see, other factors were more significant in situ- ating royal power in Ravenna. Nonetheless, it cannot be ignored that when the court is mentioned in written sources, we most often find it in Ravenna. Meetings between the bishops of Rome and the king seem to have taken place in Ravenna, for example.542 Similarly, throughout Procopius’ narrative 540 Bjornlie, ‘Governmental Administration’ is a welcome step in this direction. As he argues, one of the most fundamental functions of a regime is ‘the production of rhetoric (or an ideology) by which the governed understood themselves to be justly ruled’ (ibid., 48), thereby shifting attention from simple questions of ‘continuity or decline’ to the mobilisation of this rhetoric for the legitimisation of the administration’s institutional tools (fiscal extraction, etc.). 541 E.g. Deliyannis, Ravenna, 114. Reydellet suggests that Theoderic respected an existing distinction between Rome, as the ‘capital of the Senate’, and Ravenna, as the ‘capital of the sovereign’: Marc Reydellet, ‘La regalità teodericiana’, in Dall’età bizantina all’età ottoniana, bk. 2, Ecclesiologia, cultura e arte ed. Antonio Carile, vol. 2 of Storia di Ravenna (Ravenna: Marsilio, 1992), 10. As Gillett has shown, however, no such distinction was strictly observed during the fifth century: Andrew Gillett, ‘Rome, Ravenna and the Last Western Emperors’, PBSR 69 (2001), 131–67. Generally: Bjornlie, Politics and Tradition, 127–44. Critiquing the Ravenna-Rome dichotomy: Peter Heather, ‘A Tale of Two Cities: Rome and Ravenna under Gothic Rule’, in Ravenna: Its Role in Earlier Medieval Change and Exchange, ed. Judith Herrin and Janet Nelson (London: Institute of Historical Research, 2016), 15–38. 542 E.g. Liber Pontificalis, 54.8, 1:270: Hic papa Hormisda perrexit ad regem Theodoricum Ravenna (‘Pope Hormisdas hurried to king Theoderic at Ravenna’). During the Laurentian Schism, the papal parties met with Theoderic at Ravenna: Liber Pon- tificalis, 53.2, 1:260. Deliyannis, Ravenna, 114. 148 it is taken for granted that Ravenna was the centre of the kingdom.543 Its elevated status is confirmed, albeit in a highly rhetorical context, by Cassiodorus in one of last letters of the Variae, where he refers to it as the ‘royal city’ (urbs regia).544 Verona and Pavia. Although Ravenna may well have been the court’s main residence, it was certainly not the only one. Contemporary sources confirm that Theoderic also founded palaces in Verona and Pavia. Although evidence (archaeological or otherwise) for an Ostrogothic presence in these cities is sparse at best, they were closely associated with the regime. According to the Anonymus Valesianus, at Verona he founded both baths and a palace, as well as a colonnade stretching from the palace to the city gate, a new aqueduct to supply the city, and an addition to the wall circuit.545 At Pavia, he likewise ordered the construction of a palace, baths, amphitheatre, and new city walls.546 In both cities, the image we are confronted with is a palatial building programme of a similar nature and scale to those of the Tetrarchic sedes detailed in Chapter 1, though it is probable that many of the regime’s interventions were limited to restoration, expansion, and refurbishment rather than construc- tion ex novo.547 These were surely substantial projects and, according to Agnellus at least, were decked out accordingly. He asserts that he had himself seen an equestrian image of Theoderic executed in mosaic in Pavia, the counterpart to that which had previously existed in Ravenna.548 The Pavese palace would subsequently be adopted by the Langobard kings and, as far as can be seen, remained in use 543 Siege of Ravenna by Theoderic and his forces: Proc., Wars, V.1.15–25, 2:6–8. Procopius is explicit that the intrigues surrounding Amalasuntha’s influence over her son, Athalaric, took place at Ravenna, e.g.: ibid., V.2.29, 2:14–15. A mes- senger from Justinian meets Amalasuntha at Ravenna: ibid., V.3.16, 2:17. He further reports that the royal treasure was taken from the palace in Ravenna: ibid., VI.29.37, 2:288; VII.1.1, 2:297–98. After Ravenna fell, many other garrisons sent word to Belisarius that they also intended to surrender: ibid., VI.29.39–41, 2:288. 544 Cass., Var., XII.22.4, 378. See below for further commentary on this passage. 545 Anon. Val. (II), XII.71, 20: item Veronae thermas et palatium fecit et a porta usque ad palatium porticum addidit. aquae ductum, quod multa tempora destructum fuerat, renovavit et aquam intromisit. Muros alios novos circuit civitatem. 546 Ibid., 21: item Ticini palatium thermas amphitheatrum et alios muros civitatis fecit. 547 La Rocca, ‘Mores suos fabricae loquuntur’, 13–21. 548 Agnellus, LPRa, c. 94, 258: Post uero depredata a Longobardis Tuscia; obsiderunt Ticinum, quae ciuitas Papia dicitur, ubi et Theod- ericus palatium struxit, et eius imaginem sedentem super equum in tribunalis camerae tesellis ornatam bene conspexi. On interpreting Agenellus’ description of the Ravenna image, see above: Chapter 1. 149 through the Early Middle Ages.549 The memory of the Gothic king persisted too: sometime between 906 and 910 a placitum was given at Pavia ‘under Theoderic’, possibly a reference to the mosaic seen by Agnellus, and Wipo later confirms that Theoderic’s palace was renovated by Otto III and Henry II.550 As will be seen below, the palace at Verona also seems to have remained in use through the tenth century. The clear impression from the written evidence is that the palace complexes at Pavia and Verona were significant investments that made lasting impressions in their respective urban fabrics. Other possible palace sites. Further sites are clearly associated with the royal fisc, even if the king cannot be placed there based on surviving evidence. For example, Milan is not named as a palace site in the sources, though the Anonymus Valesianus and Ennodius both mention that Theoderic did temporarily occupy the city during his invasion of Italy in 489/490.551 Notably, its mint remained in operation and a substantial number of gold emissions bearing Theoderic’s monogram are known, and Procopius characterised it as the most significant city in the north.552 It is tempting to suppose that it remained closely tied to the regime, even if specific instances of the royal presence are now lost. Other residences were possibly smaller-scale rural retreats. In a letter from between 507 and 511, 549 Referenced also by ‘Fredegar’: Fred. Chron., II.57, 82. 550 Cesare Manaresi, ed., I Placiti del «Regnum Italiae», vol. 1 (Rome: Tipografia del Senato, 1955), no. 122, 456–57: Dum in Dei nomine civitate Papia in sacro palacio hubi domnus Berengarius rex preerat, in laubia magiore ubi sub Teuderico dicitur, in iudicio resederunt Iohannes venerabilis sancte Ticinensis Ecclesie et Adelbertus sancte Bergomate Ecclesie (‘While, in the name of God, in the city of Pavia in the sacred palace where the Lord Berengar ruled, in the great gallery known as ‘under Theoderic’, in judgment sat the venerable John of the holy Church of Ticinum and Adelbert of the holy church of Bergamo’). Wipo, Gesta Chuonradi II. imperatoris, VII, in Die Werke Wipos, ed. Harry Bresslau, 3rd ed., MGH. SS. rer. Germ. 61 (Hannover: Hahn, 1915), 29–30: Erat in civitate Papiensi palatium a Theoderico rege quondam miro opere conditum ac postea ab imperatore Ottone tertio nimis adornatum (‘There was in the city of Pavia a palace, founded long ago by king Theoderic with wonderful work- manship and afterwards excessively adorned by emperor Otto III’). The palace was likely seriously damaged in a Hungarian raid in 924 that destroyed much of the city; placita in 935 and 945 (Manaresi, Placiti, vol. 1, no. 136, 507; no. 144, 551), refer to the palace being ‘newly rebuilt’ (in palacium noviter aedificatum) by Hugh of Italy (king 924–947). The palace was again destroyed by a rising of the commune in 1024 and was never rebuilt. See: Carlrichard Brühl, ‘Das “Palatium” von Pavia und die “Honorantiae Civitatis Papiae”’, in Atti del 4° Congresso internazionale di studi sull’Alto Medioevo (Spoleto: CISAM, 1969), 189–220 (for the destruction of the palace: 195–96). The razing of the palace in 1024 is recorded by Wipo in a famous passage: Wipo, Gesta Chuonradi, VII, 30. 551 Anon. Val. (II), XI.51, 53; Ennodius, Vita beatissimi Epifani episcopi Ticinensis ecclesiae, CIX, in Magni Felicis Ennodi Opera, 97–98. 552 Michael Andreas Metlich, The Coinage of Ostrogothic Italy (London: Spink, 2004), 31–33. The gold issues are all in the name of Anastasius, whose is named ‘officially’ in the legend; Theoderic’s monogram is smaller, stamped in the left-hand field. Bjornlie, ‘Governmental Administration’, 56; Proc., Wars, VI.21, 2:241. 150 Cassiodorus, writing for Theoderic, orders the architect Aloisius to oversee the restoration of the baths at Aponus (modern Terme Abano, Veneto), including ‘the palace, shaken by long old-age’ which, Cassiodorus implies, adjoined the public bath complex.553 The implication is that Aloisius is to restore an ancient Roman structure that has fallen into disrepair (though not disuse) – another sign of Theoderic’s care for the res publica.554 A later letter in the collection, written in 535 or 536 for Theoda- had (534–36) possibly implies the continued provisioning of a royal residence in Rome, as well as others at Pavia, Piacenza, and (cryptically) in ‘other places’.555 Rome. Whether the Ostrogothic kings used the imperial palaces in Rome is a more compli- cated question. It indeed is possible that Theodahad kept a residence there, as Cassiodorus’ letter above implies, but the Eternal City is otherwise conspicuous by its absence in the sources for most of this period. Notably, Theoderic can be definitively placed in Rome only for a six month stay in 500 for the celebration of his tricennalia, when he doubtless stayed in the imperial palaces on the Palatine.556 Later, judging by the Variae, Theodahad made a late attempt to move the administration to Rome in 535, but this was ultimately short-lived on account of Theodahad’s murder and Rome’s capture by the Byzantines only a year later.557 Rome clearly held an important place in the regime’s propaganda, where it was paired with Ravenna as an equal and pendant ‘capital’ of the kingdom.558 While avoiding Rome 553 Cass., Var., II.39.10, 69: Palatium quoque longa senectute quassatum assidua reparatione corrobora. Cassiodorus goes on: ‘Clear out the space that lies between the public hall and the first of the heated baths of its wooded roughness’ (ibid.: spatium, quod inter aedem publicam et caput ignite fontis interiacet, silvestri asperitate depurga). 554 Cassiodorus refers to the ‘Neronian swimming pool’: ibid., II.39.5, 68: in piscinam Neronianum. The rationale for the renovations are given perfunctorily at the end of the letter: ‘But who would neglect to conserve that place, as much as they might be sullied by their great stinginess – especially since something that has been singularly renowned throughout the world ornaments the kingdom?’ (ibid., II.39.12, 69: Sed quis ista conservare neglegat, quamvis plurima tenacitate sordescat? siquidem ornat regnum, quod fuerit singulariter toto orbe nominatum). 555 Cass., Var., X.28.1, 315: ad urbem Romam vel ad mansionem pertinent Ravennatem, sed et eos, qui ripam Ticinensem et Placentinam sive per alia loca quicumque publicos titulos administrare noscuntur. This letter, addressed to the Praetorian Prefect, concerns the confirmation of the appointment of stewards and vendors responsible for provisioning, among others, the palace of Ra- venna (here, mansio Ravennatem). 556 Anon. Val. (II), XII.65–69, 19–20. Cassiodorus, Chronicon, a. 1338/500, in Chronica minora saec. IV. V. VI. VII., ed. Theodor Mommsen, vol. 2., MGH. AA. 11 (Berlin: Weidmann, 1884), 160. 557 Preparations to move the comitatus are described by Cassiodorus, including the repair of the Via Flaminia and bridges on the Tiber: Cass., Var., XII.18–19, 375–76. 558 Fauvinet-Ranson, ‘Capitales et residences’, 176–81. 151 in this way was conceivably a token of respect towards—or, alternatively, a means of negating—the political influence of the senatorial aristocracy, the simplest explanation is simply that Rome offered significant symbolic capital (duly exploited in 500), but little practical value for a regime whose most significant geopolitical relationships, as we will see below, were directed north- and eastwards. Even if the king himself chose to stay away from Rome, it nonetheless bore a persistent visual and institutional presence of the Ravennate government. Bricks stamped with Theoderic’s name were used in a renovation of the Palatine’s Domus Augustana, in which the ‘garden stadium’—a peristyle courtyard laid out in the characteristic elongated shape of a circus (see fig. 19)—was transformed into a private amphitheatre, a fact that tentatively supports the Anonymus Valesianus’ statement that Theod- eric promised two hundred pounds of gold annually for its upkeep. 559 Amal interest in Rome was not limited to the Palatine, however. In his Chronicle, Cassiodorus makes a point of connecting defensive works sponsored by Theoderic— ‘most fortified castles’ that were added to the walls—with renova- tions to the city’s symbolic core including, he claims, ‘admirable palaces’.560 Jonathan Arnold has also raised the possibility that members of Theoderic’s wider family, including Theodahad, may have re- sided in Rome during his reign.561 Judging from the Liber Pontificalis, official meetings between the pope and the king usually occurred in Ravenna, yet during his adjudication of the Laurentian Schism Theoderic ordered the Roman clergy to convene the synod of 1 September 502 (the Synodus Palmaris) in Santa Croce in Gerusalemme, named in the acts as the ‘Jerusalem basilica of the Sessorian Palace’.562 Though the king himself was not present, the meeting was explicitly convened under royal authority 559 Andrea Augenti, Il Palatino nel medioevo: archeologia e topografia (secoli VI–XIII), Bullettino della Commissione Archeologica Comunale di Roma, Supplementi 4 (Rome: “L’Erma” di Bretschneider, 1996), 155–62; Wulf-Rheidt, ‘Die schwierige Frage’, 139–45. Anon. Val. (II), XII.67, 19. 560 Cassiodorus, Chronica, a. 1338/500, 160: sub cuius felici imperio plurimae renovantur urbes, munitissima castella conduntur, consur- gunt admiranda palatia, magnisque operibus antiqua miracula superantur. 561 Arnold, ‘Theoderic and Rome’, 119–20. 562 Acta synhodorum habitarum Romae, VI.5, in Cassiodorus Senatoris Variae, 428: Hierusalem basilica Sessoriani palatii. Theoderic was in Ravenna at this time according to the Liber Pontificalis: LP 53.2, 1:260. For the events of the schism: John Moorhead, Theoderic in Italy (Oxford: Clarendon Press, 1992), 114–39; Thomas F. X. Noble, ‘Theodoric and the Papacy’, in Teoderico il grande e i Goti d’Italia. Atti del 13° congresso internazionale di studi sull’alto medioevo, vol. 1 (Spoleto: CISAM, 1992), 404–16. 152 and was overseen by royal agents sent from Ravenna, Gudila, Bedeulf, and the vir inlustris and comes Arigern.563 The Anonymus Valesianus records a rebellious general named Odoin was executed at the Sessorian Palace during Theoderic’s stay in 500; we should likely conclude that the Sessorian Palace was still understood as an ‘official’ residence.564 Incerta. Finally, a number of other sites have been associated with the fisc but lack external textual or material evidence demonstrating royal ownership. For example, the sixth-century villas ex- cavated at Meldola and Galeata near Forlì, and Palazzolo, north of Ravenna have been raised as pos- sible sites of royal palaces.565 A site which can plausibly be identified with the latter is associated with Theoderic by Agnellus, but there is no direct material evidence of royal ownership at any of these three villas on textual or material grounds.566 Other sites are attested in written sources but their exist- ence has not been confirmed archaeologically. The late-seventh-century Ravenna Cosmography mentions a city named after Theoderic himself, Theodoricopolis, located somewhere in the ‘land of the Alamans’ (thus probably across the Alps in Rhaetia), yet the site is not mentioned in contemporary sources.567 Lastly, Paul the Deacon states that the queen Theudelinda’s palace at Monza was located on the site 563 Acta synhodorum habitarum Romae, II.7, 422: Gudilam et Bedeulfum sublimes viros maiores domus nostrae nos de praesente misimus cum inl. v. com. Arigerno; VI.1, 426. 564 Anon. Val. (II), XII.68–69, 20: Odoin comes eius insidiabatur ei. Dum haec cognovisset, in palatio, quod apellatur Sessorium, caput eius aputari praecepit. Marius of Avrenches confirms the date of 500: Marius of Avrenches, Chronica, a. 500, in Chronica minora, 2: 234: Eo anno interfectus est Odoind Romae. Moorhead explains this in terms of the king inheriting the imperial fisc and the basilica was thus ‘at Theoderic’s disposal’: Moorhead, Theoderic, 118. 565 Sfameni, Ville residenziali, 222–29; Giovanna Bermond Montanari, ‘La zona archeologica di Palazzolo’, Corso di cultura sull’arte ravennate e bizantina 30 (1983): 17–21; Sandro De Maria, ed., Nuove ricerche e scavi nell’area della Villa di Teodorico a Galeata: atti della giornata di study, Ravenna, 26 marzo 2002 (Bologna: Ante Quem, 2004). 566 Deliyannis, Ravenna, 122, with references to further literature. Agnellus describes a ‘small palace’ built on an island a few miles to the north of Ravenna near the Lion Port and the eventual site of his mausoleum: Agnellus of Ravenna, LPRa, c. 39, 195–96. Agnellus continues that he had this palace demolished and used the materials to build his own house. 567 Ravennatis anonymi Cosmographia, IV.26, in Itineraria Romana, ed. Joseph Schnetz, vol. 2 (Leipzig: Teubner, 1940), 61: Iterum propinqua ipsius Turringie ascribitur patria Suavorum que et Alamanorum patria, confinalis existit Italiae (‘Again to the nearby land of Thuringia is ascribed the land of the Suebi and the Alamans, which constitutes the frontier of Italy’); Aldo Settia, ‘Le fortificazioni dei Goti in Italia’, in Teoderico il Grande e i Goti d’Italia, 106. Compare with the slightly later foundations of Justiniana Prima (Caričin Grad, Serbia) and Reccopolis (Spain), discussed above, Chapter 1. 153 of a Theoderican predecessor.568 While certainly plausible, the presence of a Gothic-period palace at Monza cannot be verified outside Paul’s text. How mobile was the Ostrogothic court? By any measure, this is an impressive list of properties appertaining to the king – one that should already trouble the singularity of the attention focused on Ravenna. Faced with this multiplicity of residences, it is natural to ask: when and how were they used? Reopening the question of the move- ments of the Ostrogothic court reveals a more mobile and flexible mode of rulership in Italy than the framework of Byzantine ‘emulation’ otherwise allows. It is difficult to gauge the frequency and relative extent of the court’s movements, yet it is clear that it did not approach the scale of the peripatetic, constantly mobile rulership we might associate with, say, the Ottonians in the tenth century. The written sources only attest Theoderic’s presence in his other palaces a handful of times, and without the kind of locational data provided by (for example) the Frankish capitularies it is difficult to say for certain how often the court took to the road.569 None- theless, evidence from Italy strongly suggests that the court did move, and did so with some ease. Theoderic’s entry into Rome in 500, for example, was exceptional in his reign, but it is noteworthy that the king stayed for some months before returning according to the Anonymus Valesianus – business elsewhere evidently could wait. The events surrounding and following Boethius’ imprisonment and execution are also instructive. Boethius himself recalls that he advocated for Albinus and the Senate before the king at court in Verona.570 Judgment against him was, however, made at Pavia and he was imprisoned—and eventually executed—on an extra-urban estate likely near Calvenzano in the vicinity 568 Paul. Diac., Hist. Lang., VI.21, 123: Quo in loco etiam Theudericus quondam Gothorum rex palatium construxit, pro eo quod aestivo tempore locus ipse, utpote vicinus Alpibus, temperatus ac salubris existit. 569 On this evidence, see below, Chapter 4. 570 Boethius, De consolatione philosophiae, I.4.32, in Boethius, De consolatione philosophiae, Opuscula theologica, BSGRT, 2nd ed. (Munich: Saur, 2005), 16. For sources on Boethius, see: PLRE, vol. 2, 233–37. Bjornlie notes that Boethius’ last days show the ‘peripatetic nature of the court’: Bjornlie, ‘Governmental Administration’, 56. 154 of Milan.571 These events show the court moving between Verona and Pavia in the space of around a year (523–524), but it did not stop there. The Anonymus Valesianus then implies that Theoderic had Symmachus executed in Ravenna (525?), and the Liber Pontificalis likewise suggests that Pope John was condemned in the king’s presence in the same city (526).572 This would mean that the court was lo- cated, and was conducting hardly trivial business, in three separate locations in three years – an indi- cation that we should take the mobility of the Ostrogothic court seriously. The progress of these events highlights not only the mobility of the king and the court, but also a more general flexibility of the administration in Italy. Before judgment was passed on Boethius, Theoderic had the Urban Prefect, Eusebius, brought to Pavia.573 Not only was the kingdom’s admin- istration reactive to the movements of the court, but Eusebius’ trip to Pavia also signals that the lines delimiting the remits of various court officials were becoming increasingly blurred. Why, after all, would the Urban Prefect be necessary for judicial proceedings in Liguria? It is also striking that it is nowhere suggested that the court—and thus the legal recourse available to those able to attend it in person—was inaccessible on account of its mobility or that its location was unknown. In the Anonymus Valesianus, Jews from the city of Ravenna, suffering persecution by the Christian congregation, were able to petition the king at Verona with the support of the chamberlain (praepositus cubiculi) Triwane for restitution of the synagogues razed in the violence.574 Although the deaths of Boethius, 571 Anon. Val. (II), XIV.87, 25: rex vero vocavit Eusebium, praefectum urbis, Ticinum et inaudito Boethio protulit et eum sententiam. quem mox in agro Calventiano, ubi in custodia habebatur, misere fecit occidi (‘The king indeed summoned Eusebius, the Prefect of the City, to Pavia and he pronounced sentence without Boethius even being heard. Soon, he had him put to a pitiful death on the Calventian estate, where he was being held under guard’). Though Boethius himself does not name the place of his confinement in the Consolation of Philosophy, he does remark that it was some five hundred miles from his home: Boethius, De consolatione philosophiae, I.5.36, 16. Marius of Avenches records that Boethius was executed ‘in the territory of Milan’: Marius of Avenches, Chronica, a. 524, 235: Eo anno interfectus est Boetius patricius in territorio Milanense. 572 Symmachus’ execution is placed after both Boethius’ execution and the dispatch of the embassy under Pope John to Constantinople by the AV: Anon. Val. (II), XV.92, 26; LP, LV.6, 104–107, 276. As Noble has argued, relations between Rome and Ravenna had been harmonious until now, and so the idea that Theoderic had John deliberately killed due to some anti-Catholic sentiment is to be rejected: Noble, ‘Theodoric and the Papacy’, 419–23. 573 Anon. Val. (II), XIV.87, 25: rex vero vocavit Eusebium, praefectum urbis, Ticinum. 574 Anon. Val. (II), XIV.82, 23–24. Notably, the author implies that the king finding in the Jews’ favour was unjust (as an ‘unchristian’, and thus ‘tyrannical’ act), yet they stay quiet on whether the king’s ruling was resisted or resented: they simply note that the order, ‘with the order given to Eutharic Cilliga and bishop Peter II [of Ravenna], he passed this sentence and 155 Symmachus, and John were shocking to contemporary observers, and so locational information for where they occurred is preserved, it does not necessarily follow that the movements of the court were similarly so. A letter written by Cassiodorus for Theoderic sometime between 508 and 512 provides an insight into how such frequent movements might have been possible. In this letter, Cassiodorus com- manded the saio Wiligis to ‘load all the ships you can find at the city of Ravenna with corn from the taxes’ to supply the court at Pavia and to reduce the strain on local producers.575 The presence of the court, Cassiodorus notes, had drawn ‘crowds of onlookers’ who were in need of feeding – probably a reference to people attending the court to petition the king and his comitatus.576 Cassiodorus stresses the usual state of affairs has been reversed: ‘May Ravenna return to Liguria the bounty that it is accus- tomed to accept from there. For whatever place supports my presence ought to receive aid from many people.’577 The implication is that the royal court—which, including the administration, may well have represented a significant increase in local demand on produce—was due to reside in the north for some time. From this we can infer that the movement of the court involved a proportionally large number of individuals – unsurprising perhaps when we include not only the king and the consistorium, but also their attendants, clients, and bodyguards, the notarial staff, and plausibly also accompanying soldiers. This is certainly the picture we get from one of Cassiodorus’ later letters, written for Theoda- had, in which he describes preparations for the court’s movement to Rome in around 535.578 Hence Cassiodorus’ order to Wiligis provides rare insight into the court’s mobility in action: resource it was carried out’(ibid., 24: data praecepta ad Eutharicum Cilligam et Petrum episcopum secundum hunc tenorem [praecepit] et ita adimpletum [est]). 575 Cass., Var., II.20, 57: atque ideo praesenti decernimus iussione, ut quantas in Ravennati urbe exculcatorias potueris reperire, frumentis fiscalibus oneratas ad nos usque perducas. 576 Ibid.: trahit enim comitatus noster observantium catervas et, dum ad beneficia praestanda curritur, necessaria populi scopa postulantur. Bjornlie, ‘Governmental Administration’, 56. 577 Cass., Var., II.20, 57: reddat Ravenna copiam Liguriae, quam ex ipsa consuevit accipere. nam quae praesentiam nostram sustinet, multorum debet solacia reperire. 578 Ibid., XII.18–19, 375–76 156 networks responded to the court’s movements and were able to facilitate the royal presence as easily in Pavia as in Ravenna. *** What emerges from a closer look at the sources is the mobility and flexibility of the Ostrogothic regime. Neither the king nor the comitatus were tied Earlyto any one location for the administration to function. In fact, as seen in the case of Boethius’ trial, the administration could and did move fre- quently. As Bjornlie has argued, we are likely looking at a much simplified, ‘personalised’ administra- tive staff with proximity to the king—known in other contexts as Königsnähe—now the principal factor in the distribution of political power and with the traditional boundaries between high offices steadily being eroded.579 What is striking, however, is that the places with which the court can be associated are limited to a relatively small geographical area, predominantly the Po valley and at the foot of the Alpine passes. Before the invasion of Justinian’s armies in 535, it does not seem that Gothic rulers travelled south of Rome; no royal palace or estate is attested in the Mezzogiorno. In other words, the Gothic court was mobile within a narrowly circumscribed royal landscape formed by a number of privileged royal sites, prominent among them Verona and Pavia. Ravenna may have been the most significant individual palace within this constellation, but, as we have seen, royal stays at Verona and Pavia were both lengthy and possibly also more frequent than historians have tended to allow. Mapping the northern Italian royal landscape in this way, and tracing royal mobility through it, draws our attention to the discon- tinuous, localised nature of royal authority in Italy. 579 Bjornlie, ‘Governmental Administration’, 61–68. 157 III. ‘The Fear of Neighbouring Peoples’: Militarisation and the Spatial Architectonics of the Italian Kingdom A striking feature in the preceding survey is that none of these sites, with the exception of Ravenna (and Rome, if that can be considered an Amal palace), had previously been imperial residences of any consequence. Verona and Pavia are particularly revealing here. The lack of strong ‘imperial’ associa- tions with these cities is noteworthy given the rhetoric of imperial continuity stressed by the regime during Theoderic’s reign and that former imperial palaces were located relatively nearby in Milan and, if probably in a highly damaged state, also in Aquileia.580 While the focus of studies of Ostrogothic palaces has been the performance of imitatio imperii at Ravenna, shifting our attention to Verona and Pavia presents a very different picture. The palace network manifested royal authority in strategically important locations in Northern Italy and vouchsafed control over the Po. Although the political structures of the Italian kingdom were continuous with, and were layered onto, Late Roman infra- structures, we see in the palace network the reordering of political space in the formation of a new polity. A ‘Militarised’ Landscape? The Late-Roman Background As a result of the frequent conflicts of the fourth century onwards, Northern Italy played host to a significant concentration of military resources, both during the Roman period and afterwards. The physical environment was transformed to respond to military and defensive exigencies, and social relations (with their concomitant notions of power and legitimacy) were restructured around military elites and their patronage, a process generally referred to as ‘militarisation’.581 The militarisation of the 580 Marcellinus comes records that Aquileia was destroyed by the Huns in 452: Marcellinus comes, Chronicon, a. 452, in Chronica minora, 2:84: Aquileia civitas ab Attila Hunnorum rege excisa est. Jordanes writes dramatically that the city was laid to waste by Attila to such an extant that it was scarcely possible to find traces of it: Jordanes, De origine actibusque Getarum, in Iordanis Romana et Getica, ed. Theodor Mommsen, MGH. AA 5, no. 1 (Berlin: Weidmann, 1882), 114: nex mora et invadunt civitatem, spoliant, dividunt vastantque crudeliter, ita ut vix eius vestigial ut appareat reliquerunt. 581 A concise definition is offered by: Edward James, ‘The Militarisation of Roman Society, 400–700’, in Military Aspects of Scandinavian Society in a European Perspective, AD 1–1300, ed. Anne Nørgård Jorgensen and Birthe L. Clausen (Copenhagen: 158 Italian landscape set down the resources from which the Ostrogothic kingdom would be structured. Germinating and taking root in this fertile ground, post-Roman rulership was not a like-for-like re- placement of earlier imperial practices, an unbroken continuity as Cassiodorus would imply, but an indication of new social formations emerging from the ‘interstices’ of the imperial system.582 The Notitia Dignitatum provides vivid, if not completely unproblematic, testimony of the mili- tarisation of the landscape. Produced as we have it sometime in the fifth century, the Notitia strongly suggests that by that time Northern Italy (and the Po Plain especially) played host to the lion’s share of Italy’s military infrastructure, and a proportionally large amount in terms of the Western Empire as a whole. The Notitia collects an unrivalled assortment of details and insights into the Empire’s struc- ture, setting out in turn the civil and military hierarchies of the Eastern then Western Empires accord- ing to their jurisdictions and relative status, and has consequently been accorded a prime position in examinations of the Late Roman administration.583 Despite this wealth of detail, however, the Notitia’s presentation of a well-ordered administration cannot be taken at face value. Owing to successive re- visions to the ‘original’ document until at least 419 if not later, the text is marred by inconsistencies and duplications.584 These problems mean that, in Kulikowski’s eyes, the Notitia is ‘almost worthless’ as a guide to Late Roman administration.585 While he is correct to say that the Notitia is misleading if what we desire is an objective accounting of the Empire’s personnel in any given moment, it does tell us something of how the spatial ordering of the Western Empire was conceptualised and represented National Museum, 1997), 19. This phenomenon has also been termed ‘warlordism’: Jeroen Wijnendaele, ‘Generalissimos and Warlords in the Late Roman West’, in War, Warlords and Interstate Relations in the Ancient Mediterranean, ed. Toni Ñaco del Hoyo and Fernando Lopez-Sanchez, Impact of Empire 28 (Leiden: Brill, 2017), 427–51. 582 For ‘interstices’: Mann, Sources, 1:12 583 Fundamental on the Notitia are: Peter Brennan, ‘The Notitia Dignitatum’, in Les littératures techniques dans l’Antiquité romaine : statut, public et destination, tradition, ed. Claude Nicolet (Geneva: Fondation Hardt, 1996), 147–78; Michael Kulikowski, ‘The Notitia Dignitatum as a Historical Source’, Historia 49, no. 3 (2000): 358–77. 584 Kulikowski, ‘The Notitia Dignitatum’, 360. Brennan has plausibly proposed that the extant version of the text was pro- duced for the court of Galla Placidia and Valentinian III at Ravenna: Brennan, ‘The Notitia Digntatum’, 168 585 Kulikowski, ‘The Notitia Dignitatum’: 375–76. 159 at the political centre.586 Its actual ‘administrative’ utility, in other words, was probably always marginal at best.587 As a representation, however, it contains valuable information about the organising princi- ples of Italy’s military resources at a ‘macro’ level. The evidence of the Notitia strongly suggests that northern Italy had come to be conceptualised a frontier region and was consequently highly militarised. According to the Notitia, fifth-century Italy hosted a large number of units in the field army. As many as thirty-seven separate units of the field army of the legiones palatinae (‘palatine legions’) and the auxilia palatina (‘palatine auxiliaries’) and perhaps six of the Western Empire’s ten ‘palatine vexilations’ (uexillationes palatinae) of cavalry were located in Italy by the Notitia.588 The Notitia suggests that most of the ‘barbarian’ units of the field army were billeted in northern towns, including Pavia and Turin.589 Inscriptions from Aquileia and Concordia also attest to the presence of soldiers belonging to ‘barbarian’ units in those cities in the fifth century.590 Further to this, the Notitia records three units, one part of the regular field army and two counted as pseudocomitatenses (former limitanei, or border troops, who had been drafted to the field army), were stationed to guard the passes of the Julian Alps on both the Italian and Illyrican sides.591 It is worth noting here too that Ravenna was also a military centre: in the Late Empire it had served as the head- quarters of the Praefectus militum iuniorum Italicorum and the Praefectus Classis Ravennatis.592 Its attractions 586 It can thus be thought of as a ‘representation of space’ in Lefebvre’s terminology: Lefebvre, Production, 37. Useful here also is Smith’s definition of the ‘spatial imagination’ which, he contends, ‘emerges most forcefully in the analytic domain of representations, from maps and pictorial landscapes to spatial theory and philosophy’: Smith, Political Landscape, 74. 587 Kulikowski, ‘Notitia Dignitatum’, 359; Brennan, ‘The Notitia Dignitatum’, 152–58. 588 Not. Dig. Occ., VII.2–39, in Notita Dignitatum accedunt Notitia Urbis Constantinopolitanae et Laterculi Prouinciarum, ed. Otto Seeck (Berlin: Weidmann, 1876), 133–34; ibid., VI.42–52, 130. On the organisation of the Late Roman army, see: A. H. M. Jones, The Later Roman Empire: A Social, Economic, and Administrative Survey, vol. 2 (Oxford: Blackwell, 1964), 607–86 (largely based on the evidence of the Notitia). Some duplications and errors mar these sections of the text. 589 Twelve of the fifteen (with one entry missing) Praefecti Sarmatarum gentilium were headquartered in Northern cities: Not. Dig. Occ. XLII, 49–63. It is likely the missing entry, given that it occurs between Pavia and Cremona and after the only Southern Italian entries, was also a northern city (possibly Verona?), though this is admittedly speculative. These ‘Sarmatian Prefects’ are concentrated in the Po Plain zone, with others stationed in Piedmont and the Veneto. Belonging to a unit bearing a ‘barbarian’ moniker did not necessarily suggest anything about the ethnic origin of the soldiers: Halsall, Barbarian Migrations, 106–10. 590 Neil Christie, From Constantine to Charlemagne: An Archaeology of Italy, AD 300–800 (Aldershot: Ashgate, 2006), 307. 591 Not. Dig. Occ., VII.34–35, 134; VII.60, 135. 592 Not. Dig. Occ., XLII.6–7, 215. 160 to the Amal court were undoubtedly multiple. Naturally, it is not possible to invoke any reliable sta- tistics for this period, but the Notitia certainly suggests that the army was a visible presence in Italian society at this time and that the greater part of the Italian army was quartered in northern cities. In order to supply this concentration of manpower was a concomitant concentration of pro- ductive infrastructure. Italy hosted six of the Western Empire’s twenty fabricae (arms workshops) which were counted by this stage under the authority of the magister officiorum.593 The Italian fabricae were all located in northern Italy (and, with the exception of Lucca, all on the Po Plain) and produced a variety of armaments: arrows at Concordia Sagittaria, shields and weapons and Verona, cuirasses at Mantua, shields at Cremona, bows at Pavia, and swords at Lucca (Fig. 106).594 It is notable that, according to the Notitia at least, all the other ‘state-owned’ arms factories were located in frontier provinces and other areas of military importance in both the East and the West, and this should alert us to the relative status of Northern Italy certainly by the 420s or 430s.595 The fabricae were, with the exception of Lucca, integrated into the limes road network and thereby into the Empire’s principal overland transport ar- tery, with the Via Postumia, which crosses Northern Italy between Genoa and Aquileia, and Via Claudia Augusta, which runs from the Po to the Danube in Raetia, converging at Verona.596 593 Not. Dig. Occ. IX.16–39, 145–46. By comparison, Illyricum possessed five (producing general armaments), while the Gallic provinces together are presented as having nine (producing both general armaments and some specialised products). For commentary: Jones, Later Roman Empire, 2:834–35; Simon James, ‘The Fabricae: State Arms Factories of the Later Roman Empire’, in Military Equipment and the Identity of Roman Soldiers. Proceedings of the Fourth Roman Military Equipment Conference, ed. J. C. Coulston, BAR International Series 394 (Oxford: British Archaeological Reports, 1988), 257–331. 594 Not. Dig. Occ. IX.23–29, 145. Not all of these locations are externally verified, but there is no strong reason to discount the Notitia here. Ammianus references an unnamed ‘tribune’ (presumably the praepositus) of the fabrica at Cremona, who was drawn into the politicking around the usurpation of the Frank Silvanus in 355: Amm. Marc., Res Gestae, XV.5.9–11, 1:48. Several inscriptions from Concordia relate to the fabricenses (factory workers): CIL V, 8742, 8745, 8757. One inscrip- tion, broken in two and catalogued separately in the CIL, is from the tombstone of a certain Flavius Romulianus, praepositus of the ‘arrow factory’ (FL ROMVLIANVS | PP ‧ FAB SAGIT ‧), thus also confirming the product listed in the Notitia: CIL V, 8697/8721 = ILCV 538A). On the other hand, a dedicatory inscription from the reign of Constantine in Ravenna refers to Sertorius Silanus, vir perfectissimus and praepositus of the fabrica there: CIL XI, 9. Ravenna is, however, not mentioned as possessing a fabrica in the Notitia. Several explanations are possible—perhaps the Ravenna fabrica had closed by the time the Notitia was compiled, or it was mistakenly left out—but it is impossible to say with certainty. For a summary: James, ‘Fabricae’, 261–62. 595 James, ‘Fabricae’, 262–65. 596 Ibid., 268. 161 Finally, the Notitia captures a snapshot of a changing approach to Italy’s defence. Whereas previously adversaries were repelled from Italy by a linear barrier in the Alpine passes (figured mate- rially and conceptually), if not already by the provincial forces of Noricum and Rhaetia, from the fifth century the Alpine frontier space was reimagined as a militarised ‘buffer zone’. This shift is most visible in changes in the distribution of defensive infrastructures in the Alps. In response to events such as the Alemannic incursions into the Po Plain, a new line of fortifications was constructed in the Julian Alps, the Claustra Alpium Iuliarum, and placed under the authority of the comes Italiae (‘Count of Italy’), one of seven comites rei militaris responsible for the defence of imperial provinces.597 Yet the Claustra were not in use for long. Excavations at the fort known as Ad Pirum (Hrušica) in southwestern Slove- nia, a large fort of the Claustra constructed in the second decade of the fourth century, have confirmed the fort’s abandonment already by the end of the fourth century (Fig. 107).598 In the end, Theodosius had the Claustra dismantled following his victory over Eugenius at the Frigidus (394), an act which was celebrated by Claudian.599 After the Claustra’s abandonment, it seems that the comes Italiae’s remit was expanded to the so-called Tractus Italiae circa Alpes (‘the region of Italy around the Alps’), a territo- rial responsibility depicted in his insignia (Fig. 108). By the beginning of the fifth century, the Alpine and lake zones were scattered with fortifica- tions, both freestanding and attached to urban areas. At the frontier city of Susa at the foot of the Cottian Alps, for example, a new wall circuit was constructed around 300 CE incorporating a separate 597 Comes Italiae: Not. Dig. Occ., V.127, 121. In other frontier provinces (e.g. Germania Prima or Raetia) frontier commanders have the inferior title Dux limitum: Not. Dig. Occ. V.126–143, 121. The current state of research on the Claustra, including recent Slovenian scholarship, is helpfully summarised in: Slavko Ciglenečki, ‘Claustra Alpium Iuliarum, tractus Italiae circa Alpes and the Defence of Italy in the Final Part of the Late Roman Period’, Arheološki vestnik 67 (2016): 409–24. 598 Thilo Ulbert, ‘Das spätrömische Kastell AD PIRVM – Hrušica’, in Ad pirum (Hrušica). Spätrömische Passbefestigung in den Julischen Alpen: der deutsche Beitrag zu den slowenisch-deutschen Grabungen, 1971–1973, ed. Thilo Ulbert, Münchner Beiträge zur Vor- und Frühgeschichte 31 (Munich: Beck, 1981), 42–50. 599 Claudian, Panegyricus dictus Probino et Olybrio consulibus, ll. 104–09, in Claudii Claudiani Carmina, 5; idem, Panegyricus dictus Honorio Augusto tertium consuli, ll. 89–92, in Claudii Claudiani Carmina, 56. Christie, From Constantine to Charlemagne, 324–25. For a discussion of the chronology of the fortifications of the Julian Alps: Slavko Ciglenečki, Höhenbefestigungen aus der Zeit vom 3. bis 6. Jh. im Ostalpenraum, Academia Scientiarum et Artium Slovenica: Historia et Sociologia 31 (Ljubljana: Slovenska akademika znanosti in umetnosti, 1987), 121–27 (for a catalogue of sites in the Julian Alps: ibid., 13–108). 162 castrum on the city’s eastern side, facing the Alpine road (Fig. 109). Such urban castra would become widespread between the fourth and sixth centuries – most notably at Verona, where the castrum hosted the royal palace, and Brescia, where the ducal palace was inserted into an annex in the late ancient fortifications (Fig. 110).600 Susa’s role in regulating passage through the Cottian Alps is implied by the Bordeaux Itinerary, composed sometime in the second quarter of the fourth century: the anonymous pilgrim’s journey over Montgenèvre (Matrona) from Briançon brings them to Gesdaone (Cesana Tori- nese) where the road turns north to Susa; it is there, rather than in the Alps as such, that Italy begins, according to the itinerary.601 Other defensive infrastructures, such as signal points or watchtowers and other fortified points, plausibly connected these fortified urban sites into a broader defensive network. Among these the most famous is Castelseprio (Sibrium), located in the Olona valley west of Milan (Fig. 111). While initially established, it seems, as a small outpost or sentry station, the fort quickly grew into a larger, fortified settlement that persisted into the seventh century.602 The implication is that the limes was conceptually and spatially transformed from a linear (yet discontinuous) boundary in the Alpine passes to a broader boundary region punctuated with defensive sites. Although the comes Italiae is only known from the Notitia, the political sway that such regional military commanders could exer- cise over later Roman governments was considerable. One need only consider the career of another 600 Christie, From Constantine to Charlemagne, 326–28; Brogiolo, Brescia altomedievale, 55–65, and see below. 601 Itinerarium Burdigalense, ed. Paul Geyer and Otto Cuntz, in Itineraria et alia geographica, ed. Paul Geyer et al., Corpus Chri- stianorum. Series Latina 175 (Turnhout: Brepols, 1965), 3: inde ascendis Matronam | mutatio Gesdaone milia X | mansio Ad Marte milia VIIII | ciuitas Segussione milia XVI | inde incipit Italia. The priority of this route into Italy from the west is confirmed by the Antonine Itinerary and the Tabula Peutingeriana (see, for example, Talbot’s tabulation of the routes of Italy in the Bordeaux and Antonine itineraries and the Peutinger Table: Richard Talbot, Rome’s World: The Peutinger Map Reconsidered [Cambridge: Cambridge University Press, 2010], 158). 602 For chronology and continuity, see above all: Maria Dabrowska et al., ‘Castelseprio: scavi diagnostici, 1962–63’, Sibrium 14 (1978–79): 1–140; Gian Pietro Brogiolo and Silvia Lusuardi Siena, ‘Nuove indagini archeologiche a Castelseprio’, in Atti del 6° congresso internazionale di studi sull’alto medioevo, vol. 2 (Spoleto: CISAM, 1980), 475–500; Gian Pietro Brogiolo and Sauro Gelichi, Nuove ricerche sui castelli altomedievali in Italia settentrionale (Florence: Alll’Insegna del Giglio, 1996), 119–58. The site’s first phase comprised three towers within a mural circuit, datable to the third or fourth centuries; the construction of the wall circuit with external towers probably belongs to the fifth or sixth centuries and plausibly the Ostrogothic period: Brogiolo and Lusuardi Siena, ‘Nuove indagini’, 498; Brogiolo and Gelichi, Nuove ricerche, 122. See also: Christie, From Con- stantine to Charlemagne, 330, 338–41. 163 frontier ‘count’, the comes Africae Bonifatius, to see that this was not a phenomenon limited to the magistri militum, the so-called ‘generalissimos’ of the Late Roman army.603 This reordering of the Alpine limes inevitably recalls defence analyst and sometime historian Edward Luttwak’s hypothesised transition from linear defence to a strategy of ‘defence in depth’.604 Luttwak’s work was instrumental in provoking renewed debate about the nature of the Roman frontier and its management, even if scholarly reception has been generally critical.605 Luttwak argued that Roman geopolitical and military strategy can be characterised as a progression through three distinct structural models: a ‘Julio-Claudian’ system by which the empire was defended through a system of clients rather than a fixed perimeter; a second system in which the former clients were annexed and a fixed, defended frontier was established; and finally, a strategy of ‘defence-in-depth’ put into place under Diocletian.606 The latter strategy involved a reorientation of attitudes towards the limes: rather than a linear frontier, in its place was constructed a deep line of strongholds and fortifications with mobile forces that could range between them to engage the enemy – a longstanding doctrine of mod- ern military strategy.607 We should question, however, how far such a ‘system’ could be managed from the political centre given, on the one hand, limitations on communications and logistics and the generally reactive nature of imperial government, on the other.608 Roman regimes certainly did engage in long-term mil- itary planning and were capable of running sophisticated defensive and supply infrastructures, but Luttwak overestimates the coherence of any ‘empire-wide’ system directed towards the achievement 603 Jeroen Wijnendaele, The Last of the Romans: Bonifatius, Warlord and comes Africae (London: Bloomsbury Academic, 2014). 604 Edward Luttwak, The Grand Strategy of the Roman Empire: From the First Century CE to the Third, rev. ed. (Baltimore: Johns Hopkins University Press, 2016); Christie, From Constantine to Charlemagne, 331. 605 For an overview of the debate, see: C. R. Whittaker, Rome and its Frontiers: The Dynamics of Empire (London: Routledge, 2004), 28–32. 606 Luttwak, Grand Strategy. 607 Ibid., 149–66. 608 Fergus Millar, ‘Emperors, Frontiers and Foreign Relations, 31 B.C. to A.D. 378’, Britannia 13 (1982): 1–24. 164 of defined long-term objectives.609 Despite the semblance of systematism bestowed upon it by the Notitia Dignitatum, northern Italy’s militarisation more likely originated in an accumulation of responses to various incursions rather than a single blueprint. The impression of consistency, as well as the assumed prime agency of central government, lies in the eye of the historian, not in the evidence itself. It would be reasonable to question whether we must attribute every set of defensive works to a pro- gramme directed from the centre, as alternative scenarios, such as regional responses to crisis under- taken with the centre’s encouragement, cannot be ruled out.610 All told, the militarised spatial ordering of the Alpine frontier zone was far less schematic than Luttwak allows. We are better viewing it as the result of an accumulation of different individual efforts at defining the Alps as a defensible frontier, not quite ad hoc in their motivation, but reactive to instability in the transalpine provinces and political machinations within Italy itself, rather than a conscious empire-wide plan. Verona, Pavia, and the Defence of Italy in the Sixth Century The militarised landscape was the medium of Ostrogothic authority. It was the substrate into which royal palaces were inserted and from which they drew meaning. These processes of formation inten- sified under Theoderic and his successors. Viewed in this context, evidence for the palaces at Verona and Pavia strongly suggest that strategic, military decision-making was a more significant factor in the location of political authority than, for example, ideologies of imperial continuity on its own. Their urban experiences, and the topographical shift of the political centre towards Italia annonaria that their selection as palaces communicated, alerts us to the close entanglement of royal authority with military elites, a new set of ties between regime and subject that distinguishes the post-Roman period from earlier periods. 609 Whittaker, Rome and its Frontiers, 32–37. 610 Christie, From Constantine to Charlemagne, 338. 165 Even if Verona and Pavia had not been significant as ‘imperial’ centres in previous centuries, we have seen already that the Amal court spent significant time in the north and the large material investment in the urban fabric described by the Anonymus Valesianus would indicate that they were significant places for the regime. Their elevated status as legitimate sites of political authority receives further confirmation from events following the fall of Ravenna and capture of Vitigis in 540. Proco- pius reports that the Gothic nobility remaining north of the Po resolved to elect Vitigis’ nephew Uraia as king, yet he declined in favour of Hildebad, the commander of the Veronese garrison (evidently a politically significant force), who was ultimately crowned in Pavia.611 What unites Pavia and Verona is a shared reputation for defensive solidity, replicating the urban topography witnessed at Susa and elsewhere. In Pavia’s case, the evidence is fragmentary but suggestive. Archaeology for the early medieval city’s form and appearance is sadly lacking; the location of the palace is unknown, but the eastern sector, where the Roman grid-plan is missing, seems likely as its original location (Fig. 112).612 Two topographical hints in Paul the Deacon’s Historia Langobardorum certainly point in this direction. On taking the city after a protracted siege in 572, Paul describes that Alboin entered through the ‘gate which is called Saint John’s from the eastern part of the city’ before proceeding to the palace, where he appeared before the people.613 The Porta San Giovanni is no longer extant; it was located in the vicinity of San Giovanni in Borgo, itself destroyed in 1818 (Fig. 113). Lying outside the walls in the area of the ancient necropolis, it seems likely that the original church was a sixth century foundation 611 Proc., Wars, VI.30.4–17, 2:289–91. See also the continuation to the chronicle of Marcellinus comes, which reports sub- stantially the same information: Continuatio Marcellini Comitis, a. 540, in Chronica minora 2:106. 612 Bullough concluded it was ‘doubtless an entirely new building’: Donald Bullough, ‘Urban Change in Early Medieval Italy: The Example of Pavia’, PBSR 34 (1966): 92; Ward-Perkins, on the other hand, thought it was ‘unlikely that Theoderic built wholly anew in Pavia or elsewhere’, given the presumed availability of the praetorium that could be appropriated: Bryan Ward-Perkins, From Classical Antiquity: Urban Public Building in Northern and Central Italy, AD 300–850, Oxford Historical Monographs (Oxford: Oxford University Press, 1984), 159. In the absence of archaeological evidence neither claim can be verified. The peripheral location of the south-eastern sector, abutting the walls, might mitigate against it being a pre- existing praetorium. 613 Paul. Diac., Hist. Lang., II.27, 87: In quam cum Alboin per portam quae dicitur Sancti Iohannis ab orientali urbis parte introiret. 166 by the bishop Maximus (d. 514), Ennodius’ predecessor on the Pavese episcopal throne.614 Later in his history, Paul remarks that Perctarit (661–62; 671–688) built a gate ‘adjoining’ the palace (contiguam palatio) that became known as the ‘Palatine Gate’, a firm indication of the palace’s peripheral location against the walls at that time.615 The location of the palace gate is preserved in the Via Porta Palacense, just north of the former location of San Giovanni in Borgo; Patroni also discovered the foundations of a tower or ‘robust projecting bastion’ (un robusto bastione sporgente, o grande torre) on the nearby Via Giovanni Antonio Scopoli in 1924 that he interpreted as part of the Porta Palacense, though this attribution is unclear.616 Despite the paucity of the evidence, the suggestion is that Pavia reproduces a topographical pattern familiar from Ravenna and, earlier, the late Roman imperial sedes of the palace being located on the periphery of the urban area in its own sector. Why Pavia may have been chosen for such investment is suggested in other sources. Ennodius points the way when he writes that Theoderic, fearing treachery from Odovacar’s surrendered general Tufa during the invasion of Italy, wisely chose to fortify himself in the smaller (and thus, we are given to understand, more easily defensible) Pavia instead of Milan.617 He expands on this aspect of the city’s 614 Two pieces of evidence suggest Maximus’ patronage of San Giovanni in Borgo. The first is a ‘model speech’ (dictio) written by Ennodius, in which he praises Maximus for his foundation of a church dedicated to John: Ennodius, Dictio IV, in Magni Felicis Ennodii Opera Omnia, ed. Wilhelm von Hartel, Corpus Scriptorum Ecclesiasticorum Latinorum 6 (Vienna: C. Geroldi filium, 1882), 436–37. The second is the testimony of the thirteenth century priest and mystic Opicinus de Canistris (formerly known as the Anonymus Ticinensis), who locates the church within the city’s second wall circuit (built in the tenth century and greatly expanding the intramural area). Opicinus, drawing on an earlier thirteenth-century chronicle, records that the Lombard king Rothari was the church’s patron, but also notes that Maximus was buried there: Anonymi Ticinensis Liber de laudibus civitatis ticinensis, ed. Rodolfo Maiocchi and Ferruccio Quintavalle, rev. ed., Rerum Italicarum Scriptores 11/1 (Città di Castello: Scipione Lapi, 1903), 10: Ecclesia sancti Iohannis in burgo, que olim cimiterium dicebatur, quam condidit Rotarich rex Longobardorum. In qua iacent corpora sanctorum confessorum et episcoporum Urseceni et Maximi papiensium antistitum. (‘The church of San Giovanni in Borgo, which was once called ‘the Cemetery’, which Rothari, king of the Langobards founded. In it lie the bodies of holy confessors and bishops of Pavia Ursicinus and Maximus’). The earlier chronicle (Cronica de Corporibus Sanctis Papie) is included as an appendix to Maiocchi and Quintavalle’s edition: ibid., 55–57 (San Giovanni in Borgo at 55). See also: Bullough, ‘Urban Change’, 122. 615 Paul. Diac., Hist. Lang., V.36, 156: His diebus rex Perctarit in civitate Ticinensi portam contiguam palatio, quae et Palatensis dicitur, opere mirifico construxit. 616 Giovanni Patroni, ‘Avanzi di edificio sovrapposto al pavimento romano scoperto sotto il Corso Vittorio Emanuele. Altri frammenti architettonici’, Notizie degli scavi di antichità, 1924, 268. The excavated structure has not been published beyond Patroni’s brief notice to my knowledge. Bullough hypothesised that Patroni had really found one of the bastions of the east wall: Bullough, ‘Urban Change’, 90, (with n. 25). 617 Ennodius, Vita Epifani, 111, 98. Cf. Anon. Val. (II), XI.51–52, 15; Ward-Perkins, From Classical Antiquity, 166. 167 image in the course of a meeting he stages between Epiphanius and Theoderic. In this sequence the king is impressed by the bishop and exclaims that ‘while that man [i.e. Epiphanius] lives, the city of Pavia is fortified with the strongest wall, which no attacker’s strength can overwhelm and which the shot of a Balearic sling is not even close to reaching over’.618 Ennodius here argues that the spiritual protection afforded by the bishop surpasses the already-formidable physical defences, a reference which would only make sense if Pavia was known for its sturdiness. Procopius similarly comments that the Goths trusted greatly in the city’s strong defences during Belisarius’ campaign half a century later.619 Pavia’s value surely lay in its command over a bridge over the Po and its proximity to the confluence of the Po and the Ticino. It was likely this that Procopius referred to when he described the Franks under Theudebert taking control of a bridge at Pavia in order to ambush the Gothic army camped nearby.620 More can be said of Verona. The Anonymus Valesianus records that Theoderic resided in Ve- rona after Eutharic’s consulship of 519 ‘on account of his fear of neighbouring peoples’, implicitly confirming its status as a military stronghold and the significance of control of the passes through the Julian Alps for the regime.621 The substantial construction projects undertaken in Verona—baths, a new aqueduct, porticoed streets, a new wall circuit and of course a palace—were not only to elevate the city’s status with traditional symbols of imperial euergetism, but also to reinforce the city’s defensive solidity. Later sources show that the palace was located in a separate castrum at the edge of the urban area, rather than in the city’s ancient core.622 The so-called Iconografia Rateriana (Fig. 114), an image originally contained in a tenth-century manuscript of the works of Rather of Verona (ca. 890–974) 618 Ennodius, Vita Epifani, 110, 98: fortissimo muro Ticinensis civitas incolumi isto vallatur, quos inpugnantum nulla vis possit obruere, quos nequaquam Balearis fundae transcendat excursus; Ward-Perkins, From Classical Antiquity, 166. 619 Proc., Wars, VI.12.32, 2:204. 620 Proc., Wars, VI.25.8–13, 2: 262–63. 621 Anon. Val. (II), XIV.81–82, 23: Post haec Theoderico Veronae consistente propter metum gentium facta est. 622 On Verona’s changing urban topography in the Early Middle Ages, see: Cristina La Rocca Hudson, ‘“Dark Ages” a Verona. Edilizia privata, aree aperte e strutture pubbliche in una città dell’Italia settentrionale’, Archeologia Medievale 13 (1986): 31–78. 168 and known from two eighteenth-century reproductions, is crucial evidence for the urban image of early medieval Verona.623 It has been hypothesised on iconographic grounds the Iconografia was adapted from a late ancient exemplar, though this cannot be proven.624 It is true that the axonometric view of the city’s topography conforms to the kind of ‘three-quarter perspective’ employed for the represen- tation of cities in Late Antiquity, especially in mosaic (Fig. 115), yet the illuminator having reworked a Theoderican image seems unlikely.625 The Iconografia shows a structure labelled as the palatium located across the Adige from the ancient core and linked to it by a pons marmoreus (the modern Ponte Pietra), a location which corresponds to the later-medieval castello.626 A short dedicatory poem encircles the image that picks out prominent features of the topography. It ends by praising the castrum: ‘from the top of the mountain, the fortress looks out onto the city, made with Daedalan skill and dark galler- ies’.627 The topographic features picked out by the Iconografia are also singled out by the Carolingian Versus de Verona for praise, including the castrum: ‘a great and lofty citadel [castro] provides a strong bastion against attack, there are stone bridges built over the river Adige, the heads of which link the town and the citadel’.628 Similarly, the basic topographical information contained in the Iconografia is confirmed by a description of the city in Liudprand of Cremona’s tenth-century Antapodosis: From the left side of the river, which lies to the north, the city is protected by a difficult steep hill in such a way that, if that part of the city which the aforesaid river laps on its right were seized by enemies, still Verona could be vigorously defended from there. On top of this hill a church was built with precious workmanship, consecrated in 623 Ettore Napione, ‘Breve storia di un’Iconografia perduta’, in La più antica veduta di Verona: L’Iconografia Rateriana. L’archetipo e l’immagine tramandata, ed. Antonella Arzone and Ettore Napione (Verona: Comune di Verona, 2012), 25–32. 624 Giuliana Cavalieri Manasse and Peter Hudson, ‘Nuovi dati sulle fortificazioni di Verona (III–XI secolo)’, in Le fortifica- zioni del Garda e i sistemi di difesa dell’Italia settentrionale tra tardo antico e alto medioevo, ed. Gian Pietro Brogiolo (Mantua: Società Archeologica Padana, 1999), 85; Silvia Lusuardi Siena, ‘L’origine dell’archetipo e il problema del Palatium: Una cronologia di VI secolo?’, in La più antica veduta di Verona, 59–70. 625 Such images reproduced a common ‘idea’ of a city based in the prominence of the mural circuit and the intersection of the principal streets, lined with porticoes: Dey, Afterlife, 119–20. 626 On the later history of castrum: La Rocca Hudson, ‘“Dark Ages” a Verona’, 40 (with nn. 37–41). 627 De summo montis castrum prospectat in urbem, Dedelea factum arte viisque tetris. For commentary: Marco Petolettii, ‘L’Iconografia Rateriana: Le didascale e i versi celebrativi’, in La piü antica veduta di Verona, 33–46. 628 Versus de Verona, ll. 19–21, Versus de Verona, Versum de Mediolano civitate, ed. and Italian trans. Giovanni Battista Pighi (Bologna: Nicola Zanicehlli, 1960), 152: castro magno et excelso et firma pugnacula| pontes lapideos fundatos super flumen Adesis| quorum capita pertingit in orbem in oppidum. Trans.: Godman, Poetry, 181. The Versus de Verona were composed between 796 and around 805 and was modelled on the earlier Versum de Mediolano civitate. 169 honor of the most blessed prince of the apostles Peter, and there resided Louis [III of Provence (emperor 900–905)] on account of both the pleasantness of the church and the fortification of the site.629 These later sources strongly indicate that the royal palace was located in the castrum, and it seems likely that this was the same one that had been founded by Theoderic. It was certainly the case that the hill was already fortified in the sixth century and could be defended separately to the rest of the urban area. It would not be unreasonable to hypothesise that the fortification of the castrum oc- curred at the same time as the construction of the palace, as Giuliana Cavalieri Manasse has pro- posed.630 Procopius recalls that Gothic defenders held out against the Byzantine advance in 541/542 thanks to the advantages provided by the hill, a ‘certain rock which rises to a great height facing the fortifications’, to which the Goths were able to retreat and observe their enemies before launching an attack into the city.631 Other texts also suggest that the palace was located in the castrum and continued to be used in the following centuries. Paul the Deacon, admittedly writing much later, relates that the Lombard king Alboin (c. 565–72) was murdered in the palace at Verona and buried ‘under the stairs of a certain staircase which adjoined the palace’, another topographical feature prominently repre- sented in the Iconografia.632 The location of the palatium outside of the urban core—elevated and within a separate fortified enclosure—represents a significant reorientation: away from the social space 629 Liudprand of Cremona, Antapodosis, II.40, in Liudprandi Cremonensis Opera Omnia, ed. Paolo Chiesa, Corpus Christiano- rum Continuatio Medaevalis 156 (Turnhout: Brepols, 1998), 51: A laeva autem parte fluminis, quae est aquilonem versus posita, civitas est difficili arduoque colle munita, adeo ut, si ea pars civitatis quam memoratus fluvius dexteram alluit ab hostibus capiatur ea tamen viriliter possit defendi. In huius vero collis summitate preciosi operis est ecclesia fabricata, in honore beatissimi Petri apostolorum principis consecrata, ubi et propter ecclesiae amoeniatatem locique munitionem Hulodoicus manebat. Translation: Liudprand of Cremona, ‘Retri- bution’, in The Complete Works of Liudprand of Cremona, trans. Paolo Squatriti, Medieval Texts in Translation (Washington, D.C.: Catholic University of America Press, 2007), 92–93. 630 Evidence for Verona’s wall circuit is ambiguous. It has been traditionally attributed to Gallienus on the basis of an inscription on the Porta Borsari in the south west of the walled area, which sets construction in the year 265: CIL V.3329. Cavalieri Manasse convincingly proposes a broader chronological range of strata stretching from the third to the sixth centuries embedded in the extant tracts of wall circuit, however: Giuliana Cavalieri Manasse, ‘Le mura di Verona’, in Mura delle città romane in Lombardia: atti del convegno, Como 23-24 marzo 1990 (Como: Associazione Archeologica Comense, 1993), 179–215 (for the colle San Pietro: 204); eadem, ‘Le mura teodoriciane di Verona’, in Teoderico il Grande e i Goti d’Italia, 2:633– 43; Cavalieri Manasse and Hudson, ‘Nuovi dati’, 71–91. Traces of the curtain wall encompassing the colle San Pietro were still visible in the nineteenth century but are now lost: Cavalieri Manasse, ‘Le mura teodoricane’, 638–39. 631 Proc., Wars, VII.3.14, 2:311: πέτρα δέ τις ἐς ἄγαν ὑψηλὴ πρὸ τοῦ περιβόλου ἀνέχει. 632 Paul. Diac., Hist. Lang., II.28, 89: Cuius corpus cum maximo Langobardorum fletu et lamentis sub cuiusdam scalae ascensu, quae palatio erat contigua, sepultum est. 170 characteristic of ancient ‘public’ urbanism, the forum, towards an urban topography polarised between architectures of sacred and secular authority – palace, castrum, and cathedral.633 The extent to which the wider military infrastructure that had characterised the Late Roman period persisted is hard to assess. On balance it is doubtful that arms production continued uninter- rupted through the reigns of Odovacar and Theoderic, not least due to radical changes in the forces of supply and demand acting upon production networks, but the Variae do suggest that centralised manufacturing continued in some truncated form.634 Two formulae in the seventh book of the Variae refer to the armifactores (‘arms workers’). The first addresses a (hypothetical) individual being placed in charge of the ‘soldiers and smiths of arms’, presumably of a single factory, in what should probably be taken to be equivalent to the role of praepositus.635 Interestingly, Cassiodorus is at pains to stress that the official’s work would be overseen even in the king’s absence—‘Do not let the safety of our absence deceive you. Whatever you do, we see’—and that the work of the smiths would be inspected.636 It is impossible (as with all Cassiodorus’ formulae) to know how far this reflects actual administrative reality, but it does suggest a degree of continuity of centralised arms manufacturing in Italy, plausibly at now- royal sites like Verona and Pavia.637 633 The functional transformation of the forum is widely witnessed after the fourth and fifth centuries: even where it maintained a function as an open market square, it ceased to play host to the buildings which housed secular authority in many cases. Christie From Constantine to Charlemagne, 215–17. Fora as marketplaces: Ward-Perkins, From Classical Antiquity, 182–84. In Italy, this same bipolarity can be witnessed most clearly at Brescia and Lucca (and, less dramatically, at Ravenna): Elisabetta Abela, ‘Lucca’, in Archeologia urbana in Toscana: la città altomedievale, ed. Sauro Gelichi, Documenti di archeologia 17 (Mantua: Società Archeologica Padana, 1999), 34; Brogiolo, Brescia altomedievale. Under Langobard rule, the Cidneo hill in Brescia was also fortified as an elevated castrum: ibid., 88–89. 634 Laws pertaining to the fabricae and fabricenses were issued in the name of the Western Emperors down to Anthemius (467–472), but these were issued at Constantinople and it can be questioned how much these tell us about their state in Italy at the time: James, ‘Fabricae’, 282. 635 Cass., Var., VII.18.1, 213: atque ideo ab indictione illa militibus te et fabris armorum […] praefecimus. 636 Ibid.: securitas te nostrae non inducat absentiae. quicquid feceris vos videmus. 637 On the formulae, see: Bjornlie, Politics and Tradition, 230–34. The phrase securitas nostrae absentiae is ambiguous and it does not definitively narrow down possible locations for this factory. While it would seem to suggest that it was not one fre- quented by the court, nostra absentia could conceivably also connote ‘periods of our absence’ (with a more occasional sense) and thus refer to a city like Pavia or Verona. 171 Geopolitics, the Anonymus Valesianus’ ‘fear of neighbouring peoples’, should be taken seriously as a factor in the localisation of political authority in the Italian landscape. The tractus Italiae circa Alpes seems to have been maintained and reinforced and processes of militarisation intensified during the sixth century.638 Material evidence for the maintenance and expansion of the tractus is perceptible in the region of the pre-Alpine lakes. While it is hard to definitively distinguish ‘Ostrogothic’-period remains from earlier Roman or later Byzantine or Lombard stratigraphy, the overall picture is clear that pre-existing defensive architectures continued to be inhabited and maintained without rupture. For example, Sirmione, at the head of a peninsula projecting into Lake Garda, had been radically converted from a luxury lakeside villa-resort (a character it has since reassumed) to a fortified military settlement in the mid- to late-fourth century, probably overseeing a small fleet (Fig. 116).639 In around the beginning of the sixth century, however, the town received a new, enhanced wall-circuit and the settlement continued to be occupied through the period of the Justinianic War.640 Most famously, at Monte Barro, located near Lecco in the region of Lake Como, a new fortified settlement was con- structed in the early-sixth century, surrounded by a mural circuit punctuated with towers (Fig. 117– 18).641 It was, in the end, short-lived: destruction layers indicate that it was razed sometime later in the sixth century (possibly in connection with the ongoing warfare in the region linked to the repression of Totila in the period after 540).642 The site’s hilltop location, fortified aspect, and strategic location overlooking Lake Como, the Po, and the road connecting the lakes and Verona—clearly a well- 638 Settia, ‘Le fortificazioni’, 109, 130; Brogiolo, ‘Un’enclave bizantina sul Lago di Garda?’, in Le fortificazioni del Garda, 13– 15. 639 Elisabetta Roffia, ‘Le fortificazioni di Sirmione: nuove ricerche’, in Le fortificazioni del Garda, 21–38; Christie, From Con- stantine to Charlemagne, 342–43. Como, likewise, was listed as hosting a fleet in the Notitia Dignitatum who, as at Ravenna, was also the head of the civil administration in that city: Not. Dig. Occ. XLII.9, 215: Praefectus classis Comensis cum curis eiusdem ciuitatis, Como. As we will see below, there is reason to believe that such riverine and lacustrine fleets were maintained under the Ostrogothic kings in some form. 640 Gian Pietro Brogiolo, ‘Costruire castelli nell’arco alpino tra V e VI secolo’, Archeologia Medievale 41 (special issue) (2014): 144–46; Christie, From Constantine to Charlemagne, 340. 641 Gian Pietro Brogiolo, ‘Gli scavi’, in Archeologia a Monte Barro, vol. 1, ed. Gian Pietro Brogiolo and Lanfredo Castelletti (Lecco: Stefanoni, 1991), 48–49, 55. For the defensive works: ibid., 50–55. 642 Ibid., 48–49. 172 travelled route during this period—raise the possibility that Monte Barro was a fortified enclave of the regime’s territorial administration.643 Positing a direct connection with the regime is even more tempting given the discovery of a bronze hanging crown in the north wing of the two-storey main building (Fig. 119).644 Alongside this, we see in Cassiodorus’ Variae the literary construction of a new frontier: the Alps as the barrier between Italian civilization and Gallic barbarism. Indeed, the boundaries of the Italian polity were produced less through the establishment of hard and fast boundaries at their edge than by overlaying an ideational landscape, a symbolic geography constructed by reappropriating the tropes of Roman imperialism in which a strict division is drawn between the polity’s interior and exterior.645 The acquisition of former imperial provinces beyond the Alps—the Gallic provinces chief among them—was significant in legitimising the Ostrogothic regime in the eyes of Italo-Romans, as Jonathan Arnold has recently argued, yet the rhetoric employed by Cassiodorus (and to a lesser extent by Ennodius) stresses the restoration of Roman law and customs to now-barbaric lands with remark- able uniformity, implying a definite separation between Italy and the provinces which now depended on it once more.646 In one letter, Cassiodorus directs the Praetorian Prefect, Faustus, to, provide supplies to the sixty soldiers continuously posted in the Augustana clusura with- out any delay […] For it is proper to think of the lot of the soldier who is accustomed to exert himself in furthest places for the general peace and attempts to keep out the barbarian incursions, as if from a kind of gate to the province. He will ever be in a state of preparedness, he who strives to hold back the barbarians, because only fear curbs those whom pledged faithfulness does not restrain.647 643 For the main building: Brogiolo, ‘Gli scavi’, 26–50. For the site’s function: A. Javier Martínez Jiménez, ‘Monte Barro: An Ostrogothic Fortified Site in the Alps’, Assemblage 11 (2011): 34–46. Christie goes further and ponders whether it was plausibly a rural royal palace site: Christie, From Constantine to Charlemagne, 456. 644 Paola Marina De Marchi, ‘Reperti metallici miscellanea’, in Archeologia a Monte Barro, 1:106–113; Christie, From Cosntantine to Charlemagne, 456; Martínez Jiménez, ‘Monte Barro’, 38–40. 645 Smith, Political Landscape, 182; Knapp and Ashmore, ‘Archaeological Landscapes’, 12. 646 Jonathan Arnold, ‘Ostrogothic Provinces: Administration and Ideology’, in Companion to Ostrogothic Italy, 73–97. 647 Cass., Var., II.5.1, 49–50: quapropter illustrem magnificentiam tuam praesenti auctoritate praecipimus sexaginta militibus in Augustanis clusuris iugiter constitutis annonas, sicut aliis quoque decretae sunt, sine aliqua dubitatione praestare […] Decet enim cogitare de militis transactione, qui pro generali quiete finalibus locis noscitur insudare et quasi a quadam porta provinciae gentiles introitus probatur excludere. in procinctu semper erit, qui barbaros prohibere contendit, quia solus metus cohibet, quos fides promissa non retinet. 173 Cassiodorus probably refers here to the Alpine passage at Bard in the Val d’Aosta. Tellingly, in his Eucharisticum de vita sua, Ennodius lingers on the impotence of Italy’s defences in the face of food shortages, ‘when famine ripped through the lofty citadels of the mountains and forts, and a more savage destitution besieged those located on the summits’.648 Even if sixty soldiers is not in itself an impressive number, Cassiodorus is leaving us in no doubt of the conceptual significance of maintain- ing the Alpine frontier in his description: Italy’s gate was to be held fast by the ‘bolt’ (clusura) of the fort at Aosta. In another letter, a formula for the Dux Raetiarum (‘Duke of the Rhaetian Provinces’), Cassiodorus sets out the stakes with a characteristically rhetorical flourish: it seems that much is entrusted to those appointed as frontiersmen, because there is not so much the dispensing of justice, as in peaceful regions, as much as besieging suspect peoples, where wars are suspected more than crimes, and not only does the voice of the herald sound but the roar of trumpets frequently rings out. For the Rhae- tian provinces are the rampart of Italy and the barrier of the province (claustra provin- ciae). We adjudge that they have been called this not without merit, since they are arranged just like obstacles of nets against the most bestial wild peoples.649 Procopius too recognised northern Italy’s role as the bulwark against what, in Roman ethnographic imagination, lurked beyond the limits of civilisation. In a speech given by a Roman, Paulus, to the Roman army sent to relieve the siege of Milan in 539/539, he characterises the city as not only greater than all others, but also as a significant place in the militarised landscape: ‘apart from these qualities, it is an outpost against the Germans and the other barbarians and, so to speak, a projection that 648 Ennodius, Eucharisticum de vita sua, in Opera Omnia, ed. Hartel, 398–99: cum excelsa montium castrorumque arces penuria per- rumperet et in culminibus locatos armis saeuior egestas obsideret. 649 Cass., Var., VII.4.1–2, 203–04: Quamvis spectabilitatis honor unus esse videatur nec in his aliquid aliud nisi tempus soleat anteferri, tamen rerum qualitate perpensa multum his creditum videtur quibus confinales populi deputantur, quia non est tale pacatis regionibus ius dicere, quale suspectis gentibus assidere, ubi non tantum vitia quantum bella suspecta sunt nec solum vox praeconis insonat, sed tubarum crepitus frequenter insultat. Raetiae namque munimina sunt Italiae et claustra provinciae: quae non immerito sic appellata esse iudicamus, quando contra feras et agrestissimas gentes velut quaedam plagarum obstacula disponuntur. Obstacula plagarum is an apt wordplay here: plaga can mean both a ‘region, zone, tract of land’, but also a ‘hunting-net’, or ‘snare’, thus punning on the name Raetia, similar to the noun rete, ‘net’. 174 protects the whole Roman empire.’650 Without Milan and the other fortified cities of Liguria and Ve- netia, Procopius tells us, northern Italy was untenable. *** Under Theoderic and his successors we see defensive exigency assume primary significance in the location of palaces in Italy. The ‘fear of neighbouring peoples’ encouraged the location of palaces in defensively-significant locations in connection with the Po and at the base of the Alpine passes. As Bryan Ward-Perkins remarked, ‘the choice of the sites of Theoderic’s three main residences illustrates a darker side to Gothic rule: the need in troubled times for militarily powerful bases’.651 While Verona and Pavia have been conspicuous in their absence from most treatments of palaces in Ostrogothic Italy, it has been argued here that they were consequential poles of authority in northern Italy in the sixth century and beyond. These anxieties about Italy’s defence, however, expressed broader developments in the evolu- tion of the Italian political landscape. The preference shown towards Italia annonaria, and especially the lands north of the Po, was not unprecedented, but a result of processes of militarisation in north- ern Italy that had begun already in the fourth century. The network of Ostrogothic palatia was thus inextricably connected with defensive infrastructures that were maintained, restored, and iterated upon in the sixth century. These palaces, far from an unambigious expression of imitatio imperii, are far more equivocal. They point us towards a new form of rulership that developed out of the ‘warlord’ politics of the fifth century. In the process, the interior spatial organisation of the kingdom was reimagined. Rome was once more peripheral to localisation of authority in Italy, now concentrated on the frontier. 650 Proc., Wars, VI.21, 2:241: Μεδιόλανος γὰρ ἥδε, πόλεων τῶν ἐν Ιταλίᾳ πασῶν μάλιστα μεγέθει τε καὶ πολυανθρωπίᾳ καὶ τῇ ἄλλῃ εὐδαιμονίᾳ παρὰ πολὺ προὔχουσα, χωρὶς δὲ τούτβν προς τε Γερμανοὺς καὶ τοὺς ἄλλους βαρβάρους ἐπιτείχισμά τε οὖσα καὶ πάσης, ὡς εἰπεῖν, προβεβλημένη τῆς Ῥωμαίων ἀρχῆς. 651 Ward-Perkins, From Classical Antiquity, 160. 175 We will witness the magnetic, generative effect of Roman frontier architectures again in the next chap- ter. IV. Feeding the People, Forming the Polity: Food Supply and Fiscal Administration in Os- trogothic Italy Let us return momentarily to the ships of Sant’Apollinare Nuovo and Cassiodorus’ letters to Abun- dantius, with which we began. As we saw earlier, in neither case was the deployment of naval imagery an idle boast of military strength, but an assertion of the safety of the ‘public grain’. The placement of the Classe mosaic adjacent to the main portal of the palace church and opposite the palatium only underscores the significance of this image in the rhetoric of the Amal regime. In fact, it seems that the provision of the annona and, more generally, questions of taxation and the burdens facing landowners, became volatile issues in the context of sixth century Italy. Short-term food shortage was not uncom- mon in Late Antiquity, but in the Italian kingdom it seems to have become a critique levied at the regime. In part in response to the lurking threat of crisis, the patterns of movement of people and goods through the landscape—the ‘spatial experience’ of authority—were reoriented to centre on the Po, now regulated by the palatial network.652 Whether the king was present at any given time or not, the palaces of Ravenna and Pavia, alongside other fiscal sites on the Po plain, extended the regime’s hand over the fiscal apparatus of the Italian kingdom and its principal commercial artery. Subsistence crisis as political discourse in Ostrogothic Italy The palace was inextricably tied to the kingdom’s fiscal structures and thereby to agrarian production. We have seen already that the mobility of the court between Ravenna and Pavia had the potential to tip local producers in Liguria into crisis, as Cassiodorus’ letter to Wiligis recounts.653 This relationship has, however, gone largely unexplored in scholarship of this period. Social and political upheaval since 652 See: Smith, Political Landscape, 169–80. For spatial experience: ibid., 73. 653 See above, 157. 176 the early fifth century, including the capture of North Africa to the Vandals after 429 and the severance of Italy from the transalpine provinces in 461, had significantly realigned economic and productive relationships in Italy and in the Mediterranean more generally.654 Viewed in this context, the ships of Sant’Apollinare Nuovo were a pointed message in a prominent location about the regime’s care for the health of the Italian body politic. Instances of food shortage as a result of a combination of over-taxation and wartime pillage are ubiquitous in the sources for this period, to such an extent that we must infer that they formed the basis of a frequent critique of the regime.655 Most explicitly, Boethius stresses in his Consolation of Philosophy that he had prevented the institution of compulsory purchase, coemptio, of grain in Campania since it would have brought hardship to the people.656 Boethius’ objection is not an unusual one. The volatile discursive potential of the public food supply and its mismanagement is also in evidence in Procopius’ Secret History, especially in terms of corruption in the collection of taxes and the burdens that compulsory purchase, in Greek synone, placed on landowners and producers.657 Elsewhere, we see traces of this critique in the responses of figures associated with the regime. Ennodius, for example, stresses Theoderic’s care for the people in his Life of Epiphanius. He writes that Epiphanius intervened not only to secure a tax exemption for Pavia after its sack by Odovacar’s troops, but also sought to ameliorate the shortages caused by the quartering of Theoderic’s army in the city and the pillaging of 654 E.g. Jason Linn, ‘The Roman Grain Supply, 442–455’, Journal of Late Antiquity 5, no. 2 (2012): 298–321; Wickham, Framing, 711–13. 655 Still essential on the provision of the annona is: Lellia Cracco Ruggini, Economia e società nell’Italia annonaria: rapporti fra agricoltura e commercio dal IV al VI secolo d. C. (Milan: A. Giuffrè, 1961). 656 Boethius, De consolatione philosophiae, I.4.11–12, 13: Provincialium fortunas tum privatis rapinis tum publicis vectigalibus pessumdari non aliter quam qui patiebantur indolui. Cum acerbae famis tempore gravis atque inexplicabilis indicta coemptio profligatura inopia Campa- niam provinciam videretur, certamen adversum praefectum praetorii communis commodi ratione suscepi, rege cognoscente contendi et ne coemptio exigeretur evici. (‘When the fortunes of the provincials was ruined by both private theft and public taxation, I grieved no less than the victims. When in a time of harsh famine it seemed that a grave and indefensible order of coemptio would have brought the province of Campania poverty, I took up the fight against the Praetorian Prefect for the common good. I argued the case before the king and I won completely: the coemptio would not be instituted.’) 657 Procopius, Secret History (Ἀπόκρυφη Ἱστορία), XXI.1–19, in Opera Omnia, 3.1:128–31 (on unjust taxation); XXII.14–20, 136–38 (demands of compulsory purchase and corruption in its collection). 177 its environs by the enemy during the invasion of 489.658 Later, the king laments the desolation of Liguria after a Burgundian invasion and, apparently, the flight of its farmers: It aggrieves me that the fertile plain brings forth thorns and unwanted bushes, and that the mother of such a bounty of men—Liguria, to whom a countless progeny of farm- ers used to belong, but now is she childless and sterile—presents to our gaze only barren soil. She calls out to me as she mourns, left uncombed by the plough, wherever I consider her vine-rich face. Oh grief!659 Ennodius’ presentation of Theoderic’s care for Liguria indicates that the food supply and agrarian production had become a source of political tension during this time. Most interestingly, the Anonymus Valesianus—a text sensitive to the ideologies of the court, but which ultimately turns to polemic when narrating Theoderic’s later years—comments on the plenty that Theoderic’s thirty-year reign brought to Italy. Not only was he ‘generous with gifts and supplies (annonas)’, but he also restored Italy’s fi- nances despite discovering ‘only hay in the public treasury’, a metaphor again pointing to the close connection between perception of the kingdom’s prosperity and agrarian production.660 In fact, the Anonymus continues, in his time sixty modii of wheat or thirty amphorae of wine could be purchased for a single solidus – a rather improbable number.661 Efforts to counter the appearance of a rapacious and burdensome tax system are especially evident in Cassiodorus’ Variae. A group of letters written probably in late-537 against the backdrop 658 Ennodius, Vita Epifani, CVI–CVIII, 97 (legation to Odovacar); CXIII–CXV, 98 (patron to the city during Theoderic’s war with Odovacar). 659 Ennodius, Vita Epifani., CXXXVIII, 101: in tristitiam meam segetum ferax spinas atque iniussa plantaria campus adportat, et illa mater humanae messis Liguria, cui numerosa agricolarum solebat constare progenies, orbata atque sterilis ieiunum cespitem nostris monstrat obtutibus. interpellat me terra , quocumque respicio uberem vinetis faciem, cum aratris inpexa contristat. o dolor! 660 Anon. Val. (II), XII.60, 17: dona et annonas largitus, exhibens ludos circensium et amphitheatrum, quamquam aerarium publicum ex toto faeneum invenisset, suo labore recuperavit et opulentum fecit. 661 Anon. Val. (II), XII.73, 21: sexaginta modios tritici in solidum ipsius tempore emerunt et vinum triginta amphoras in solidum. To contextualise this figure, Durliat has suggested that the average price for wheat during this period was around thirty modii of wheat per solidus: Jean Durliat, De la ville antique à la ville byzantine: le problème des subsistances, Collection de l’École française de Rome 136 (Rome: École française de Rome, 1990), 502 (for Italy specifically: ibid., 501–02, n. 33); see also the discussion in: Dionysios Stathakopoulos, Famine and Pestilence in the Late Roman and Early Byzantine Empire: a Systematic Survey of Subsistence Crises and Epidemics, Birmingham Byzantine and Ottoman Monographs 9 (Farnham: Ashgate, 2004), 389–94. By contrast, prices could inflate quickly in times of famine according to the sources. Procopius, for example, in describing the siege of Rome in 545–46, notes that wheat reached the price of seven solidi per modius: Proc., Wars, VII.17.10, 2:371. Procopius does not seem to be uninformed here: his account correlates with figures given in other sources, such as the six solidi per modius given by the Anonymus Valesianus for the price of wheat during Theoderic’s siege of Ravenna in 489/490: Anon. Val. (II), XI.53, 15. See also the tabulated prices in: Stathakopoulos, Famine and Pestilence, 390–91. 178 of the Byzantine advance into southern Italy, highlight the administration reacting as if in real time to reorient supply networks and to justify increased and unusual demands with the enemy approaching and cities already falling into Byzantine hands.662 The first, to the ‘Provincials of Istria’, deals with the province’s taxation at some length. He comments that the province has enjoyed a good harvest that year and so, in addition to its usual taxes, Cassiodorus also mandates a compulsory purchase (coemp- tio)—exactly the measure that Boethius had earlier opposed—of additional supplies to the tune of an unspecified number of solidi.663 To ensure compliance, Cassiodorus says he has dispatched his deputy, Laurentius, to assess the state of the harvest personally.664 Amidst these more directly ‘administrative’ concerns, Cassiodorus writes at some length about Istria’s special connection with Ravenna: For Istria is the region located closest to us across the span of the Ionian sea, stuffed with olives, ornamented by its wheatfields, abundant in its vines, where every crop flows forth in hoped-for fecundity, as if from three teats famous for their fertility. Not undeservedly is it known as the Campania of Ravenna, the storehouse of the royal city, an exceedingly pleasant and delightful retreat. […] Clearly does it restore the watchful guards, it ornaments the imperium of Italy, it feeds the nobles with delicacies, the mid- dle-ranking men with its tributes of foodstuff, and whatever grows there is nearly to- tally consumed in the royal city.665 That Istria was to Ravenna as Campania was to Rome is a striking image, and one Cassiodorus is content to linger on for some time. Behind his purple prose lurks crisis: with the loss of the fertile fields of the south, shortage was near at hand in the north. His letter thus betrays the regime’s attempts to justify new and increased demands on a provincial population in response to anxieties about the fulfilment of the annona. The other letters of this group highlight the reorientation of the fiscal net- works to provide for the army. In one, Cassiodorus writes to his agent, a certain ‘industrious man’ 662 On this period, see: Cracco Ruggini, Economia e società, 321–49. 663 Cass., Var., XII.22.2, 378. 664 Ibid., XII.22.6, 379: Laurentium virum experientissimum et magnis nobis in re publica laboribus comprobatum cum praesenti auctoritate direximus. 665 Ibid., XII.22.3–5, 378: est enim próxima nobis regio supra sinum maris Ionii constituta, olivis referta, segetibus ornata, vite copiosa, ubi quasi tribus uberibus egregia ubertate largatis omnis fructus optabili fecunditate profluxuit quae non immerito dicitur Ravennae Campania, urbis regiae cella penaria, voluptuosa nimis et deliciosa digressio. […] reficit plane comitatenses excubias, Italiae ornat imperium, primates deliciis, mediocres victualium pascit expensis et quid illic nascitur, paene totum in urbe regia possidetur. 179 Paul, directing him to remit the wine and corn to be levied for the army from Concordia, Aquileia, and Cividale del Friuli (Foroiuliense), and instead to request wine from Istria.666 The frequency by which we encounter anxieties over subsistence crisis, agrarian production, and the demands of the public food supply mitigates against it being simply a rhetorical trope. We hear from sources with as diverse agendas as Ennodius, Procopius, and Gregory of Tours that North- ern Italy suffered greatly in the late-fifth and sixth centuries.667 Mortality, it seems, was great. According to Ennodius, ‘whatever survived swords, hunger killed’.668 Starvation was not the only factor in the vulnerability of rural communities at this time. Epidemic disease, a threat that so often accompanies subsistence crisis, seems to have been a frequent occurrence.669 Natural events, too, including earth- quakes and, in 536 or 537, even a volcanic eruption that both Cassiodorus and Procopius claim dimmed the light of the sun, may have caused long-term disruption to agrarian production like that recognised by Ennodius in Liguria.670 It is hard to say what strategies and tactics were available to rural farmers and other non-elite actors in these situations, but the sources are unanimous that the situation in Italy was delicately balanced throughout the sixth century. Against this background, provisioning the ‘royal city’—the palace, the court, and the army—took on additional political salience. 666 Ibid., XII.26.2, 382: Et ideo tanti viri allegatione permoti vinum et triticum, quod vos in apparatum exercitus ex Concordiense, Aqui- leiense et Foroiuliense civitatibus colligere feceramus, praesenti auctoritate remittimus, carnes tantum, sicut brevis vobis datus continet, exinde providentes. 667 Greg. Tur, Hist., III.32, 128: Theudobertus vero in Italia abiit et exinde multim adquisivit. Sed quia loca illa, ut fertur, morbida sunt, exercitus eius in diversis febribus corruens vexabatur; multi enim ex his in illis locis mortui sunt (‘But Theudebert went to Italy and acquired much from there. But since those places are, it is said, disease-ridden, his ailing army was harassed by different fevers; many of them died there.’) Procopius offers a chilling description of a famine afflicting Picenum, Emilia, Tuscany, and Venetia in 539. In Picenum alone he estimated as many as 50 000 farmers died, though one suspects this is an exag- gerated figure: Proc., Wars, VI.20.15–33, 2:238–40. 668 Ennodius, Eucharisticum, 398: quod superesset gladiis fames necaret. 669 Stathakopoulos, Famine and Pestilence, 156. 670 Proc., Wars, IV.14.5, 1:482–83; Cass. Var., XII.25, 381; Cam Grey, ‘Landowning and Labour in the Rural Economy’, in Companion to Ostrogothic Italy, 285–89. 180 Palaces and resource networks in Italia annonaria The supply of the annona was facilitated through public granaries and storage sites associated with fiscal sites throughout northern Italy, among which palaces figured prominently. We should expect that most major settlements and military installations possessed granaries in which the tax yield was stored, though archaeology is not much of an aid on this front.671 Even though we cannot locate urban and fiscal granaries with any certainty outside of Rome, written sources indicate that ‘public’ granaries, in the sense of being managed by the regime and not by private landowners, were widespread. From Cassiodorus we hear of granaries located at Pavia, Milan, Tortona, Tarvisio, and Trento (Fig. 120).672 In a further letter, datable to around 510 or 511, he writes to the Urban Prefect, Argolicus, to grant a senatorial petition to take control of two horrea, ‘long ruined by age, to which Antiquity attached the name such-and-such, if now they are shown not to be necessary for the public good and there is nothing collected there that belongs to the fisc’.673 By implication, while some of the imperial horrea may have been in a dilapidated state by the sixth century, the condition that they were only to be transferred if no longer necessary implies that other public granaries were still in use. Procopius attests further such horrea in Rome, as we might expect, and at Ravenna, the razing of which was a major factor in the 671 Not only have few structures that could be identified as horrea been excavated in urban areas, but the grounds for identification themselves are challenging. For a start, urban and public granaries identified in the sources were likely con- tinuous from the Roman period, rather than new constructions of the Ostrogothic regime (cf. Christie, From Constantine to Charlemagne, 357). In other sites, we might not necessarily be dealing with separate storage structures at all. An example is the main building at Monte Barro. No large structure that could be interpreted as a separate ‘granary’ was identified in the excavation. On the contrary, it seems that grain was stored on the ground floor of the building’s north wing, beneath the presumed representational suites: Lanfredo Castelletti and Elisabetta Castiglioni, ‘Resti vegetali’, in Archeologia a Monte Barro, 1:186, 202–03. Remains of a cocciopesto-like floor and of painted plaster indicate that the representational spaces were located on the upper floor: Brogiolo, ‘Gli scavi’, 33; Juliette Hanselmann, ‘Intonaci dipinti’, in Archeologia a Monte Barro, 1:141–43. Likewise, the east wing, undivided except for the central line of pilasters supporting the upper floor and postholes indicat- ing temporary wooden partitions, was probably used in part for storage, but the presence of several hearths indicate that this space was multifunctional: Brogiolo, ‘Gli scavi’, 32; Castelletti and Castiglioni, ‘Resti vegetali’, 188. The division be- tween a piano nobile-esque upper storey and multifunctional service spaces on the ground floor have parallels elsewhere in the Mediterranean at this time, including in the suspected palatium at Reccopolis: see above, Chapter 1. 672 Cass., Var., X.27, 314. 673 Ibid., III.30, 94: horrea longi temporis vetustate destructa, quibus illud atque illud vocabulum praefixit antiquitas, si nunc usui publico minime necesaria esse monstrantur nec aliqua ibidem est species quae ad fiscum pertinet congregata. 181 city’s capitulation to Belisarius.674 Writing in his capacity as Praetorian Prefect to his deputy, Ambro- sius, Cassiodorus summarises the importance of such structures: let your prudence conquer future hunger with past crops, since last year’s harvest was so great that its provisions will also suffice for the coming months. Let anything re- quired for food be put away. A private citizen will easily find what they need once the public stores have been filled.675 It is possible that these sites were fortified. In a letter addressed to the citizens of Tortona, a city with an attested public granary, the inhabitants are commanded to relocate their homes inside the walled castrum—which was simultaneously to be strengthened—so that the site could be better defended.676 While he notes that the citizens will enjoy the advantages of greater security by doing so, despite it being peacetime, the command is framed in terms of the public good and thus, one suspects, with the public granaries located there.677 Palaces were integrated into the system of fiscal sites, like that at Tortona, through their con- nection to the Po. The Po, with Ravenna at its head on the Adriatic coast, was the principal artery for the transport of goods and people through the kingdom. Previous examples from the Variae have already highlighted that riverine shipping seems to have been a normal means for conveying supplies between royal sites. We have already seen, for example, the saio Wiligis being tasked with loading ships with grain at Ravenna which were to be sent to Pavia; elsewhere Cassiodorus has noted the economic dependence of Ravenna on Istria across the Adriatic. Transport infrastructure along the Po seems either to have been continuous with the Late Roman period, if in reduced form, or was re-established early in the reign of Theoderic. In a letter written for Theoderic between 507 and 511, Cassiodorus writes to the ‘boatmen’ (dromonarii), a collegium of sailors in public service, instructing them to gather at 674 Proc., Wars, V.14.17, 2:78 (Belisarius storing grain in public granaries); VI.28.25, 2:280 (granaries in Ravenna razed). 675 Cass., Var., XII.25, 381: Atque ideo de veteribus frugibus prudentia tua futuram vincat inopia , quia tanta fuit anni praeteriti felix ubertas , ut et venturis mensibus provisa sufficiant. reponatur omne quod ad victum quaeritur. facile privatus necessaria reperit, cum se publicus apparatus expleverit. 676 Ibid., I.17.3, 23: Et ideo praesenti auctoritate decernimus, ut domos vobis in praedicto castello alacriter construatis. 677 Ibid., I.17.1, 23: Publicae utilitatis ratione commoniti, quae nos cura semper libenter oneravit, castrum iuxta vos positum praecipimus communiri, quia res proeliorum bene disponitur, quotiens in pace tractatur. 182 Hostilia (modern Ostiglia, between Mantua and Ferrara on the Po) to assist the public post.678 Later, as Praetorian Prefect, Cassiodorus wrote to the ‘Tribune of the Mariners’ (tribunus maritimorum), in- structing him to see to the transport of the supplies requisitioned from Istria. This long text is note- worthy not only because it attests to an official in charge of a dedicated transport fleet in Venetia, but also for the lengthy, rhetorical digression on the nature of the tribune’s ships. Cassiodorus writes that, when the sea lanes are closed by inclement weather, the tribune is fortunate since ‘another tranquil route is open to you with perpetual security’ through the ‘most pleasant of riverways’ through Vene- tia.679 Cassiodorus describes at length how the tribune’s ships do not fear running aground as they pass through the riverine landscape (suggesting that Cassiodorus has in mind a kind of dual-purpose, shal- low-keeled galley) and that they are drawn by cables: Your keels do not fear harsh gales: they touch the earth with the greatest joy and are ignorant of harm, however frequently they run onto it. From a distance they seem as if they are carried through the fields, since where they meet the channel cannot be seen. Drawn by cables they move, those which are accustomed to be held still with ropes, and with the usual situation changed men help their ships with their feet: they drag their carriers without effort, and instead of fear under sail they use the more fa- vourable step of the sailors.680 The palace at Ravenna was especially prominent within this system. As Shane Bjornlie has suggested, it served as an ‘entrepot for resources (taxes) travelling to the various points of the Po River valley that served as residences for the comitatus of Amal rulers.’681 While Ravenna now lies some ten kilometres inland, the postclassical city was far closer to the water and until as late as the eighth century 678 Ibid., II.31, 64. Cassiodorus comments that these men were performing militia, which by the fourth century had become the official term for state service in both the civil and military spheres: Cass., Var., II.31, 64: Publicis debent utilitatibus insudare qui nomen dedere militiae (‘Those who have enlisted in civil service ought to sweat for the public good”). The Roman annona system had long relied to the collegia for functioning, though it ultimately constituted a public monopoly. On the organisa- tion of the grain trade, see: Paul Erdkamp, The Grain Market in the Roman Empire: A Social, Political and Economic Study (Cambridge: Cambridge University Press, 2005), esp. 258–316. 679 Cass., Var., XII.24.2, 380: accedit etiam commodis vestris, quod vobis aliud iter aperitur perpetua securitate tranquillum. nam cum ventis saevientibus mare fuerit clausum, via vobis panditur per amoenissima fluviorum. 680 Ibid.: carinae vestrae flatus asperos non pavescunt: terram cum summa felicitate contingunt et perire nesciunt, quae frequenter inpingunt. Putantur eminus quasi per prata ferri, cum eorum contingit alveum non videri. tractae funibus ambulant, quae stare rudentibus consuerunt, et condicione mutata pedibus iuvant homines naves suas: vectrices sine labore trahunt, et pro pavore velorum utuntur passu prosperiore nautarum. 681 Bjornlie, ‘Governmental Administration’, 57. 183 the city continued to be linked to the Po riverine system (Fig. 121).682 Sidonius Apollinaris, for exam- ple, narrated his journey from Pavia all the way to Ravenna on his way to Rome – noting that a branch of the Po flowed through the city’s centre.683 In the sixth century, though the cumulative effects of deltaic coastal progradation meant that the city itself was no longer easily accessible by ship. Procopius, for example, claims that the sea formed unnavigable shoals for as much as thirty stades from the shore.684 Jordanes offers a memorable description of the city’s changing terraqueous topography. Whereas in the past, he tells us, the harbour could hold a fleet of two hundred and fifty ships, in his day he had heard that ‘where there had once been a port, it now displays a very spacious garden full of trees, but now apples hang from them, not sails.’685 Despite this changing topography, the city still retained its maritime orientation, and its continued capabilities as a port-city have been vividly demon- strated from archaeological evidence. Excavations beginning in 1974 at Podere Chiavichetta, located south of Ravenna on the former site of the channel leading from the older Augustan interior harbour to the Adriatic, have shown that the port stayed in use into the seventh century (Figs. 122–23).686 The 682 The paths of ancient waterways have wandered over the centuries. While the Po’s current course discharges into the Adriatic in a riverine delta, extending from north of Comacchio and terminating south of Chioggia, the southern extent of the late-ancient delta lay closer to Ravenna: See, for example: Marco Bondesan, ‘L’area deltizia padana: caratteri geo- grafici e geomorfologici’, in Il Parco del Delta del Po: studi ed immagini, ed. Carlo Bassi, vol. 2 (Ferrara: Spazio Libri, 1990), 9– 48, esp. 16–19. Ravenna had been linked to the Po’s main course by a canal commissioned by Augustus (the Fossa Augusta), probably a canalised stretch of the Padenna, a minor tributary: Cirelli, Ravenna, 19–20. While this had evidently suffered the effects of silting by the sixth century, Procopius states explicitly that rivercraft were able to travel inland from the city when the inlet becomes navigable at high tide: Proc., Wars, V.1.19–21, 2:7. Likewise, Jordanes notes that Ravenna was still connected to the Po via two of its tributaries: Jordanes, De origine actibusque Getarum, XXIX, 96–97. 683 Sidonius Apollinaris, Epistulae, I.V.5–6, 7. Compare Jordanes, who similarly notes that this channel ran through the centre of the city to the harbour and that it was a branch of the Po: Jordanes, De origine actibusque Getarum, XXIX, 97: a meridie item ipse Padus, quem Italiae soli fluviorum regem dicunt, cognomento Eridanus, ab Augusto imperatore latissimi fossa demissus, quo septima sui alvei parte per mediam influit civitatem, ad ostia sua amoenissimum portum praebens (‘To the south is the same river, the Po, surnamed Eridanus, which in Italy alone they call the king of rivers. It was sunk by the emperor Augustus in a very wide canal, in which a seventh part of its current flows through the middle of the city toward its mouth, where it offers a most pleasant harbour’) 684 Proc., Wars, V.1.17, 2:6. The mean rate of coastal progradation in the Po Delta between 1000 BCE and 1200 CE has been estimated to be as much as four metres a year: James Syvitski et al., ‘Distributary Channels and Their Impact on Sediment Dispersal’, Marine Geology 222–23 (2005): 83–84. 685 Jordanes, De origine actibusque Getarum, XXIX, 97: qui nunc, ut Favius ait, quod aliquanto portus fuerit, spatiosissimus ortus ostendit arboribus plenus, verum de quibus non pendeat vela, sed poma . 686 For Classe’s archaeology, see: Maria Grazia Maioli, ‘Classe, podere Chiavichetta, zona portuale’, in Ravenna e il porto di Classe: venti anni di richerche archeologiche tra Ravenna e Classe, ed. Giovanna Bermond Monatanari (Imola: Santerno, 1983), 65– 78; Maria Grazia Maioli and Maria Luisa Stoppioni, Classe e Ravenna fra terra e mare – città, necropoli, monumenti: un’avventura della archeologia, gli scavi nella zona archeologica di Classe (Ravenna: Sirri, 1987); Luigi Malnati et al., ‘Nuovi scavi archeologici a 184 channel’s quayside was lined with large magazines, fronted by sturdy covered porches or porticoes, which stayed in use until its later sixth- and seventh-century phases after which point several of the former warehouses seem to have been converted for domestic use.687 The excavations at Classe have covered but a small fraction of the harbour zone (around 1.2 hectares) but the results are nonetheless suggestive: the sheer scale of ceramic evidence indicates that Classe was a thriving port both under the Amals and the exarchs for imports from both north Africa and the Eastern Mediterranean.688 Classe was of sufficient capacity, in other words, to act as a ‘centre of redistribution’ in the Adriatic, integrated into local, regional, and transregional exchange networks.689 At Classe, therefore, we witness the chorogenetic power of palaces. That is to say, the localisation of authority in the palace created a landscape around them: they formed the key structuring elements in networks that traversed the Italian landscape for the extraction and redistribution of resources. These networks were not limited to Italy, either. As the excavations at Podere Chiavichetta have shown, Classe was as much a trading hub for Mediterranean trade as an exchange for infra-regional traffic.690 It is in this context that the ships of Sant’Apollinare Nuovo, and those commissioned from Abundantius, take on additional significance. The reliance of the Italian regime on riverine and coastal Classe: campagne 2004–2005’, in Felix Ravenna. La croce, la spada, la vela: l’alto Adriatico fra V e VI secolo, ed. Andrea Augenti and Carlo Bertelli (Milan: Skira, 2007), 33–38. 687 Andrea Augenti and Enrico Cirelli, ‘From Suburb to Port: The Rise (and Fall) of Classe as a Centre of Trade and Redistribution’, in Rome, Portus and the Mediterranean, ed. Simon Keay, Archaeological Monographs of the British School at Rome 21 (London: British School at Rome, 2012), 207–08. 688 The excavated structures had the capacity to store large quantities of goods. One warehouse, that seems to have burnt down in the final years of the fifth century, contained an assemblage of 180 intact North African amphorae for oil and 100 spatheia. Augenti and Cirelli, extrapolating from this assemblage for the total area of the structure, estimated that the total capacity of just these rooms could be as much as 1440 modii of grain or its equivalent; the total for the ground floor in total plus the upper storey could be as much as three times greater at full capacity, enough to serve the needs of as much as 18 000 individuals in their calculation. Augenti and Cirelli, ‘From Suburb to Port’, 216. 689 Ibid. 690 Stella Patitucci Uggeri has theorised that new harbour settlements on the Adriatic littoral and the mouth of Po delta, including Comacchio, may have originated in response to the reorientation of resource networks by the regime, if not its direct encouragement: Stella Patitucci Uggeri, ‘La politica navale di Teoderico: Riflessi topografici nel Ravennate’, in Teod- erico il grande e i Goti d’Italia, 2:785–86. Comacchio’s name has been interpreted as a degradation of conventus navium (‘gath- ering of the ships’): ibid., 773. Could it in fact have been in origin the mustering site of the fleet described by Cassiodorus? (Cass., Var., V.18, 154 and V.19, 155 simply note that sailors and shipmasters were to convene at Ravenna on the Ides of June). 185 shipping was hopelessly exposed during the opening stages of the invasion of Byzantine forces under Belisarius. Procopius describes in detail the Byzantine blockade of Italian ports throughout the inva- sion’s early stages causing sufficient shortages that cities like Naples, unable to resupplied from the sea, capitulated quickly; similarly, he notes an unsuccessful attempt to relieve the besieged Ravenna in 539 or 540 via the Po from Liguria.691 Cassiodorus too seems to have recognised this. While applauding Abundantius’ hard work, he writes that the kingdom’s enemies would take note: ‘envious, they watch what they used to fulfil their desires only at great cost flourish also for us’.’692 In the end it would seem efforts to provide for a fleet of dromones were well-founded, but possibly too little too late. *** It is hard to ignore the transformations that the upheavals of the fifth century caused in the form of the Italian polity and its administration. While significant continuities can be traced between the Ro- man and Ostrogothic polities in terms of bureaucratic institutions and the rhetoric used to justify the demands placed on local landowners, the regime’s efforts to sustain the public food supply highlight the degree to which the infrastructures of power were reordered in the decades of Ostrogothic rule. The regime still raised taxes and levied tolls on goods, yet overall the mechanisms of surplus extraction were smaller-scale and more directly under royal purview than previously, a sign of a simplified ad- ministration engaging in changed social relations of power. The movement of goods and people through the polity—in Smith’s terminology the ‘experience’ of the landscape—defined how it fit to- gether, the architectonics connecting the discontinuous spaces of royal authority. In this respect, Ra- venna’s attractiveness to the Amal regime was as much its economic potential as the impression of imperial continuity. 691 E.g. Proc., Wars, V.8.6, 2:39–40 (siege and coastal blockade of Naples in 536); VI.12.17, 2: 202 (Byzantine soldiers boast that they controlled the sea); VI.16.23, 2:223 (Byzantine forces marching close to the coast, supported by a fleet); VI.28.6, 2:276 (Ravenna cut off from supplies arriving from both the Po and the Adriatic). Note also however: VII.13.6, 2:350 (blockade of outbound traffic from Naples by Totila). 692 Cass., Var., V.17.3, 153: illud apud nos invidi vigere respiciunt, unde illi per magna pretia sua vota complebant. 186 V. Conclusion In the foregoing, we have at times left palaces behind and focused our attention instead on the con- nective tissues that bound them together. This function—of palaces operating together as a network— was not only essential for the formation of a new polity in the Roman imperial landscape but also, it has been argued, for the reuse of its resources to create new political and social structures. It has been shown that the Ostrogothic royal landscape was a further layer in the sedimentation of different con- ceptual and material strata that had been accumulating from the Roman period and earlier, but at the same time the landscape that supported the royal regime was not a passive continuation from earlier centuries. These transformations were threefold. Firstly, we saw that multiplication of the places of power in the Italian landscape shifted the political centre of gravity northwards, to the Po plain and the Alpine foothills. Contrary to usual scholarly presentations, the royal regime was not static in Ra- venna, its majestic showpiece, but could and did travel throughout the north, spending significant time at the palaces of Verona and Pavia. These palaces have often been deemed secondary to Ravenna, or even ignored entirely, yet there is every indication that they were the products of significant invest- ments of resources and labour. Secondly, this movement of the political centre was motivated by short-term defensive exigency but also, more significantly, the long-term militarisation of the Italian landscape. The palaces of Pavia and Verona were reputed for their defensive solidity and their prom- inent locations close to the Po (Pavia), the Alpine passes (Verona), and the major routeways that crossed northern Italy extended the regime’s surveillance over the infrastructures of the kingdom, a function they shared with a broader system of defensive architectures throughout the Alpine and Laghi prealpini regions. Simultaneously, we witnessed the creation of a new conceptual frontier, dividing ‘Ro- man’ Italy from the barbaric north, a reassertion of Roman ethnographic spatial categorisation. Finally, palaces became essential components in the collection, storage, and redistribution of the taxes for the 187 purposes of the annona, the public food supply. The annona seems to have been a volatile political issue in the sixth century that formed the basis of critique and opposition by provincial possessores, and the Ravennate court took care to emphasise the justice and wisdom of its demands. The palatial network vouchsafed royal control over the Po and extended royal surveillance over the principal arteries of exchange in the kingdom. In each of these transformations, the pre-existing Roman landscape was a resource from which new social and political structures were built. In pursuing such an analysis, there is a danger of making the Italian polity seem too ‘neat’, of presenting the forces that shaped it in an overly schematic form. To be sure, these processes were ‘messy’ – faltering, piecemeal, and unpredictable. But a common thread has nonetheless emerged, one that has been underestimated in most scholarly literature on the transition of Roman to post-Roman Italy: the role that palaces played in the reorientation of the landscape. Throughout this chapter, we have seen that palaces formed the basis of new networks through which social power was exercised. In a broader sense, we have also seen the generative possibilities of the Roman frontier in the for- mation of new political units. It has been argued here that the Ostrogothic regime arose out of the context of the ‘warlord’ politics of the fifth century, a phenomenon that itself left an indelible mark on the Italian landscape. While in Ostrogothic Italy we are only able to glimpse these processes at work over a short, if intensely documented, period, by turning our attention north of the Alps we will once again see how palaces reordered the Roman political landscape. 188 CHAPTER 4 Mobility, Itinerancy, and Political Integration in the Regnum Francorum I. Introduction On the Tabula Peutingeriana, the Roman world becomes no more than a series of routes, waypoints, and distances (Figs. 124–25). Produced as we have it in Germany in the early-thirteenth century, though likely dependent on a late-Roman original of the fourth or fifth, the map’s attenuated form displays less interest in details of natural topography than with the civitates, represented sparingly in two dimensions, and the road network that connected them.693 Unfurling the map, one is confronted by the Roman world not as a bounded ‘territory’ as such, but as the sum of its infrastructures spreading outwards from the two nuclei of political authority, Rome and Constantinople (Fig. 126).694 That ‘all roads lead to Rome’ is undoubtedly a cliché, but one nonetheless powerfully asserted by the map. What the Tabula ultimately represents is the spatial ideology of Roman imperialism: the exten- sion of the political centre’s authority over its periphery through infrastructural networks. This fact was captured neatly by Michel Rouche’s description of the Roman cursus publicus as a ‘colossal hand placed over Gaul’ (une gigantesque main posée sur la Gaule), pulling the transalpine provinces into an Italian orbit.695 Across the green-blue channel representing the Rhine along the map’s upper edge, is the great 693 For the circumstances of production and an attempted recovery of the original map: Talbot, Rome’s World, 73–85, 123– 61. The extant copy is a mere 33cm tall but 672cm long, excluding the missing leftmost sections. Emily Albu, however, has stressed that we should view the map principally in its medieval context: Emily Albu, The Medieval Peutinger Map: Imperial Roman Revival in a German Empire (Cambridge: Cambridge University Press, 2014). The emphasis of a Carolingian interme- diary (or intermediaries: ibid., 48) is useful, but hypothetical. Overall, Albu’s efforts to reject a late ancient origin to the map’s visual form is not convincing, though her call to view the map in the context of the work it did in the thirteenth century is a useful one. 694 A significant component of the manuscript’s source material were itineraries (such as the Antonine Itinerary), shown clearly by linguistic features such as the retention of ablative and accusative declensions of proper nouns: Talbot, Rome’s World, 136–41. 695 Michel Rouche, ‘L’héritage de la voirie antique dans la Gaule du Haut Moyen Âge’, in L’homme et la route en Europe occidentale au Moyen Âge et aux temps modernes, ed. Charles Higounet, Flaran 2 (Toulouse: Presses universitaires du Midi, 1982), 15. 189 unknown, barbaricum, at which the represented routes abruptly halt. The sliver of visible land between the empire and the Ocean is labelled according to the groups reputedly inhabiting each area: Alaman- nia, Suevia, and in prominent red capitals, ‘Francia’ (Fig. 127). Already by the end of the fifth century, however, ‘Francia’, or more properly the regnum Francorum (‘the kingdom of the Franks’), referred not to any politico-spatial categorisation of the lands beyond the frontier in the Roman ethnographic im- agination, but a significant polity established within the Empire’s former limits. The Tabula Peutingeriana, though not concerned with palaces as such, directs our attention to the central theme of this chapter: the re-evaluation of political space and the role of Roman frontier infrastructures in that process. The creation of new nodes of political authority, palaces, was an act of negotiation and transformation, precipitating the reorientation of social and economic networks and, as a result, new landscapes of political action. In the previous chapter, we saw palaces’ dual role as both representational apparatus and nodes in the material networks through which royal authority was actioned. In the Frankish kingdom, we similarly see the part palaces played in the formation—and continual reformation—of new conceptual geographies out of the Roman military-political system. The Frankish ‘palatial system’ and the problem of itinerancy A dizzying number of royal residences are attested in Frankish written sources between the sixth and ninth centuries. Narrative and hagiographical sources are replete with kings lodging at sites variously termed palatium, aula regis, villa, or sedes regni located both in former Roman cities and on rural fiscal properties. The kingdom’s periodic division among members of the Merovingian line into sub-king- doms (generally known in scholarship by the German term, Teilreiche) meant that, by the ninth century, it was strewn with a dense patchwork of royal residences.696 Even during periods of unification under 696 The boundaries of the subkingdoms had solidified by the seventh century into three distinct units: Neustria (based around Paris and the Aisne-Oise valleys), Austrasia (centred on the Middle Rhine), and Burgundy (in the Rhone and Saône valleys). Fundamental are: Eugen Ewig, ‘Die fränkischen Teilungen und Teilreiche (511–613)’, in Eugen Ewig, Spätantikes und fränkisches Gallien. Gesammelte Schriften (1952–1973), ed. Hartmut Atsma, vol. 1, Beihefte zu Francia 3 (Munich: Artemis, 190 a single ruler, the norm was that separate courts were maintained with their own hierarchies, aristoc- racies, and palaces in each Teilreich.697 The accumulation of successive strata in these landscapes, the ‘chronological sedimentation’ of a Frankish palatial ‘system’, profoundly shaped the imagined geogra- phies of the Frankish regnum into the Carolingian period and beyond.698 North of the Alps the limits of the evidence are keenly felt, and we are dependent on (often fleeting) mentions of palace sites in texts, particularly in the subscription and dating clauses of charters. Most of these have not been identified archaeologically and, in some cases, the Latin toponyms have not been definitively associated with modern sites at all. A slew of archaeological projects undertaken during post-war reconstruction vastly increased our knowledge of numerous French and German cit- ies, but very few brought structures to light that could possibly belong to a royal palace.699 The evi- dence’s textual bias, while unfortunate, has also significantly shaped the historiography of palaces in northern Europe. The preponderance of diplomatic texts, and the relative absence of archaeology, pushed the study of palaces into the sphere of legal and economic relations, a methodological priority that stands in stark contrast to the study of late-Roman palaces.700 As we saw in the previous chapter, the multiplication of royal residences has traditionally been taken to signal a transition from an ancient, centralised mode of rulership to one characterised by polyfocality and itinerancy. The basic narrative of the stable, city-based Merovingians giving way to the ‘itinerant’ Carolingians is well-established (and often repeated) but its guiding methodological assumptions and consequences have seldom been the 1976), 114–71; idem, ‘Die fränkischen Teilreiche im 7. Jahrhundert (613–714)’, in Spätantikes und fränkisches Gallien, 1:172– 230. 697 Wickham, Framing, 103–04. 698 Josiane Barbier, ‘Le système palatial franc : genèse et fonctionnement dans le nord-ouest du regnum’, Bibliothèque de l’École des Chartes 148, no. 2 (1990): 245–99. N.B. that Gregory of Tours, the seventh-century chronicle of ‘Fredegar’, as well as seventh-century hagiographers all had a concept of the overall Frankish realm, which they referred to as the regnum Fran- corum (only seldomly as Francia, which was usually used in reference to Neustria): Ewig, ‘Die fränkischen Teilungen’, 158– 59. 699 An exception are the excavations undertaken in Cologne underneath the city hall that revealed the fourth-century praetorium, a structure that was possibly reused as an Austrasian royal palace (see below). The possibilities of urban archae- ology are amply demonstrated by, for example: Guy Halsall, Settlement and Social Organisation: The Merovingian Region of Metz (Cambridge: Cambridge University Press, 1995). 700 See above, Introduction. 191 subject of scrutiny.701 Whereas in Italy our objective was to show that the focus on Ravenna as a singular ‘capital’ obscured a more flexible series of practices by which Italy was ruled, in this chapter we must reappraise the notion of ‘itinerancy’ itself. Itinerancy refers to a strategy of rule ‘in which a king carries out all the functions and symbolic representations of governing by periodically or constantly travelling throughout the areas of his do- minion’.702 Arising in post-war German scholarship, itinerancy provided a framework for explaining how the small-scale, ‘personalised’ administration represented by the royal court could extend author- ity over a large geographical area and overcome the vested interests of aristocratic independence.703 Others called back to Tacitus’ account of legal practice in first-century Germania, in which he described chieftains (principes) elected at the annual assembly who ‘administered law throughout the districts and villages’, and thus posited a continuous inheritance of a distant Germanic past in the Early Middle Ages.704 Originally applied to the frequent movements of the Ottonian and Salian emperors and sub- sequently projected backwards onto the Carolingian period, Itinerarforschung has become a standard methodology of medieval history in Germany.705 As we saw above, it is closely entwined with the historiography of capital cities and their disappearance: with the ruler not able to commit to a single, stable seat, the political centre moved throughout the territory, preventing the formation of large- 701 It is sometimes claimed that the Merovingians were also itinerant. See for example: Wood, Merovingian Kingdoms, 152– 53; McCormick, Eternal Victory, 332. As Barbier has explained, this idea has a long pedigree in historiography: Josiane Barbier, ‘Les rois mérovingiens entre fausse immobilité et fausse itinérance’, in Le gouvernement en déplacement. Pouvoir et mobilité de l’Antiquité à nos jours, ed. Sylvain Destephen et al. (Rennes : Presses Universitaires de Rennes, 2019), 189–90. 702 John William Bernhardt, Itinerant Kingship and Royal Monasteries in Early Medieval Germany, c. 936–1075, CSMLT, 4th Series, 21 (Cambridge: Cambridge University Press, 1993), 45. 703 For an overview: Schieffer, ‘Von Ort zu Ort’. That the king and a ‘free’ aristocracy possessed authority that was both 1. similar in quality (if unequal it its extent) and 2. opposed in its interests has been a powerful force in German scholarship. For an overview and critique, see: Steffen Patzold, ‘Der König als Alleinherrscher? Ein Versuch über die Möglichkeit der Monarchie im Frühmittelalter’, in Monarchische Herrschaft im Altertum, ed. Stefan Rebenich, Schriften des Historischen Kol- leges 94 (Berlin: De Gruyter, 2017), 605–633. 704 Tacitus, Germania, ed. Wilhelm Reeb (Leipzig: Teubner, 1930), c. 12, 33: eliguntur in iisdem conciliis et principes, qui iura per pagos vicosque reddunt. See e.g.: Berges, ‘Das Reich ohne Hauptstadt’, 8–9; Hans Conrad Peyer, ‘Das Reisekönigtum des Mittelalters’, Vierteljahrschrift für Sozial- und Wirtschaftsgeschichte 51 (1964): 3. For the deeper origins of the idea of a ‘Germanic’ constitution, see: Ian Wood, The Modern Origins of the Early Middle Ages (Oxford: Oxford University Press, 2013), 37–54. 705 It has had more limited impact elsewhere. See e.g.; Levi Roach, ‘Hosting the King: Hospitality and the Royal iter in Tenth-Century England’, Journal of Medieval History 37 (2011): 34–46. 192 scale, centralised administrative apparatuses.706 In the tradition of Verfassungsgeschichte, royal move- ments also express economic obligations—and thus ‘constitutional’ relationships—between the king and the aristocracy whose development can be traced from the Merovingian period through to the Central Middle Ages.707 As a mode of rulership, therefore, itinerancy expresses three characteristics: 1. A personalisation of political authority, such that it resided in the king’s person and not in broader administrative structures; 2. ‘Charismatic’ relations of power that could be interpreted in symbolic or sacral terms; and 3. A set of economic obligations by which the peripatetic court was provisioned. From a methodological perspective, ‘itinerancy’ is not a neutral concept; it represents the pe- riod through a series of distinctive narrative lenses. The first is a function of the evidence. The dra- matic increase in extant (authentic) charters from the eighth centuries onwards, as well as the absence of any equivalent to the Carolingian annalistic sources during the Merovingian period, means that reconstructing a recognisable ‘itinerary’ for Merovingian royal movements between palaces is essen- tially impossible.708 In a basic sense, this lends the perception of a break occurring with the rise of the Carolingians that we would be wise to treat with caution. In other quarters, however, the changing source base is taken to merely demonstrate the maturation of a longer-term, continuous development, despite the sparseness of the evidence for any such formal arrangements in the earlier period.709 In any case, Rosamond McKitterick has recently questioned whether the subscriptions of Charlemagne’s di- plomas always truthfully testified to his actual presence, and it is possible that the degree of Carolingian 706 Berges, ‘Das Reich ohne Hauptstadt‘, 8–9, 12. 707 Especially: Carlrichard Brühl, Fodrum, Gistum, Servitium Regis. Studien zu den wirtschaftlichen Grundlagen des Königtums im Frankenreich und in den fränkischen Nachfolgestaaten Deutschland, Frankreich und Italien vom 6. bis zur Mitte des 14. Jahrhunderts, 2 vols., Kölner Historische Abhandlungen 14 (Cologne: Böhlau, 1968). 708 Of the 196 Merovingian diplomas included in the new edition led by Kölzer (DD Merov. 1), 118 (sixty percent) are designated ‘inauthenthic’ (unecht); 77, in Kölzer’s estimation are authentic or ‘at least predominantly authentic’ (zumindest überwiegend echt), with one remainder of doubtful status (zweifelhaft): Theo Kölzer, introduction to DD Merov 1, xii. On spuria and the criteria adopted by the editors: Alexander Callander Murray, ‘The New MGH Edition of the Charters of the Merovingian Kings’, Journal of Medieval Latin 15 (2005): 263–68. On the at-times poor transferability of the methodologies of Pfalzen- and Itinerarforschung: Airlie, ‘The Palace Complex’, 260–61. 709 E.g. Georg Waitz, Deutsche Verfassungsgeschichte, vol. 2 (Kiel: Schwers’sche Buchhandlung, 1847), 544 (and similarly: ibid., vol. 4 [Kiel: Ernst Homann, 1861], 13); Brühl, Fodrum, Gistum, Servitium Regis, 1:18. 193 royal mobility is liable to be exaggerated.710 While the findings of Itinerarforschung usefully direct our attention to the institutional, legal, and economic mechanisms that supported Frankish kingship, there is a danger of making the exercise of royal authority seem overly transactional, and reducing palaces’ role therein to no more than ‘pieces of logistical support’.711 More generally, as we saw previously, positing a transition from ancient stability to medieval itinerancy is invested in a social evolutionary narrative centred on the State. To illustrate this, we can turn to Hans Conrad Peyer’s influential 1964 article that established the contours of ‘itinerant kingship’ in much subsequent scholarship.712 Peyer presented itinerancy as a widespread phenomenon in the Middle Ages and beyond, especially on Europe’s politically ‘primitive’ (urtümlich) periphery.713 He con- tinued, however, by noting that itinerant kingship is not unique to Europe, or indeed the Middle Ages, and goes on to describe rulership practices in various historical contexts, from Morocco and Ethiopia to the indigenous peoples of Micronesia and Polynesia.714 In Europe, Peyer contends that itinerancy died out with the transition to city-based, centralised administration – in other words, with the genesis of the State.715 In this, Peyer echoes Weber’s opposition of law-based, (modern) bureaucratic govern- ment with (implicitly non-modern) charismatic authority.716 To that end, Peyer concludes by position- ing itinerant rule as one of the ‘globally-distributed basic forms [Grundformen] of state-based rulership’.717 Itinerancy is thus situated in Peyer’s account within a social evolutionary progression 710 McKitterick, Charlemagne, 178–213. 711 Timothy Reuter, ‘Regemque quem in Francia pene perditit, in patria magnifice recepit: Ottonian Ruler Representation in Syn- chronic and Diachronic Comparison’, in Medieval Politics and Modern Mentalities, 141. Airlie, ‘The Palace Complex’, 260. 712 Peyer, ‘Reisekönigtum’. 713 Ibid., 11. For further critique, see: McKitterick, Charlemagne, 173–74. 714 Peyer, ‘Reisekönigtum’, 16–19. Peyer’s ethnographic turn here distantly echoes Geertz’s exposition of Weberian cha- risma through comparative examinations of royal mobility in sixteenth-century England, fourteenth-century Java, and Morocco in the late nineteenth century: Clifford Geertz, ‘Centers, Kings, and Charisma: Reflections on the Symbolics of Power’, in idem, Local Knowledge: Further Essays in Interpretive Anthropology (New York: Basic Books, 1983), 121–46. 715 Peyer, ‘Reisekönigtum’, 20. 716 Weber, Economy and Society, 3.4.10, 374–76. Weber remarks that, charismatic rulership is ‘bluntly opposed to all other forms of rule – rational, especially bureaucratic, as much as traditional, especially patriarchal or patrimonial, or rule based on social rank’ (ibid., 376). 717 Peyer, ‘Reisekönigtum’, 21: In jedem Falle aber zeigt dieser Überblick, daß Umritt, Herrscherreise und Gastung zu den weltweit verbreiteten Grundformen staatlicher Herrschaft gehören. 194 between primitive societies and modernity. The alliance of ‘itinerancy’ with an institutional history of the State primitivizes early medieval society—establishing Modernity as the telos toward which all things inexorably moved—but also abstracts politics from the particularities of the spaces and places where it found meaning.718 Yet this is not to say that there are not useful points to be taken from the ‘itinerancy’ framework. An attention to royal mobility reorients our attention in the moment of en- counter between ruler and power brokers in society and in the negotiation of obedience between centre and elites. It thus has the potential to indicate, as Itinerarforscher have shown, the establishment of zones of royal authority and the extension of royal authority over space – that is to say, the pro- duction of new ‘royal landscapes’. Overview This chapter considers the political landscapes of the two main Frankish subkingdoms, Austrasia and Neustria, and their renegotiation under the Carolingians in the formation of a new, ‘imperial’ political geography. As we will see, the palatial network of the Frankish regnum were not just waypoints in royal itineraries—places where a weary king could overnight—but sources of integration and intensification of royal authority. The Merovingian and, later, Carolingian kings were certainly mobile, but this does not at all compare to Reisekönigtum of the Ottonian or Salian empires. Unlike in Italy, however, the Frankish regnum provides the opportunity to trace the fortunes and fluctuations of the palace network and their attendant landscapes over a longer time span. Beginning in the eastern kingdom of Austrasia in the sixth century, we will once again witness the generative potential of the Roman frontier in the formation of new political units as we did in northern Italy under Theoderic. But the Roman inher- itance played a limited role in dictating the future course of political formation. Steadily from the seventh century, we see the ‘resources’ of the Roman military-political landscape becoming less 718 Smith, Political Landscape, 79, 90–94. 195 relevant, and by the ninth-century zenith of Carolingian imperial ambition, we are able to discern the coalescence of new political landscapes that owed little to the Late Roman Empire. II. Authority on the Frontier: Austrasia The death of Chlothar I in 561 is a watershed in Gregory of Tours’ Histories. Gregory tells us that Chlothar’s son, Chilperic, immediately seized his father’s treasury from the villa of Berny-Rivière and distributed gifts to the ‘more advantageous’ of the Franks to win support for his cause.719 Feeling confident in his position, he invaded the territory of his half-brother, Childebert, and occupied Paris, the former seat of the great king Clovis and still the symbolic heart of the kingdom.720 Unfortunately for him, however, victory would prove illusory. Gregory goes on to narrate that, he was not able to hold onto it [Paris] for long, however. For his brothers in alliance drove him out, and thus the four of them—that is, Charibert, Guntram, Chilperic, and Sigibert—made a lawful division among themselves. The agreement granted to Char- ibert the kingdom of Childebert and he made Paris his seat, to Guntram the kingdom of Chlodomer and he had his seat at Orléans, to Chilperic the kingdom of Chlothar, his father, and he had his throne at Soissons, and to Sigibert the kingdom of Theuderic and he had his seat at Reims.721 These events seem to have recalled a similar divisio made between Clovis’ sons on his death half a century earlier, but their meaning for Gregory principally derived from contemporary political circum- stances.722 The settlement of 561 marked for the bishop of Tours the disintegration of the unified kingdom of Chlothar and thus heralded the political order of Gregory’s own day, the subject of a 719 Greg. Tur., Hist., IV.22, 154: Chilpericus vero post patris funera thesaurus qui in villa Brannacum erant congregati accepit et ad Francos utiliores petiit ipsusque muneribus mollitus sibi subdidit. 720 On Paris, see below, 203–09. 721 Greg. Tur., Hist., IV.22, 155: non diu ei hoc licuit possedere; nam coniuncti fratres eius eum exinde repulerunt, et sic inter se hii quattuor, id est Chariberthus, Gunthramnus, Chilpericus atque Sigiberthus, divisionem legitimam faciunt. Deditque sors Charibertho regnum Childeber- thi sedemque habere Parisius, Gunthramno vero regnum Chlodomeris ac tenere sedem Aurilianensem, Chilperico vero regnum Chlothari, patris eius, cathedramque Sessionas habere, Sygibertho quoque regnum Theuderici sedemque habere Remensim. Sors can denote the drawing of lots, yet here it clearly refers to the outcome of the settlement rather than some kind of random allocation. 722 Gregory says little about the 511 partition, only that Clovis’ sons ‘took possession of his kingdom and divided it equally amongst themselves’: Greg. Tur., Hist., III.1, 97: regnum eius accipiunt et inter se aequa lantia dividunt. A divisio legitima probably refers to a legal framework for the allocation of civitates to each brother sealed by a treaty, rather than having a sense of ‘equal’ as such, fitting given Chilperic’s attempts to usurp the entirety of the kingdom at the expense of his brothers. Cf. the 511 division, done ‘fairly’ (aequa lantia; literally ‘with the scales equal’): Hist. III.1, 97. Marc Widdowson has pointedly argued that Gregory’s description of the 511 and 561 settlements was related to the political situation at his time of writing: Marc Widdowson, ‘Merovingian Partitions: A “Genealogical Charter”?’, EME 17 (2009): 1–22. 196 lengthy polemic in the preface to the fifth book of his Histories.723 Although the sources for this period are scarce at best, just discernible are the contours of new political landscapes taking shape. Concen- trating first on the eastern kingdom, known already by the end of the sixth century as Austrasia, reveals the first stage of this process. 724 In Austrasia, the erstwhile Roman frontier provided the ‘resources’ for a new political landscape, formed through the localisation of power in new palaces in former imperial and military centres. Military infrastructure and political space: northern Gaul from the third to fifth centuries The Frankish regnum was formed out of the militarised society of late-Roman Gaul. Before turning to the establishment of new political centres on the ‘barbarian’ frontier, we must first look to this context. From the late-third century, following incursions across the Rhine and the abandonment of the trans- Rhenan agri decumates, the provinces of Gaul north of the Loire were steadily militarised, producing a wide border region in which the society and the economy was turned towards the frontier.725 These transformations were visible in the physical landscape both on the frontier and behind it. Most evi- dently, as in northern Italy in the same period, an extensive network of fortifications were constructed from the fourth century onwards.726 The fortress of Divitia (Deutz), for example, was constructed during the reign of Constantine on the Rhine’s right bank and connected to Cologne opposite by a bridge to provide a defensible bridgehead for sorties beyond the river and to protect a river fleet (Figs. 723 Greg. Tur., Hist., V.praef. 193–94; Guy Halsall, ‘The Preface to Book V of Gregory of Tours’ Histories: Its Form, Context, and Significance’, English Historical Review 122, no. 496 (2007): 297–317. 724 As a group name for inhabitants of the Reims area (Austrasiis): Greg. Tur., Hist., V.14, 213; V.18, 224. As a territorial denomination, referring to the area of Metz (Austria): Greg. Tur., De virtutibus S. Martini, IV.29, in Gregorii episcopi Turonensis miracula et opera minora, 206. From the seventh century the form Auster would become common. See also: Ewig, ‘Die fränkischen Teilungen’, 166–71. 725 Simon Esmonde Cleary, The Roman West, AD 200–500: An Archaeological Study (Cambridge: Cambridge University Press, 2013), 42–96, 376–86; Laury Sarti, ‘Die spätantike Militärpräsenz und die Entstehung einer militarisierten „Grenzgesell- schaft“ in der nordwesteuropäischen limes-Region’, in Militärische Migration vom Altertum bis zur Gegenwart, ed. Christoph Rass, Studien zur Historischen Migrationsforschung 30 (Paderborn: Ferdinand Schöningh, 2016), 43–56. 726 For an overview: Hans Schönberger, ‘The Roman Frontier in Germany: An Archaeological Survey’, JRS 59, no. 1–2 (1969): 144–97; and especially: Harald von Petrikovits, ‘Fortifications in the North-Western Roman Empire from the Third to the Fifth Centuries A.D.’, JRS 61 (1971): 178–218. 197 129–30).727 The hinterland was punctuated by a range of fortified sites: newly-walled urban areas, outposts overlooking highly-trafficked routes (known as burgi), and other installations of variable size and typology.728 It is important to note, however, that not all of these installations were the result of central planning. As we established in the previous chapter, correlating provincial militarisation in the late Empire with the adoption of a strategy of ‘defence in depth’ overestimates the singular coherency of what were, by 461, two centuries of developments executed with varying levels of consistency and consideration.729 The burgus at Liesenich, for example, completed in 268 or 269 and located on a road running south-east from the Moselle across the Hunsrück (Rheinland-Pfalz) was built at local initiative rather than with explicit imperial sponsorship.730 The militarisation of the physical and urban environment was matched in society. Following the collapse of the Rhine frontier and the disintegration of central authority over the Gallic provinces, a number of small-scale political formations coalesced in the north out of provincial Roman military structures. The Franks of Childeric, whose lavishly furnished grave was discovered in Tournai in 1653 (Fig. 131), were but one group active in this area.731 Importantly, these groups were not composed exclusively of newcomers arriving from outside Roman territory, but out of the structures of Roman society itself. Procopius, for example, describes Roman soldiers surrendering to the Franks (whom he terms, in proper classicising fashion ‘Germans’, Γερμανοί) and another group known as the Arborychoi (Ἀρβόρυχοι), who he claims had once also served Rome.732 In return for their submission, these soldiers 727 Maureen Caroll-Spillecke, ‘Das Militärlager Divitia in Köln-Deutz’, Kölner Jahrbuch 26 (1993): 321–444; Werner Eck, Köln in römischer Zeit: Geschichte einer Stadt im Rahmen des Imperium Romanum, Geschichte der Stadt Köln 1 (Cologne: Greven, 2004), 605–10. Similar ‘bridgeheads’ were built opposite Mainz (Kastell) and Kaiseraugst (Whylen). 728 Esmonde Cleary, Roman West, 49–79; von Petrikovits, ‘Fortifications’, 193–203. 729 Above, Chapter 3, 160–61. 730 CIL XIII, 11976: Qui burgum (a)edificaverunt Lup(ulinius) Am|minus pr(a)efectus Sab(inius) Acceptio Vid(ucius) | Perpetu(u)s Flavius Tasgillus CO() Lepidus | Min(ucius) Luppus cum C(a)es(ius) Ursulus paratus | est Victorino Augusto et | Sa(n)cto co(n)s(ulibus) X Kal(endas) Iunias; Schönberger, ‘Roman Frontier’, 178. 731 Halsall has argued that Childeric’s power was centred not in the region of Tournai, but probably in Loire valley and the Paris basin: Halsall, Barbarian Migrations, 270. Most of the treasure discovered in Childeric’s grave was stolen and destroyed in 1831. 732 Proc., Wars, V.12.7–11, 2:64. 198 were permitted to maintain their own customs and, when called upon for battle, to march under their own standards.733 This episode demonstrates two important aspects of the post-Roman transition in Gaul. Firstly, it highlights the sheer localisation of social action: these soldiers were essentially or- phaned from any larger-scale political structures after the collapse of the Roman frontier. But it also indicates the persistence of essentially ‘Roman’ military identities as the basis for new, albeit small scale, groupings, as highlighted by Procopius’ focus on the soldiers retaining their standards, distinc- tive clothing, and customs. The involvement of the Arborychoi in this episode, probably a corruption of Armoricani (that is, soldiers of the former tractus Armoricanus), similarly demonstrates the lack of a definitive boundary between an ex-Roman military unit and a ‘barbarian warband’ in a post-imperial context.734 There are echoes in Gregory of Tours’ Histories of similar groups on the Rhine frontier. For example, Gregory relates that the Franks elected Aegidius, the magister militum, as king in the time of Majorian; his son, Syagrius, is later named ‘king of the Romans’ (rex Romanorum) and apparently ruled a small kingdom from Soissons.735 Likewise, a comes Arbogast ruled from Trier, who was praised by Sidonius Apollinaris for his excellent Latin despite his kinship with barbarians, and elsewhere Gregory reports that an independent Frankish polity ruled by a king Sigibert—a cousin of Clovis—was based at Cologne and subsequently brought under Clovis’ authority shortly after his victory at Vouillé in 507.736 733 Ibid., V.17–18, 65. 734 Stefan Esders ‘Nordwestgallien um 500. Von der militarisierten spätrömischen Provinzgesellschaft zur erweiterten Mi- litäradministration des merowingischen Königtums’, in Chlodwigs Welt: Organisation von Herrschaft um 500, ed. Mischa Meier and Steffen Patzold, Roma Aeterna 3 (Stuttgart: Steiner, 2014), 345–50. The jurisdiction of the dux tractus Armoricani, like the comes Italiae discussed in Chapter 3, is reported in the Notitia Dignitatum: Not. Dig. Occ. XXXVII, 204–06. The area under the dux’s authority was large, taking in the three provinces of Lugdunensis between the Seine and Loire and the two of Aquitania in the southwest: ibid., XXXVII.24–29, 205: Extenditur tamen tractus Armoricani et Neruicani limitis per prouinicas quinque: per Aquitanicam primam et secundam, Lugdunensem Senoniam, Secundam, et tertiam. 735 Greg. Tur., Hist., II.12, 61–62; II.18, 65; II.27, 71. On the question of whether this geography preserved the limits of late-Roman military districts: Ewig, ‘Die fränkische Teilungen’, 118–20. 736 Sidonius Apollinaris, Epistulae, IV.17, 68: tertia urbanitas, qua te ineptire facetissime allegas et Quirinalis impletus fonte facundiae potor Mosellae Tiberim ructas, sic barbarorum familiaris, quod tamen nescius barbarismorum, par ducibus antiquis lingua manuque, sed quorum dextera solebat non stilum minus tractare quam gladium. Cologne: Greg. Tur., Hist., II.40, 90–91; Eugen Ewig, Trier im Merowingerreich: Civitas, Stadt, Bistum (Trier: Paulinus, 1954), 22, 62. 199 The key point here is that the new political groupings of fifth-century Gaul, absorbed into the growing Frankish regnum in the time of Clovis, originated out of the militarised, ‘warlord’ politics of the Roman frontier – a development we also witnessed in northern Italy. As scholars such as Esders, Halsall, and others have emphasised, early Frankish rule in Gaul took the form of an expanded Roman military administration far more than any kind of ‘barbarian’ collapse.737 With the collapse of Roman hegemony over Gaul, new, often fissiparous social groupings came together as earlier frameworks of power and legitimacy were fractured. Palaces and the Austrasian political landscape The consequent transformation of Gaul’s political geography is most visible in the case of Austrasia, centred on the Middle Rhine and Moselle. The establishment of palaces in communication with the Rhine limes highlights the generative potential of the Roman frontier in the formation of new political units. As we saw in Italia annonaria under the Ostrogoths in the last chapter, in the sixth century Roman military-political infrastructures functioned as a resource from which a new landscape was produced. Additionally, however, a closer attention to the location of Frankish palaces in the sixth century will also call the common perception of the early Merovingians as ruling from ‘capital-like’ residences in direct continuity with Roman urban politics into question (Fig. 132). Reims. In the settlement of 561 the eastern kingdom of Theudebert fell to Sigibert and, in Gregory’s telling, he established his seat at Reims. Reims had been a significant node in Roman limes system. As the provincial capital of Belgica II, the headquarters of military prefects of barbarian laeti and ‘Sarmatian’ cavalry, and a site of arms production and other industrial products, Reims and its hinterland represented a significant concentration of military and political resources.738 Clovis’ baptism 737 Esders, ‘Nordwestgallien’, 354. 738 Not. Dig. Occ. IX.36, 146: Remensis spatharia; XLII.42, 217: Praefectus laetorum gentilium, Remo et Siluanectas Belgicae secundae; XLII.67, 219: Praefectus Sarmatarum gentilium, inter Renos et Tambianos, prouinciae Belgicae secundae. It also housed one of the four praepositi thesaurorum of Gaul, and factories producing wool, and gold-cloth weavers: Not. Dig. Occ. XI.34, 56, 76, pp. 150–51. Brühl, Palatium und Civitas, 1:51–55. 200 by Remigius in 496 no doubt invested the city with additional significance, but also reflects the city’s existing status as a site of ecclesiastical and secular power.739 In the context of the political geography of the kingdom in the sixth century, Reims was oriented towards the regnum’s centre as were the other seats mentioned by Gregory (Paris, Orléans, and Soissons). Reims’ importance as a palace site seems to have been waning already by 561, however. It is commonly understood that it had served as a royal seat since the partitioning of Clovis’ kingdom in 511, though it is noteworthy that Gregory does not specify this in his cursory remarks (and we should suspect that Gregory did not himself know much about this earlier event).740 It is notable that, unlike other centres, no contemporary source physically places the Austrasian kings in Reims. While not meaningful on its own given the general paucity of sources from this period, there are other indications that the city was no longer the pre-eminent royal seat by Gregory’s time of writing if not earlier.741 This would not be unprecedented: the strong impression given by Gregory’s narrative is that, although Orléans was Guntram’s designated regnal seat according to the 561 divisio, he evidently preferred Cha- lon-sur-Saône for much of his reign.742 Already by the 570s, for example, the Champagne region was governed by a dux based at Reims. Venantius Fortunatus dedicated three poems to Lupus, dux of Champagne (d. after 590), a magnate at the court of Sigibert and, later, Childebert II (r. 575–96).743 In a lengthy poem in praise of his patron, Fortunatus dwells on Lupus’ central role at the court: In the presence of our masters you make the august palace complete, and its glory is magnified by the glory that enters with you. With your coming the venerable halls shine with a new brightness, and the royal dwelling reacquires its true nature. For when it sees you return the palaces receives back its eyesight, since the eyes of the dukes 739 Greg. Tur., Hist., II.31, 76–78. 740 E.g. Dierkens and Périn, ‘Les sedes regiae mérovingiennes’, in Sedes regiae (ann. 400–800), 284–85. Gregory’s assigning Paris and Soissons to Childebert and Chlothar respectively at least is credible given their foundation of the privileged churches of Saint-Vincent and Saint-Médard respectively: Ewig, ‘Die fränkischen Teilungen’, 116, n. 12; idem, ‘Résidence et capitale’, 49–50. 741 Ewig, ‘Résidence et capitale’, 49. 742 E.g. Greg. Tur., Hist., VIII.1, 370–71. 743 PLRE, 3B:798–99. 201 depend on your, their common light, right hand of the ruler, glory of your country, bulwark of your kin, consolation for others, for all their one love.744 His praise of Lupus, as with his similar letters to the magnate Gogo, seem to suggest that Lupus resided at least temporarily at the royal court. As we will see below, however, insofar as he provides specific details, Fortunatus associates the court with Metz and the Moselle rather than Reims. While admittedly conjectural, it seems that already by the early-570’s Champagne had been entrusted to a dux while the court resided elsewhere. Reims evidently remained an important political node but left unspoken by Gregory is that the kingdom’s centre of gravity was already shifting eastwards. Cologne. We encounter additional palaces of the Austrasian kings in other sources. Located in the kingdom’s eastern half in communication with the Moselle and Rhine river systems, these sites underscore the importance of the militarised Roman frontier as the basis of Frankish authority. Co- logne’s importance, for example, can doubtless be connected to its commanding position on the Rhine. Cologne certainly possessed a palace of some kind in the sixth century, possibly as early as the reign of Theuderic I (511–534), and both ‘Fredegar’ and the Liber Historiae Francorum state clearly that it was the location of Theudebert II’s treasury (king of Austrasia 595–612) during his conflict with his brother Theuderic II (king of Burgundy 595–613, king of Austrasia 612–13).745 Although Cologne is not mentioned by name in the Notitia Dignitatum, the city’s praetorium—newly rebuilt at the end of the fourth century and clearly intended to have a strong visual impact over the nearby Rhine—probably served as the seat of the consularis Germaniae secundae and later of the Austrasian kings.746 744 Ven. Fort., Carm. VII.7.65–72, 2:96: Occurrens dominis ueneranda palatia complens | et tecum ingrediens multiplicatur honor. | Te ueniente nouo domus emicat alma sereno | et reparent genium regia tecta suum. | Nempe oculos recipit cum te uidet aula redire | quem commune ducum lumina lumen habent : | prinicipis auxilium, patriae decus, arma parentum, | consultum reliquis, omnibus unus amor. Trans. Rob- erts 439–41. 745 Greg. Tur., Vita Patrum, VI.2, 231; Fred., Chron., IV.38, 139; Liber Historiae Francorum, c. 38, 307–08. See above, Chapter 2, 113–14. 746 Not Dig. Occ. I.72, 105; Eck, Köln, 664–66. For continuity with the Frankish period: Brühl, Palatium und Civitas, 2:37; Dierkens and Périn, ‘Les sedes regiae mérovingiennes’, 278; Otto Doppelfeld, ‘Das praetorium unter dem Kölner Rathaus’, in Neue Ausgrabungen in Deutschland, 313–21. 202 Trier. The other significant imperial site taken over as a Merovingian palace was Trier. There are indications in Gregory of Tours’ writings that the former provincial and imperial seat served as a royal residence under Theuderic and his son, Theudebert, yet it goes unmentioned in other sources.747 As the metropolis of Gallia Belgica, in the Diocletianic Empire Trier became the seat of the Praetorian Prefecture of the Gauls (until it was moved to Arles in 407) as well as a favoured imperial residence for much of the fourth century. Along with the imperial and prefectural administrations, Trier hosted shield, ballista, cloth-weaving and armour-gilding manufactories, making it the most important city in Gaul for outfitting the army, and we should likely expect a significant troop presence billeted in its hinterland.748 Even if royal presence in Trier is hard to detect, it nonetheless remained a political Schwerpunkt. The imperial fiscal lands lying between the Moselle and the Rhine made the king the most important landowner in the region, and the prominence of Nicetius, bishop of Trier (d. 566) points to close links between the former imperial sedes and the royal court.749 Metz. The favoured location of the Austrasian court from the 560’s, however, had less obvi- ous imperial associations, though it was still implicated in the arterial routes of the Roman imperial machine. Located on the banks of the Moselle, Metz occupied a significant location in the road net- work of the Roman limes, at a crossroads between the routes leading from the central regions of Lug- dunensis and Aquitania, Strasbourg to the East, and the Rhine frontier via Trier to the North. While Metz does not feature prominently in the Notitia Dignitatum, it was garrisoned by the pseudocomi- tatensan legion I Flavia Metis possibly under the authority of the dux Belgicae secundae at Reims.750 The 747 Greg. Tur. Hist., X.29, 522. 748 Not. Dig. Oc. IX.37–38, 146; XI.58, 77, p. 152. Ewig, Trier, 21. 749 Ewig, Trier, 64. Greg. Tur., Vita patrum, XVII.1, 277. Nicetius was the author or recipient of seven letters in the Episulae Austrasiacae, with correspondents including the emperor Justinian and Chlodosuinth, daughter of Sigibert and wife of the Lombard king Alboin. It is worth noting that the collection seems to have been based at some level on royal archives and all the individuals involved were close to the royal court. For a recent discussion of the collection’s date and place of compilation see: Andrew Gillett, ‘Telling Off Justinian: Theudebert I, the Epistolae Austrasicae, and Communication Strat- egies in Sixth-Century Merovingian-Byzantine Relations’, EME 27, no. 2 (2019): 171–80. 750 Listed under the command of the magister peditum praesentalis, I Flauia Metis was associated with Metz by A. H. M. Jones: Not. Dig. Oc., V.269, 127; Jones, Later Roman Empire, 3:365, 373. By analogy with the pseudocomitatensan unit I Flavia 203 urban area contracted significantly after 400 but it seems to have enjoyed renewed intensity in social and economic activity from around 550, possibly on account of a regular royal presence.751 Although Gregory gives little detail about the city, his narrative regularly places Childebert II (r. 575–595) in Metz and he clearly understood it to be the usual royal residence.752 ‘Fredegar’ concurs that Metz had been Childebert’s regular seat, which was inherited by his son Theudebert II (r. 596–612).753 On the basis of a later charter of Charles Martel, witnessed in 743, the royal palace was still in use in the eighth century.754 The Austrasian kings were associated, therefore, with four civitates closely linked by the Mo- selle–Rhine river systems and the Roman frontier road network. Strikingly, however, it seems that the ‘imperial’ symbolism offered by Trier were of limited significance in the localisation of power in the Austrasian landscape. As we saw earlier in Ostrogothic Italy, it rather seems that sites of military sig- nificance—here associated with the navigable river systems of the Rhine, Meuse, and Moselle— formed the basis for a new royal landscape that emerged out of the infrastructures of Roman imperi- alism. What we are also able to perceive is that political authority in Austrasia was not tied to a single stable centre, but was altogether more flexible in its spatial setting, a development we can see even more clearly by examining evidence for royal mobility in the sixth century. ‘The whole of the royal retinue was afloat on the waves’: mobility on the Austrasian frontier The Italian poet Venantius Fortunatus provide a tantalising insight into the life of Austrasian court. While he does not provide anywhere near enough information to reconstruct a coherent ‘itinerary’ of Gallicana Constantia, who are listed under the authority of the dux Tractus Armoricani (in the Notitia the commander of forces of Gallia Lugdunensis) and headquartered in the territory of their eponymous city, Coutances, this seems likely: Not. Dig. Oc., XXXVII.20, 205: Praefectus militum primae Flauiae, Constantia. 751 Halsall, Settlement, 214–41. 752 E.g. Greg. Tur., Hist., VIII.36, 404; Gregory also recounts that he visited Childebert at Metz in around 587: Greg. Tur., Hist., IX.13, 428. On the possible location of the palace: Brühl, Palatium und Civitas, 2:43–44, 59–62. 753 Fredegar, Chron., IV.16, 127 : Teudebertus sortitus est Auster sedem habens Mittensem. 754 Die Urkunden der Arnulfinger, ed. Ingrid Heidrich, MGH. DD Arnulf (Hannover: Hahn, 2011), no. 17, 41: in civitate Metis in palatio regio. 204 the Austrasian kings per se, two of his works nonetheless illustrate the thought world they inhabited— the social assumptions and norms that governed their behaviour—and suggest a more mobile, flexible mode of authority than a singular concentration on the sedes regiae would suggest.755 In what must doubtlessly be interpreted as a humorous anecdote, in one poem Fortunatus relates a fraught riverine journey from Metz to a royal estate named Nauriacum, possibly Norroy-le- Veneur outside Metz (Fig. 132).756 After his boat was appropriated by the king’s cook before he could board it, he suffered an uncomfortable journey to Nauriacum to join the court. There he remained as an honoured guest of Sigibert while a new boat was sought for him by Gogo and the comes Papulus (Fortunatus does not report the fate of the cook).757 Amusing as Fortunatus’ hyperbolic protestations are, the vignette hinges on the mobility of the court as a group. He laments: ‘They looked, but were unable to find any vessel, while the whole of the royal retinue was afloat on the waves’.758 Another poem describes his travelling again on the Moselle, this time with Childebert II to another royal resi- dence at the castellum of Andernach, in which echoes of Ausonius’ Mosella give way to something more resembling panegyric as the poem reaches its climax of the royal banquet in the hall.759 Andernach is notable for its former role as a legionary fortress and port for the Rhine patrol fleet in the fifth century according to the Notitia Dignitatum; archaeological investigations have indicated that the site was 755 Barbier, ‘Les rois mérovingiens’, 199–200. 756 Nauriacum has not been definitively identified. Norroy-le-Veneur, some seven kilometres outside Metz and close to the Moselle’s left bank, is the likeliest location. Though first textually attested as a fiscal estate in a charter of Otto I (Diplomata Ottonis I, no. 210, in Die Urkunden Konrad I. Heinrich I. und Otto I., ed. Gesellschaft für ältere deutsche Geschichtskunde, vol. 1, MGH. Diplomatum Regum et Imperatorum Germaniae 1 [Hannover: Hahn, 1879], 290: in comitatu Scarponensi villam Nogaredus dictam cum aecclesia et manso indomicato et ombinibus ad eum aspicientibus [in the county of Scarponna the estate named Nogaredus along with its church and the demesne manse and everything that pertains to it]), the site was clearly of earlier date, as indicated by the large cemetery of 298 burials excavated in 2005, datable to between the fifth and seventh centuries (study of which is ongoing). 757 Ven. Fort., Carm., VI.8, 2:77–79. Gogo was important at the court of Sigibert and was the addressee of four poems by Fortunatus (Carm., VII.1–4). He was later the tutor to the young Childebert before his death in 581. See: PLRE, 3A:541. Papulus: PLRE, 3B: 967. 758 Ven. Fort., Carm., VI.8.35, 2:79: Quaerunt, nec poterant aliquam reperire carinam, | donec cuncta cohors regia fluxit aquis. Trans. Roberts, 407. 759 Ven. Fort., Carm., X.9.63–64; 3:85; Michael Roberts, ‘The Description of Landscape in the Poetry of Venantius Fortu- natus: The Moselle Poems’, Traditio 49 (1994): 12–20. See also: Paul Dräger, ‘Zwei Moselfahrten des Venantius Fortunatus (carmina 6,8 and 10,9)’, Kurtrierisches Jahrbuch 39 (1999): 67–89. 205 continuously occupied into the sixth century.760 Andernach seems, in fact, to have been a site of some importance. Capitularies of Childebert appended to the Pactus legis salicae in 596 were given at Ander- nach, as well as Maastricht and Cologne, in a meeting of the Austrasian elite held each year on the 1 March.761 The meeting of the Austrasian elite at these sites represents a striking reassertion of the Roman frontier in the imaginative construction of new boundaries between ‘insider’ and ‘outsider’ – between members of the Austrasian political community and the Saxons and Frisians that lay outside it.762 Significant in Austrasian royal mobility was the importance of the Roman road and supply networks for supporting royal power, and especially riverine routes. Venantius Fortunatus’ poems praising the Moselle’s landscape has as much to do with its status as an axis of royal and aristocratic activity as his reverence for Ausonius. His borrowings from the fourth-century poet accentuate the nobility of his subjects by elevating the landscape over which they held sway to a paradisiacal idyll; the beauty and fecundity of the Moselle reflects those who ruled it.763 The case of Andernach also under- lines the importance of river travel for connecting locality and political centre. The Moselle allowed direct travel between them in as little as three days, as opposed to likely around a week needed to travel by road.764 Riverine travel also offered the potential for travel further afield. Fortunatus also 760 Not. Dig. Occ. XLI, 13 and 25, 213–14; Josef Röder, ‘Neue Ausgrabungen in Andernach’, Germania 39 (1961): 208–13; Eveline Saal, ‘Forschungen zum antiken Rheinhafen in Andernach’, in Der Rhein als europäische Verkehrsachse: Die Römerzeit, ed. Heike Kennecke (Bonn: Vor- und Frühgeschichtliche Archäologie, Rheinische Friedrich-Willhelms-Universität, 2014), 63–77. 761 Capitula legi Salicae addita VI.1, in Pactus legis Salicae, ed. Paul August Eckhardt, MGH. LL nat. Germ. 4/1 (Hannover: Hahnsche, 1962), 267. Maastricht, like Cologne and Andernach, was implicated in limes supply network and was later an episcopal centre. See: Frans Theuws, ‘Maastricht as a Centre of Power in the Early Middle Ages’, in Topographies of Power in the Early Middle Ages, ed. Mayke de Jong et al., Transformation of the Roman World 6 (Leiden: Brill, 2001), 155–216. 762 Stefan Esders, ‘Man-like Discipline and Loyalty Against the “Enemies of God”: Some Observations on the Militarised Frontier Society of Eastern Francia Around 600’, in Early Medieval Militarisation, ed. Ellora Bennett et al. (Manchester: Manchester University Press, 2021), 181–95. I offer my sincere thanks to Prof. Dr. Esders for sharing an advanced copy of his essay with me. 763 See e.g. Ven. Fort., Carm., VII.4, 2:88–89, where a long list of animals hunted by Gogo by the Moselle and Rhine (as well as in the Ardennes and Vosges) is paired with the size of his retinue in the royal palace: sive palatina residet modo laetus in aula,| cui scola congrediens plaudit amore sequax? Roberts, ‘Description of Landscape’, 6–7, 9, 18. 764 Estimates aided by ORBIS: The Stanford Geospatial Network Model of the Roman World, orbis.stanford.edu, accessed 15/12/2019. I would suggest that the royal party would break their journey probably at Trier (mentioned by Fortunatus in his poem), Gondorf on the Moselle, where there was a ‘fortress of ancient nobility’ (Ven. Fort., Carm., X.9.45–46, 3:85: 206 describes the route between Italy and Gaul in his metrical Life of Saint Martin as one largely undertaken by boat. For him Gaul and Italy were linked by two overlapping geographies: a natural one criss- crossed with navigable rivers either side of the Alps, and a conceptual one structured by the cults of important saints which lay in his way.765 Despite their relative compactness of the Austrasian royal landscape, therefore, we should not lose sight of the superimposition and interconnectedness of many such conceptual orderings of space in this period. *** In Austrasia we again witness the decisive role of the Roman frontier in the shaping of early medieval political formations. The overall pattern is like that of northern Italy: Roman frontier infrastructures providing the ‘resources’ through which the authority of new groups could be constituted and articu- lated. The emplacement of palaces into this landscape reflects the new political underpinnings of the Austrasian kingdom, a transformed notion of political authority originating out of the ‘warlord’ politics of post-imperial Gaul. We have also seen that, while not ‘itinerant’, there are hints in sources such as Venantius Fortunatus that the rulership of the Austrasian kings was marked by a high degree of mo- bility that both spun new webs of social relations and inscribed the new territory into the landscape. III. From civitas to pagus: Neustria In the 561 settlement, the kingdom of Childebert, with its seat at Paris, fell to his grandson, Chari- bert.766 In addition to the largest contiguous territory, Charibert’s receipt of Paris marked him out with distinction. Paris was the symbolically the most significant city in the Frankish kingdom from Clovis’ quo fuit antiquum nobilitate caput. Trans. Roberts, 679) and possibly also at Koblenz before reaching Andernach. By road Andernach is around 262km (ca. 163 miles) from Metz. A manageable pace of 30 miles per day by horse would thus have the royal party arrive during the sixth day, but depending on the size of the party, the amount of goods brought along, the condition of the roads, the number of petitions heard en route (etc.) one could easily imagine this taking longer. 765 Ven. Fort., Carm., praef.4, 1:4; Vita S. Martini, IV.630–80, in Venantius Fortunatus, Oeuvres, ed. and French trans. Solange Quesnel, vol. 4, Collection des universités de France (Paris: Les Belles Lettres, 2002), 98–100. 766 Chilperic had to make do with a much-reduced territory in the north-east, though would receive much of his elder half- brother’s possessions on his death six years later.; Ewig, ‘Die fränkischen Teilungen’, 138–40. 207 reign, and control over it was understood by contemporaries to symbolise a legitimate right to rule. Strikingly, however, by the seventh century the king was seldom to be found in Paris itself, but rather in a corona of suburban villa-estates located in the pagus. The central region of Neustria, therefore, highlights a significant transformation in the settings where royal authority was apprehended in the Early Middle Ages. From the later-sixth century, royal power was increasingly displaced from cities into rural estates, a function of a transition from primarily urban to non-urban modes of political representation.767 The Neustrian kings—and, significantly, those of the unified regnum—were mobile among a constellation of palaces located between the Seine and Oise, a sign not only of the continued gravitational pull of Clovis’ cathedra regni but also of the entanglement of royal and aristocratic power in this landscape. This constellation of palaces was not permanent, however, and the political geogra- phy of the Neustrian Teilreich was continually renegotiated and reformed over the course of the sixth and seventh centuries. Concentrating our attention on Paris and Neustria, therefore, not only high- lights a fundamental transformation in the environments of political action after 500, but also the sedimentation of the borders (if not the centre) of a stable royal landscape. Paris as Symbolic Centre Unlike Austrasia, whose social and political structures coalesced in close dialogue with the infrastruc- tures of the Rhine frontier, what would by the seventh century be known as Neustria possessed a distinct structure and character.768 Formed out of the combination of the kingdoms of Paris and Sois- sons in the divisiones of the sixth century, Neustria stretched from the Silva Carbonaria in the East to the Breton March and the Loire in the West and South.769 Tellingly, the name ‘Neustria’ is not itself attested in contemporary sources before the middle of the seventh century, and even then principally 767 See above, Chapter 2. 768 We need not distinguish between a ‘Roman’ Neustria and ‘Germanic’ Austrasia, however: e.g. Waitz, Deutsche Verfas- sungsgeschichte, 2:77–79. 769 Ewig, ‘Die fränkischen Teilungen’, 115, 154. 208 in Austrasian and Burgundian sources; authors from Neustria itself referred to themselves simply as Franci, reflecting a political identity rooted in a perceived direct lineage from a distant past.770 Unlike Cologne, Trier, or even Reims, Paris had been a Roman civitas of only middling rank. According to the Notitia Dignitatum its land had quartered a division of ‘Sarmatian’ cavalry as well as a riverine fleet, and its integration into good communication routes made it attractive as a site to winter campaigning armies, as Julian did in 358.771 The provincial capital of Lugdunensis IV was not Paris, however, but Sens, and there is no evidence in the written or archaeological corpora that the city was marked out for especial distinction by the fourth century emperors that would reflect any kind of ‘imperial’ status.772 The Parisian palace, which Brühl identified with the Roman praetorium, seems to have lain within the walled castrum of the Île de la Cité, possibly on the site of the later Palais de Justice (Fig. 133), though this has not been proven archaeologically.773 Ecclesiastical patronage underlines that the walled urban area remained of interest to the royal court through the sixth and the seventh centu- ries and, as fiscal land, lay within the royal gift.774 770 Jonas of Bobbio’s Vita Columbani, completed in 642 is the first attestation of Neustria (as Neustrasii): Jonas of Bobbio, Vita Columbani abbatis, I.24, in Ionae vitae sanctorum Columbani, Vedastis, Iohannis, ed. Bruno Krusch, MGH. SS rer. Germ. 37 (Hannover: Hahn, 1905), 206–07: Moratus ergo ibi paululum, post ad Chlotharium, Chilerpici regis filium, qui Neustrasis Francis regnabat, extrema Gallia ad Oceanum positis, pergit. (So having tarried there for a little while, he then hurried to Chlothar, the son of king Chilperic, who ruled over the Neustrian Franks, located at the edge of Gaul next to the Ocean’). See also: Fred. Chron. IV.60, 150 (Neptreco); IV.62, 151, (Neptrasiis). This vocabulary takes on an explicitly political edge in, for ex- ample, the Liber Historiae Francorum, which distinguishes Franci from Austrasii. On the origins of the name: Ewig, ‘Die fränkischen Teilungen’, 151–60. 771 Not. Dig. Occ. XLII.23, 216: Praefectus classis Andertianorum, Parisius, XLII.66, 219: Praefectus Sarmatarum gentilium, a Chora Parisios usque; Julian at Paris: Amm. Marc., Res Gestae, XVII.8, 117. He was later acclaimed as augustus by the troops stationed there in 360: ibid., XX.4, 188–92. 772 Notitia Galliarum IV.7, in Chronica Minora, 1:556. Paris highlights the disjuncture between the secular and ecclesiastical geographies in the early Frankish kingdoms: the bishop of Paris had been a suffragan of the bishop of Sens until 1622. Brühl, Palatium und Civitas, 1:7. Cf. Meier, Völkerwanderung, 623. 773 The two poles of authority inside the walls, the cathedral and the palace, are dimly illuminated in Gregory’s Histories. E.g.: Greg. Tur., Hist., VII.4, 328: Interea Fredegundis regina iam viduata Parisius advenit et cum thesauris quos infra murorum septa concluserat, ad aeclesiam confugit adque a Ragnemodo fovetur episcopo. (‘Meanwhile the now-bereft Queen Fredegund came to Paris and with her treasures which she stored inside the circuit of the walls, she fled to the church and she received support from the bishop, Ragnemod’). Brühl, Palatium und Civitas, 1:21–22; Patrick Périn, ‘Paris mérovingien, sedes regia’, Klio 71, no. 2 (1989): 490–91; Dierkens and Périn, ‘Les sedes regiae mérovingiennes’, 286. 774 Evidence summarised in: Noël Duval et al., ‘Paris’, in Province écclesiastique de Sens (Lugdunensis Senonia), ed. Jean-Charles Picard, vol. 8 of Topographie chrétienne des cités de la Gaule des origines au milieu du VIIIe siècle, ed. Nancy Gauthier and Jean- Charles Picard (Paris: De Boccard, 1992), 107, 109–12. The monastery of Saint-Martial (later Saint-Éloi), founded by Eligius of Noyon in the seventh century, originally stood close to the palace on the present site of the Préfecture de Police. 209 Paris’ prominence as the undisputed centre of the Frankish regnum owes little to a direct Roman heritage as such, but to the evolving political geography of the kingdom from the time of Clovis onwards. Gregory reports that Clovis established it as his cathedra regni, a cryptic phrase that underlines Gregory’s understanding of Paris’ superiority to other royal residences in the kingdom, and later that the church of St Peter, commissioned by Clovis and his wife Clotild, would provide his final resting place.775 We get a sense of Paris’ significance throughout Gregory of Tours’ Histories. For example, in one episode, Gundovald, in conversation with Bishop Magnulf of Toulouse, states that he will march on Paris and establish it as his sedes and thereby assert his legitimate heirship to Chlothar’s royal power.776 So potent were these symbolic affordances that, according to Gregory, Charibert’s brothers had, on his death, ‘agreed in a pact that none of them should enter Paris without the consent of the others’, a ‘curse’ (maledictum) that was evidently threatening enough for Chilperic to send relics ahead before entering himself for the baptism of his son by the bishop, Ragnemod.777 The contemporary evidence is unanimous that Paris constituted the regnum’s symbolic heart, and during periods of unified rule after 613, it would be ruled from its environs. Paris’ symbolic importance has frequently been explained in terms of dialectic of ‘capital’ and ‘residence’. For Ewig, for example, Paris belonged in the same bracket as Visigothic Toledo and Pavia under the Langobards as the unification of the symbolic centre of the kingdom with its royal seat, thus reproducing an ancient tradition patterned on Constantinople or Ravenna as its proxy – in other words, an expression of imitatio imperii.778 Others are more definitive, arguing that, for example, the The land was granted to Eligius from the fisc (de regio munere/ex munere regis): Jonas of Bobbio, Vita Columbani, II.10, 255; Vita Eligii, I.17, 682; Duval et al., ‘Paris’, 114. 775 Cathedra regni: Greg. Tur., Hist., II.38, 89. St Peter/‘the basilica of the Holy Apostles’: ibid., II.43, 93. 776 Ibid., VII.27, 345–46. 777 Ibid., VI.27, 295: Et ut maledictum, quod in pactione sua vel fratrum suorum conscriptum erat, ut nullus eorum Parisius sine alterius voluntate ingrederetur, career possit, reliquias sanctorum multorum praecidentibus, urbem ingressus est. 778 Ewig, ‘Résidence et capitale’, 53, cf. Brühl, ‘Rémarques’, 200–01. Elsewhere Brühl raises the possibility that Paris was intended to recall Constantinople, but concludes we cannot be sure from the available evidence: Brühl, Palatium und Civitas, 1:7. 210 dedication of Clovis’ new church dedicated to the Holy Apostles was a deliberate reference to Con- stantine’s burial place in Constantinople.779 Yet this church was already known by its dedication to Peter by the 570s (rather than to all twelve Apostles together, as at Constantinople), and it seems likely that Gregory here commits a sleight of hand, eliding its dedication with the Constantinopolitan resting place of the first Christian emperor to elevate Clovis, the ‘new Constantine’ of his history.780 That is to say, the impression of imitatio imperii in Paris’ importance lies as much in Gregory’s authorial inten- tions as in Clovis’ actions. As Ewig has underlined, the choice of Paris as the centre of the united regnum in the early decades of the seventh century was primarily down to its central location in the geography of the Frankish kingdom at that time, offering a means of uniting Neustria with Austrasia and the rich lands of Burgundy, rather than any ‘Roman’ symbolic affordances.781 Paris’ evolution from a provincial military-administrative station to the symbolic centre of the Merovingian regnum underscores trends we witnessed earlier in Austrasia. While Constantinople pro- vided a host of possible images for the legitimation of political power in the former western provinces, these were of limited significance ‘on the ground’ in the localisation of political authority in the land- scape. Rather, Paris’ elevation highlights the renegotiation of existing political geographies to establish new central places and, with them, new landscapes of political action. A place in the country: the palaces of the Pagus Parisiacus In 613, the series of explosive conflicts that had wracked the Frankish regnum came to a bloody con- clusion with the successive deaths of the brothers Theudebert II of Austrasia and Theuderic II of Burgundy in the space of a year. Resentful of the continued influence of the royal grandmother, 779 E.g. Dierkens and Périn, ‘Les sedes regiae mérovingiennes’, 276; Meier, Völkerwanderung, 646–47. 780 Gregory nowhere else refers to the church as the ‘Holy Apostles’, only by its dedication to Peter. Greg. Tur., Hist., V.18, 217: apud Parisius in basilica sancti Petri apostoli. One of the two MSS of the acta of the 614 Parisian council state that it took place in basilica domni Petri: ‘Concilium Parisiense (614), in Concilia aevi Merovingici, 185. For Clovis as a ‘new Constan- tine’: Greg. Tur., Hist., II.31, 77: Procedit novos Constantinus ad lavacrum. 781 Ewig, ‘Die fränkischen Teilreiche’, 175. 211 Brunechild, the Austrasian and Burgundian aristocracy pledged their allegiance to the last king stand- ing, Chlothar II of Neustria. The elderly Brunechild, having lost the support of the Austrasian elite, was brutally executed; Chlothar ascended to the throne of the unified regnum.782 While Chlothar’s reign heralded (in the eyes of the ‘Fredegar’ chronicler at least) a welcome period of peace and prosperity, it also marked a watershed in the formation of the Frankish political landscape.783 The Neustrian king would rule the unified regnum from Paris, yet he was seldom to be found within its walls. Instead, a series of villae in its environs would serve as the regnum’s centre in both a social and geographic sense for much of the seventh century. This preference for the villas outside Paris highlights a revolutionary shift in spatial patterns by which authority was represented in this period: the palace had now finally achieved ‘spatial autonomy’ from the city.784 Between Chlothar’s accession to the thrones of Austrasia and Burgundy in 613 and the death of his son Dagobert in 637, the regnum was essentially ruled from the villas of the Paris basin.785 A number of estates within this landscape are attested in written sources, including Chelles, Rueil, Bonneuil, and Épinay-sur-Seine, but Clichy, some six kilometres outside Paris’ walls, was pre-eminent among them (Fig. 134).786 Its centrality is discernible in the convocation of assemblies of magnates, both ecclesiastical and lay. In October 614, Chlothar brought together lay magnates and bishops from across the Frankish hegemony (including as far afield as Kent) to legislate on general matters of secular 782 Fred., Chron., IV.42, 141–42; LHF, c. 40, 310–11. Brunechild’s treatment reflects both her long success in exercising independent political agency in the Frankish regnum, but also the misogynistic reaction of aristocratic society when her position became untenable, reflected too in the sources who recorded it. On Brunechild’s life and the position of royal women in the Frankish kingdoms, see: Janet Nelson, ‘Queens as Jezebels: Brunhild and Balthild in Merovingian History’, in Janet Nelson, Politics and Ritual in Early Medieval Europe (London: Hambledon, 1986), 1–48. 783 Fred., Chron., IV.42, 142. 784 Barbier, ‘Le système palatial franc’, 268. 785 Clovis II would likewise rule both Neustria and Burgundy in the same fashion until his death in 657. 786 Chelles: Greg. Tur., Hist., V.39, 245; VI.46, 319; VII.4, 328. The Queen Balthild founded a convent there in the mid- seventh century: Vita Balthildis, cc. 7, 18, 489, 506. Rueil: Greg. Tur., Hist., X.28, 521; Vita Eligii, I.6, 678. Bonneuil: Fred., Chron., IV.44, 143. Épinay-sur-Seine: Fred., Chron., IV.79, 161; LHF, c. 43, 315. Generally: Ewig, ‘Die fränkischen Teil- reiche’, 180–81. 212 and ecclesiastical concern.787 They met in the church of St Peter, as we have seen a location rich in early Merovingian historical memory, probably in an adjoining audience chamber (secretarium).788 Fur- ther councils of the Frankish episcopate summoned by Chlothar in 626/627 and by his son, Dagobert, in 636 did not meet in Paris, however, but at the extra-mural church of Saint-Denis.789 As we saw in Chapter 1, the establishment of Saint-Denis as a royal church heralds a new development in the lan- guages of royal authority, one which reflects both a growing liturgification of political action and a greater stress being placed on the authority of the king as an agent in religious affairs. Thus, the pro- logue to the acts of the 626/627 council states that it was not only convoked at the king’s instruction, but also that it took place in a Marian chapel located in the atrium of Saint-Denis, ‘joined to the manor named Clichy’.790 We see this also in the social make-up of the court. The social milieu of the Frankish palace had significant overlap with the episcopate in the seventh century, a trend exhibited in the network of prominent bishops educated and active at court (such as Audoin of Rouen, Eligius of Noyon, or Aunemund of Lyon), as well as the numerous Vitae written for contemporary religious leaders which detail their roles in factional conflicts.791 Clichy—and the network of villas in the Paris 787 Capitularia regum Francorum, ed. Alfred Boretius, vol. 1, MGH. Capit. 1 (Hannover: Hahn, 1883), no. 9, 20–23; Concilium Parisiense (614), 185–92. Justus, the bishop of Rochester, is a signatory of the acts, as is Peter, abbot of SS. Peter and Paul, Canterbury: ibid., 192: Ex civitate Castro ultra mare Iustus episcopus. […] Peter abba de Dorouerno. 788 In addition to Clovis and Clothild, Saint Genovefa would later also be interred there. Gregory of Tours, Hist., II.42, 93; IV.1, 135. Vita Genovefae virginis Parisiensis, c. 56, in Passiones vitaeque sanctorum aevi Merovingici et antiquiorum aliquot, ed. Bruno Krusch, MGH SS rer. Merov. 3 (Hannover: Hahn, 1896), 237–38. The bishops of the 577 council of Paris convened in the secretarium: Greg. Tur., Hist., V.18, 217. 789 Concilium Clippiacense (626/627), in Concilia aevi Merovingici, 196–201. The acta of the 636 council are not preserved, but the council is referenced in the Vita of Agilus, the first abbot of the monastery founded by Audoin in the same year at Rebais: Vita Agili abbatis Resbacensis, ed. Johannes Stilting, in Acta Sanctorum, August vol 6, edited by Joannes Pinio et al. (Paris: Victor Palmé, 1868), 584: Proinde in episcopali synodo, quae Calendis Maii in Clypiaco eo anno est habita, B. Agilum praecelentis- simus rex in memorato coenobio praefecit abbatem (‘Consequently at the episcopal council, which took place on the Kalends of May that year in Clichy, the most distinguished king appointed the blessed Agilus abbot of the renowned monastery’). 790 Concilium Clippiacense (626/627), 196: iuxta predium, quod Clipiaco dicitur. Dagobert issued a charter granting rights of asylum to Saint-Denis shortly after this council at Clichy, in which it is referred to as palacio nostro: DD Merov. 1, no. 29, 79. 791 Paul Fouracre, ‘Francia in the Seventh Century’, in c. 500–c. 700, ed. Paul Fouracre, vol. 1 of The New Cambridge Medieval History (Cambridge: Cambridge University Press, 2005), 383–84; idem, ‘Merovingian History and Merovingian Hagiog- raphy’, Past & Present 127 (1990): 3–38. Audoin (Saint Ouen) died at Clichy in 686 while on royal business: Liber Historiae Franorum, c. 47 321–22; Vita Audoini episcopi Rotomagensis, c. 15, in Passiones vitaeque sanctorum aevi Merovingici, ed. Bruno Krusch and Wilhelm Levison, vol. 3 MGH. SS rer. Merov. 5 (Hannover: Hahn, 1910), 563–64. According to his Vita, Audoin served as refendarius under Dagobert: Ibid., c. 2, 555. Only a few authentic charters subscribed by Audoin survive 213 basin more generally—was thus located at the centre of social networks that spanned the Merovingian hegemony across both ‘secular’ and ‘religious’ spheres. Clichy’s pre-eminence in this constellation is discernible in other contexts. In 625, Dagobert’s marriage to Gomentrude, the sister of Chlothar’s wife Sichilde was celebrated at Clichy, presumably in Saint-Denis – an important royal event that must have re-emphasised the young Dagobert’s subor- dination to his father to those present.792 That the extent of Dagobert’s independence was at issue is hinted at by the fact that, according to ‘Fredegar’, a disagreement broke out over Clothar having re- tained Austrasian lands in the Ardennes and Vosges lying on the Neustrian border; they were restored after a meeting of lay and secular magnates, probably indicating a formal placitum.793 Two years later, Chlothar again held an assembly of the military and ecclesiastical magnates of Neustria and Burgundy at Clichy, apparently to legislate on matters concerning the relationship of those two kingdoms.794 Likewise, during Dagobert’s reign over the regnum as a whole, the Gascon nobility travelled to Clichy in 636 to give their oaths of allegiance to the Frankish king after a punitive campaign.795 The meetings of various sections of the aristocracy at Clichy in all these cases established it as the regnum’s effective centre. The ascendancy of Clichy illuminates a radical break in the apprehension of political power from the Late Roman Empire. As we saw in Chapter 2, while in the Late Empire, political power was legitimised—however theoretically—principally before an urban citizenry in ar- chitectonic settings designed and curated for this purpose, from the sixth century political ritual was increasingly transferred into the palace’s interior among a smaller clique of elites. The villas of the (e.g. DD Merov. 1, no. 41, 74). Eligius was likewise prominent at the court of Dagobert (e.g. Vita Audoini, c. 4, 556), and they both belonged to a network of aristocratic bishops connected to the court with sees located throughout the kingdom. See generally: Gregory Halfond, Bishops and the Politics of Patronage in Merovingian Gaul (Ithaca, NY: Cornell University Press, 2019). 792 Fred., Chron. IV.53, 147–48. 793 These lands were retained by Chlothar after he seated Dagobert in Austrasia: Fred., Chron., IV.47, 144: retinens sibi quod Ardinna et Vosaco versus Neuster et Burgundia excludebant. The resolution of the argument: ibid., IV.53, 147. 794 Fred., Chron., IV.55, 148. 795 Ibid., IV.78, 159–60. 214 pagus Parisiacus, therefore, highlight a broader transition in concepts of political legitimacy and the spatial distribution of political power. While cities remained socially and economically important cen- tres, not least due to the prominence of bishops as political actors, the ‘public’ before whom political authority was negotiated had been redefined. Regional Politics and Political Integration The territorial cementation of the Teilreiche by the seventh century led to a political system in which (theoretically, at least) access to power and patronage was distributed between regional centres over which the Merovingian king ruled by consensus of the aristocracy.796 Interruptions in the access to that power and patronage provoked the unrest and violence, the ‘tyranny’ and subsequent overthrow of the mayor of the palace Ebroin in 673 being a prime example.797 Practically, this led to the Teilreiche becoming the primary frames of reference for aristocratic identity and political allegiance. Hence after 613 Chlothar, though established in the Neustrian royal landscape in and around Paris, ruled in Bur- gundy and Austrasia through their respective courts, represented by the powerful mayors of the pal- aces Warnachar and Rado.798 The centralisation of the regnum in Neustria highlights not only the spatial transformation of the royal sedes, but also of the regnum as a whole. That is to say, the establishment of the pagus Parisiacus as the centre of the unified regnum also highlights palaces’ role as engines of spatial transformation and political integration on a larger scale. As we saw in Chapter 2, aristocratic society was by the seventh century dependent on the political centre, the palace and court, as it supplied the 796 Paul Fouracre, ‘The Nature of Frankish Political Institutions in the Seventh Century’, in Franks and Alamanni in the Merovingian Period: An Ethnographic Perspective, ed. Ian Wood, Studies in Historical Archaeoethnology 3 (Rochester, NY: Boydell Press, 1999), 285–316, esp. 295–96. 797 E.g. Passio Leudegarii I, cc. 4–6, 286–89, in which Ebroin forbids the Burgundian aristocracy from going to the palace and unlawfully confiscates property, raises Theuderic III without calling a general assembly (contrary to custom), and is driven out. 798 Fred., Chron., IV.42, 142: Warnacharius in regnum Burgundiae substituetur maior domi, sacramentum a Choltarhium acceptum, ne umquam vitae suae temporebus degradaretur. In Auster Rado idemque hoc gradum honoris adsumpsit. (‘Warnachar served as mayor of the palace in the kingdom of Burgundy after Chlothar made an oath that he would not remove him during his lifetime. In Austrasia Rado assumed the same office’). 215 social arena for patronage and intra-elite competition.799 Here, we witness the effects of this process in the ordering of political space: the role of the palace in anchoring aristocratic social networks and thereby creating both intensive and extensive relations of power.800 An aspect of this tension can be glimpsed in the ‘Edict of Paris’ of 614. Issued by Chlothar II following the meeting of ecclesiastical and lay magnates in October of that year, the Edict of Paris seeks to correct abuses in the kingdom’s administration—‘with Christ as our protector, through the content of our edict we have arranged to generally correct what has been done and instituted contrary to reason, lest it should happen in the future that God turns from us’—as well as problems that had arisen from the kingdom’s divisions.801 Thus in addition to, for example, legislating on episcopal suc- cession (an issue of some consequence, given the political power wielded by bishops), the edict also instructs that royal officers (iudices) be chosen only from within the region of their jurisdiction, and that land lost by any of the king’s leudes in a division or transfer or royal territory was to be reinstated, among other issues.802 In earlier scholarship these provisions were seen as concessions of political autonomy to the regional aristocracies of Austrasia and Burgundy in exchange for their co-operation and support – a kind of Merovingian Magna Carta.803 There is no evidence that the Edict caused a serious diminution of authority at the political centre, however. As Alexander Callander Murray has pointed out, for example, even if Chlothar proclaimed that iudices were to be appointed to serve in their region of 799 Wickham, Framing, 154. On Gaul/Francia specifically: ibid., 169–203. For the ritual components to this ‘competitive’ culture, see above, Chapter 2. 800 See especially the discussion of gift-giving: above, Chapter 2. For intensive/extensive relations of power: Mann, Sources, 1:7–8. 801 Capit. 1, no. 9, 20: et quod contra rationis ordinem acta vel ordinata sunt, ne inantea, quod avertat divinitas, contingat, disposuimus Christo praesole per huius edicti nostri tenorem generaliter emendare. 802 Ibid., cc. 1–2, 12, 17, pp. 21–23. 803 Explicitly: Heinrich Mitteis, Der Staat des hohen Mittelalters: Grundlinien einer vergleichenden Verfassungsgeschichte des Lehnzeital- ters, 4th ed. (Weimar: Hermann Böhlaus Nachfolger, 1953 [first pub. 1940]), 52. That the Edict of Paris reflected the victory of an autogenous ‘Germanic’ nobility over the Merovingian monarchy was a prevalent view in postwar German scholar- ship. For a critical reassessment: Alexander Callander Murray, ‘Immunity, Nobility, and the Edict of Paris’, Speculum 69, no. 1 (1994): 18–39. 216 origin, that appointment was still made by the king and could still benefit members of palace society.804 What we can say is that the 614 Edict was part of a broader effort to determine the formal relationships between the royal centre and regional elites under a new political arrangement.805 A year earlier Chlo- thar had convened the Burgundian aristocracy at Bonneuil, where ‘hearing all their just petitions, he confirmed them with rescripts’.806 The Edict therefore speaks not of a constitutional struggle between the king and the aristocracy, but rather the inherent tensions that existed between the magnetic pull of the palace as a political centre and regional aristocratic interest. The ‘Fredegar’ chronicler’s treatment of Dagobert’s reign is revealing in this respect. After Warnachar’s death in 626, the Burgundian magnates chose at an assembly at Troyes that they not be represented by another Mayor of the Palace and that instead the king deal with them directly.807 In 629/630 Dagobert travelled to Burgundy to give judgments in various locales, yet ‘Fredegar’ com- ments that this caused ‘great fear’ among the Burgundian aristocracy – a curious comment given their request of only a few years earlier.808 One suspects that Dagobert’s Burgundian tour was motivated by abuses which had taken hold in the absence of a nearby royal court. In the next year, Dagobert also ‘went around Austrasia in the royal fashion’, perhaps suggesting a similar purpose.809 The chronicler contrasts this mobility with Dagobert’s later conduct while he was settled at Clichy, when he ‘forgot the justice that he had once loved. He longed for ecclesiastical property and for the goods of his subjects and he greedily sought by every means to amass fresh treasure.’810 804 Capit. 1, no. 9, 22: Et nullus iudex de aliis provinciis aut regionibus in alia loca ordinetur ; ut, si aliquid mali de quibuslibet condicionibus perpetraverit, de suis propriis rebus exinde quod male abstolerit iuxta legis ordine debeat restaurare; Callander Murray, ‘Immunity’, 27. 805 Fouracre, ‘Francia in the Seventh Century’, 374–75. 806 Fred., Chron., IV.44, 143: Anno 33. Regni Chlothariae Warnacharium maioris domus cum universis ponificibus Burgundiae seo et Burgundaefaronis Bonogillo villa ad se venire precepit; ibique cunctis illorum iustis peticionibus annuens, preceptionebus roboravit. 807 Fred., Chron., IV.54, 148: Eo anno Chlotharius cum procerebus et leudibus Burgundiae Trecassis coningetur […] omnes unianimiter delignates si, nequicquam si velle maiorem domus elegere, regis graciam obnexe petentes cum rege transagere. 808 Namely Langres, Saint-Jean-de-Losne, and Chalon before returning to Paris via Auxerre, Autun, and Sens: Fred., Chron., IV.58, 149–50. ‘Great fear’: ibid., 149: Tanta timore ponteficibus et procerebus in regnum Burgundiae consistentibus seo et citeris leudibus adventus Dagoberti concusserat, ut a cunctis esset mirandum. 809 Fred., Chron., IV.59, 150: Anno 8. regni sui cum Auster regio cultu circuerit 810 Ibid.., IV.60 150–51: cum omnem iustitiam quem prius dilixerat fuisset oblitus, cupiditates instincto super rebus ecclesiarum et leudibus sagace desiderio uellit omnibus undique expoliis nouos implere thinsauros. Trans.: Wallace-Hadrill, 50. 217 Dagobert’s lengthier journeys into the Teilreiche were not usual to seventh-century ruling prac- tices, but exceptional events responding to specific social tensions, underlining that direct royal au- thority in the seventh century was still circumscribed within relatively small landscapes. We witness this same phenomenon in 632, when he was compelled to establish his son Sigibert III (d. ca. 656) as king in Austrasia in response to Wendish incursions.811 Sigibert’s was not to be simply a puppet regime of his father in Neustria, however. In a ceremony at Metz the new king’s court was provided with a hierarchy of courtiers (headed by Bishop Chunibert of Cologne and the dux Adalgisel) as well as a sufficient treasury to fund it, gifts legally confirmed with charters.812 There was precedent for this. Dagobert had himself been installed in Austrasia in 622/623, though the motivations behind this are not explained by ‘Fredegar’.813 The presence of the court nearby apparently strengthened the Austra- sians’ resolve against their assailants.814 Two explanations suggest themselves: that the proximity of the court aided in military organisation—possible, but why would this require specifically a separate court hierarchy?—or that the establishment of satellite courts was the result of internal political manoeuvring. After Dagobert’s death, Sigibert enjoyed the continued service of Chunibert of Cologne along with the Mayor of the Palace Pepin of Landen who, according to ‘Fredegar’, ‘drew the Austrasian notables into their orbit, ruled them generously, won their support and knew how to keep it’, an arresting remark that might suggest continued restiveness on the part of the Austrasian elite during 811 Ibid.., IV.75, 158–59. 812 Ibid.: Dagobertus Mettis orbem ueniens, cum consilio ponteuecum seo et procerum, omnesque primatis regni sui consencientebus, Sigybertum filium suum in Auster regem sublimauit sedemque ei Mettis ciuitatem habere permisit. Chunibertum Coloniae urbis ponteuecum et Adalgyselum ducem palacium et regnum gobernandum instetuit. Tinsaurum quid suffecerit filium tradens, condigne ut decuit eum uius culmine sublimauit ; seo quodcumque eidem largitus fuerat sigillatem praeceptionebus roborandum decreuit The Liber Historiae Francorum states that Pepin of Landen was installed to ‘direct’ Sigibert’s court: Liber Historiae Francorum, c. 43, 315: Sighibertum vero, maiorem filium suum, in Auster una cum Pippino duce direxit in regno statuto. 813 Fred., Chron., IV.47. In the Liber Historiae Francorum, Chlothar sent Dagobert, who was then acclaimed by the assembled Austrasian elite, underscoring not only the consensus-based nature of Merovingian politics, but also that Dagobert’s ele- vation was the result of exchanges between Chlothar and the Austrasians: Liber Historiae Francorum, c. 41, 311: Quem rex adultum una cum Pippino duce in Auster regnaturum direxit. Austrasii vero Franci superiors congregate in unum, Dagobertum super se regem statuunt. 814 Fred. Chron., IV.75, 159. 218 the young king’s minority.815 We are reminded here also of Gregory of Tours’ account of the leading citizens of Soissons and Meaux petitioning Childebert II in 589 to dispatch one of his sons to establish a satellite court, arguing that ‘If we have a member of your family resident among us as a pledge, we shall resist your enemies with all the greater zeal and have more reason to protect the lands round your city’.816 It would seem that the establishment of a satellite court with its own opportunities for financial advancement was a means of placating the Austrasian nobility, resentful at being ruled from Neustria. Josiane Barbier has similarly suggested that the favour shown to Compiègne in the latter half of the century at Clichy’s expense, as well as the foundation of new palaces at Crécy-en-Ponthieu and possibly Noyon, similarly points at efforts to maintain cohesion in the face of friction between Neus- tria and Austrasia by shifting the kingdom’s centre of gravity eastwards.817 In all these examples, we see the role of palaces as engines of aristocratic integration into the political system. Without access to the power and patronage of the political centre, an aristocratic culture that revolved around competing for wealth and status at court could be put into crisis. What is more, we see in the case of Dagobert that kings leaving their immediate palatial landscape was a rare and exceptional event by the seventh century, limited to occasions of significant social tension.818 In general, the Merovingian kings ruled through consensus between centre and locality and by delega- tion.819 We should here bear in mind Stuart Airlie’s argument that even the empty palace—one that seldom received the king personally, or staffed only by subordinates—still served as a persistent re- minder of royal power in the landscape and created economic, social, and imaginative ties between 815 Fred., Chron., IV.85, 164 : omnesque leudis Austrasiorum secum uterque prudenter et cum dulcedene adtragentes, eos benigne gobernantes, eorum amiciciam constringent semperque servandum. Trans. Wallace-Hadrill, 72. Sigibert was born in 629/630 (Fred., Chron., IV.59, 150), making him at most ten years old on his father’s death. 816 Greg. Tur., Hist., IX.36, 457: Da nobis unum de filiis tuis, ut serviamus ei, scilicet ut de progenie tua pignus retenentes nobiscum, facilius resistentes inimicis, terminus urbis tuae defensare studeamus. Trans. Thorpe, 524. 817 Barbier, ‘Le système palatial franc’, 269–76. The principal zones of royal activity after ca. 650 were in the Oise and Aisne regions – to the east and northeast of Paris. 818 Barbier, ‘Les rois mérovingiens’, 195–96. 819 Ibid., 194. See for example the formulae for the appointment of a comes, dux, or patricius in the Formulary of Marculf: Marculfi Formulae, I.8, 47–48. 219 regional elites and central authority.820 It was the double exchange of people, objects, and ideas that took place at the palace, not the court’s mobility, that created extensive networks of power and in- scribed royal authority into the landscape. *** Although Clichy stood at the kingdom’s centre for a relatively short period, the prominence accorded to it in the seventh century illustrates three significant and interrelated developments. Firstly, the pref- erence shown to the palaces in the pagus Parisiacus exhibits the final transference of royal authority out of the city and into rural fiscal estates. Secondly, the close relationship between Clichy and Saint-Denis highlights the elaboration of new spatial patterns for the presentation of royal authority which re- sponded to new ideals for its legitimate practice. Although precisely the same spatial arrangement was not repeated elsewhere, that Saint-Denis functioned as the centre of gravity of the Parisian palatial landscape reflects a change in how political authority was expressed in this period. But the hints of an increasingly ‘sacral’ accoutrements of Merovingian in the sources preserve only one side of the story. Thus, thirdly, the reigns of Chlothar and Dagobert also demonstrate the interplay of two superim- posed landscapes: the regional royal landscapes of the Teilreiche and the supra-regional landscape of the regnum. Tensions caused by the concentration of a singular royal presence in Neustria underscore that, at this time, the governmental superstructure of Francia at this time was not (yet) sufficiently intensive to exercise authority over the combined area of all three Teilreiche from a single centre. IV. Forging a New Empire: Carolingian Royal Landscapes In both Moduin of Autun’s Ecloga and the anonymous epic known as Karolus magnus et Leo papa, the Carolingian palace was a new Rome, a restoration of ancient glory under the government of 820 Stuart Airlie, ‘The Palace of Memory: The Carolingian Court as Political Centre’, in Courts and Regions in Medieval Europe, ed. Sarah Rees Jones et al. (Woodbridge: York Medieval Press, 2000), 10–12. Though writing with reference to the Caro- lingian period, Airlie’s point is more generally applicable. 220 Charlemagne and his son, Louis the Pious.821 Yet while these authors called back to Antiquity as the lodestar of the court’s ambitious programme of moral and cultural reform, the ascension of the Car- olingians is simultaneously, in both its contemporary rhetorical and modern scholarly presentation, a break with the immediate past.822 In much previous scholarship the Carolingian correctio was bound up with sweeping changes in the structures of authority across Europe, not only in terms of a new rhetoric of Christian kingship, but more generally in the simultaneous extension and intensification of rule from a single, stable centre – a new imperial ‘capital’ at Aachen.823 ‘Rome’ was, in this sense, both cultural-religious aspiration and political claim. Certainly, the texts that issued from the palace society at Aachen often present it as the kingdom’s unrivalled centre. But the singularity of scholarly attention on Aachen has flattened out temporal and geographical distinctions in the use of the extensive palatial network that the Carolingians inherited.824 The political landscape of the Carolingian polity was not static, nor was it radically disconnected from that of previous centuries. Rather, it reflects the elabo- ration of previous practices and experimentation in the context of a new and evolving set of political circumstances. 821 Moduin, Ecloga, I.24–26 in Poetae latini aevi Carolini, ed. Ernst Dümmler, vol. 1, MGH. Poetae 1 (Berlin: Weidmann, 1881), 385: Prospicit alta novae Romae meus arce Palemon | Cuncta suo impero consistere regna triumpho | Rursus in antiquos mutataque secula mores. | Aurea Roma iterum renovate renascitur orbi! (Trans. Godman, 193: ‘My Palaemon looks out from the lofty citadel of the new Rome and sees all the kingdoms forged into an empire through his victories. Our times are transformed into the civilisation of Antiquity. Golden Rome is reborn and restored anew to the world!’). Karolus magnus et Leo papa, ll. 94– 96, 66: Augustus, sed et urbe potens, ubi Roma secunda | Flore novo, ingenti, magna consurgit ad alta, | Mole, tholis muro praecelsis sidera tangens. (Trans. Godman, 203: ‘emperor, and lord too of the city where a second Rome flowers anew, its mighty mass rising up to the great heights, the lofty cupolas on its walls touching the sky.’) The poet continues with a description of the new palace under construction – probably that at Paderborn, the site of Leo and Charles’ meeting, rather than Aachen, as is often supposed: Karolus magnus et Leo papa, ll. 97–136, 66–68. 822 On the memory of Rome in the Early Middle Ages, see: Smith, Europe after Rome, 253–92. Einhard performs a similar manoeuvre through his combination of Roman literary models with an insistence on Charlemagne’s adherence to ‘Frank- ish’ customs and norms of consent: David Ganz, ‘Einhard’s Charlemagne: The Characterisation of Greatness’, in Charle- magne: Empire and Society, ed. Joanna Story (Manchester: Manchester University Press, 2005), 38–51. 823 For Aachen as ‘capital’, e.g.: François Louis Ganshof, ‘Charlemagne’s Programme of Imperial Government’, in The Carolingians and the Frankish Monarchy: Studies in Carolingian History, trans. Janet Sondheimer (London: Longman, 1971), 55– 85; Janet Nelson, ‘Was Charlemagne’s Court a Courtly Society?’, in Court Culture in the Early Middle Ages: The Proceedings of the First Alcuin Conference, ed. Catherine Cubitt, Studies in the Early Middle Ages 3 (Turnhout: Brepols, 2003), 39–57. 824 Jennifer Davis’ recent work is a useful step in this direction, elucidating the ‘patterns of rulership’ in Carolingian sources from an empirical perspective: Davis, Charlemagne’s Practice of Empire. 221 Carolingian Royal Mobility In the year 790, the Annales regni Francorum (‘Annals of the Kingdom of the Franks’, henceforth ARF) record that Charlemagne took a break from campaigning. For the annals’ later reviser, however, this inactivity required further explanation. ‘So it would not seem that he had grown idle and soft with leisure,’ the new, longer entry continues, ‘the king sailed on the River Main to his palace at Salz, which he had constructed on the River Saale in Germany. From Salz he returned again downstream on the same river to Worms’, where he celebrated Christmas and lodged until the following campaigning season.825 The equation the reviser draws here between the king’s mobility between palaces, his mas- culine identity as a warrior-king, and thereby his authority—it would be politically damaging, we are led to believe, that Charlemagne be perceived as ‘idle’—is a connection that we have not seen in such stark terms in earlier periods.826 Einhard’s ridicule of the last Merovingian king, Childebert III, was predicated on the Merovingian king’s inactivity, his humble mode of transport contrasting with the spectacle of the palace and assembly at which he was, essentially, wheeled out: ‘When he had to travel about, he went in a cart pulled by yoked oxen and led by a plowman in the country fashion. Thus he used to go to the palace or to the public meeting of his people that was held annually for the benefit of the kingdom.’827 When we contrast the ARF or Einhard with the sources of the seventh century, it is hard to ignore the impression of a sea change in the structures of political authority. Mobility, many scholars 825 ARF, a. 791(R), 87: Rex autem, ne quasi per otium torpere ac tempus terere videretur, per Moenum fluvium ad Saltz palatium suum in Germania iuxta Salam fluvium constructum navigavit atque inde iterum per eundem amnem secunda aqua Wormiaciam reversus est. Trans. Scholz, 69. 826 On the dating and composition of the ARF, see: McKitterick, Charlemagne, 31–43. The revised version reworks the ARF’s original entries for 741–801 and continues the text to 829, where it breaks off. There has been considerable debate about the date of the reviser’s work, with many favouring the early years of Louis the Pious’ reign, others ca. 829. For a reassessment, see: Roger Collins, ‘The “Reviser” Revisited: Another Look at the Alternative Version of the Annales regni francorum’, in After Rome’s Fall: Narrators and Sources of Early Medieval History. Essays Presented to Walter Goffart, ed. Alexander Callander Murray (Toronto: University of Toronto Press, 1998), 191–213; McKitterick, Charlemagne, 27–39. It is worth noting that this entry occurs between the revolts of Hardrad (785/86) and Pippin the Hunchback (792), evidently a time of political turmoil, which may give some credence to the anxieties over appearing ‘idle and soft’. 827 Einhard, Vita Karoli imperatoris, c. 1, 3: Quocumque eundum erat, carpento ibat, quod bubus iunctis et bubulco rustic more agente trahebatur. Sic ad palatium, sic ad publicum populi sui conventum, qui annuatim ob regni utilitatem celebratur, ire, sic domum redire solebat. 222 have argued, was an essential strategy of ruling the rapidly expanding Frankish hegemony after the rise of the Carolingians. This is, in part a trick of the eye, however: a product of a vastly different basis of sources (particularly annalistic texts and charters) that are simply unavailable for earlier periods. Before turning to the specific royal landscapes of the Carolingian hegemony and their fluctuation over time, therefore, it is necessary to tackle this question head-on. Did the rise of the Carolingians bring about a transition to itinerant rulership? Itinerancy has been a significant force in Carolingian historiography. Charlemagne’s long, forty-eight-year reign has traditionally been split into two distinct periods: one military protagonism, in which the king led the army on campaign in person in a series of campaigns which rapidly expanded Francia’s borders, and one of consolidation, in which the previously active king ‘settled’ at Aachen.828 This picture essentially reproduces the structure of Einhard’s Vita Karoli imperatoris in which the mili- tary exploits of the first half are set alongside a portrait of his personal habits and private life – his vita et conversatio as well as his res gestae.829 François Louis Ganshof distinguished between early expansionist activity and late ‘imperial’ stability, yet was dismissive of the lasting effect the Carolingians had on institutions of government which, in his eyes, only began to show signs of growing complexity in the twelfth century and later.830 For both Ewig and Brühl, the Carolingian period ushered in a transfor- mation of earlier patterns of authority, both marking Charlemagne’s reign as the final departure from city-based ancient practice.831 Yet Brühl elsewhere avers that Pippin’s ascent to the throne in 751 changed earlier practices little and only brought the legal character of royal itinerancy more clearly to 828 E.g. Ewig, ‘Résidence et capitale’, 61; François Louis Ganshof, ‘The Institutional Framework of the Frankish Monarchy: A Survey of its General Characteristics’, in The Carolingians and the Frankish Monarchy, 90; Janet Nelson, ‘Aachen as a Place of Power’, in Topographies of Power, 219; Donald Bullough, ‘Aula Renovata: The Carolingian Court before the Aachen Palace’, in Carolingian Renewal, 142; Ludwig Falkenstein, ‘Charlemagne et Aix-la-Chapelle’, Byzantion 61, no. 1 (1991): 231–32, 235– 36, 275–82; Johannes Fried, Charlemagne, trans. Peter Lewis (Cambridge, MA: Harvard University Press, 2016), 327–31. 829 Einhard, Vita Karoli imperatoris, prologue, 1: Vitam et conversationem et ex parte non modica res gestas domini et nutritoris mei Karoli; Ganz, ‘Einhard’s Charlemagne’, 41. 830 Stated most explicitly in: François Louis Ganshof, ‘The Last Period of Charlemagne’s reign: A Study in Decomposition’, and ‘Charlemagne’s Failure’, in The Carolingians and the Frankish Monarchy, 248–50, 259. 831 Ewig, ‘Résidence et capitale, 71; Brühl, ‘Remarques’, 207. 223 the fore.832 For him, Charlemagne and Louis the Pious had possessed similar dreams to their Mero- vingian predecessors of basing themselves in a stable ‘capital’, though he clarifies that ‘favourite resi- dence’ (Lieblingsresidenz) would be more precise.833 Adolf Gauert likewise asserted that an examination of Charlemagne’s royal itinerary revealed the overall structure of the Carolingian polity in theory and practice. In his eyes, the regularity of Charlemagne’s wintering at Aachen from the 790s onwards indicates, for example, that Aachen had been designated the principal royal seat (though not, he says, a ‘capital’ as such), part of a royal landscape in which the king’s presence ‘made the reality of the empire visible’.834 Given the undercurrent that a transition to itinerancy represented the final disinte- gration of the ancient past, we would do well to treat these claims cautiously. In any case, positing a stark binary opposition between periods of ‘itinerancy’ and ‘stability’ is unhelpful in the face of the far more equivocal evidence. Rather, a re-examination of the mobility of the Carolingian kings reveals the spatial forces by which the Carolingian polity was constituted and underlines that intensifications in political power necessarily re-ordered spaces of political action in the regnum. The ARF took an uncommon interest in royal movements. Most entries are structured around a basic set of information, occasionally augmented with additional details: the destination of the army’s campaign, where the annual Spring assembly was held, and where the king celebrated Easter and Christmas.835 Therefore, while inevitably only a snapshot of a broader range of activities and ceremo- nies, the Frankish royal annals do indicate the strategic spatial choices made by Charlemagne and Louis the Pious for the observance of the most important occasions of the royal calendar, but also the 832 Brühl, Fodrum Gistum, Servitium Regis, 18. 833 Ibid., 21–22. 834 Adolf Gauert, ‘Zum Itinerar Karls des Großen’, in Persönlichkeit und Geschichte, ed. Helmut Beumann, vol. 1 of Karl der Große. Lebenswerk und Nachleben, 321. 835 This basic structure breaks down in Louis’ reign. The entries become longer with a more detailed narrative but cease to consistently record the locations of Easter and Christmas. On changes in vocabulary and style as possibly indicative of changing authorship: McKitterick, Charlemagne, 39–43. The ARF’s connection to the palace is clear, but their presentation as a singular, ‘official’ version of events, added to over time, is misleading. See e.g.: Davis, Charlemagne’s Practice of Empire, 179–98; Collins, ‘The “Reviser” Revisited’. 224 reception of these choices by the elite: the events and locations that were chosen to be recorded and remembered. Charlemagne and Louis the Pious’ movements between palaces, primarily as presented in the ARF and its continuations, are tabulated in Appendix I (see also Fig. 135). If we exclude Char- lemagne’s expeditions outside the regnum’s borders, we are presented with a much more focused sense of royal movements.836 Four interlocking zones located principally in connection with the Rhine are immediately discernible in Charlemagne’s reign: the regions of the Meuse and Lower Rhine; around the confluence of the Rhine and Main; the lands of the Moselle and Middle Rhine; and finally the Neustrian lands between the Aisne and Oise. That Charlemagne’s (and, after him, Louis’) regular presence at Aachen coincided with efforts towards more intensive governance—a topic to which we will return—is not unimportant, but a singular focus on Aachen flattens any sense of chronological development as well as obscuring the importance of other spatial relationships in the performance and enaction of Carolingian authority in this period. The Carolingians certainly were mobile, regularly visiting a half-dozen or more palaces in a given year, but this cannot be characterised as ‘itinerant kingship’ in any real sense.837 Assemblies. Court-adjacent sources, from the ARF and its successors to Thegan and the Astronomer regularly record the location of the annual assembly. Hincmar establishes their im- portance to the Carolingian polity. According to him, the general assembly, determined the status of the entire realm for the remainder of the year. No turn of events, saving only the greatest crisis which struck the entire realm at once, could change what had been established. All the important men, both clerics and laymen, attended this general assembly. The important men came to participate in the deliber- ations, and those of lower station were present in order to hear the decisions and oc- casionally also to deliberate concerning them, and to confirm them not out of coercion but by their own understanding and agreement.838 836 McKitterick, Charlemagne, 179–80, critiquing Gauert, ‘Zum Itinerar’. 837 E.g. a. 821 (Aachen, Nijmegen, Thionville, Worms, Trier, Metz, and Remiremont) 838 Hincmar, De ordine palatii, c. 29, 82–83: Unum, quando ordinabatur status totius regni ad anni vertentis spatium; quod ordinatum nullus eventus rerum, nisi summa necessitas, quae similiter toti regno incumbebat, mutabatur. In quo placito generalitas universorum maiorum, tam clericorum quam laicorum conveniebat, seniores propter consilium ordinandum, minores propter idem consilium suscipiendum et interdum pariter tractandum et non ex potestate, sed ex proprio mentis intellectu vel sententia confirmandum. Trans.: Herlihy, 222. 225 Though Hincmar’s text constructs a theoretical, idealised vision of the palace as a microcosm of the Carolingian polity—a ‘deft exercise in political sociology’, in Nelson’s words—the importance he ac- cords to the assembly is not without reason.839 The assembly was the meeting of the political commu- nity, in Hincmar’s words the generalitas universorum maiorum, the ‘the community of all the more important men’, both lay and clerical, and its convening on royal estates thus represented the king- dom’s foremost ‘public’ political arena.840 As Table 1 shows, in Charlemagne’s reign the Spring as- semblies often doubled as the mustering point of the army on the frontier.841 Given that a mark of membership of the maiores was military service, there was significant overlap between the personnel of the army and the assembly, and the army could thus form the audience of (and be a participant in) spectacles of royal authority. In Karolus Magnus et Leo papa, the meeting of Charlemagne and the Pope at Paderborn in 799 occurred with the army arrayed around them.842 The king’s role in legitimising the proceedings of the assembly was indicated already by Einhard: as we saw, Childebert III was of little use, except to be carted to and from the assembly when needed. As Timothy Reuter underlined, it is hard to imagine how many of the functions one associates with ‘government’ would have been possi- ble without the instrument of the assembly: ‘it was through embodying itself as an assembly that […] the “political community” was empowered and enabled to practice politics.’843 Notably, and in contrast to his father, Louis held not one annual assembly in the Spring, but two or even three spaced out across the year, often in January or February, the Spring, and sometimes also the Autumn in different palaces. According to Hincmar, writing in around 882, these assemblies 839 Nelson, ‘Settings of the Gift’, 140. 840 Hincmar, De ordine palatii, c. 29, 82. 841 Particularly the palaces such as Worms, Düren, and Paderborn, during the Saxon campaigns (e.g. aa. 772, 775, 776, 779, 782, 783, 784, 785), or Regensburg in preparation for campaigning against the Avars (791). 842 Karolus magnus et Leo papa, ll. 487–98, 94; Stuart Airlie, ‘Talking Heads: Assemblies in Early Medieval Germany’, in Political Assemblies in the Earlier Middle Ages, ed. Paul Barnwell and Marco Mostert, Studies in the Early Middle Ages 7 (Turnhout: Brepols, 2003), 34. 843 Timothy Reuter, ‘Assembly Politics in Western Europe from the Eighth Century to the Twelfth’, in Medieval Politics and Modern Mentalities, 207. 226 had distinct purposes. The first was a general meeting of the realm’s magnates to decide on ‘the status of the entire realm for the remainder of the year. No turn of events saving only the greatest crisis which struck the entire realm at once could change what had been established,’ a gathering probably corresponding to the Marchfield, held annually on 1 March (or, from the mid-eighth century, the Mayfield, on 1 May).844 The second assembly, Hincmar tells us, was composed only of the more im- portant men of the kingdom (seniores) who would meet to discuss specific issues, including laws, trea- ties, and warfare, that would be raised publicly at the next general assembly.845 Hincmar claims to have based his work on an earlier treatise by Adalhard of Corbie (d. 818), and scholars have generally ac- cepted that, at its core, Hincmar’s text accurately reflected practices that dated back to Louis’ reign.846 It is unlikely that a precise functional distinction between ‘main’ and ‘supplementary’ assemblies (if you will) was uniformly and consistently maintained, and one suspects that Hincmar/Adalhard gives a semblance of systematicity to an altogether more flexible set of arrangements and conventions. As we saw above, the multiplication of assemblies also speaks to the general instability of Louis’ reign and the consequent need to continually reinforce his legitimacy through displays of general consensus and rituals of gift-giving.847 The convocation of assemblies had a visible effect on the localisation of power in the Caro- lingian polity, and particularly striking are those regions conspicuous in their absence. Assemblies were very infrequently called by Charlemagne and Louis at Neustrian palaces, for example. None are rec- orded for Charlemagne’s reign, and only Compiègne (four), Quierzy (two), and Attigny (two) hosted 844 Hincmar, De ordine palatii, c. 29, 82–84: Unum, quando ordinabatur status totius regni ad anni vertentis spatium; quod ordinatum nullus eventus rerum, nisi summa necessitas, quae similiter toti regno incumbebat, mutabatur. Trans. Herlihy, 222. Reuter, ‘Assembly Politics’, 197–98. As Reuter notes, this does not mean that the gathering took place precisely on 1 March/1 May every year without fail: it was a ‘general, other-things-being-equal assumption within the political community’ rather than an absolute (ibid., 197). 845 Hincmar, De ordine palatii, c. 30, 84. 846 Adalhard was a cousin of Charlemagne, advisor to his son Pepin, and tutor to his grandson Bernard of Italy. On Hincmar’s use of Adalhard’s text: Gross and Schieffer, introduction to De ordine palatii, 10–11, and e.g.: Nelson, ‘Was Charlemagne’s Court a Courtly Society?’, 43; eadem, ‘Aachen as a Place of Power’, 227–28. 847 See above, Chapter 2. 227 assemblies multiple times in Louis the Pious’ reign.848 Displays of legitimacy were only valuable with the participation of an audience that mattered in political terms, and the infrequency that assemblies were called in Neustria reflects the marginalisation of the Neustrian aristocracy in the empire’s politics. This contrasts with the situation under Pepin III, whose reign by and large reproduced the spatial patterns of the Merovingian kings and Charles Martel as Mayor of the Palace. The evidence of the ARF—sparse for Pepin’s reign, but able to be augmented somewhat with his few extant charters— suggest that royal dealings with the aristocracy were concentrated in the traditional heartland of the Neustrian aristocracy and thus before primarily Neustrian audiences (see Table 3). Not only do we most often find Pepin at the palaces of Quierzy and Compiègne, but he was crowned by the Frankish aristocracy at Soissons in 751 and anointed by Pope Stephen at Saint-Denis in 754.849 Other considerations must also have informed where assemblies took place. The crowd of over one hundred thousand men present at Paderborn for the meeting of Charlemagne and Pope Leo III in Karolus magnus et Leo papa is without question poetic hyperbole, but the capacity to house and feed the gathering of what must have been a substantial number of people was surely a factor.850 Paderborn’s wide plain and plentiful water thanks to the proximity of both the Pader and Lippe rivers, must have made it attractive from this perspective; we should imagine that most of the participants were lodged in tents.851 On the other hand, both Einhard and Notker tell us that the seniores kept their 848 The other Neustrian palaces attested are Paris (possibly St-Denis?) and Langres in 834, during the campaign westwards against Lothar, and Lyon (or Tramoyes outside Lyon) in 835. 849 Soissons: ARF, a. 750, 9–10. The Clausula de unctione Pippini locates Pepin’s second anointing in Saint-Denis: Clausula de unctione Pippini, in Gregorii episcopi Turonensis miracula et opera minora, 15. Pepin met Stephen at Quierzy, however: ARF, a. 754, 13. 850 Karolus magnus et Leo papa, ll. 448–54, 90: Obvius ire parat genitoris iussa facessans Pippinus ; centum laetus cum milibus ibit. […] Utque vidit patulo adversum se tendere campo | Pastor apostolicus centum cum milibus altum | Pippinum, geminas extendit ad aethera palmas, | Pro populoque preces effundens pectore largas. (‘Pippin eagerly prepares to go to meet him at his father’s command; joyful will he go with a hundred thousand alongside him. […] When the apostolic shepherd sees the lofty Pepin approaching him on the plain with a hundred thousand men he spreads both arms to heaven, pouring out abundant prayers for the people from his heart’) 851 See e.g. the assembly known as the ‘Field of Lies’, in which Louis was usurped by his son Lothar, which took place ‘in a great field between Strasbourg and Basel’. The participants, according to Thegan, were lodged in tents, which they aban- doned to join the camp of Louis’ rebellious sons: Thegan, Gesta Hludowici imperatoris, c. 42, 228: Qui congregavit exercitum, 228 own houses at Aachen, probably for the same purpose – a reflection of Aachen’s centrality as the usual site of the assembly from late in Charlemagne’s reign (a topic to which we will return momen- tarily).852 The palace chosen to host the assembly must also, on a practical level, have possessed suffi- cient interior space to facilitate the meeting. At the end of our period, in Nithard’s account of the negotiations between Lothar, Pepin of Aquitaine, and Louis the German at Koblenz in 842, the legates tasked with dividing the imperial domains between the three brothers numbered one hundred and twenty (all of whom presumably brought their own followers, being lay and ecclesiastical magnates), who met each day together in the basilica of St Kastor.853 These retinues could be sizeable in their own right: the Annales Xantenses state that the Norse king Harald Klak brought ‘more than four hundred followers, both men and women’ with him to his meeting with Louis the Pious in Ingelheim and his baptism in Mainz in 826.854 Aachen presents a special case. The frequency with which assemblies were called at Aachen from Charlemagne’s later reign onwards does imply the general acceptance—or, at least, the court’s continual insistence—on Aachen as a stable centre for the political community of the whole Empire. The location of the assembly in the years it is not given by the ARF might be explained by a passing comment made in the entry for 817: ‘On his return Nijmegen he held the general assembly of the perrexit obviam eis usque in magnum campum, qui est inter Argentoriam et Basilam, qui usque hodie nominatur Campus-Mendacii, ubi plurimorum fidelitas extincta est. […] Quadam nocte maxima pars dimisit eum, et tentoria eorum reliquentes pervenerunt ad filios. 852 Notker, Gesta Karoli, I.30, 41: mansiones omnium cuiusdam dignitatis hominum, quę ita circa palatium peritissimi Karoli eius disposi- tione constructę sunt. In his account of brining the relics of SS Marcellinus and Peter to Francia, Einhard tells us that he possessed his own house at Aachen that was separate to the royal palace: Einhard, Translatio et miracula sanctorum Marcellini et Petri, ed. Georg Waitz, I.1, II.1, II.3, in Scriptorum Supplementum 1, MGH. SS 15/1 (Hannover: Hahn, 1887), 239, 245, 246. 853 Nithard, Historiae, IV.4–5, in Nithardi Historiarum Libri IIII, ed. Georg Heinrich Pertz, rev. Ernst Müller, 3rd ed., MGH. SS rer. Germ 44 (Hannover: Hahn, 1907), 46. As another example, at the meeting of Lothar II, Louis the German, and Charles the Bald at Savonnnières in 862, there were two hundred consiliarii present, including lay magnates, abbots, and bishops: Capitularia regum Francorum II, ed. Alfred Boretius and Victor Krause, MGH. Capit. 2 (Hannover: Hahn, 1897), no. 243, 165: postquam coram omnibus, qui adfuerunt, trium regum consiliariis fere ducentis, tam epsicopis quam abbatibus et laicis, relectas penitus reiecerunt. Airlie, ‘Talking Heads’, 38, and generally on this issue: Brühl, Fodrum, Gistum, Servitium Regis, 70–72. 854 Annales Xantenses, a. 826, in Annales Xantenses et Annales Vedastini, ed. Bernhard von Simson, MGH. SS rer. Germ. 12 (Hannover: Hahn, 1909),6–7: Ludewivus imperator habuit sinodum Episcoporum ad Ingulunheim, et illic venit multitude ad eum Nord- mannorum, et princeps eorum nominee Herioldus baptizatus est et uxor eius, et cum eis plus quam CCCC homines promiscui sexus. Brühl, Fodrum, Gistum, Servitium Regis, 71. 229 people as usual [more solito, ‘in the usual manner’] at Aachen’.855 This assembly was a major event: not only was it concerned with monastic and ecclesiastical reform, but at its close Louis raised his son Lothar as co-emperor, and named Pepin and the younger Louis kings of Aquitaine and Bavaria re- spectively.856 The ARF’s laconic more solito, ‘as usual’ refers not to succession practices, or even the timing of the assembly (which was somewhat later than usual, in July), but to its having taken place at Aachen.857 Looking at the spatial patterns of Charlemagne and Louis’ reigns, it is hard to resist the impression that Aachen quickly became the political centre of the expanded regnum, if not a ‘capital’ per se.858 Christian festivals. As noted above, sources such as the ARF and Annales Bertiniani are re- markable for their preservation of where the king celebrated the two main festivals of the Christian calendar, Christmas and Easter. A clear trend presents itself on Tables 1 and 2. Firstly, for most of Charlemagne’s reign, the palaces where the king celebrated Christmas and Easter rotated frequently. There was not a single preferred location, though certain palaces (for example, Herstal and Worms) hosted the court with some frequency. From 794, however, with the construction of the new chapel, Christmas and Easter were celebrated nearly exclusively at Aachen, a trend continued by Louis. That this represents a ‘centralisation’ of the court’s religious life is at once indicated by the ARF, which states that in 825 Louis held the Easter celebrations at Aachen again ‘as usual’, and generally we must 855 ARF, a. 817, 146: Unde reversus generalem populi sui conventum Aquisgrani more solito habuit. Trans. Scholz, 102. A similar phrase, more sollempni (lit. ‘in a solemn manner’, but also with the sense of ‘customarily’ or ‘as usual’), is used for the assembly in 825, then held in August but again at Aachen: ARF, a. 825, 168: ac peracta venatione Aquasgrani rediens generalem populi sui conventum more sollempni mense Augusto habuit (‘and with the hunt complete, returning to Aachen he held the general assembly of his people in the usual manner in August’). 856 Astronomer, Vita Hludovici, cc. 28–29, 372–84. Thegan comments only that Lothar was raised as co-emperor and omits the simultaneous coronations of Pepin and Louis: Thegan, Gesta Hludovici, c. 21, 210. A single exemplar of the capitulary promulgating this division, the Ordinatio imperii, is extant: Capitularia I, no. 136, 270–73. 857 The date given by the Ordinatio imperii: Capitularia I, n. 136, 270. 858 Cf. Davis, Charlemagne’s Practice of Empire, 325, 333–34, and see below. 230 surmise that the consistency by which these palaces were recorded was due to their importance in the politics of the kingdom.859 Assessing the nature of that importance is more challenging, however. It is a longstanding trope of Carolingian historiography that Christmas and Easter were opportunities for ‘crown wear- ings’, and that these rituals were introduced with Charlemagne, or possibly even the accession of Pepin III.860 At the root of the question is Einhard’s claim in the Vita Karoli imperatoris that only on high feast days (in festivitatibus) would Charlemagne eschew ‘Frankish’ clothing, appearing in ‘clothes woven with gold thread, gemmed shoes, a cloak fastened with a gold pin, and a golden crown with jewels’.861 It must be said, however, any evidence that the king was regularly ‘crowned’ on high feast days in the presence of their magnates is thin for this period, deriving mostly from brief anecdotes in, for example, the biographies of Einhard (as above) and Thegan, rather than the annalistic sources.862 That having been said, the practice of recording where the king celebrated Christmas and Easter is a decidedly new practice from Pepin III’s reign onwards, and it is likely that it did reflect new ritual practices.863 With the palace chapel taking on the role of the centre of the court’s religious life, the celebrations of Easter and Christmas must have offered important opportunities for staging royal piety. A group of palace chapels were evidently favoured as sites of the Christmas and Easter cele- brations.864 In Charlemagne’s early reign Herstal, Worms, and Quierzy emerge as preferred locations, but from the 790s onwards it is clear that religious life was being focused into the Marian chapel at 859 ARF, a. 825, 167: Sacro paschali festo sollempniter Aquisgrani celebrato (‘With the holy paschal feast celebrated at Aachen as usual’). As above, sollempniter has the sense of annual regularity as well as the meaning of its English cognate, ‘solemnly’. 860 E.g. Georg Waitz, Deutsche Verfassungsgeschichte, vol. 3 (Kiel: Homann, 1860), 212–13; Carlrichard Brühl, ‘Fränkischer Krönungsbrauch und das Problem der “Festkrönungen”’, Historische Zeitschrift 194, no. 2 (1961): 265–326. 861 Einhard, Vita Karoli imperatoris, c. 23, 28: In festivitatibus veste auro texta et calciamentis gemmatis et fibula aurea sagum adstringente, diademate quoque ex auro et gemmis ornatus incedebat. Trans. Noble, 40. 862 Brühl, ‘Fränkischer Krönungsbrauch’, 274. Thegan adapts Einhard’s praise of Charlemagne’s clothing (above) for Louis: Thegan, Gesta Hludovici, c. 19, 202–04. 863 Fleckenstein, Hofkapelle, 37. We should be cautious of pushing this too far, however, since beginning of the ARF provides a year-by-year view of the royal calendar that we sorely lack for the later Merovingian period. 864 Fleckenstein’s Festtagsitinerar: Ibid., 96–97. 231 Aachen, a centralisation of the polity’s religious life in the court. In the final two decades of his life, for example, Charlemagne celebrated Christmas a recorded fourteen times at Aachen. Formally and symbolically, the capella at Aachen was the foremost of a network of churches connected with the court – both the private capellae of other royal fiscal estates interspersed throughout the regnum and monastic foundations connected with the court through institutional and social ties of kin, patronage, and personnel.865 Yet in a broader sense still, the chapel was understood to be metonymic of the regnum’s moral health. When Einhard and Notker write of Charlemagne taking a personal interest in correcting the performance of the mass in his capella, this is metonymic of a legislative programme that identified the Carolingian polity with the ecclesia in the minds of its subjects.866As Ildar Garipzanov has argued, the development of a ‘liturgy of authority’ was an important tool in constructing an image of legitimate Carolingian kingship based in the logics of Christian ministerium.867 The constitution of such ‘networks of prayer’ were not one-sided propaganda, however, but a form of symbolic communication connecting centre and periphery by both legitimising Carolingian rule and establishing common stand- ards of expected Christian behaviour.868 Hunting. As we saw in the examination of palace ritual and its development, hunting grounds became a favoured location for displays of royal virtus late in Charlemagne’s reign and especially under Louis the Pious.869 It consequently played an important role in the localisation of political power in the Carolingian polity. As Table 2 shows, the Annales Bertiniani (as well as, to a lesser extent, the As- tronomer) record when and where Louis went hunting with remarkable consistency. Most frequently, 865 In the eighth and ninth centuries capella indicated private oratories (Eigenkirchen) available for royal use and staffed by clergy attached permanently to the court. See especially: Fleckenstein, Hofkapelle 95–109. On royal monasteries: Mayke de Jong, ‘Carolingian Monasticism: The Power of Prayer’, in c. 700–c. 900, ed. Rosamond McKitterick, vol. 2 of The New Cambridge Medieval History, (Cambridge: Cambridge University Press, 1995), 623–27; Fleckenstein, Hofkapelle, 103–04. 866 Einhard, Vita Karoli imperatoris, c. 26, 30–31; Notker, Gesta Karoli, I.8, 10–11; I.10, 12–15; de Jong, ‘Ecclesia and the Early Medieval Polity’, 119. 867 Garipzanov, Symbolic Language, 43–46, 96–100. See above, Chapter 2. 868 Ibid., 50–52. For ‘networks of prayer’: de Jong, ‘Carolingian Monasticism’, 650; McKitterick, Charlemagne, 170–71. 869 See above, Chapter 2. Goldberg, ‘Louis the Pious and the Hunt’, 625–26. 232 these were the lands in the Ardennes, the Vosges, plus the landed estates of the palaces of Nijmegen and Frankfurt.870 Some palaces, notably Remiremont in the Vosges, likely did have the character of a lodge to house the royal party during their expedition, but the notion of specialised Jagdpfalzen (‘hunt- ing palaces’) is of limited utility.871 We should note that other palaces associated with hunting grounds, including Nijmegen, Compiègne, and Aachen itself, were used for other important functions, includ- ing assemblies, receiving ambassadors, and ecclesiastical synods. During the reign of Louis the Pious, the royal hunt was the principal stimuli of royal mobility other than the assembly. Even where there is no direct and explicit evidence for royal hunting, the pattern which emerges from sources is that favoured palaces were frequently located in a liminal po- sition on the edge of a highland zone, at the boundary of civilisation and wilderness.872 For example, not only Aachen, but also Herstal, Düren, Attigny, and Longlier among others lie at the edge of the large highland plateau formed by the Ardennes and Eifel.873 The close relationship of royal hunting, landowning, and zones of intensive authority are most clearly underlined by the Vosges. Alsace had, like Aquitaine, drifted from royal control by the late-Merovingian period to become a semi-independ- ent dukedom under the Etichonid family; it was brought back under Frankish authority (peacefully) only in the 740s.874 Charters from Charlemagne’s reign show the monasteries of Alsace being brought more closely into a royal orbit.875 For example, in a charter issued at Aachen on 13 January 769, Char- lemagne granted the monastery of Saint-Dié-des-Vosges, which the text claims had formerly been 870 Ingelheim, Salz, Compiègne and Aachen itself also seem to have been endowed with ample hunting grounds. As dis- cussed in Chapter 2, it is likely that hunting grounds on palatial landed estates became de rigueur in the Carolingian period. 871 Such terminology imposes a false sense of singular use to these palaces (see also e.g. Taufpfalzen, Winterpfalzen). Moreover, as we saw in Chapter 2, hunting was a common function of Carolingian palaces generally. On terminology, see: Thomas Zotz, ‘Pfalzen zur Karolingerzeit. Neue Aspekte aus historischer Sicht’, in Splendor palatii, 13–23 872 A feature stressed more generally by Rollason, ‘Forests, Parks, Palaces’, 432–34. 873 The Ardennes and the Eifel are both extensions of the large geological zone known as the Rhenish Massif/Rheinisches Schiefergebirge, which extends on both sides of the Rhine from the Ardennes to the Sauerland. 874 Hans Hummer, Politics and Power in Early Medieval Europe: Alsace and the Frankish Realm, 600–1000, Cambridge Studies in Medieval Life and Thought, 4th Series, 65 (Cambridge: Cambridge University Press, 2005), 56–75. 875 Ibid., 61–63. 233 under Pippin III’s protection, to Saint-Denis.876 The monastery’s proximity to the royal house is un- derlined by a condition of the grant being that ten or fifteen monks guard it and ‘constantly in psalms and masses and other prayers of supplication, as well as special orations for us and for the glorious lord our father, not cease day or night from raising prayers to God’.877 Efforts to incorporate the region more closely into the Carolingian hegemony were not limited to the politics of landownership and establishing ties with monastic institutions, however, but also the spatial praxis of hunting. One of the rare mentions of Charlemagne hunting was in July 805 in Vosges, detailed in the ARF and the Annales Mettenses priores.878 The ARF records that Charlemagne did not himself accompany the army that year, but instead went hunting in the Vosges, reaching the estate of Champs via Thionville and Metz.879 Once reunited with his son, he moved to nearby Remiremont, later a favoured hunting estate of Louis’, possibly suggesting that the Autumn too was devoted to the chase.880 Hunting was not just a representational practice, but also inherently spatial. As we saw previ- ously, the designation of land on palatial estates for the sole purpose of hunting—and thus restricting economics rights to its resources—went hand in hand with the growth in the importance of hunting as political ritual. By extension, however, hunting also shaped the localisation of power in the Caro- lingian polity, both physically and institutionally. The concentration of palace estates on the fringes of the Ardennes, the Eifel, and the Vosges (and, by extension, the prominence of these areas in reports 876 DD Karol. 1, no. 55, 81: hoc est monasteriolo aliquo qui nuncupatur Ad sancto Deodato infra Uosago silva, sicut eum domnus et genitor noster Pippinus in sua vestitura tenuisse conprobatum est. Saint-Dié had been founded by Deodatus on fiscal land in around 669: Alain Stoclet, Autour de Fulrad de Saint-Denis (v. 710–784), Hautes Études Médiévales et Modernes 72 (Geneva: Librarie Droz, 1993), 86–93 (on the 769 charter: ibid., 93–97). The connection to Pepin made supervision of the monastery an appropriate donation to Saint-Denis, since the former king had been interred there the previous year. 877 DD Karol. 1, no. 55, 81: ea videlicet ratione semper ipsi fratres decem aut quindecim per vices ibidem ipsum locum custodire debeant et ibi assiduae in psalmis et missas et ceteris obsecrationum orationibus vel peculiares orations pro nobis et pro domno adque pro glorio[so] genitore nostro deum preces exorare die et nocte non desistant. 878 Goldberg, ‘Louis the Pious and the Hunt’, 642. 879 ARF, a. 805, 120: et inde regressus in Vosego silva ad imperatorem venit in loco, qui dicitur Camp. Nam imperator Iulio mense de Aquisgrani profectus Theodonis villam atque per Mettis transiens Vosegum petiit. Ibique venationi operam dans post reversionem exercitus ad Rumerici castellum profectus ibique aliquantum temporis moratus ad hiemandum in Theodonis villa palatio suo consedit. The entry in the Annales Mettenses priores contains the same information: Annales Mettenses priores, a. 805, , ed. Bernhard von Simson, MGH. Rer. Germ. 10 (Hannover: Hahn, 1905), 94–95. 880 Louis went hunting at Remiremont in 821 and 825: ARF, aa. 821; 825. 234 of royal mobility) highlights the significance of the forest as a royal landscape overlaying and existing alongside the patterns produced by Carolingian landholding. Categorising the motivations for royal mobility in this way reveals two important trends. As shown plainly by Tables 1 and 2, Charlemagne and Louis the Pious most certainly were mobile and, at least as far as we can tell from the evidence, on a far greater scale than the later Merovingians. If we reframe our focus, however, and exclude the radical displacement caused by Charlemagne’s military campaigning, this mobility was contained within closely circumscribed landscapes for the most part. From this perspective, the reproduction of existing zones of royal authority represents not a revolu- tion, but a continuation and intensification of existing Merovingian practices. In Josiane Barbier’s words, the rise of the Carolingians ‘reveals a rather different tonality: less the instigator of a radical caesura than the heir of the kingdom’s spatial organisation and political attitudes, which it made ex- plicit and amplified’.881 The revolution lay not in the extent of royal mobility as such, but in the con- stitution of the political landscapes of the Frankish regnum in the context of a dramatically expanded polity. In fact, as we saw with the convocation of assemblies in Neustria, these landscapes were not necessarily stable and persistent, but rather (de)activated according to contemporary political circum- stance. Secondly, we have also seen that the establishment of Aachen in the later eighth century set in train a symbolic and institutional centralisation of the Frankish empire – new supra-regional networks overlaying the traditional landscapes of royal authority. Austrasia as a Carolingian royal landscape In his Vita Karoli imperatoris, Einhard presents Aachen as Charlemagne’s crowning architectural achievement, ahead of other palaces he began at Ingelheim and Nijmegen.882 Similarly, for Notker, the chapel at Aachen was the axis of the regnum’s moral and political universe, constructed ‘following the 881 Barbier, ‘Le système palatial franc’, 297. 882 Einhard, Vita Karoli imperatoris, c. 17, 20. 235 example of wise old Solomon as a resting place for God and for himself, for all his bishops, abbots, counts, and all the rest who came from all over the world.’883 For an observer in the ninth century, Aachen was unquestionably the empire’s symbolic centre, the ‘first seat of Francia’ in Nithard’s words.884 Before 794, however, it is seldom mentioned in the ARF and other contemporary sources.885 In scholarship too Aachen has occupied a central place in historical imaginations of the period, being presented as the material expression of a programmatic imperial vision, the ‘capital’ of a new realm. We saw above that royal mobility in the Carolingian period should be seen as at most an intensification of pre-existing Merovingian practices, but the central place that Aachen came to occupy in the mindset of the architects of the Carolingian imperial project immediately marks it out as distinct. Essential questions present itself, therefore: to what extent was Aachen intended as a central place (or, even, a ‘capital’), and what factors informed the choice of this previously insignificant place? Replacing Aa- chen into the wider context of the political landscapes of the Carolingian realm throw the dynamics of Carolingian ‘centralisation’ into sharper relief. Owing to their rise as mayors of the Austrasian palace in the seventh century, the Carolin- gian/Pippinid family possessed a dense concentration of estates in the region between the Meuse and the Rhine, to which were added the substantial holdings of the royal fisc after Pepin III’s accession. When the Pippinids first came to prominence during the seventh century, their social and economic sway seems to have been predominantly regional, in contrast to rival family groupings including the so-called ‘Faronids’ and the Agilolfings, both of whom possessed lands and connections across the regnum and beyond.886 Regional territorialisation of power can be glimpsed through what is known in 883 Notker, Gesta Karoli, I.27, 38: De quibus mox docebo, si prius de edificiis, quę cesar augustus imperator Karolus apud Aquasgrani iuxta sapientissimi Salemonis exemplum Deo vel sibi vel omnibus episcopis, abbatibus, comitibus et cunctis de toto orbe venientibus hospitibus mirifice construxit. Trans. Noble, 83. 884 Nithard, Historiae, IV.1, 40: Aquis palatium, quod tunc sedes prima Frantiae erat. 885 Pepin III wintered there in 765, showing that there was already a royal estate and oratory there in the mid-eighth century. Charlemagne celebrated Christmas there in 768 and wintered there in 788/89. 886 Paul Fouracre, The Age of Charles Martel, The Medieval World (London: Pearson, 2000), 34–35. The reconstructed family group known in scholarship as the ‘Faronids’ possessed extensive holdings in western Austrasia and probably counted 236 scholarship as Klosterpolitik – the endowment of, and struggles to control, monasteries and other ec- clesiastical institutions as a function of ‘secular’ politics. The installation of a favoured supporter in a bishopric or monastery would anchor a family’s power in a local landscape; on the other hand, donat- ing land to a monastery tied to the family by charter provided a means of keeping landed holdings undivided.887 The evidence of charters and diplomas, therefore, can reveal both concentrations of landed wealth but also, by extension, its limits. The charter evidence suggests that the focus of Pippinid landholding was originally on the Meuse, in the region of modern Belgium. This base was augmented, however, through marriage alli- ances to add significant holdings on the Moselle and Middle Rhine and in the region of Maastricht and Liège.888 On the other hand, direct Pippinid influence in Neustria seems to have been limited outside of their control over the political instruments of the palace. For example, attempts to extend influence over Neustrian ecclesiastical institutions were halting and, by and large, unsuccessful. The marriage of Pepin II’s son Drogo to Anstrud, the daughter of the former mayor (and possibly count of Rouen) Waratto, allowed the Pippinids to gain a foothold on the Seine. Audoin’s successor as bishop of Rouen, Ansbert, was replaced with a Pippinid supporter, Gripho, and in the first decade of Rado, Chlothar’s Austrasian mayor of the palace between 613 and 617, as well as Audoin of Rouen, among their number. The Agilolfings held the ducal title of Bavaria and a line of the family also ruled in the Lombard kingdom intermittently in the seventh and early-eighth centuries. 887 Fouracre, Charles Martel, 45–50. 888 According to the early ninth-century Annales Mettenses priores, Pippin of Landen ‘governed the people living in the wide area between the Silva Carbonaria and the river Meuse and all the way to the border with the Frisians with just laws’ (i.e. roughly the Western half of modern Wallonia): Annales Mettenses priores, a. 688, 2: populum inter Carbonariam silvam et Mosam fluvium et usque ad Fresionum fines vastis limitibus habitantem iustis legibus gubernabat. The Pippinid monastic foundations of Nivelles and Fosses, which date to the mid-seventh century, lie in this region. Werner saw Liège as part of the original possessions of the Pippinids: Matthias Werner, Der Lütticher Raum in frühkarolingischer Zeit. Untersuchungen zur Geschichte einer karolinischer Stammlandschaft, Veröffentlichungen des Max-Planck-Instituts für Geschichte 62 (Göttingen: Vandenhoeck & Ruprecht, 1979), 347–354 (largely confirming the report of the Annales Mettenses priores, but also hypothesising landholding around Liège). Gerberding has convincingly argued, however, that Liège was not part of an established Pippinid patrimony but was only brought into a Carolingian orbit under Charles Martel: Gerberding, Rise, 120–29. Most of the family’s landed wealth in the Metz and Rhine regions (as evidenced by e.g. the charters of the monastery of Echternach) came via Pepin II’s marriage to Plectrud: ibid., 95. Martel’s victories over the Neustrians 716–718 brought the the Austrasian elite and the important sees of Metz and Trier, as well as Liège (which would shortly afterwards become the seat of the bishops of Maastricht), under Charles’ influence: ibid., 133–40. Theuws stresses that Maastricht and its hinterland were fiscal land and possibly a northern ‘power base’ of the Merovingian kings, possibly explaining Liège being developed as a competitor by the Pippinids: Theuws, ‘Maastricht’, 182–85. 237 the eighth century the monasteries of Saint-Wandrille and Fleury were taken under Pepin’s protec- tion.889 But elsewhere Pippinid influence is conspicuously lacking. Paul Fouracre has argued, for ex- ample, that charters and royal placita show a continued absence of Pippinid influence in the Neustrian heartlands around Paris and that the elite successfully barred them from influence over monasteries and other significant ecclesiastical positions in that region, including Saint-Denis.890 These patterns strongly suggest that the frequency with which Charlemagne wintered at Aa- chen were as much for economic motivations as exploiting the symbolic resources of a new ‘capital’. The demand the court placed on local producers during the winter months were clearly substantial. We get a sense of this in the ‘Astronomer’s account of Louis’ reign in Aquitaine (781–814) before succeeding his father as king of the Franks and emperor. Seeking to emphasise Louis’ wisdom and generosity, he writes that the Louis ordained that he would spend each winter in [one of] four places, so that after three years had passed, each one of these places would only have supported him for one winter. These were the palaces of Doué, Chasseneuil, Angeac, and Ebreuil, and each of these places, when the fourth year had come around, would provide adequate re- sources for the royal service. With these arrangements most prudently ordained he forbade that military expenses, commonly called fodder, be met any longer by the common people.891 The area between the Meuse, Moselle, and Rhine was marked by a strong concentration of royal ma- norial estates, and the dependent peasants who worked them paid tribute to the court (or, more pre- cisely, its officers: the iudices) in kind.892 The Capitulare de villis details responsibilities of the iudices for collecting the foodstuffs, money, and other goods that were to be supplied to the court at regular intervals: ‘Each and every iudex is to ensure that produce is to always come to the court in plentiful 889 Gerberding, Rise, 97–98. 890 Fouracre, Charles Martel, 49–50. 891 Astronomer, Vita Hludowici, c. 7, 304: Nam ordinavit, qualiter in quattuor locis hiberna transigeret, ut tribus annis exactis, quarto demum anno hiematurum se quisque eorum susciperet locus, Theotuadum scilicet palatium, Cassinogilum, Andiacum et Eurogilum; que loca, quando quartum redigebatur ad annum, sufficientem regio servitio exibebant expensam. Quibus prudentissime ordinatis, inhibuit a plebeis ulterius annonas militares, quas vulgo foderum vocant, dari. Trans. Noble, 233–34. 892 Especially: Wolfgang Metz, Das karolingische Reichsgut: eine verfassungs- und verwaltungsgeschichtliche Untersuchung (Berlin: De Gruyter, 1960), 91–195. 238 supply, and what is more let them undertake their visits three or four times, or more.’893 Darryl Camp- bell has argued that the Capitulare de villis most likely dates to the final decade of the eighth century and reflects the logistical challenges of supplying the court at Aachen on the basis of the demands that the stabilisation of a formerly ‘peripatetic’ court at Aachen would place on agrarian production and the overseers of royal estates.894 Campbell is likely correct in his dating the Capitulare de villis to this period, but not in attributing its motivation to a transition from itinerancy to stability as such. Given Aachen’s emplacement in a landscape of royal estates—hardly an isolated locale—it would be strange to sup- pose that supplying the court presented new logistical difficulties. Rather, the Capitulare de villis ex- presses both the new, centralising impulses of the court and the closely connected rhetoric of moral order and correct governance that was distinctive of the regime from the 790s onwards, of which the foundation of Aachen formed a part. Life on the limes: palaces beyond the Rhine While Aachen took centre stage in Einhard’s account of Charlemagne’s accomplishments, he notes also that he also constructed—or, more precisely, ‘began’ (inchoavit)—other palaces, specifically at Nij- megen and Ingelheim.895 As we saw in Chapter 1, the programme of palaces constructed under Char- lemagne and Louis are a distinctive feature of this period and the landscapes they produced reshaped the Frankish domain. In particular, the palace of Paderborn in Saxony and the palaces connected to the Main highlight the role of palaces as tools of exerting royal authority in newly incorporated areas (Fig. 135). 893 Capit 1, no. 32.20, 84: Unusquisque iudex fructa semper habundanter faciat omni año ad curtem venire, excepto visitationes eorum per vices tres aut quattuor seu amplius dirigant. 894 Darryl Campbell, ‘The Capitulare de Villis, the Brevium exempla, and the Carolingian Court at Aachen‘, EME 18, no. 3 (2010): 260–64. 895 Einhard, Vita Karoli imperatoris, c. 17, 20: Inchoavit et palatia operis egreii, unum haud longe a Mogontiaco civitate, iuxta villam cui vocabulum est Ingilenheim, alterum Noviomagi super Vahalem fluvium. choice to note that Charlemagne only ‘began’ these palaces is another hint that his outlook (the centrality of Aachen, etc.) reflects the reign of Louis, when Ingelheim and Nijmegen were more-frequently attested in the ARF and other sources, rather than that of his father. 239 Implicit in Charlemagne’s concern not to appear ‘idle’ or ‘soft’ on his sojourn to Salz in 790 is that the palace lay in a liminal area for Carolingian authority. Located on the Franconian Saale, a tributary of the Main, the palace at Salz was located close to Saxony and the regnum’s expanded eastern border.896 From the annalist’s comments we can infer two things. The first is that to appear in person on the periphery in this way was risky, but also that it was politically expedient, a means of interfacing a liminal region to the political centre.897 Salz itself was already a thriving settlement in the late-Mero- vingian period, but under Charlemagne it was expanded with a new palace.898 Salz would continue to be an important regional royal centre, exerting control over a productive agrarian landscape, into the Ottonian period and beyond.899 Ingelheim, located close to Mainz and the confluence of the Rhine and Main, alongside Frankfurt and, further east, Salz, created a chain of sites projecting royal authority into a frontier region that was, since the early eighth century, only loosely incorporated into the Frank- ish hegemony.900 896 Charlemagne returned in 793 according to the Astronomer in the aftermath of Pepin the Hunchback’s rebellion (not reported in the ARF): ‘Proceeding rapidly, they [Louis and Pepin of Italy] came to their father in the region of Bavaria, at a place called Salz, and were received by him most warmly. Louis spent whatever was left of the summer, the fall, and the winter with his father, the king (Astronomer, Vita Hludowici imperatoris, c.6, 302: Concite ergo pergentes, in partibus ad patrem Baoarie venerunt in loco, cuius est vocabulum Salz, et ab eo gratissimę sunt recepti. Quicquid autem superfuit aestatis, autumni et himeis, cum patre rege rex Hludouuicus exegit. Trans. Noble, 233). Christmas that year was celebrated at Würzburg (which was noted in the ARF), the nearest ecclesiastical centre to Salz. Charlemagne would again return to Salz in 803, where he met a Byzantine delegation and then set off to bring order to Pannonia: ARF, a. 803, 118. 897 Particularly important in the wake of Hardrad’s rebellion in Thuringia in 785/786. See: Davis, Charlemagne’s Practice of Empire, 140–49. 898 On Salz, see especially: Heinrich Wagner, ‘Zur Topographie von Königsgut und Pfalz Salz’, in Pfalzen – Reichsgut – Königshöfe, ed. Lutz Fenske, Deutsche Königspfalzen 4 (Göttingen: Vandenhoeck & Ruprecht, 1996), 149–83. 899 Salz and its landscape is the subject of an ongoing research project based at Friedrich-Schiller-Universität Jena, with particular emphasis on settlement archaeology. See: Peter Ettel et al., ‘Vorbericht zu den Untersuchungen 2009 bis 2012 im Königsgutbezirk und Pfalzgebiet Salz, Lkr. Bad Neustadt a. d. Saale’, Beiträge zur Archäologie in Ober- und Unterfranken 8 (2013): 213–48; Lukas Werther, ‘Der Königsgutkomplex Salz und das Neustädter Becken – ein frühmittelalterlicher Zent- ralraum im Wandel der Zeit’, in Zentrale Orte und Zentrale Räume des Frühmittelalters in Süddeutschland, ed. Peter Ettel and Lukas Werther, RGZM-Tagungen 18 (Mainz: Verlag des Römisch-Germanischen Zentralmuseums, 2013), 89–112. 900 Thuringia had also been the object of missionizing by Boniface and, after its establishment in 742, the bishops of Würzburg; while there was likely a strong Christian presence in Thuringia, it was probably institutionally thin. Missionary activities were also a tool of its firmer incorporation into the Frankish hegemony. See: John-Henry Clay, ‘Boniface in Hessia and Thuringia’, in A Companion to Bomiface, ed. Michel Aaij and Shannon Godlove, Brill’s Companions to the Chris- tian Tradition 92 (Leiden: Brill, 2020), 270–98. 240 The regime’s efforts to connect these landscapes institutionally and imaginatively to the polit- ical centre was acted out in the physical landscape. The priority given by Einhard (and later, Notker) to the stone bridge Charlemagne ordered built at Mainz highlights a contemporary sensitivity to the close alliance of infrastructure and the spatial extension of the polity east of the Rhine.901 Even more strikingly, for the year 793 the ARF records that Charlemagne travelled by ship from Regensburg ‘to the great trench between Altmühl and Rednitz, and there emissaries of the pope appeared with large presents.’902 The later reviser expands on this and reports the disappointing results: The king was persuaded by self-styled experts that one could travel most conveniently from the Danube into the Rhine if a navigable canal was built between the rivers Red- nitz and Altmühl, since one of these rivers flow into the Danube and the other into the Main. So he went at once with his entire following to the place, gathered a large number of people, and spent the whole fall on this project. A ditch was dug between these two rivers, two thousand paces long and three hundred feet wide. But it was in vain; for due to continuous rain and because the swampy ground as such contained too much water, the work that was done did not hold. Whatever the diggers dug out during the day would sink back into its former place during the night.903 The fossatum magnum, the ‘great channel’, clearly made an impression on contemporaries and both its construction and Charles’ visit is mentioned in many of the so-called ‘minor’ annals.904 Its significance, however, lies not (only) in its engineering ambition, but the evidence it provides for the regime’s efforts to connect the distinct royal landscapes of the Carolingian polity. In the event, according to the ARF Charlemagne travelled to the fossatum from Regensburg on the Danube but continued 901 Einhard places the bridge second only to the palace chapel at Aachen. Einhard, Vita Karoli imperatoris, c. 17, 20; Notker, Gesta Karoli, I.30, 40. Notker states that ‘all of Europe’ (tota Europa) contributed to its building – in the context of the passage referring to levies imposed on lay and ecclesiastical lords under Carolingian authority. 902 ARF, a. 793, 92–94: Rex autumnali tempore de Reganesburg iter navigio faciens usque ad fossatum magnum inter Alemania et Radantia pervenit, ibique apostolici cum magnis muneribus praesentati sunt. Trans. Scholz, 71. 903 ARF (R), a. 793, 93: Et cum ei persuasum esset a quibusdam, qui id sibi compertum esse dicebant, quod si inter Radantiam et Alemonam fluvios eiusmodi fossa duceretur, quae esset navium capax, posse percommode a Danubio in Rhenum navigari, quia horum fluviorum alter Danubio, alter Moeno miscetur, confestim cum omni comitatu suo ad locum venit ac magna hominum multitudine congregata totum autumni tempus in eo opere consumpsit. Ducta est itaque fossa inter praedictos fluvios duum milium passuum longitudine, latitudine tecentorum pedum; sed in cassum. Nam propter iuges pluvias et terram, quae palustris erat, nimio humore naturaliter infectam opus, quod fiebat, consistere non potuit; sed quantum interdiu terrae a fossoribus fuerat egestum, tantum noctibus humo iterum in locum suum relabente subsidebat. Trans. Scholz, 71. 904 On the archaeological evidence of the fossa Carolina: Lukas Werther et al., ‘792 or 793? Charlemagne’s Canal Project: Craft, Nature and Memory’, EME 28, no. 3 (2020): 444–65 (esp. 455–63). 241 onwards down its planned course to the Main and Frankfurt, illustrating the connection that the chan- nel would form between royal centres. Paderborn similarly highlights the importance of palaces in the creation of new royal land- scapes. Founded in 776, Paderborn was inextricably tied with the conquest and conversion of Saxony. The forced baptism of defeated Saxons at Paderborn in 776 and again at the annual assembly the following year also held at Paderborn emphasises the interconnection of the wars’ violence with their religious objectives and the spectacle must have left a lasting psychological scar on Saxon communities involved.905 The anonymous Carmen de conversione Saxonum, for example, closely associates Charle- magne’s majesty on the battlefield as he ‘dragged the hosts of wood-dwellers to the kingdom of heaven’ with his personal involvement in the baptism of the survivors – he ‘sent the uncultivated new- Christians to the stars of heaven’ before leading them into the palace hall.906 The palace’s destruction in 778 and 794 and subsequent rebuilding (as well as the punitive campaign that followed) underlines its symbolic importance for asserting Carolingian authority over the region already at that time.907 That Paderborn lay at the centre of an evolving strategy of palatial representation for a Frankish (non-local) audience is signalled in the Carolingian minor annals. In the otherwise brief and laconic Annales Mo- sellani, the annalist remarks that the king ‘built a city across the river Lippe, which he named Karlesburg’, an attribution confirmed by the Annales Petaviani referring to Paderborn as the urbs Karoli.908 The meet- ing of Charlemagne and Leo III, dramatised in Karolus magnus et Leo papa, further emphasises the 905 ARF, aa. 776–777, 42–49. On the Saxon Wars and their connection to Frankish missionising, see especially: Ingrid Rembold, Conquest and Christianization: Saxony and the Carolingian World, 772–888, CSMLT, 4th Series, 108 (Cambridge: Cam- bridge University Press, 2017), 39–84. 906 Carmen de conversione Saxonum, ll. 47, 59, 62, in Poetae I, 381: Traxit silviculas ad caeli regna phalanges […] Chrosticolasque rudes ad caeli sidera misit […] Progeniemque novam Christi perduxit in aulam; Rembold, Conquest and Christianization, 78–79. 907 ARF, a. 778, 50–53. The ARF does not specify that the palace was destroyed. This detail is given by: Annales Petaviani, a. 778, in Scriptores, ed. Georg Heinrich Pertz, MGH. SS. 1 (Hannover: Hahn, 1826), 16: incenderunt oppida, et igne cremaverunt civitatem, quae Franci construxerunt infra flumen Lipiam. 908 Annales mosellani, a. 776, ed. Johann Martin Lappenberg, in Annales aevi Suevici I, ed. Georg Heinrich Pertz, MGH. SS. 16 (Hannover: Hahn, 1889), 496: Et aedificavit civitatem super fluvio Lippiae, que appellatur Karlesburg; Annales Petaviani, a. 776, 16: aedificaverunt Franci in finibus Saxanorum civitatem quae vocatur Urbs Karoli. Paderborn’s self-consciously ideological for- mation is captured neatly by Rembold, who describes it as Charlemagne’s ‘showcase Saxon village’: Rembold, Conquest and Christianization, 154. 242 intended significance of the site for symbolic displays of royal power before an audience of the Frank- ish exercitus. Importantly, however, this symbolic power also carried institutional heft. While initially the ecclesiastical organisation of Paderborn was entrusted to the bishops of Würzburg, in 806 Paderborn itself became a bishopric. The genesis of the Paderborner stemmed not only from royally sponsored missionary efforts, but also the royal chapel. That Paderborn’s basilica, dedicated by the Pope with relics of St Stephen in 799, was a statement of intent is signalled already by its ‘wonderous size’ in the words of the Annales Laureshammenses, far outstripping any realistic congregation it may have had at that time.909 The palace and its church linked Saxony to the network of capellae interspersed throughout the Carolingian domain—at the centre of which stood Aachen from the 790s—and thereby brought it under the extended supervisory and administrative reach of the court.910 Put simply, the establish- ment of Paderborn not inscribed a lasting mark of royal authority into the landscape, but interfaced it with the broader palace network of the Frankish regnum, integrating it in both an institutional and moral sense. *** The ninth-century Empire was a conglomeration of landscapes, at the centre of which stood Aachen. We have seen that the traditional narrative, in which early medieval itinerancy gave way to the stability of a new (quasi-)capital at Aachen, misrepresents and delocalises the transformation of the Frankish political landscape under the Carolingians. Rather, we have seen that the first Carolingian kings were mobile within landscapes that were continuous in significant ways with earlier centuries. In particular, the region of western Austrasia in which Aachen was situated was densely packed with fiscal estates 909 Annales Laureshamenses, a. 799, 38: ecclesia mirae magnitudinis. The three-aisled basilica, terminating in a triple apse, was in its initial phase (ca. 799) 42.7m long by 21m wide. A large western transept and apsidal chapel was added early in the first half of the ninth century, probably coinciding with the translation of St Liborius to Paderborn in 836, alongside an expan- sion of the palace: Balzer, ‘Paderborn’, 41–43, 53–57; Gai, ‘Pfalz Karls des Großen’, 187–90. 910 Fleckenstein, Hofkapelle, 41. On the capella as chancellery: ibid., 74–86. 243 and properties belonging to the Pippinid line. At the same time, however, these landscapes were acti- vated and deactivated according to contemporary politics. Neustria, for example, while favoured under Pepin III, was not a significant zone of royal activity under Charlemagne and Louis. This is not to say, however, that Aachen was not distinctive. Aachen took on the status of the empire’s symbolic and administrative centre in a fashion not shared by other Carolingian palaces. Ultimately, this centrality was manifested less in the movement of the king and court, but of the missi and iudices who oscillated between it and the empire’s localities. As Stuart Airlie memorably character- ised it, the Aachen court ‘was like a great railway junction shunting personnel all over the realm’.911 Equally remarkable are the efforts to interconnect this new central landscape with others throughout the regnum—particularly Saxony and the ‘eastern Frankish’ regions on the Rhine’s left bank—through methods at once ideological and administrative, but also material: the foundation of new palaces and even, as with Charlemagne’s fossatum magnum, physical interventions in natural topography. The Caro- lingian regnum was, very literally, inscribed in the landscape. V. Conclusion The royal landscapes of the Frankish regnum were the basis of political action and they exerted a ‘cen- tripetal’ force on elite social networks. The Frankish palatial network was the ‘architecture’ on which the Frankish political system was built. Tracing its evolution over three centuries, from the mid-sixth to the mid-ninth, we have seen a different side to the transition between ‘Late Antiquity’ and ‘the Early Middle Ages’. At root, the landscapes that underwrote the Frankish regnum’s distinctive political system were formed out of the militarised infrastructures of the late Roman frontier. These landscapes, though emerging from the ‘sedimentation’ of numerous layers over time, were not wholly static, how- ever. As we saw in the Carolingian period, new localisations of political power could both form new 911 Airlie, ‘Palace of Memory’, 8. 244 landscapes (as in Saxony and Franconia) and reform existing ones, shifting their boundaries or reori- enting them towards a new centre. We have also seen the role of palaces in the fluctuating dynamics of intensification and extension of political authority in north-western Europe. Rather than a smooth evolutionary curve from small-scale post-Roman societies towards the centralised, bureaucratic gov- ernments of the central Middle Ages and later (and by implication, the modern State), palaces created elite social networks and integrated elites into ‘central’ political structures that were not necessarily persistent, as we saw clearly in the period of unification in Neustria in the seventh century. 245 CONCLUSION To bring the various threads of this study together, it is worth returning to Cassiodorus’ letter to the curator palatii one final time.912 On first reading, Cassiodorus might seem to endorse a ‘common sense’ interpretation of the palace’s function in the political system. The ‘fitting face’ of Theoderic’s imperium was designed to impress, to overawe, to reflect an authority that already existed. When we take the time to interrogate the connections between palaces and power in the Early Middle Ages more care- fully, however, a more nuanced picture emerges of palaces’ roles in early medieval political systems. A sub-text of Cassiodorus’ depiction of the amazement of the legates, adjudging the king to be ‘as great as his dwelling’, is that his authority was not inherent but produced in the encounter of ruler and elite at the palace.913 That is to say, palaces were not just a stage against which the ruler’s power was paraded, but were also actively implicated in its negotiation, legitimatisation, and action. Lurking between the lines of Cassiodorus’ prose, therefore, is an indication of the palace’s essential duality: both architec- tural object and a social formation. In fact, we can a step further: as we saw in the preceding chapters, the early medieval palace was simultaneously a class of building, a focus for ideas surrounding author- ity and its legitimate action, a hub of social relations, and a node in the material networks by which power was actualised. In contrast to the prevailing historiographical situation, one compartmentalised between prac- titioners of different disciplines, methodological schools, and contextual specialisms, palaces have here been treated as part of a continuous history stretching from Antiquity to the Middle Ages. In Chapter 1, for example, we saw that palaces (as a species of elite domestic architecture) were transformed in response to changing ideals of political authority and material display. A coherent ‘grammar’ of Late Roman palaces, one that related the new imperial sedes to an authorising prototype on the Palatine, 912 Above, Introduction, 1. 913 Cass., Var., VII.5, 204: prima fronte talis dominus esse creditur, quale eius habitaculum comprobatur. 246 precipitously broke down in the sixth century. In its place coalesced a new configuration attuned to new symbolic languages of political authority in a post-Roman world. Although archaeological evi- dence for palaces is lacking for the ‘long seventh century’, it is nonetheless possible to replace palace architecture into a continuous history of architectural change between the fourth and ninth centuries CE. In the midst of these changes, it was shown that the ‘palace chapel’ as an architectural concept was not to a constant feature from Late Antiquity onwards, but one that originated in the specific context of the ‘liturgification’ of authority in the Early Middle Ages. But what did palaces ‘do’ in early medieval society? Chapter 2 concentrated on the encounters of the ruler with the elite that took place within the palace. Focusing on the ‘social spaces’ which together constituted the palace—a term drawn from the theorisations of space pioneered by Henri Lefebvre and others—we saw that rituals were not one-sided expressions of royal ideology as derived from court-adjacent written sources, but a medium of symbolic communication, a means of negotiat- ing and defining ideals of political authority and legitimacy. In the audience hall, the hunting grounds, the palace church, and rituals of gift exchange mediated through the treasury, the political community negotiated norms of legitimacy and authority that affected the ruler as much as the aristocracy. It was the performance of these rituals that designated the palace as a ‘central place’, the focus of early me- dieval political life. The construction of an image of royal authority was not separate to the economic and material structures by which that power was actioned, however. Part II investigated the palace’s ‘systemic’ role: how the localisation of political power in palaces generated elite social networks, established spatial control over infrastructures, and established extensive relations between the political centre and pe- riphery. In the Ostrogothic kingdom of Italy and in the Frankish regnum, we saw that palaces were instrumental in the creation of new polities from fragments of the Roman imperial infrastructure and, in broader terms, the reshaping of Late Empire’s ‘political landscape’. While the methodologies of 247 Pfalzenforschung provide a useful basis for situating politics in local negotiations of power, it was argued in Chapters 3 and 4 that the underlying investment in questions of Staatlichkeit, ‘statehood’, and insti- tutional continuity between the late Roman and early medieval worlds risks abstracting authority from the spaces in which it was produced and found meaning. A focus on ‘royal landscapes’, I argue, re- places palaces at the centre of early medieval political culture, providing a vocabulary for a holistic analysis sensitive both to ‘ideology’ and to the material mechanisms of power’s action. Located at the intersection of the varied material, ideological, and social changes that occurred between the fragmen- tation of the Western Roman Empire and the rise of the Carolingian hegemony, palaces redirect our attention from the traditional terms of debate—catastrophe and collapse, or continuity and transfor- mation?—towards a renewed focus on the creativity and agency of early medieval people in the nego- tiation of the Roman inheritance. Palaces were, in this respect, the essential technology by which the transition from a ‘Roman’ to ‘early medieval’ world was negotiated. 248 APPENDIX I Royal Mobility Under the Carolingians (751–840 CE) The following tables visualise the movements of Pepin III, Charlemagne, and Louis the Pious between royal estates and palaces, principally on the basis of the ARF and its continuations, the Annales Bertin- iani (‘Annals of St-Bertin’, AB) and the Annales Fuldenses (‘Annals of Fulda’, AF).914 Important also for Louis’ reign are the histories authored by Thegan and the ‘Astronomer’, which in places expand upon information derived from other sources. The objective here is not a comprehensive reconstruction of these kings’ movements—for which, see above all the Regesta imperii—but to illustrate their principal zones of activity and the reception of royal mobility by the elite.915 Note: the ARF mostly, but not with total consistency, groups Christmas of one year and Easter of the next together at the end of an entry (e.g. that Charlemagne celebrated Christmas 768 at Aachen and Easter 769 in Rouen is reported under 768). This has been adjusted for in the tables. Table 1. Palaces visited by Charlemagne according to the ARF (768–814) Year Campaign Assembly Easter Christmas Other Destination 768 - - - Aachen Noyon 769 Aquitaine - Rouen Düren - 770 - Worms Liège Mainz - 771 - Valenciennes Herstal Attigny Corbény 772 Saxony Worms Herstal Herstal 773 Italy Geneva Herstal (in camp) Thionville 774 Italy; army - Rome Quierzy Ingelheim sent to Sax- ony 914 The Annales Bertiniani and Annales Fuldenses present distinctively ‘West Frankish’ and ‘East Frankish’ narratives of the tumultuous events of Louis’ later reign. The Seligenstadt Annales Fuldenses narrative for the period 829–37 is brief and adds little to the more-detailed entries in the Annales Bertiniani. 915 Johann Friedrich Böhmer, Die Regesten des Kaiserreichs unter den Karolingern, 751–918, rev. Engelbert Mühlbacher and Johann Lechner, 2nd ed., vol. 1, bk. 1, Regesta Imperii 1/1 (Innsbruck: Verlag der Wagner’schen Universitäts-Buchhand- lung, 1908). 249 775 Saxony; Friuli Düren Quierzy Schlettstadt - 776 Friuli; Saxony Worms Treviso Herstal - 777 - Paderborn Nijmegen Douzy - 778 Spain; army - Chasseneuil Herstal - sent to Sax- ony 779 Saxony Düren Herstal Worms Compiègne; Verzenay 780 - - Worms Pavia - 781 - - Rome Quierzy Milan(?); Worms 782 Saxony "At the source Quierzy Thionville Cologne of the River Lippe" (ubi Lip- pia consurgit/ad fontem Lippiae) = nr. Pader- born 783 Saxony Paderborn Thionville Herstal Worms 784 Saxony (x2) Worms(?) Herstal Lügde, near - Schieder in the Weissgau 785 Saxony Paderborn Eresburg Attigny - 786 Army sent to Worms Attigny Florence - Brittany 787 Benevento; Worms Rome Ingelheim - Bavaria 788 vs. the Avars Ingelheim Ingelheim Aachen Regensburg 789 vs. the Wilzi - Aachen Worms Cologne 790 - - Worms Worms Salz 791 vs. the Avars Regensburg Worms Regensburg Ingelheim (Astrono- mer) 792 (No Cam- - Regensburg Regensburg - paign) 793 - - Regensburg St. Killian's, Salz (Astronomer) Würzburg 794 Saxony - Frankfurt Aachen - 795 Saxony Kostheim, nr. Aachen Aachen - Mainz 796 Saxony; army - Aachen Aachen - sent vs. Avars 797 Saxony - Aachen (Saxony - Her- Aachen stelle) 798 Saxony - (Saxony - Aachen - Herstelle) 250 799 Saxony - Aachen Aachen Paderborn 800 Benevento Mainz St.-Riquier Rome Aachen, Rouen (Centulum) 801 - - Rome Aachen Pavia 802 Army sent to - - Aachen - Saxony 803 - - - Aachen Salz 804 Saxony - - Quierzy Cologne, Aachen, Reims, 805 Army sent to - - Thionville Champ (Vosges), Aa- Bohemia chen, Metz, Remire- mont 806 Army sent vs. Thionville Nijmegen Aachen Seilles (Meuse) the Sorbs 807 Army sent to - Aachen Aachen - Corsica 808 Army sent vs. - Nijmegen Aachen - the Wilzi 809 - - Aachen - - 810 vs. the Danes - (Aachen?) - - 811 Armies sent Aachen - Aachen - vs. the Li- nones, to Pannonia, and to Brittany 812 vs. the Wilzi; Aachen - Aachen - army sent to defend Italy vs. Muslim raiders 813 - Aachen - Aachen - 814 - - - - - Table 2. Palaces visited by Louis the Pious as king of the Franks/emperor (814–840) Year Campaign Assembly Easter Christmas Other Destination 814 - Aachen - - Doué 815 vs. Danes Paderborn - Aachen Frankfurt (Thegan) 816 vs. the Sorbs - Aachen Reims, Compiègne 817 - Aachen (July) - (Aachen?) Nijmegen (hunting), Ingelheim, Chalon- sur-Saône 251 818 Brittany Vannes Aachen Aachen Angers, Rouen, Amiens, Cambrai, Herstal 819 Army sent Aachen (Janu- - Aachen Kreuznach, Bingen, from Italy to ary), Ingelheim Koblenz Pannonia; Pip- (July) pin of Aqui- taine sent to Gascony 820 vs. Ljudovit, Aachen (Janu- - Aachen - duke of Pan- ary), Quierzy nonia 821 vs. Ljudovit, Aachen (Febru- Aachen Worms Trier, Metz, duke of Pan- ary), Nijmegen Remiremont (hunt- nonia (May), Thion- ing) ville (October) 822 Army sent Attigny (Au- - Frankfurt - from Italy to gust); Frankfurt Pannonia 823 - Frankfurt - - Pavia, Worms (May), Com- piègne (No- vember) 824 Brittany Compiègne (24 - Aachen Rouen June) 825 - Aachen (May), Aachen (‘as Aachen Nijmegen (hunting), Aachen (Au- usual’) Remiremont (hunt- gust) ing), Nijmegen (hunt- ing) 826 - Ingelheim Aachen916 - Salz (June), Ingel- heim (October) 827 Army sent to Nijmegen - - Compiègne, Spanish March (Win- Quierzy 917 ter/Spring), Compiègne (Autumn) 828 - Aachen (Febru- - Aachen918 Frankfurt, Worms, ary); Ingelheim Thionville, Aachen (June) 916 The ARF annalist states that Louis wintered at Aachen 825/826, was still there in February 826, and only left in mid- May – making it likely that Easter had also been celebrated there. 917 ‘He himself [Louis] stayed between Compiègne, Quierzy, and other palaces neighbouring them until the beginning of winter’: ARF, a. 827, 173: ipse inter Compendium et Carisiacum caeteraque his vicina palatia usque ad hiberni temporis initium conversatus est. 918 ‘The emperor came to Aachen around Martinmas [11 November] and, once there, spent the entire winter in various assemblies convoked concerning urgent matters of the kingdom’: ARF, a. 828, 176: Imperator circa missam sancti Martini 252 829 vs. Danes Worms (Au- Aachen Aachen Frankfurt (Autumn (cancelled) gust) hunting) 830 Brittany Aachen (Febru- Unknown (on Aachen Compiègne (Louis ary), Nijmegen campaign) usurped) (October) 831 - Aachen (Febru- Aachen(?) Aachen - ary), Ingelheim (May), Thion- ville 832 vs. Louis 'the Orléans - Le Mans Tribur (camped with German' (planned, can- army), Augsburg(? celled), Mainz Received Louis' sub- (April), Orléans mission), Salz, Frank- (September) furt (Thegan), Limoges, 833 (Louis de- Worms Worms Aachen Aachen. Louis taken posed by Lo- (Spring), Com- to Soissons via Metz, thar with the piègne (Octo- and then Compiègne. help of his ber, held by Lothar meets Louis brothers) Lothar) the German at Mainz 834 vs. Lothar Paris (St.- Aachen Metz Orléans, Paris, Thi- Denis?), onville (wintered) Langres (Au- gust), Attigny (November) 835 - Thionville (Metz?) Aachen Ardennes (Autumn (February), AB: hunting) Tramoyes, nr Lyon, AF: Lyon (June) 836 - Thionville - Aachen Remiremont (hunt- (May), Worms ing), Frankfurt (hunt- (September) ing), Ingelheim (AF) 837 vs. Danes; vs. Aachen (Febru- - Aachen Paris Bretons ary, ecclesiasti- cal council), Thionville (May), Nijme- gen(?), Aachen 838 - Nijmegen - Frankfurt; Ver, Compiègne, and (May), Quierzy Mainz "other places in the (August) vicinity" (hunting); Attigny; Mainz Aquasgrani ad hiemandum venit ibique positus totum hiberni temporis spatium in diversis conventibus ob necessaria regni negotia congregatis inpendit. 253 839 vs. Louis 'the Worms (May), Bodman Poitiers Mainz; Frankfurt; German' Chalon-sur- Ingelheim; Kreuz- Saône (Septem- nach (castrum; ber) hunting) 840 - - - - - Table 3. Palaces visited by Pepin III as king (751–768) Year Campaign Assembly Easter Christ- Oth er Destination mas 751 - - - - Soissons 752 - - - - 1. Verberie (March: DD Ka- rol. 1, no. 1) 2. Herstal (April: DD Karol. 1, no. 2) 3. Werestein (May: DD Ka- rol. 1, no. 3) 753 Saxony - - Thion- Verberie (May: DD Karol., ville919 no. 4) 754 - Berny-Rivière Quierzy - 1. Ponthion (Continuationes (Continuationes Fredegarii, c. 36) Fredegarii, c. 37) 2. Verberie (DD Karol. 1, no., 7) 755 Italy - - - Compiègne (July: DD Karol. 1, no. 8) 756 Italy - - - - 757 - Compiègne - Cor- Attigny (August: DD Karol. beny920 1, no. 9) 758 Saxony - Corbeny - Düren (September: DD Ka- rol. 1, no. 10) 759 - - - Longlie Compiègne (October: DD r Karol. 1, no. 12) 760 Aquitaine - Jupille Quierzy 1. Attigny (June: DD Karol. 1, no. 13) 2. Verberie (June: DD Karol. 1, no. 14) 761 Aquitaine Düren Quierzy Quierzy - 762 Aquitaine - Quierzy Gentilly 1. Sinzig (Sentiaco palatio) (July: DD Karol. 1, no. 15) 919 That Christmas in 753 was celebrated in Thionville and Easter 754 in Quierzy is supplied in the D manuscript of the ARF (ARF, a. 753, 10); confirmed by Continuationes Fredegarii, c. 36, 183: Theudone villa publica super Mosella. 920 Reported in the C and D MSS: ARF, a. 757, 16. 254 2. Trisgodros villa puplica921 (DD Karol. 1, no. 16) 763 Aquitaine (x2) Nevers Gentilly Longlie Mellier (August: DD Karol. r 1, no. 18) 764 (no campaign) Worms Longlier Quierzy - 765 (no campaign) Attigny Quierzy Aachen - 766 Aquitaine (Orléans)922 Aachen Samous - sy 767 Aquitaine (x2) Bourges, ‘in the Vienne (or Bourges Gentilly field’ (in campo) Gentilly?)923 768 Aquitaine - ‘In the for- - Saintes?924 tress which is called Sels’ (in castro, qui dici- tur Sels) 921 Not definitively identified. Historians have traditionally associated with the pagus Trigorium and the municipality of Treis, on the Moselle. This has been recently questioned on linguistic grounds, with Triguères (Loiret) proposed as an alternative. Martina Pitz and Roland Puhl, 'Trisgodros = Triguères (Loiret) ? Pour une nouvelle localisation d'une villa publica énig- matique mentionnée dans une charte de Pépin le Bref', Nouvelle revue d'onomastique 49-50 (2008): 55-81. 922 Given as location of the assembly in the ARF. Two charters given by Pepin at Orléans are extant (DD Karol. 1, nos. 22-23), and they refer to Orléans not as palatium (vel. sim.) but civitas: Aurelianis civitate publice. Had the Merovingian palace in Orléans been deactivated? 923 The entry for 766 states that Pepin celebrated both Christmas and Easter (i.e. Easter 767) at Gentilly, but the next entry states that he celebrated Easter that year in Vienne after an ecclesiastical council in Gentilly. 924 The ARF states that Pepin ‘returned in triumph to Saintes’ (cum triumpho victoriae ad Sanctones reversus est. ARF, a. 768, 26), but it is unclear whether there was an existing royal palace there. 255 FIGURES Introduction Fig. 1. View of the so-called ‘Peristyle’ at the Palace of Diocletian, Split (Croatia). Late-third to early-fourth century CE. Photo: Wikimedia Commons. 256 Fig. 2. Axonometric plan of the Villa Romana del Casale outside Piazza Armerina, Sicily. Fourth century CE. Wulf- Rheidt, ‘Die Bedeutung der severischen Paläste für spätere Residenzbauten’, fig. 5. 257 Fig. 3. Axonometric reconstruction of the excavated villa at Cercadilla (Córdoba, Spain). Fourth century CE. Hidalgo Prieto, Espacio publico y espacio privado, fig. 5. 258 Chapter 1 Fig. 4. Plan of the fourth-century palace in the western quadrant of late ancient Milan. Shown in red are excavated structures associated with the ‘palace quarter’. In light blue are the remains of the circus. Società Lombarda di Ar- cheologia, http://milanoarcheologia.beniculturali.it. 259 Fig. 5. The so-called ‘Torre di Ansperto’, San Maurizio al Monastero Maggiore, originally a tower of the third-/fourth-century city walls. Photo: Wikimedia Commons. Fig. 6. Tower of San Maurizio al Monastero Maggiore, originally a tower of the carceres of the third-/fourth-century circus. Photo: Wikimedia Commons. 260 Fig. 7. Plan of the excavated remains of the representational group on the Via Brisa, Milan, probably belonging to the late-imperial palace. Mirabella Roberti, Milano romana, fig. 79. 261 Fig. 8. The so-called ‘Temple of Minerva Medica’, Rome. Fourth century CE. Photo: author. 262 Fig. 9. Plan of the excavations at the site at 2–4 Via Gorani, Milan. Società Lombarda di Archeologia, http://mila- noarcheologia.beniculturali.it. 263 Fig. 10. Plan of late-ancient Trier. Trier: Kaiserresidenz und Bischofssitz, 351, with labels added. A: Imperial baths. B. Constantinian aula/’Basilica’. C. Hypothesised location of the circus. D. Cathedral group. 264 Fig. 11. Exterior of the Constantinian aula. Early fourth century CE. Photo: Wikimedia Commons. 265 Fig. 12. Detail of the late-Roman ceiling painting discovered under the Trierer Dom. Early fourth-century CE. Mu- seum am Dom, Trier, Germany. Photo: ArtStor. 266 C B A D Fig. 13. Schematic plan of the palace-quarter in Thessaloniki. Dey, Afterlife, fig. 2.4, with labels added. A: Peristyle and representational spaces. B. The Arch of Galerius. C. Mausoleum. D. Circus. 267 A B D C F E Fig. 14. Plan of the Great Palace at Constantinople in the sixth century. Dey, Afterlife, fig. 3.4, with labels added. A. Hagia Sophia. B. Augustaion. C. Route of the Mese. D. Chalke Gate. E. Hippodrome. F. Palace 268 F A C B E D Fig. 15. Plan of the area of the Sessorian Palace in its fourth-century phase. Colli, ‘Il palazzo sessoriano’, fig. 23, with labels added. A. ‘Baths of Helena’ B. aula/Santa Croce in Gerusalemme. C. Apsidal aula. D. Amphiteatrum Castrense. E. Circus Vari- anus. F. fourth-century domus. 269 Fig. 16. Remains of the carceres of the Circus Varianus, by Santa Croce in Gerusalemme, Rome. Photo: author. 270 Fig. 17. The terminal wall (exterior) of the ‘Civil Basilica’ of the Sessorian Palace, Rome. Fourth century CE. Photo: author. 271 Fig. 18. View of the Palatine from the southwest across the Circus Maximus. Visible are the ‘arcades’ of the Severan and Maxentian periods. At centre left are the remains of baths attributed to Maxentius. Photo: author. 272 A B C D E Fig. 19. Plan of the Palatine residences (upper level). Architekturferrat des DAI. Ulrike Wulf-Rheidt, ‘The Palace of the Roman Emperors on the Palatine in Rome’, fig. 3, with labels added. Flavian remains, highlighted in orange, comprise the majority of the extant fabric. Later phases (dark green: early- Antonine, light green: Antonine; blue: Severan, pink: late ancient) are mostly additions and alterations of the existing structures. The Circus Maximus is located in the valley to the south west A: Aula Regia. B: triclinium/‘Cenatio Iovis’. C: ‘Garden stadium’. D: Severan Baths. E. Baths of Maxentius. 273 Fig. 20. Fragments 8f and 8g of the Severan Marble Plan. On the right-hand side is the pulvinar of the Circus Maximus. Photo: Stanford Digital Forma Urbis Romae Project, https://formaurbis.stanford.edu/fragment.php?record=31. 274 Fig. 21. Remains of the Severan baths on the southwestern platform of the Palatine (so-called Domus Severiana). Late-second/third century CE. Photo: author. 275 276 A D E B C Fig. 22. Plan of the ‘Villa of Maxentius’ on the Via Appia. Ioppolo and Pisani Sartorio, eds., La villa di Massenzio sulla Via Appia, vol. 2, fig. 1, with labels added. A. Villa. B. Circus. C. Mausoleum enclosure. D. Pulvinar. E. Judge’s tribunal. Fig. 23. The carceres of the Circus of Maxentius, Rome. 306–312 CE. At left is the wall of the mausoleum enclosure. Photo: author. Fig. 24. Elevation of the remains of the pulvinar at the Villa of Maxentius on the Via Appia. Pisani Sartorio and Calza, La villa di Massenzio sulla Via Appia, vol. 1, fig. 132. 277 Fig. 25. Apse of the aula in the villa of Maxentius, Rome. 306–312 CE. Photo: author. 278 279 Fig. 26. Plan of the centre of Rome in the fourth century. Ziemssen, ‘Die Kaiserresidenz Rom in der Zeit der Tetrarchie’, fig. 4. Fig. 27. Reconstruction of the southwestern Palatine in its Maxentian phase. Wulf-Rheidt, ‘Die schwierige Frage der Nutzung des römischen Kaiserpalastes auf dem Palatin in Rom in der Spätantike’, fig. 6. Fig. 28. View of the Basilica of Maxentius from the Palatine, Rome. 306–312 CE. Photo: author. 280 Fig. 29. Plan of the excavated structures on the Byrsa Hill, late-fifth or sixth century CE. Lézine, ‘Le “Palais” de Byrsa’, in Carthage-Utique, fig. 6, after Ferron and Pirard, ‘Les fouilles’, pl. 2. Fig. 30. Schematic plan of the representational structures on the Byrsa Hill: Lézine, ‘Le “Palais” de Byrsa’, in Carthage-Utique, fig. 7. 281 Fig. 31. Map of Ravenna, c. 530. Deliyannis, Ravenna, fig. 28. 282 Fig. 32. The so-called ‘Palace of Theoderic’ in its fourth phase, dated by Andrea Augenti to the early-sixth century CE. Andrea Augenti, ‘Archeologia e topografia a Ravenna’, fig. 8. 283 Fig. 33. Central field of the mosaic pavement of the triconchos, showing Bellerophon slaying the Chimera. Early sixth- century CE(?). Photo: Berti, Mosaici antichi in Italia, cat. n. 60. 284 Fig. 34. Palatium mosaic on the south nave wall of Sant’Apollinare Nuovo, Ravenna. Late-fifth/early-sixth century CE. Photo: author. Fig. 35. Detail of the Palatium mosaic, showing the severed hand of an excised figure. Sant’Apollinare Nuovo, Ra- venna. Photo: author. 285 Fig. 36. General plan of Reccopolis. Mid-sixth century CE. Henning et al., ‘Reccopolis Revealed’, fig. 3. 286 Fig. 37. Plan of the ‘palace compound’ of Reccopolis. Mid-sixth century CE. Henning et al., ‘Reccopolis Revealed’, fig. 4. 287 Fig. 38. View towards the basilica at Reccopolis. Mid-sixth century CE, with later additions. In the foreground are the remains of the narthex, with spolia column bases. Photo: author. Fig. 39. The central piers of the lower storey of the residence at Reccopolis, looking towards the east. Mid-sixth century CE. Photo: author. 288 Fig. 40. General plan of Caričin Grad (Justiniana Prima), with detail of the acropolis. Robert Ousterhout, Eastern Medieval Architecture, fig. 7.26. 289 Fig. 41. Chronological plan of the palace of Aachen superimposed on the modern street plan. Müller et al, ‘Pfalz und vicus Aachen’, fig. 25. 290 Fig. 42. Elevation of the palace chapel of Aachen. Late-eighth/early-ninth century CE. Kreusch, ‘Kirche, Atrium und Portikus der Aachener Pfalz’, figs. 4 and 5. 291 Fig. 43. Interior view of San Vitale, showing the rich mosaics of the presbytery (centre) and the column-screened exedrae to either side. San Vitale, Ravenna. 526–547 CE. Photo: author. 292 Fig. 44. Interior of the palace chapel of Aachen, looking upwards into the dome. Architecture of the late- eighth/early-ninth century, with later additions and restoration. Photo: Lessing/Art Resource, NY (ArtStor). 293 Fig. 45. Plan of the standing and excavated structures at Ingelheim. Lobbedey, ‘Carolingian Royal Palaces’, fig. 19b. Fig. 46. Isometric representation of the standing and excavated structures at Ingelheim. Christian Rauch/Forschung- stelle Kaiserpfalz Ingelheim. www.heritage-route.eu/ingelheim 294 Fig. 47. Plan of the villa at Montmaurin (Haute-Garonne, France). Fourth century CE. Sfameni, Ville residenziali, fig. 12. 295 Fig. 48. Detail of the palatium mosaic in Sant’Apollinare Nuovo, Ravenna, showing the clerestory of the palace façade. Photo: author. Fig. 49. Reconstruction of the villa at San Giovanni di Ruoti (Basilicata) in its late-fifth-century phase (Phase 3B). Small and Buck, Excavations, fig. 133. 296 Fig. 50. Plan and elevation of the audience chamber in the episcopium of Poreč in its sixth-century phase (under bishop Euphrasius). After Matejčić and Chevalier, ‘The Episcopium of Poreč’, fig. 3. 297 Figs. 51–52. Plan and axonometric reconstruction of the Merovingian ‘nobleman’s house’, Camp de Larina, Lyon. Samson, ‘The Merovingian nobleman’s home’, figs. 1–2. 298 Fig. 53. An early medieval domus on the Forum Transitorium, Rome. Ninth century CE. Photo: author. 299 Fig. 54. Plan of the palace of Paderborn ca. 799, overlain on the modern urban topography (shaded). Sveva Gai, ‘Die Pfalz Karls des Großen’, fig. 6. Fig. 55. Plan of the excavated palace structures in the region of Frankfurt Cathedral. Stamm, ‘Zur karolingischen Kö- nigspfalz in Frankfurt am Main’, fig. 3. 300 Figure 56. Reconstructed plan of the palace of Arechis in Salerno. Peduto et al., eds., Salerno. Una sede ducale della langobardia meridionale, plate 3, fig. 1. 301 Figs. 57. Arcade of the palace incorporated into later structures on the Via Dogana Vecchia, Salerno Photo: author Figs. 58–59. Details of the Corinthian capitals of the Arechian portico, Salerno. Photo: author. 302 Figure 60. Exterior of San Pietro a Corte, Salerno. Late eighth-century CE, with later additions. Photo: author. Fig. 61. Longitudinal section of San Pietro a Corte, Salerno. Paolo Peduto, 'Paolo Diacono e la Cappella Palatina di Salerno', fig. 5. 303 Fig. 62. View of the eastern façade of Santa María del Naranco, Oviedo. 842–850 CE. Photo: author. Fig. 63. Isometric view of Santa María del Naranco, Oviedo. Arbeiter and Noack-Haley, Christliche Denkmäler, fig. 101. 304 Fig. 64. Detail of a roundel located in the spandrel of the interior arcade of Santa María del Naranco. The image depicts a dragon with a whipping tail, surrounded by a vine motif. 842–850 CE. Photo: author. Fig. 65. Yeavering, in Hope-Taylor’s phase IIIc. Mid-seventh century CE. Hope-Taylor, Yeavering, fig. 77. 305 Fig. 66. Digital reconstruction of the so-called cuneus at Yeavering. Past Perfect. pastperfect.org.uk. Fig. 67. Reconstructed axonometric plan of Santa Croce in Gerusalemme in the fourth century. Krautheimer, Corpus basilicarum, fig. 117. 306 Fig. 68. Christ enthroned, flanked by angels, at the east end of the nave of Sant’Apollinare Nuovo, Ra- venna. Saint Martin, leading a procession of male saints into Christ’s presence, is visible at right. Late- fifth/early-sixth century CE. Photo: author. Figure 69. Reconstruction of the palace group of Alfonso II, 791–842 CE. Arbeiter and Noack-Haley, Christliche Denkmä- ler, fig. 6. 307 Figure 70. The ‘Torre Vieja de San Miguel’, 791–842 CE, on the south side of the Cathedral of San Salvador, Oviedo. Photo: author. Figure 71. The exterior of the Cámara Santa, 791–842 CE. Cathedral of San Salvador, Oviedo. Photo: author. 308 Chapter 2 Figs. 72–73. Two silver missoria with images of Constantius II, found near Kerch. 350–361 CE. Hermitage Museum, St. Petersburg. Photo: hermitagemuseum.org/ 309 Fig. 74. The Colonne di San Lorenzo, Milan. Erected 376–378 CE. Photo: Scala/ArtStor. Fig. 75. The submission of barbarians in the Hippodrome. Obelisk Base of Theodosius (north-western face), Istanbul. 390 CE. Photo: ArtStor. 310 Fig. 76. Fourth-century mosaic of Christ with the Apostles, handing the text of the law to Peter (traditio legis), located in the chapel of S. Aquilino. San Lorenzo, Milan. Photo: Giovanni Dall’Orto/Wikimedia Commons. Fig. 77. Three-dimensional reconstruction of the Septizodium, Rome. The Septizodium screened the conjunction of the palace with the Circus Maximus (see above, e.g. fig. 27). Franck Devedjian/Wikimedia Commons. 311 Fig. 78. Nunburnholme Cross, Nunburnholme, East Yorkshire. Tenth century CE. Photo: Corpus of Anglo-Saxon Stone Sculpture, vol. 3, pl. 709. 312 Fig. 79. Consular portrait of Constantius II in the Chronography of 354, f. 13r. 1620 CE. This MS derives from a Carolin- gian copy, known as the Codex Luxemburgensis, now lost. Vatican City, Biblioteca Apostolica Vaticana, Barb. lat. 2154. Photo: digi.vatlib.it 313 Fig. 80. Missorium of Theodosius, 388 CE. Real Academia de la Historia, Madrid. Photo: Real Academia de la Historia. 314 Fig. 81. Christ and Barabbas before Pilate. Codex Purpureus Rossanensis, f. 16v. Sixth century CE. Museo Diocesano, Rossano. Photo: Archive for Research on Archetypal Symbolism/ArtStor. 315 Fig. 82. Charles the Bald receiving his Bible from the monks of Tours. The First Bible of Charles the Bald, f. 423r. 845 CE. Paris, Bibliothèque nationale de France, MS. lat. 1. Photo: BnF Gallica. 316 Fig. 83. Commemorative gold triple-solidus of Theoderic (The ‘Senigallia Medallion’), c. 500 CE(?). Museo Nazionale Romano, Palazzo Massimo alle Terme. Photo: fotosar.it 317 Fig. 84. The Repton Stone. Seventh century CE? Derby Museum, Derby, UK. Photo: Wikimedia Commons. Fig. 85. The ‘Hornhausen Rider’. Seventh century CE? Photo: Landesamt für Denkmalpflege und Archäologie-Sachsen Anhalt, Halle, https://lda.sachsen-anhalt.de. 318 Fig. 86. The Worcester Hunt Mosaic, discovered at the villa at Daphne, near Antioch. 500-550 CE. Worcester Art Museum, Worcester, MA. Photo: Kondoleon, ed., Antioch: The Lost Ancient City, 66, fig. 2. 319 Fig. 87. The Little Hunt mosaic from the Villa del Casale Romana, Piazza Armerina, Sicily. Early fourth century CE. Photo: Andrew Malone/Wikimedia Commons. 320 Fig. 88. The so-called ‘Throne of Charlemagne’. Early-ninth century CE(?). Aachen Cathedral. Photo: Lessing/Art Resource, NY/ArtStor. 321 Fig. 89. View upwards into the cupola of the Palatine Chapel, Aachen. The viewer is oriented facing East. The original mosaics dated to the first decade of the ninth century; the Carolingian tesserae were replaced in 1880/1881. Photo: Wiki- media Commons. Fig. 90. Engraving representing the cupola mosaic of the Palatine Chapel, Aachen, from Giovanni Giusto Ciampini’s Vetera monimenta in quibus praecipua (1690). Photo: ArtStor. 322 Fig. 91. Exterior of the standing fabric of San Miguel de Lillo, Oviedo. 848 CE. Photo: author. Fig. 92. Isometric reconstruction of San Miguel de Lillo, 848 CE. Arbeiter and Noack-Haley, Christliche Denkmäler, fig. 83. 323 Fig. 93. Plan of San Julián de los Prados, Oviedo. Arbeiter and Noack-Haley, Christliche Denkmäler, fig. 58. 324 Fig. 94. Interior view towards the eastern apses in San Julián de los Prados, Oviedo. Photo: Alamy. 325 Figs. 95–97. Details of the ‘basilica’ motif of the frescoes of San Julián de los Prados, set within niches. From, respec- tively, the north and west walls of the transept, and from the western wall of the nave. Arbeiter and Noack-Haley, Christliche Denkmäler, pl. 17b–d. 326 Fig. 98. Monumental cross in the upper register of San Julián de los Prados, located over the apse-arch. Photo: Alamy. 327 Fig. 99. Plan of Santa Sofia, Benevento, showing the reconstructed phases of interventions in its structure. Belting, Studien zur beneventanischen Malerei, fig. 12. 328 Fig. 100. Southeastern exterior view of Santa Sofia, Benevento, showing the original fabric of the ‘zig-zag’ walls. Mid- eighth century. Photo: author. Fig. 101. View across the nave of Santa Sofia, Benevento, towards the eastern triple apse group. A vertex of the ‘zig- zag’ walls are just visible at right. Photo: author. 329 Fig. 102. Arechis enthroned and crowned, ordering the construction of Santa Sofia in the Chronicon Sanctae Sophiae. Early- twelfth century CE. Bibliotheca Apostolica Vaticana, Vat. Lat. 4939, f. 26r. Photo: Digital Vatican Library. 330 Chapter 3 Fig. 103. Mosaic of the CIVI(TAS) CLASSIS. Sant’Apollinare Nuovo, Ravenna. Late-fifth/early-sixth century CE. Photo: author. 331 Fig. 104. Detail of the ships in Classe’s harbour. Sant’Apollinare Nuovo, Ravenna. Photo: author. 332 Fig. 105. Palaces of the Ostrogothic rulers in Italy (493–553). Map produced with QGIS. 333 Fig. 106. Insignia of the (western) magister officiorum in the Munich manuscript of the Notitia Dignitatum (Not. Dig. Occ. IX), 1542/1550–1551. Bayerische Staatsbibliothek 10291, fol. 207v. Photo: MDZ Digitale Bibliothek. 334 Fig. 107. Plan of the fourth-century fort of Ad pirum (Hrušica): Ulbert, ed., AD PIRVM (Hrušica), appendix 1. 335 Fig. 108. Insignia of the comes Italiae in the Munich manuscript of the Notitia Dignitatum (Not. Dig. Occ. XXIV). Bayerische Staatsbibliothek 10291, fol. 214r. Photo: MDZ Digitale Bibliothek. 336 Fig. 109. Plan of late ancient Susa. Christie, From Constantine to Charlemagne, fig. 61. 337 Fig. 110. Plan of Brescia, mid-fourth century to sixth century, including the palatium in a separate wall enclosure in the western sector (12). Brogiolo, Brescia altomedievale, fig. 22. 338 Fig. 111. Plan of the archaeological area of Castelseprio. Brogiolo and Gelichi, Nuove ricerche, 121 fig. 38. 339 Fig. 112. Plan of Pavia in the seventh century CE. The palace of Theoderic, later that of the Lombard kings, was likely located in the eastern sector, adjacent to the Porta Palacense. Progetto Monasteri imperiali di Pavia. www.monasteriimperialipavia.it 340 Fig. 113. The façade of San Giovanni in Borgo (Pavia) before its destruction in 1818. Ferdinand de Dartein. Musei Civici di Pavia. 341 342 Fig. 114. The Iconografia Rateriana. Eighteenth century copy of a tenth century manuscript (based on a sixth century exemplar?). Verona, Bibliotheca capitolare MS CXIV. Photo: Wikimedia Commons. Fig. 115. Jerusalem on the Madaba Map Mosaic (Agios Georgios, Madaba, Jordan). Sixth century CE. Photo: Lessing/Art Resource NY. 343 Fig. 116. Plan of the Sirmione peninsula in Late Antiquity. Christie, From Constantine to Charlemagne, fig. 65. 344 Fig. 117. Plan of the site of the main building at Monte Barro (sector B). Brogiolo and Castelletti, Archeologia a Monte Barro, vol. 1, fig. 9. Fig. 118. Reconstruction of the grande edificio at Monte Barro (sector B). Brogiolo and Castelletti, Archeologia a Monte Barro, vol. 1, fig. 27. 345 Fig. 119. Bronze hanging crown from Monte Barro (Sector B) and a reconstruction. Brogiolo and Castelletti, Archeologia a Monte Barro, vol. 1, figs. 52 and 54. 346 Fig. 120. Map of public granaries in the Ostrogothic Kingdom of Italy. Map produced with QGIS. 347 Fig. 121. Reconstruction of the late-ancient topography of Ravenna and Classe, superimposed on a modern sat- ellite map. Augenti and Cirelli, ‘From Suburb to Port’, fig. 10.1. 348 Fig. 122. Plan of the excavations at Podere Chiavichetta, Ravenna. Augenti and Cirelli, ‘From Suburb to Port’, fig. 10.2. Fig. 123. Buildings 1 and 2 at Podere Chiavichetta, Ravenna, viewed from the south west. The course of the harbour channel is visible behind the remains of the excavated structures. Photo: author. 349 Chapter 4 Figs. 124–25. Sections I and II of the Tabula Peutingeriana. Österreichische Nationalbibliothek, MS Cod. 324, first half of the thirteenth century, Colmar/Reichenau?. Photo: ONB Digital. 350 351 Fig. 126. Gaul on the Tabula Peutingeriana (Sections I and II), from Konrad Miller’s 1888 reproduction. Wikimedia Commons. Fig. 127. Rome and Ostia on the Tabula Peutingeriana (Section IV, here from Miller’s 1888 reproduction). Wikimedia Commons. Fig. 128. ‘Francia’ on the Tabula Peutingeriana (Section II, from Miller’s 1888 reproduction). Wikimedia Commons. 352 Fig. 129. Plan of the Roman fort at Divitia (Deutz). Plan: Gundolf Precht. Caroll-Spilleke, ‘Das Militärlager Divitia in Köln-Deutz’, fig. 1. Fig. 130. Reconstruction of Cologne in the fourth century. Deutz is visible in the lower right. Römisch-Germanisches Museum, Cologne. Wikimedia Commons. 353 Fig. 131. Facsimile of the signet ring of Childeric I. The inscription reads CHILDERICI REGIS. Bibliothèque nationale de France, département Monnaies, médailles et antiques, inv.56.460. Photo: BnF Gallica. 354 Fig. 132. The palaces of Austrasia in the sixth- and early-seventh century. Map produced with QGIS. 355 Fig. 133. Paris in Late Antiquity and the Early Middle Ages. Dierkens and Périn, ‘Les sedes regiae mérovingiennes’, fig. 3. 356 Fig. 134. Palaces of the Pagus Parisiacus in the seventh century. Map: author. Produced with QGIS. 357 Fig. 135. Palaces of the Carolingian Empire mentioned in text. Map produced with QGIS. 358 BIBLIOGRAPHY I. Primary sources in modern editions Acta synhodorum habitarum Romae. In Cassiodori Senatoris Variae, edited by Theodor Mommsen, 395–455. MGH. AA 12. Berlin: Weidmann, 1894. Adomnán. De locis sanctis. Edited by Denis Meehan. Scriptores Latini Hiberniae 3. Dublin: Dublin Institute for Advanced Studies, 1958. Agnellus of Ravenna. Liber pontificalis ecclesiae Ravennatis. Edited by Deborah Mauskopf Deliyannis. Cor- pus Christianorum. Continuatio Mediaevalis 199. Turnhout: Brepols, 2006. al-Rāzī, Ahmad ibn Muhammad. Cronica del Moro Rasis. Edited by Diego Catalan and Maria Soledad de Andres. Fuentes cronísticas de la Historia de España. Madrid: Ediciones Gredos, 1975. Ambrose of Milan. Epistulae. In Sancti Ambrosii Opera. Edited by Michaela Zelzer. Corpus Scriptorum Ecclesiasticorum Latinorum 82/3. Vienna: Verlag der Österreichischen Akademie der Wis- senschaften, 1982. Ambrosio de Morales. Viaje de Ambrosio de Morales por orden del rey D. Felipe II, a los reinos de León y Galicia, y Principado de Asturias, para reconocer las reliquias de santos, sepulcros reales y libros manuscritos de las catedrales y monasterios. Edited by Enrique Florez. Gran Biblioteca Historica Asturiana: Parte Eclesiástica 1. Oviedo: Regareda y Comp., 1866. Ammianus Marcellinus. Ammiani Marcellini Rerum gestarum libri qui supersunt. Edited by Wolfgang Sey- farth. 2 vols. BSGRT. Leipzig: Teubner, 1978. Annales Bertiniani. Edited by Georg Waitz. MGH. SS Rer. Germ. 5. Hannover: Hahn, 1883. Annales Laureshamenses. In Annales et chronica aevi Carolini, edited by Georg Heinrich Pertz, 22–39. MGH. SS. 1. Hannover: Hahn, 1826. Annales Mettenses priores. Edited by Bernhard von Simson. MGH. SS rer. Germ. 10 (Hannover: Hahn, 1905. Annales mosellani. Edited by Johann Martin Lappenberg. In Scirptorum Tomus XVI ed. Georg Heinrich Pertz, 491–99. MGH. SS. 16. Hannover: Hahn, 1889. Annales Petaviani. In Scriptores, edited by Georg Heinrich Pertz, 7–18. MGH. SS. 1. Hannover: Hahn, 1826. Annales regni Francorum inde a. 741 usque ad 829, qui dicuntur Annales Laurissenses maiores et Einhardi. Edited by Georg Heinrich Pertz. Revised by Friedrich Kurze. MGH. SS rer. Germ. 6. Hannover: Hahn, 1895. Annales Xantenses. In Annales Xantenses et Annales Vedastini, edited by Bernhard von Simson, 1–39. MGH. SS rer. Germ. 12. Hannover: Hahn, 1909. 359 Anonymi Ticinensis Liber de laudibus civitatis ticinensis. Edited by Rodolfo Maiocchi and Ferruccio Quinta- valle. Rev. ed. Rerum Italicarum Scriptores 11/1. Città di Castello: Scipione Lapi, 1903. Anonymus Valesianus, pars posterior. In Excerpta Valesiana, edited by Jacques Moreau, 10–27. Revised by Velizar Velkov. BSGRT. Leipzig: Teubner, 1968. Anthologia Latina. Edited by David Shackleton-Bailey. Vol. 1, bk. 1. BSGRT. Stuttgart: Teubner, 1982. Astronomer. Vita Hludowici imperatoris. In Thegan, Die Taten Kaiser Ludwigs. Astronomus, Das Leben Kaiser Ludwigs, edited and translated by Ernst Tremp, 279–555. MGH. SS rer. Germ. 64. Hannover: Hahnsche Buchhandlung. Aurelius Victor. Sexti Aurelii Victoris Liber de Caesaribus. Edited by Franz Pichlmayr. BSGRT. Leipzig: Teubner, 1901. Ausonius. De herediolo. In Decimi Magni Ausonii Burgidalensis Opuscula, edited by Sextus Prete, 89–90. BSGRT. Leipzig: Teubner, 1978. ———. Ordo nobelium urbium. In Decimi Magni Ausonii Burdigalensis Opuscula, edited by Sesto Prete, 193– 201. BSGRT. Leipzig: Teubner, 1978. Bede. The Ecclesiastical History of the English People. Edited and translated by Bertram Colgrave and R. A. B. Mynors. Oxford Medieval Texts. Oxford: Clarendon Press, 1969. Boethius. De consolatione philosophiae, Opuscula theologica. BSGRT, 2nd ed. Munich: Saur, 2005. Boniface. Die Briefe des heiligen Bonifatius und Lullus. Edited by Michael Tangl. MGH. Epistulae selectae 1. Berlin: Weidmann, 1916. Boretius, Alfred. Capitularia regum Francorum. Vol. 1. MGH. Capit. 1. Hannover: Hahn, 1883. Boretius, Alfred, and Victor Krause. Capitularia regum Francorum. Vol. 2. MGH. Capit. 2. Hannover: Hahn, 1897. Carmen de conversione Saxonum. In Poetae latini aevi Carolini, edited by Ernst Dümmler, 1:380–381. MGH. Poetae 1 Berlin: Weidmann, 1881. Cassiodorus. Cassiodori Senatoris Variae. Edited by Theodor Mommsen. MGH. AA 12. Berlin: Weidmann, 1894. ———. Chronica. In Chronica minora, ed. Mommsen, 2:109–61. Chronica Adefonsi tertii regis. In Cronicas asturianas edited by Juan Gil Fernández, translated by Jose L. Moralejo, 114–49. Publicaciones del Departamento de Historia Medieval, Universidad de Oviedo 11. Oviedo: Universidad de Oviedo, 1985. Chronicon Albeldensis. In Cronicas asturianas, edited by Juan Gil Fernández, translated by Jose L. Moralejo, 153–88. Publicaciones del Departamento de Historia Medieval, Universidad de Oviedo 11. Oviedo: Universidad de Oviedo, 1985. 360 Chronicon Sanctae Sophiae (cod. Vat. Lat. 4939). Edited by Jean-Marie Martin. 2 vols. Fonti per la storia dell’Italia medievale 3. Rome: Istituto storico italiano per il Medio Evo, 2000. Chronography of 354. In Chronica minora saec. IV, V, VI, VII, edited by Theodor Mommsen, 1:13–196. MGH AA 9. Berlin: Weidmann, 1882. Claudian. De consulatu Stilichonis. In Claudii Claudiani Carmina, edited by John Barrie Hall, 190–238. BSGRT. Leipzig: Teubner, 1985. ———. Panegyricus dictus Honorio Augusto tertium consuli. In Claudii Claudiani Carmina, edited by John Barrie Hall, 52–85. BSGRT. Leipzig: Teubner, 1985. ———. Panegyricus dictus Probino et Olybrio consulibus. In Claudii Claudiani Carmina, edited by John Barrie Hall, 1–11. BSGRT. Leipzig: Teubner, 1985. Codex Carolinus. In Epistolae Merowingici et Karolini aevi, edited by Ernst Dümmler, 1:469–657. MGH. Epistolae 3. Berlin: Weidmann, 1892. Continuationes Fredegarii. In Fredegarii et aliorum chronica. Vitae sanctorum, edited by Bruno Krusch, 168– 93. MGH. SS rer. Merov. 2. Hannover: Hahn, 1888. Diehl, Ernst. Inscriptiones latinae christianae veteres. Vol. 1. Rev. ed.. Berlin: Weidmann, 1961. Duchesne, Louis, ed. Le Liber ponitificalis: texte, introduction et commentaire. 2 vols. Bibliothèque des écoles franc ̧aises d'Athènes et de Rome, 2nd series, 3. Paris: E. Thorin, 1886–92. Einhard. Vita Karoli magni. Edited by Georg Heinrich Pertz. Revised by Georg Waitz. 6th ed. Hannover: Hahn, 1911. Ennodius of Pavia. Dictiones. In Magni Felicis Ennodii Opera omnia. Edited by Wilhelm von Hartel. Cor- pus Scriptorum Ecclesiasticorum Latinorum 6. Vienna: C. Geroldi filium, 1882. ———. Eucharisticum de vita sua. In Magni Felicis Ennodii Opera omnia, edited by Wilhelm von Hartel, 393–401. Corpus Scriptorum Ecclesiasticorum Latinorum 6. Vienna: C. Geroldi filium, 1882. ———. Panegyricus dictus Clementissimo Regi Theoderico. In Magni Felicis Ennodi Opera, edited by Friedrich Vogel, 203–14. MGH. AA 7. Berlin: Weidmann, 1885. ———. Vita beatissimi Epifani episcopi Ticinensis ecclesiae In Magni Felicis Ennodi Opera, edited by Friedrich Vogel, 84–109. MGH. AA 7. Berlin: Weidmann, 1885. Eipstolae Austrasiacae. Edited by Wilhelm Gundlach. In Epistolae Merowingici et Karolini aevi, edited by Ernst Dümmler, 1:110–53. MGH. Epistolae 3. Berlin: Weidmann, 1892. Erchempert. Ystoriola Langobardorum beneventum degentium. In Erchempert, Piccola storia dei Longobardi di Benevento. Edited and translated by Luigi Andrea Berto. Nuovo Medioevo 94. Naples: Liguori, 2013. 361 Ermoldus Nigellus. In honorem Hludowici. In Poetae latini aevi Carolini, edited by Ernst Dümmler, 2:4–79. MGH. Poetae 2. Berlin: Weidmann, 1884. ———. In laudem Pippini regis. In In Poetae latini aevi Carolini, edited by Ernst Dümmler, 2:79–91. MGH. Poetae 2. Berlin: Weidmann, 1884. Fantuzzi, Marco ed. Monumenti ravennati de’ secoli di mezzo per la maggior parte inediti. 6 vols. Venice: Fran- cesco Andreola, 1801–1804. Fredegar. Chronicon. In Fredegarii et aliorum chronica. Vitae sanctorum, edited by Bruno Krusch, 1–168. MGH. SS rer. Merov. 2. Hannover: Hahn, 1888. Gesellschaft für Ältere Deutsche Geschichtskunde, ed. Die Urkunden der Karolinger. Vol. 1. MGH. DD Karol. 1. Hannover: Hahn, 1906. ———. Die Urkunden Konrad I. Heinrich I. und Otto I. Vol. 1. MGH. DD. Reg. et Imp. Germ. 1. Han- nover: Hahn, 1879. Gregory of Tours. De virtutibus S. Martini. In Gregorii episcopi Turonensis miracula et opera minora, edited by Bruno Krusch, 134–211. MGH. SS rer. Merov. 1/2. Hannover: Hahn, 1885. ———. Gregorii episcopi Turonensis libri Historiarum X. Edited by Bruno Krusch and Wilhelm Levison. 2nd ed. MGH. SS rer. Merov. 1/1. Hannover: Hahn, 1951. ———. Vita patrum. In Gregorii episcopi Turonensis Miracula et opera minora, edited by Bruno Krusch, 211– 94. MGH SS rer. Merov. 1/2. Hannover: Hahn, 1885. Heidrich, Ingrid, ed. Die Urkunden der Arnulfinger. MGH. DD Arnulf. Hannover: Hahn, 2011. Herren, Michael, ed. and trans. ‘The “De imagine Tetrici” of Walahfrid Strabo: Edition and Transla- tion’. Jorunal of Medieval Latin 1 (1991): 118–39. Hincmar of Reims. De ordine palatii. Edited and translated by Thomas Gross and Rudolf Schieffer. MGH. Fontes iuris 3. Hannover: Hahn, 1980. Historia corporis sancti Mercurii. In Acta passionis et translationes sanctorum martyrum Mercurii ac XII Fratrum, edited by Vittorio Giovardi, 55–62. Rome: Typis Joannis Baptistae a Caporalibus, 1730. Isidore of Seville. Isidori Hispalensis Episcopi Etymologiarum siue Originum Libri XX. Edited by W. M. Lindsay. 2 vols. Oxford Classical Texts. Oxford: Clarendon, 1911. ———. Historia Gothorum, Vandalorum et Sueuorum. In Chronica minora, ed. Mommsen, 2:267–303. Itinerarium Burdigalense. Edited by Paul Geyer and Otto Cuntz. In Itineraria et alia geographica, edited by Paul Geyer, Otto Cuntz, Ezio Francheschini, Robert Weber, Jean Fraipont, and Francois Glo- rie, 1–26. Corpus Christianorum. Series Latina 175. Turnhout: Brepols, 1965. 362 John of Biclaro. Chronicon. In Victor Tunnunensis Chronicon cum reliquiis ex Consularibus Caesaraugustanis et Iohannis Biclarensis Chronicon, edited by Carmen Cardelle de Hartmann, 59–83. Corpus Chris- tianorum. Series Latina 173A. Turnhout: Brepols, 2001. Jonas of Bobbio. Vita Columbani abbatis. In Ionae vitae sanctorum Columbani, Vedastis, Iohannis, edited by Bruno Krusch, 1–294. MGH. SS rer. Germ. 37. Hannover: Hahn, 1905. Jonas of Orleans. Le métier de roi. Edited and translated by Alain Dubreucq. Sources chrétiennes 407. Paris: Éditions du Cerf, 1995. Jordanes. De origine actibusque Getarum. In Iordanis Romana et Getica, edited by Theodor Mommsen, 53– 138. MGH. AA 5, no. 1. Berlin: Weidmann, 1882. Julian of Toledo. Historiae Wambae regis. In Passiones vitaeque sanctorum aevi merovingici, edited by Bruno Krusch and Wilhelm Levison, 3:486–535. MGH. SS rer. Merov. 5. Hannover: Hahn, 1910. Karolus magnus et Leo papa. Edited and translated by Franz Brunhölzl. In Helmut Beumann, Franz Brunhölzl, and Wilhelm Winkelmann, Karolus magnus et Leo papa. Ein Paderborner Epos vom Jahre 799, 55–98. Studien und Quellen zur Westfälischen Geschichte 8. Paderborn: Verein für Ge- schichte und Altertumskunde Westfalens, 1966. Kölzer, Theo. Die Urkunden der Merowinger. 2 vols. MGH. DD Merov. 1–2. Hannover: Hahn, 2001. Lactantius. De mortibus persecutorum. Edited and translated by J. L. Creed. Oxford Classical Texts. Ox- ford: Clarendon Press, 1984. Leges Alamannorum. Edited by Karl Lehmann. 2nd ed. MGH. LL. 5/1. Hannover: Hahn, 1966. Lex Ribuaria. Edited by Franz Beyerle and Rudolf Buchner. MGH. LL nat. Germ. 3/ 2. Hannover: Hahn, 1954. Libanius, Orationes. In Libanii Opera, edited by Richard Förster, vols. 1–4. BSGRT. Leipzig: Teubner, 1903. Liber Historiae Francorum. In Fredegarii et aliorum Chronica. Vitae sanctorum, edited by Bruno Krusch, 215– 328. MGH. SS rer. Merov. 2. Hannover: Hahn, 1888. Liudprand of Cremona. Antapodosis. In Liudprandi Cremonensis Opera Omnia, edited by Paolo Chiesa, 3– 150. Corpus Christianorum Continuatio Medaevalis 156. Turnhout: Brepols, 1998. Livy. T. Livi Ab urba condita libri. Edited by Wilhelm Weissenborn. Revised by Hermann Johannes Müller. Vol. 1. Berlin: Weidmann, 1885. Maassen, Friedrich, ed. Concilia aevi Merovingici. MGH. Conc. 1. Hannover: Hahn, 1893. Manaresi, Cesare, ed. I Placiti del «Regnum Italiae». Vol. 1. Rome: Tipografia del Senato, 1955. Marcellinus comes. Chronicon. In Chronica minora, ed. Mommsen, 2:37–108. 363 Marculfi Formulae. In Formulae Merowingici et Karolini aevi, edited by Karl Zeumer, 32–112. MGH. For- mulae. Hannover: Hahn, 1886. Marius of Avrenches. Chronica. In Chronica minora, ed. Theodor Mommsen, 2:225–39. Moduin. Ecloga. In Poetae latini aevi Carolini, edited by Ernst Dümmler, 1:382–92. MGH. Poetae 1 Ber- lin: Weidmann, 1881. Mommsen, Theodor, ed. Chronica minora saec. IV. V. VI. VII. 3 vols. MGH. AA. 9, 11, 13. Berlin: Weidmann, 1884–1898. Nithard. Nithardi Historiarum Libri IIII. Edited by Georg Heinrich Pertz. Revised by Ernst Müller. 3rd ed. MGH. SS rer. Germ 44. Hannover: Hahn, 1907. Notitia Galliarum. In Chronica minora., ed. Mommsen, 1:552–612. Notker. Die Taten Kaiser Karls des Grossen. Edited by Hans Haefele. Berlin: Weidmann, 1959. Olympiodorus. Historiae fragmenta. In Photius, Bibliothèque, edited and translated by René Henry. Paris: Les Belles Lettres, 1959. Pactus legis Salicae. Edited by Paul August Eckhardt. MGH. LL nat. Germ. 4/1. Hannover: Hahn, 1962. Passio Leudegarii episcopi et martyris Augustoddensis. Edited by Bruno Krusch. In Passiones vitaeque sanctorum aevi Merovingici, edited by Bruno Krusch and Wilhelm Levison, 3:249–362. MGH. SS rer. Merov. 5. Hannover: Hahn, 1910. Paul the Deacon. Historia Langobardorum. Edited by Ludwig Bethmann and Georg Waitz. In Scroptores rerum Langobardicarum et Italicarum, saec. VI–IX, 12–187. MGH. SS rer. Lang 1. Hanover: Hahn, 1878. ———. Die Gedichte des Paulus Diaconus: kritische und erklärende Ausgabe. Edited by Karl Neff. Quellen und Untersuchungen zur lateinischen Philologie des Mittelalters 3/4. Munich: Beck, 1908. Paulinus. Vita Ambrosii. Edited by A. A. R. Basiansen. Translated by Luca Canali. In Vite dei Santi, edited by Christine Mohrmann, 3:51–126. Scrittori greci e latini. Fondazione Lorenzo Valla: A. Mondadori, 1975. Philostorgius. Kirchengeschichte. Edited by Joseph Bidez. Revised by Friedhelm Winkelmann. Die Grie- chischen Christlichen Schriftsteller. Berlin: Akademie Verlag, 1981. Pighi, Giovanni Battista, ed. and trans. Versus de Verona, Versum de Mediolano civitate. Bologna: Nicola Zanicehlli, 1960. Preger, Theodor, ed. Scriptores originum Constantinopolitanarum. 2 vols. BSGRT. Leipzig: Teubner, 1901– 1907. Procopius. Procopii Caesariensis opera omnia. Edited by Jakob Haury. BSGRT. Leipzig: Teubner, 1905. 364 Ravennatis anonymi Cosmographia. In Itineraria Romana, edited by Joseph Schnetz, 21–110. Leipzig: Teub- ner, 1940. Sidonius Apollinaris. Carm. 22: Bvrgvs Pontii Leontii. Edited by Norbert Delhey. Untersuchungen zur antiken Literatur und Geschichte 40. Berlin: de Gruyter, 1993. ———. Epistulae. In Sidoine Apollinaire, edited and translated by André Loyen, vols. 2–3. Paris: Les Belles Lettres, 1970. Symmachus. Epistulae. In Q. Aurelii Symmachi quae superseunt, edited by Otto Seeck, 1–278. MGH. AA 6/1. Berlin: Weidmann, 1883. Tacitus. Germania. Edited by Wilhelm Reeb. Leipzig: Teubner, 1930. Thegan of Trier. Gesta Hludowici imperatoris. In Thegan, Die Taten Kaiser Ludwigs. Astronomus, Das Leben Kaiser Ludwigs, edited and translated by Ernst Tremp, 167–277. MGH. SS rer. Germ. 64. Han- nover: Hahnsche Buchhandlung. Theophanes. Chronographia. Edited by Carl de Boor. Leipzig: Teubner, 1883. Tjäder, Jan Olof, ed. Die nichtliterarischen lateinischen Papyri Italiens aus der Zeit 445–700. 3 vols. Skrifter utgivna av Svenska Instituet i Rom, 4th series, 19. Lund: C. W. K. Gleerup, 1954–82. Translatio sancti Mercurii. Edited by Georg Waitz. In Scriptores rerum Langobardicarum et Italicarum saec. VI– IX, 576–80. MGH. SS rer. Lang. 1. Hannover: Hahn, 1878. Venantius Fortunatus. Poèmes. Edited and translated by Marc Reydellet. 3 vols. Collection des univer- sités de France. Paris: Les Belles Lettres, 1994–2004. ———. Vita S. Martini. In Venantius Fortunatus, Oeuvres, edited and translated by Solange Quesnel, 4:1–102. Collection des universités de France. Paris: Les Belles Lettres, 2002. Vita Agili abbatis Resbacensis. Edited by Johannes Stilting. In Acta Sanctorum Augusti, edited by Joannes Pinio, Guilielmo Cupero, and Johannes Stelting, 6:574–87. Paris: Victor Palmé, 1868. Vita Audoini episcopi Rotomagensis. In Passiones vitaeque sanctorum aevi Merovingici, edited by Bruno Krusch and Wilhelm Levison, 3:536–67. MGH. SS rer. Merov. 5. Hannover: Hahn, 1910. Vita Balthildis. In Fredegarii et aliorum chronica. Vitae sanctorum, edited by Bruno Krusch, 476–508. MGH. SS rer. Merov. 2. Hannover: Hahn, 1888. Vita Eligii episcopi Noviomagensis. In Passiones vitaeque sanctorum aevi Merovingici, edited by Bruno Krusch, 2:634–742. MGH. SS. Rer. Mer. 4. Hannover: Hahnsche, 1902. Vita Genovefae virginis Parisiensis. In Passiones vitaeque sanctorum aevi Merovingici et antiquiorum aliquot, edited by Bruno Krusch, 204–38. MGH SS rer. Merov. 3. Hannover: Hahn, 1896. 365 Vitruvius. De l’Architecture. Edited and translated (French) by Philippe Fleury, Pierre Gros, Catherine Saliou, Louis Callebat, Bernard Liou, Michel Zuinghedau, and Jean Soubiran. 10 vols. Paris: Les Belles Lettres, 1969–2009. Walser, Gerold, ed. Die Einsiedler Inschriftensammlung und der Pilgerführer durch Rom: Codex Einsidlensis 326. Wiesbaden: Steiner, 1987. Wipo. Gesta Chuonradi II. imperatoris. In Die Werke Wipos, edited by Harry Bresslau, 1–62. 3rd ed. MGH. SS. rer. Germ. 61. Hannover: Hahn, 1915. Westerbergh, Ulla. Chronicon Salernitanum: A Critical Edition with Studies on Literary and Historical Sources and on Language. Studia Latina Stockholmiensia 3. Stockholm: Almsqvist & Wiksell, 1956. XII Panegyrici Latini. Edited by R. A. B. Mynors. Oxford Classical Texts. Oxford: Clarendon, 1964. II. Primary sources in translation Agnellus of Ravenna. The Book of Pontiffs of the Church of Ravenna. Translated by Deborah Mauskopf Deliyannis. Medieval Texts in Translation. Washington, DC: Catholic University of America Press, 2004. Ammianus Marcellinus. The Later Roman Empire, A.D. 354–378. Translated by Walter Hamilton. Pen- guin Classics. Harmondsworth: Penguin, 1986. Aurelius Victor. Liber de Caesaribus of Sextus Aurelius Victor. Translated by H. W. Bird. TTH 17. Liver- pool: Liverpool University Press, 1994. Ausonius. The Order of Famous Cities. In Ausonius, translated by Hugh G. Evelyn White, 269–86. Loeb Classical Library 96. Cambridge, MA: Harvard University Press, 1919. Baxter Wolf, Kenneth, trans. Conquerors and Chroniclers of Early Medieval Spain. 2nd ed. TTH 9. Liverpool: Liverpool University Press, 1990. Blockley, R. C., trans. The Fragmentary Classicising Historians of the Later Roman Empire: Eunapius, Olympi- odorus, Priscus and Malchus. Vol. 2. ARCA Classical and Medieval Texts, Papers and Monographs 10. Liverpool: Francis Cairns, 1983. Cassiodorus. Selected Variae of Magnus Aurelius Cassiodorus Senator. Translated by Samuel J. B. Barnish. TTH 12. Liverpool: Liverpool University Press 1992. ———. The Variae: The Complete Translation. Translated by M. Shane Bjornlie. Berkeley, CA: University of California Press, 2019. Davis, Raymond, trans. The Book of Pontiffs (Liber Pontificalis): The Ancient Biographies of the First Ninety Roman Bishops to AD 715. Revised ed. TTH 6. Liverpool: Liverpool University Press, 2010. ———. The Lives of the Eighth-Century Popes (Liber Pontificalis): The Ancient Biographies of Nine Popes from AD 715 to AD 817. Revised ed. TTH 13. Liverpool: Liverpool University Press, 2007. 366 ———. The Lives of the Ninth-Century Popes (Liber Pontificalis): The Ancient Biographies of Ten Popes from AD 817–891. TTH 20. Liverpool: Liverpool University Press, 1995. Fouracre, Paul, and Richard Gerberding, trans. Late Merovingian France: History and Hagiography, 640– 720. Manchester Medieval Sources. Manchester: Manchester University Press, 1996. Godman, Peter, trans. Poetry of the Carolingian Renaissance. London: Duckworth, 1985. Herlihy, David, trans. The History of Feudalism. The Documentary History of Western Civilization. Lon- don: Macmillan, 1970. Isidore of Seville. The Etymologies of Isidore of Seville. Translated by Stephen A. Barney, W. J. Lewis, J. A. Beach, and Oliver Berghof. Cambridge: Cambridge University Press, 2006. Liudprand of Cremona. The Complete Works of Liudprand of Cremona. Translated by Paolo Squatriti. Me- dieval Texts in Translation. Washington, D.C.: Catholic University of America Press, 2007. Nelson, Janet, trans. The Annals of St. Bertin. Ninth-Century Histories 1. Manchester: Manchester Uni- versity Press, 1991. Nixon, C. E. V. and Barbara Saylor Rodgers, trans. In Praise of Later Roman Emperors: The Panegyrici Latini. The Transformation of the Classical Heritage 21. Berkeley, CA: University of California Press, 1994. Noble, Thomas F. X., trans. Charlemagne and Louis the Pious: Lives by Einhard, Notker, Ermoldus, Thegan, and the Astronomer. University Park, PA: Pennsylvania State University Press, 2009. Procopius. The Wars of Justinian. Translated by H. B. Dewing. Revised by Anthony Kaldellis. Indian- apolis: Hackett, 2014. Rio, Alice. The Formularies of Angers and Marculf: Two Merovingian Legal Handbooks. TTH 46. Liverpool: Liverpool University Press, 2008. Vitruvius. Ten Books on Architecture. Translated by Ingrid Rowland. Cambridge: Cambridge University Press, 1999. Venantius Fortunatus. Poems. Translated by Michael Roberts. Dumbarton Oaks Medieval Library 46. Cambridge, MA: Cambridge University Press, 2017. III. Secondary Sources Abela, Elisabetta. ‘Lucca’. In Archeologia urbana in Toscana: la città altomedievale, edited by Sauro Gelichi, 23–44. Documenti di archeologia 17. Mantua: Società Archeologica Padana, 1999. Airlie, Stuart. ‘The Aristocracy in the Service of the State in the Carolingian Period’. In Staat im frühen Mittelalter, ed. Airlie et al., 93–111. ———. ‘Charlemagne and the Aristocracy: Captains and Kings’. In Charlemagne: Empire and Society, edited by Joanna Story, 90–102. Manchester: Manchester University Press, 2005. 367 ———. ‘The Palace Complex’. In Diverging Paths? The Shapes of Power and Institutions in Medieval Chris- tendom and Islam, edited by John Hudson and Ana Rodriguez, 255–90. Leiden: Brill, 2014. ———. ‘The Palace of Memory: The Carolingian Court as Political Centre’. In Courts and Regions in Medieval Europe, edited by Sarah Rees Jones, Richard Marks, and A. J. Minnis, 1–20. Wood- bridge: York Medieval Press, 2000. ———. ‘Talking Heads: Assemblies in Early Medieval Germany’. In Political Assemblies in the Earlier Middle Ages, edited by Paul Barnwell and Marco Mostert, 29–46. Studies in the Early Middle Ages 7. Turnhout: Brepols, 2003. Airlie, Stuart, Walter Pohl, and Helmut Reimitz, eds. Staat im frühen Mittelalter. Forschungen zur Ge- schichte des Mittelalters 11. Vienna: Verlag der Österreichischen Akademie der Wissenschaf- ten, 2006. Albu, Emily. The Medieval Peutinger Map: Imperial Roman Revival in a German Empire. Cambridge: Cam- bridge University Press, 2014. Allsen, Thomas. The Royal Hunt in Eurasian History. Philadelphia: University of Pennsylvania Press, 2006. Althoff, Gerd. Spielregeln der Politik im Mittelalter: Kommunikation in Frieden und Fehde. Darmstadt: Primus, 1997. Alto Bauer, Franz. Stadt, Platz und Denkmal in der Spätantike: Untersuchungen zur Austattung des öffentlichen Raums in den spätantiken Städten Rom, Konstantinopel und Ephesos. Mainz: Phillipp von Zabern, 1996. Anderson, Benjamin. ‘In Search of Late Roman Porticoes’. Review of The Afterlife of the Roman City: Architecture and Ceremony in Late Antiquity and the Early Middle Ages, by Hendrik Dey. Journal of Roman Archaeology 29, no. 2 (2016): 971–74. Anderson, Glaire. The Islamic Villa in Early Medieval Iberia. Farnham: Ashgate, 2013. Arbeiter, Achim, and Sabine Noack-Haley. Christliche Denkmäler des frühen Mittelalters vom 8. bis ins 11. Jahrhundert. Hispania Antiqua. Mainz: Phillipp von Zabern, 1999. Arce, Javier. ‘The so-called Visigothic “Palatium” of Recópolis (Spain): An Archaeological and His- torical Analysis’. in The Emperor’s House, ed. Featherstone et al., 63–70. Archer, John. Architecture and Suburbia: from English Villa to American Dream House, 1690–2000. Minne- apolis, MN: University of Minnesota Press, 2005. Arnold, Felix. Islamic Palace Architecture in the Western Mediterranean: A History. Oxford: Oxford Univer- sitz Press, 2017. 368 Arnold, Jonathan. ‘Ostrogothic Provinces: Administration and Ideology’. In Companion to Ostrogothic Italy, ed. Arnold et al., 73–97. ———. Theoderic and the Roman Imperial Restoration. Cambridge: Cambridge University Press, 2014. ———. ‘Theoderic’s Invincible Mustache’. JLA 6, no. 1 (2013): 152–83. Arnold, Jonathan, M. Shane Bjornlie, and Kristina Sessa, eds. A Companion to Ostrogothic Italy. Brill’s Companions to European History. Leiden: Brill, 2016. Audollent, Auguste. Carthage romaine, 146 avant Jésus-Christ–698 après Jésus-Christ. Paris: Thorin et Fils, 1901. Augenti, Andrea. ‘Archeologia e topografia a Ravenna: il Palazzo di Teoderico e la Moneta Aurea’, Archeologia Medievale 32 (2005): 7–23. ———. Il Palatino nel medioevo: archeologia e topografia (secoli VI–XIII). Bullettino della Commissione Ar- cheologica Comunale di Roma, Supplementi 4. Rome: “L’Erma” di Bretschneider, 1996. ———. ‘The Palace of Theoderic at Ravenna: A New Analysis of the Complex’. In Housing in Late Antiquity: From Palaces to Shops, edited by Luke Lavan, Lale Özgenel, and Alexander Sarantis, 425–53. Late Antique Archaeology 3/2. Leiden: Brill, 2007. Augenti, Andrea, and Enrico Cirelli. ‘From Suburb to Port: The Rise (and Fall) of Classe as a Centre of Trade and Redistribution’. In Rome, Portus and the Mediterranean, edited by Simon Keay, 205– 21. Archaeological Monographs of the British School at Rome 21. London: British School at Rome, 2012. Ball, Larry. The Domus Aurea and the Roman Architectural Revolution. Cambridge: Cambridge University Press, 2003. Balmaseda Muncharaz, Luis J. ‘La escultura de Recópolis’. in Recópolis y la ciudad, ed. Olmo Enciso, 142–57. Balzer, Manfred. ‘Paderborn als karolingischer Pfalzort’. In Deutsche Königspfalzen 3, 9–85 Bandmann, Günter. Early Medieval Architecture as Bearer of Meaning. Translated by Kendall Wallis. New York: Columbia University Press, 2005. ———. ‘Die Vorbilder der Aachener Pfalzkapelle’. In Karolingische Kunst, edited by Wolfgang Braunfels and Hermann Schnitzler, 424–62. Vol. 3 of Karl der Große. Lebenswerk und Nachleben, edited by Wolfgang Braunfels. Düsseldorf: L. Schwann, 1965. Barbier, Josiane. ‘Le système palatial franc : genèse et fonctionnement dans le nord-ouest du regnum’. Bibliothèque de l’École des Chartes 148, no. 2 (1990): 245–99 369 ———. ‘Les lieux du pouvoir en Gaule franque: l’exemple des palais’. In Places of Power – Orte der Herrschaft – Lieux du Pouvoir, edited by Caspar Ehlers, 227–46. Deutsche Königspfalzen 8. Göt- tingen: Vandenhoeck & Ruprecht, 2007. ———. ‘Les rois mérovingiens entre fausse immobilité et fausse itinérance’. In Le gouvernement en dé- placement. Pouvoir et mobilité de l’Antiquité à nos jours, edited by Sylvain Destephen, Josiane Barbier, and François Chausson, 189–212. Rennes : Presses Universitaires de Rennes, 2019. ———. ‘Un rituel politique à la cour mérovingienne : l’audience royale’. In L’Audience. Rituels et cadres spatiaux dans l’Antiquité et le haut Moyen Âge, edited by Jean-Pierre Caillet and Michel Sot, 241– 63. Textes, images et monuments de l’Antiquité au Moyen Âge. Paris: Editions Picard, 2007. Barry, Fabio. ‘Walking on Water: Cosmic Floors in Antiquity and the Middle Ages’. The Art Bulletin 89, no. 4 (2007): 627–56. Basić, Ivan. ‘Diocletian’s villa in Late Antique and Early Medieval Historiography: A Reconsideration’. Hortus Artium Medievalium 20, no. 1 (2014): 63–76. Bell, Catherine. Ritual Theory, Ritual Practice. Oxford: Oxford University Press, 1992. Bell, Peter. Social Conflict in the Age of Justinian: Its Nature, Management, and Mediation. Oxford: Oxford University Press, 2013. Belting, Hans. ‘Studien zum Beneventanischen Hof im 8. Jahrhundert’. DOP 16 (1962): 141–93. ———.Studien zur beneventanischen Malerei. Forschungen zur Kunstgeschichte und christlichen Archä- ologie 7. Wiesbaden: Franz Steiner, 1968. Ben Abed, Aïcha, and Noël Duval. ‘Carthage, la capitale du royaume et les villes de Tunisie à l’époque vandale’. In Sedes regiae ann. 400–800), ed. Ripoll and Gurt, 163–218. Berges, Wilhelm. ‘Das Reich ohne Hauptstadt’. In Das Hauptstadtproblem in der Geschichte. Festgabe zum 90. Geburtstag Friedrich Meineckes gewidmet vom Friedrich-Meinecke-Institut an der Freien Universität Berlin, edited by the Mitarbeiter des Friedrich-Meinecke-Instituts, 1–29. Tübingen: Niemeyer, 1952. Bermond Montanari, Giovanna. ‘La zona archeologica di Palazzolo’. Corso di cultura sull’arte ravennate e bizantina 30 (1983): 17–21. Bernhardt, John William. Itinerant Kingship and Royal Monasteries in Early Medieval Germany, c. 936–1075, CSMLT, 4th Series, 21. Cambridge: Cambridge University Press, 1993. Berti, Fede. Mosaici antichi in Italia. VIII: Aemilia: Ravenna. Roma: Istituto Poligrafico e Zecca dello Stato, 1976. Biddle, Martin, and Birthe Kjølbye-Biddle. ‘The Repton Stone’. Anglo-Saxon England 14 (1985): 233– 92. 370 Bijsterveld, Arnoud-Jan A. ‘The Medieval Gift as Agent of Social Bonding and Political Power: A Comparative Approach’. In Medieval Transformations: Texts, Power, and Gifts in Context, edited by Esther Cohen and Mayke de Jong, 123–56. Cultures Beliefs and Traditions 11. Leiden: Brill, 2001. Binon, Stéphanie. Essai sur le cycle de Saint Mercure: Martyr de Dèce et meurtrier de l’empereur Julien. Biblio- thèque de l’École des Hautes Études. Sciences Religieuses 53. Paris: Librarie Ernest Leroux, 1937. Bjornlie, M. Shane. ‘Governmental Administration’. In Companion to Ostrogothic Italy, ed. Arnold et al., 47–72. ———. Politics and Tradition between Rome, Ravenna, and Constantinople: A Study of Cassiodorus and the ‘Var- iae’, 527–554. CSMLT, 4th series 89. Cambridge: Cambridge University Press, 2013. Blair, John. Building Anglo-Saxon England. Princeton, NJ: Princeton University Press, 2018. ———. ‘Palaces or Minsters? Northampton and Cheddar Reconsidered’. Anglo Saxon England 25 (1996): 97–121. Bockmann, Ralf. Capital continuous. A Study of Vandal Carthage and Central North Africa from an Archaeo- logical Perspective. Spa ̈tantike—Fruḧ es Christentum—Byzanz. Kunst im ersten Jahrtausend. Se- ries B, 37. Wiesbaden: Reichert, 2013. Bondesan, Marco. ‘L’area deltizia padana: caratteri geografici e geomorfologici’. In Il Parco del Delta del Po: studi ed immagini, edited by Carlo Bassi, 2:9–48. Ferrara: Spazio Libri, 1990. Bonnet, Charles, and Alain Peillex. Les fouilles de la cathédrale Saint-Pierre de Genève. Les édifices chrétiens et le groupe épiscopal. Mémoires et documents 65. Geneva : Société d'histoire et d'archéologie de Genève, 2012. Borgia, Esliabetta, Donato Colli, Sergio Palladino, and Claudia Paterna. ‘Horti spei veteris e Palatium Sessorianum: nuove acquisizioni da interventi urbani 1996–2008’. In Roma. Scavi archeologici e sco- perte degli ultimi 10 anni, edited by Helga Di Giuseppe and Elizabeth Fentress, 101–42. Fold&Raccolte 1. Rome: Scienze e Lettere, 2018. Bosl, Karl. ‘Pfalzen und Forsten’. In Deutsche Königspfalzen 1, edited by Adolf Gauert, 1–29. Göt- tingen: Vandenhoeck & Ruprecht, 1963. Bovini, Giuseppe. ‘Antichi rifacimenti nei mosaici di S. Apollinare Nuovo di Ravenna’. Corso di cultura sull’arte ravennate e bizantina 13 (1966): 51–81. Bowes, Kim. ‘Christianization of Villas’. In The Roman Villa in the Mediterranean Basin: Late Republic to Late Antiquity, edited by Annalisa Marzano and Guy Métraux, 453–64. Cambridge: Cambridge University Press, 2018. ———. Houses and Society in the Later Roman Empire. Duckworth Debates in Archaeology. London: Duckworth, 2012. 371 Böhmer, Johann Friedrich. Die Regesten des Kaiserreichs unter den Karolingern, 751–918. Revised by Engel- bert Mühlbacher and Johann Lechner. 2nd ed. Regesta Imperii 1 Innsbruck: Verlag der Wag- ner’schen Universitäts-Buchhandlung, 1908. Böhner, Kurt. ‘Die Reliefplatten von Hornhausen’. Jahrbuch des Römisch-Germanischen Zentralmuseums 23- 24 (1976–1977): 89–138. Brandenburg, Hugo. Ancient Churches of Rome from the Fourth to the Seventh Century: The Dawn of Christian Architecture in the West. Bibliothèque de l’Antiquité Tardive 8. Turnhout: Brepols, 2005. Braunfels, Wolfgang, ed. Karl der Grosse. Lebenswerk und Nachleben. 5 vols. Düsseldorf: L. Schwann, 1965–1968. Brenk, Beat. The Apse, the Image, and the Icon: An Historical Perspective of the Apse as a Space for Images. Spätantike –Frühes Christentum – Byzanz 26. Wiesbaden: Reichert, 2010. Brennan, Peter. ‘The Notitia Dignitatum’. In Les littératures techniques dans l’Antiquité romaine : statut, public et destination, tradition, edited by Claude Nicolet 147–78. Geneva: Fondation Hardt, 1996. Brogiolo, Gian Pietro. Brescia altomedievale. Urbanistica ed edilizia dal IV al IX secolo. Documenti di ar- cheologica 2. Mantua: Società Archeologica Padana, 1993. ———. ‘Costruire castelli nell’arco alpino tra V e VI secolo’. Archeologia Medievale 41 (special issue) (2014): 144–56. ———. ed. Le fortificazioni del Garda e i sistemi di difesa dell’Italia settentrionale tra tardo antico e alto medioevo. Mantua: Società Archeologica Padana, 1999. ———. ‘Un’enclave bizantina sul Lago di Garda?’ In Le fortificazioni del Garda, ed. Brogiolo, 13–20. Brogiolo, Gian Pietro, and Lanfredo Castelletti, eds. Archeologia a Monte Barro. 2 vols. Lecco: Stefa- noni, 1991–2001. Brogiolo, Gian Pietro, and Silvia Lusuardi Siena. ‘Nuove indagini archeologiche a Castelseprio’. In Atti del 6° congresso internazionale di studi sull’alto medioevo, 2:475–500. Spoleto: CISAM, 1980. Brogiolo, Gian Pietro, and Sauro Gelichi. Nuove ricerche sui castelli altomedievali in Italia settentrionale. Flor- ence: Alll’Insegna del Giglio, 1996. Brooks Hedstrom, Darlene. The Monastic Landscape of Late Antique Egypt: An Archaeological Reconstruc- tion. Cambridge: Cambridge University Press, 2017. Bruno, Daniela. ‘Region X. Palatium’. In The Atlas of Ancient Rome: Biography and Portraits of the City, edited by Andrea Carandini and Paolo Carafa, 215–80. Translated by Andrew Campbell Ha- lavais. Princeton: Princeton University Press, 2017. Brühl, Carlrichard. ‘Das “Palatium” von Pavia und die “Honorantiae Civitatis Papiae”’. In Atti del 4° Congresso internazionale di studi sull’Alto Medioevo, 189–220. Spoleto: CISAM, 1969. 372 ———. Fodrum, Gistum, Servitium Regis. Studien zu den wirtschaftlichen Grundlagen des Königtums im Franken- reich und in den fränkischen Nachfolgestaaten Deutschland, Frankreich und Italien vom 6. bis zur Mitte des 14. Jahrhunderts. 2 vols. Kölner Historische Abhandlungen 14. Cologne: Böhlau, 1968. ———. ‘Fränkischer Krönungsbrauch und das Problem der “Festkrönungen”’. Historische Zeitschrift 194, no. 2 (1961): 265–326. ———. Palatium und Civitas: Studien zur Profantopographie spätantiker Civitates vom 3. bis zum 13. Jahrhundert. 2 vols. Cologne: Böhlau, 1975–1990. ———. ‘Remarques sur les notions de «capitale» et de «résidence» pendant le haut Moyen Âge’. Journal des savants, 1967, 193–215. ———. ‘Zum Hauptstadtproblem im Mittelalter’. In Festschrift für Harald Keller: zum sechzigsten Ge- burtstag, dargebracht von seinen Schülern, edited by Hans Martin von Erffa and Elisabeth Herget, 45–70. Darmstadt: E. Roether, 1963. Buc, Philippe. The Dangers of Ritual: Between Early Medieval Texts and Social Scientific Theory. Princeton, NJ: Princeton University Press, 2001. Bullough, Donald. ‘Aula Renovata: The Carolingian Court before the Aachen Palace’. In Donald Bullough, Carolingian Renewal: Sources and Heritage, 123–60. Manchester: Manchester University Press, 1991. ———. ‘Imagines regum and the Early Medieval West’. In Donald Bullough, Carolingian Renewal: Sources and Heritage, 39–96. Manchester: Manchester University Press, 1991. ———. ‘Urban Change in Early Medieval Italy: The Example of Pavia’. PBSR 34 (1966): 82–130. Cagiano de Azevedo, Michelangelo. ‘I palazzi tardoantichi e altomedievali’. In Atti del XVI congresso di storia dell’architettura, Athene, 29 settembre – 5 ottobre 1969, 303–18. Rome: Centro di Studi per la Storia dell’Architettura, 1977. Callander Murray, Alexander. ‘Immunity, Nobility, and the Edict of Paris’. Speculum 69, no. 1 (1994): 18–39. ———. ‘The New MGH Edition of the Charters of the Merovingian Kings’. Journal of Medieval Latin 15 (2005): 263–68. Cameron, Alan. Circus Factions: Blues and Greens at Rome and Constantinople. Oxford: Clarendon Press, 1976. ———. ‘Flavius: A Nicety of Protocol’. Latomus 47 (1988): 26–33. Campbell, Darryl. ‘The Capitulare de Villis, the Brevium exempla, and the Carolingian Court at Aachen’. EME 18, no. 3 (2010): 243–64. 373 Cannadine, David, and Simon Price, eds. Rituals of Royalty: Power and Ceremonial in Traditional Societies. Past and Present Publications. Cambridge: Cambridge University Press, 1987. Caporusso, Donatella. ‘La zona di corso di Porta Romana in età romana e medioevale’. In Scavi MM3. Ricerche di archeologia urbana a Milano durante la costruzione della linea 3, edited by Donatella Capo- russo, 1:251–57. Milan: Edizioni ET, 1991. Carandini, Andrea, Andreina Ricci, and Mariette de Vos. Filosofiana. La villa di Piazza Armerina: immagine di un aristocratico romano al tempo di Costantino. Palermo: Flaccovio, 1982. Carioli, Cristina. ‘Note sul palatium e la Moneta Aurea a Ravenna’. Felix Ravenna 107/108 (1974): 131– 50. Carroll-Spilleke, Maureen. ‘Das Militärlager Divitia in Köln-Deutz’. Kölner Jahrbuch 26 (1993): 321–444. Carter, Paul. Road to Botany Bay: An Exploration of Landscape and History. London: Faber and Faber, 1987. Repr. Minneapolis, MN: University of Minnesota Press, 2010. Cavalieri Manasse, Giuliana. ‘Le mura di Verona’. In Mura delle città romane in Lombardia: atti del convegno, Como 23-24 marzo 1990, 179–215. Como: Associazione Archeologica Comense, 1993. ———. ‘Le mura teodoriciane di Verona’. In Teoderico il Grande e i Goti d’Italia. Atti del 13° congresso internazionale di studi sull’alto medioevo, 2: 633–43. Spoleto: CISAM, 1992. Cavalieri Manasse, Giuliana, and Peter Hudson. ‘Nuovi dati sulle fortificazioni di Verona (III–XI se- colo)’, in Le fortificazioni del Garda, ed. Brogiolo, 71–91. Ceresa Mori, Anna. ‘Riflessioni sul palazzo imperiale di Milano alla luce delle recenti indagini’. In Milano e la chiesa di Milano prima di Ambrogio, edited by Raffaele Passarella, 95–120. Studia Am- brosiana. Milan: Biblioteca Ambrosiana, 2018. Ceresa Mori, Anna, Laura Lodovici, and Carla Pagani. ‘Via Gorani 2–4. Resti di domus di età romana e del Palatium imperiale’. Notiziario della Soprintendenza Archeologica della Lombardia (2010–2011): 241–47. Chalon, Michel, Georges Devallet, Paul Force, Michel Griffe, Jean-Marie Lassère, and Jean-Noël Mi- chaud. ‘Memorabile factvm. Une celebration de l’évergétisme des rois vandals dans l’Antho- logie Latine’. Antiquités africaines 21 (1985): 207–62. Chazelle, Celia. ‘“Romanness” in Early Medieval Culture’. In Paradigms and Methods in Early Medieval Studies, edited by Celia Chazelle and Felice Lifshitz, 81–98. The New Middle Ages. New York: Palgrave Macmillan, 2007. Christie, Neil. From Constantine to Charlemagne: An Archaeology of Italy, AD 300–800. Aldershot: Ashgate, 2006. Ciglenečki, Slavko. ‘Claustra Alpium Iuliarum, tractus Italiae circa Alpes and the Defence of Italy in the Final Part of the Late Roman Period’. Arheološki vestnik 67 (2016): 409–24. 374 ———. Höhenbefestigungen aus der Zeit vom 3. bis 6. Jh. im Ostalpenraum. Academia Scientiarum et Artium Slovenica: Historia et Sociologia 31. Ljubljana: Slovenska akademika znanosti in umetnosti, 1987. Cirelli, Enrico. Ravenna: archeologia di una città. Contributi di Archeologia Medievale 2. Florence: All’In- segna del Giglio, 2008. Clay, John Henry. ‘Boniface in Hessia and Thuringia’. In A Companion to Bomiface, edited by Michel Aaij and Shannon Godlove, 270–98. Brill’s Companions to the Christian Tradition 92. Leiden: Brill, 2020. Clover, Frank. ‘Felix Karthago’. DOP 40 (1986): 1–16. Coarelli, Filippo. Palatium: il Palatino dalle origini all’impero. Rome: Quasar, 2012. Cohen, Meredith. The Sainte-Chapelle and the Construction of Sacral Monarchy: Royal Architecture in Thirteenth- Century Paris. Cambridge: Cambridge University Press, 2015. Conant, Kenneth. Carolingian and Romanesque Architecture, 800–1200. 2nd ed. The Pelican History of Art. Harmondsworth, UK: Penguin Books, 1979. Colli, Donato. ‘Il palazzo Sessoriano nell’area archeologica di S. Croce in Gerusalemme : ultima sede imperiale a Roma?’ Mélanges de l’École française de Rome. Antiquité 108, no. 2 (1996): 771–815. Collins, Roger. ‘The “Reviser” Revisited: Another Look at the Alternative Version of the Annales regni francorum’. In After Rome’s Fall: Narrators and Sources of Early Medieval History. Essays Presented to Walter Goffart, edited by Alexander Callander Murray, 191–213. Toronto: University of To- ronto Press, 1998. ———. Visigothic Spain, 409–711. Oxford: Blackwell, 2004. Conlin, Diane A., Anne E. Haeckle, Gianni Ponti, Christopher Gregg and Paola Virgili. ‘The Villa of Maxentius on the Via Appia: Report on the 2005 Excavations’. Memoirs of the American Academy in Rome 51/52 (2007): 347–70. Corcoran, Simon. ‘Maxentius: A Roman Emperor in Rome’. Antiquité Tardive 25 (2017): 59–74. Cracco Ruggini, Lellia. Economia e società nell’Italia annonaria: rapporti fra agricoltura e commercio dal IV al VI secolo d. C. Milan: A. Giuffrè, 1961. James Lang, ed. York and Eastern Yorkshire. Vol. 3 of The Corpus of Anglo-Saxon Stone Sculpture in England. Edited by Rosemary Cramp. Oxford: Oxford University Press, 1991. Crosby, Sumner McKnight. The Royal Abbey of Saint-Denis from its Beginnings to the Death of Sugar, 475– 1151. Edited by Pamela Blum. Yale Publications in the History of Art 37. New Haven, CT: Yale University Press, 1987. Cullhed, Mats. Conservator urbis suae: Studies in the Politics and Propaganda of the Emperor Maxentius. Skrifter utgivna av Svenska institutet i Rom, 8th series, 20. Stockholm: P. Aströms Förlag, 1994. 375 Curta, Florin. ‘Merovingian and Carolingian Gift Giving’. Speculum 81, no. 3 (2006): 671–99. Ćurčić, Slobodan. ‘Late-Antique Palaces: The Meaning of Urban Context’. Ars Orientalis 23 (1993): 67–90. Dabrowska, Maria, Lech Leciejewicz, Eleonora Tabaczynska, and Stanislaw Tabaczynski. ‘Castelse- prio: scavi diagnostici, 1962–63’, Sibrium 14 (1978–79): 1–140. David, Massimiliano. ‘“Palatinaeque arces…” Temi di architettura palaziale a Milano tra III e X secolo.’ In ‘Ubi palatio dicitur’. Residenze di re e imperatori in Lombardia, edited by Massimiliano David, 9– 46. Milan: Cinisello Balsamo, 1999. Davis, Jennifer. Charlemagne’s Practice of Empire. Cambridge: Cambridge University Press, 2015. De Capitani d’Arzago, Alberto. Il circo romano. Istituto di Studi Romani – Sezione Lombarda. Ricerche per la Forma Urbis Mediolani. Milan: Ceschina, 1939. De Francovich, Géza. Il Palatium di Teodorico a Ravenna e la cosiddetta architettura di potenza: Problemi d’in- terpretazione di raffigurazioni architettoniche nell’arte tardoantica e altomedioevale. Quaderni di commen- tari 1. Rome: De Luca, 1970. Deichmann, Friedrich. Ravenna: Hauptstadt des spätantiken Abendlandes. 3 vols. Wiesbaden: Steiner, 1969–1989. De Jong, Mayke. ‘Bride Shows Revisited: Praise, Slander, Exegesis in the Reign of the Empress Judith’. In Gender in the Early Medieval World: East and West, 300–900, edited by Leslie Brubaker and Julia Smith, 257–77. Cambridge: Cambridge University Press, 2004. ———. ‘Carolingian Monasticism: The Power of Prayer’. In c. 700–c. 900, edited by Rosamond McKitterick, 622–653. Vol. 2 of The New Cambridge Medieval History. Cambridge: Cambridge University Press, 1995. ———. ‘Charlemagne’s Balcony: The Solarium in Ninth-century Narratives’. In The Long Morning of Medieval Europe: New Directions in Early Medieval Studies, edited by Jennifer Davis and Michael McCormick, 277–89. Aldershot: Ashgate, 2008. ———. ‘Ecclesia and the Early Medieval Polity’, in Staat im frühen Mittelalter, ed. Airlie et. al., 113–32. De Maria, Sandro, ed. Nuove ricerche e scavi nell’area della Villa di Teodorico a Galeata: atti della giornata di study, Ravenna, 26 marzo 2002. Bologna: Ante Quem, 2004. Deliyannis, Deborah Mauskopf. Ravenna in Late Antiquity. Cambridge: Cambridge University Press, 2010. Delogu, Paolo. Mito di una città meridionale: Salerno, secoli VIII–XI. Nuovo Medioevo 2. Naples: Liguori Editore, 1977. 376 Deneauve, Jean. ‘Les structures romaines de Byrsa: Historique des recherches’. In Byrsa I: Rapports préliminaires des fouilles (1974–1976), edited by Serge Lancel, 51–66. Collection de l’École fran- çaise de Rome 41. Rome: École française de Rome, 1979. Dey, Hendrik. The Afterlife of the Roman City: Architecture and Ceremony in Late Antiquity and the Early Middle Ages. Cambridge: Cambridge University Press, 2014. ———. The Aurelian Wall and the Refashioning of Imperial Rome, AD 271–855. Cambridge: Cambridge University Press, 2011. ———. ‘From “Street” to “Piazza”: Urban Politics, Public Ceremony, and the Redefinition of platea in Communal Italy and Beyond’, Speculum 91, no. 4 (2016): 919–44. Dierkens, Alain, and Patrick Périn. ‘Les sedes regiae mérovingiennes entre Seine et Rhin’. In Sedes regiae (ann. 400–800), ed. Ripoll and Gurt, 267–304. Doppelfeld, Otto. ‘Das praetorium unter dem Kölner Rathaus’, in Neue Ausgrabungen in Deutschland, ed. Krämer, 313–21. Dräger, Paul. ‘Zwei Moselfahrten des Venantius Fortunatus (carmina 6,8 and 10,9)’. Kurtrierisches Jahr- buch 39 (1999): 67–89. Drijvers, Jan Willem. ‘Helena Augusta and the City of Rome’. In Monuments & Memory. Christian Cult Buildings and Constructions of the Past: Essays in honour of Sible de Blaauw, edited by Mariëtte Verhoeven, 147–53. Architectural Crossroads 3. Turnhout: Brepols, 2016. Dunbabin, Katherine. The Mosaics of Roman North Africa: Studies in Iconography and Patronage. Oxford Monographs on Classical Archaeology. Oxford: Clarendon Press, 1978. ———. ‘The Use of Private Space’. In La ciudad en el mondo romano. Atti del XIV congresso internazionale di archeologia classica (Tarragona, 1993), volume 1, 165–66. Tarragona: Consejo Superior de In- vestigaciones Cientificas-Institut d’Etudis Catalans, 1994. Durliat, Jean. De la ville antique à la ville byzantine: le problème des subsistances. Collection de l’École française de Rome 136. Rome: École française de Rome, 1990. Dutton, Paul. Charlemagne’s Mustache, and Other Cultural Clusters of a Dark Age. The New Middle Ages. London: Palgrave Macmillan, 2004. Dutton, Paul, and Herbert Kessler. The Poetry and Paintings of the First Bible of Charles the Bald. Ann Arbor, MI: University of Michigan Press, 1997. Duval, Noël. ‘Comment reconnaitre un palais imperial ou royal? Ravenne et Piazza Armerina’, Felix Ravenna 115 (1978): 27–62. ———. ‘La place de Split dans l’architecture aulique du Bas-Empire’. Urbs 4 (1961): 67–95. ———. ‘Que savons-nous du palais de Théodoric à Ravenne ?’, Mélanges d’archéologie et d’histoire 72 (1960) : 337–371. 377 ———. ‘La représentation du palais dans l’art du Bas-Empire et du Haut Moyen Age d’après le Psau- tier d’Utrecht’. Cahiers Archéologiques 15 (1965): 207–54. Duval, Noël, and Vladislav Popović, eds. Caričin Grad, 4 vols. Collection de l’École française de Rome 75. Rome: École française de Rome, 1984–2019. Duval, Noël, Patrick Périn, and Jean-Charles Picard. ‘Paris’. In Province écclesiastique de Sens (Lugdunensis Senonia), edited by Jean-Charles Picard, 97–127. Vol. 8 of Topographie chrétienne des cités de la Gaule des origines au milieu du VIIIe siècle, edited by Nancy Gauthier and Jean-Charles Picard. Paris: De Boccard, 1992. Dyggve, Ejnar. Ravennatum palatium sacrum: La basilica ipetrale per ceremonie. Studii sull’architettura dei palazzi della tarda antichità. Arkæologisk-kunsthistoriske Meddelelser 3, no. 2. Copenhagen: E. Munksgaard, 1941. Easterling, Keller. Extrastatecraft: The Power of Infrastructure Space. London: Verso, 2014. Eck, Werner. Köln in römischer Zeit: Geschichte einer Stadt im Rahmen des Imperium Romanum. Geschichte der Stadt Köln 1. Cologne: Greven, 2004. Ehlers, Caspar, ed. Orte der Herrschaft. Mittelalterliche Königspfalzen. Göttingen: Vandenhoeck & Ruprecht, 2002. ———. ‘Pfalzenforschung Heute. Eine Einführung in das Repertorium der deutschen Königspfal- zen’. In Orte der Herrschaft, ed. Ehlers, 25–54. Eiden, Hans. ‘Ausgrabungen im spätantiken Trier’. In Neue Ausgrabungen in Deutschland, ed. Krämer, 340–67. Elias, Norbert. The Court Society. Translated by Edmund Jephcott. New York: Pantheon, 1983. Ellis, Simon. ‘Power, Architecture, and Decor: How the Late Roman Aristocrat Appeared to His Guests’. In Roman Art in the Private Sphere: New Perspectives on the Architecture and Decor of the Domus, Villa, and Insula, edited by Elaine Gazda, 117–34. 2nd ed. (Ann Arbor: University of Michigan Press, 2010. ———. Roman Housing. London: Duckworth, 2000. Ennabli, Liliane. Carthage: une métropole chrétienne du IVe à la fin du VIIe siècle. Paris: CNRS Editions, 1997. Erdkamp, Paul. ‘The Food Supply of the Capital’. In The Cambridge Companion to Ancient Rome, edited by Paul Erdkamp, 262–77. Cambridge: Cambridge University Press, 2013. ———. The Grain Market in the Roman Empire: A Social, Political and Economic Study. Cambridge: Cambridge University Press, 2005. Esders, Stefan. ‘Nordwestgallien um 500. Von der militarisierten spätrömischen Provinzgesellschaft zur erweiterten Militäradministration des merowingischen Königtums’. In Chlodwigs Welt: 378 Organisation von Herrschaft um 500, edited by Mischa Meier and Steffen Patzold, 339–61. Roma Aeterna 3. Stuttgart: Steiner, 2014. ———. ‘Man-like Discipline and Loyalty Against the “Enemies of God”: Some Observations on the Militarized Frontier Society of Eastern Francia Around 600’. In Early Medieval Militarisation, edited by Ellora Bennett, Guido Berndt, Stefan Esders, and Laury Sarti, 181–95. Manchester: Manchester University Press, 2021. Esmonde Cleary, Simon. The Ending of Roman Britain. London: Batsford, 1989. ———. The Roman West, AD 200–500: An Archaeological Study. Cambridge: Cambridge University Press, 2013. Ettel, Peter, Lukas Werther, and Petra Wolters. ‘Vorbericht zu den Untersuchungen 2009 bis 2012 im Königsgutbezirk und Pfalzgebiet Salz, Lkr. Bad Neustadt a. d. Saale’. Beiträge zur Archäologie in Ober- und Unterfranken 8 (2013): 213–48. Ewald, Björn, and Carlos Noreña, eds. The Emperor and Rome: Space, Representation, and Ritual, Yale Clas- sical Studies 35. Cambridge: Cambridge University Press, 2010. Ewig, Eugen. ‘Die fränkischen Teilungen und Teilreiche (511–613)’. In Eugen Ewig, Spätantikes und fränkisches Gallien. Gesammelte Schriften (1952–1973), edited by Hartmut Atsma, 1:114–71. Bei- hefte zu Francia 3. Munich: Artemis, 1976.. ———. ‘Die fränkischen Teilreiche im 7. Jahrhundert (613–714)’. In Eugen Ewig, Spätantikes und fränkisches Gallien. Gesammelte Schriften (1952–1973), edited by Hartmut Atsma, 1: 172–230. Bei- hefte zu Francia 3. Munich: Artemis, 1976 ———. ‘Résidence et capitale pendant le haut Moyen Âge’. Revue historique 230, no. 1 (1963): 25–72. ———. Trier im Merowingerreich: Civitas, Stadt, Bistum. Trier: Paulinus, 1954. ———. ‘Zum christlichen Königsgedanken im Frühmittelalter’. In Das Königtum, seine geistigen und recht- lichen Grundlagen, edited by Theodor Mayer, 7–73. Vorträge und Forschungen 3. Lindau: J. Thorbecke 1956. Fafinski, Mateusz. Roman Infrastructure in Early Medieval Britain: The Adaptations of the Past in Text and Stone. Amsterdam: Amsterdam University Press, 2021. Falkenstein, Ludwig. ‘Charlemagne et Aix-la-Chapelle’. Byzantion 61, no. 1 (1991): 231–89. ———. ‘Pfalz und vicus Aachen’. In Orte der Herrschaft, ed. Ehlers, 131–182. Fauvinet-Ranson, Valérie. ‘Capitales et residences du roi Théodoric à l’orée du VIe siècle’. In Le gou- vernement en déplacement: Pouvoir et mobilité de l’Antiquité à nos jours, edited by Sylvan Destephen, Josiane Barbier, and François Chausson, 173–88. Rennes: Presses Universitaires de Rennes, 2019. 379 ———. Decor Civitatis, Decor Italiae. Monuments, travaux publics et spectacles au VIe siècle d’après les Variae de Cassiodore. Munera 23. Bari: Edipuglia, 2006. Featherstone, Michael, Jean-Michel Spieser, and Ulrike Wulf-Rheidt, eds. The Emperor’s House: Palaces from Augustus to the Age of Absolutism. Urban Spaces 4. Berlin: De Gruyter, 2015. Fenske, Lutz. ‘Jagd und Jäger im früheren Mittelalter. Aspekte ihres Verhältnisses’. In Jagd und höfische Kultur im Mittelalter, edited by Werner Rösener, 29–93. Veröffentlichungen des Max-Planck- Instituts für Geschichte 135. Göttingen: Vandenhoeck & Ruprecht, 1997. Ferron, Jean, and Maurice Pirard. ‘Les fouilles de Byrsa, 1953–1954’. Cahiers de Byrsa 5 (1955): 31–264. Fichtenau, Heinrich. ‘Byzanz und die Pfalz zu Aachen’. Mitteilungen des Instituts für Österreichische Ge- schichtsforschung 59, no. 1 (1951): 1–54. Flach, Dietmar. Untersuchungen zur Verfassung und Verwaltung des Aachener Reichsgutes. Veröffentlichungen des Max-Planck-Instituts für Geschichte 46. Göttingen: Vandenhoeck & Ruprecht, 1976. Fleckenstein, Josef. Die Hofkapelle der deutschen Könige. 2 vols. Schriften der Monumenta Germaniae historica 16. Stuttgart: Anton Hiersemann, 1959. Flower, Richard. ‘TAMQVAM FIGMENTVM HOMINIS: Ammianus, Constantius II and the Por- trayal of Imperial Ritual’. The Classical Quarterly 65, no. 2 (2015): 822–35. Foucault, Michel. Discipline and Punish: The Birth of the Prison. Translated by Alan Sheridan. 2nd ed. New York: Vintage, 1995. ———. ‘Questions on Geography’. In Power/Knowledge: Selected Interviews and Other Writings, 1972–1977. Translated by Colin Gordon. New York: Pantheon, 1980. Fouracre, Paul. The Age of Charles Martel. The Medieval World. London: Pearson, 2000. ———. ‘Francia in the Seventh Century’. In c. 500–c. 700, edited by Paul Fouracre, 371–96. Vol. 1 of The New Cambridge Medieval History. Cambridge: Cambridge University Press, 2005. ———. ‘Merovingian History and Merovingian Hagiography’. Past & Present 127 (1990): 3–38. ———. ‘The Nature of Frankish Political Institutions in the Seventh Century’. In Franks and Alamanni in the Merovingian Period: An Ethnographic Perspective, edited by Ian Wood, 285–316. Studies in Historical Archaeoethnology 3. Rochester, NY: Boydell Press, 1999. Fried, Johannes. Charlemagne. Translated by Peter Lewis. Cambridge, MA: Harvard University Press, 2016. Friedrich, Mattias, and James Harland, eds. Interrogating the ‘Germanic’: A Category and its Use in Late Antiquity and the Early Middle Ages. Ergänzungsbände zum Reallexicon der Germanischen Al- tertumskunde 123. Berlin: De Gruyter, 2021. 380 Frova, Antonio. ‘Trovamenti e scavi dal 1950 al 1953’. In Ritrovamenti e scavi per la “Forma Urbis Medio- lani” IV, edited by Aristide Calderini and Carla Gerra, 5–17. Quaderni di Studi Romani 5. Milan: Ceschina, 1955. Gai, Antonella Sveva. ‘Die Pfalz Karls des Großen in Paderborn: Ihre Entwicklung von 777 bis zum Ende des 10. Jahrhunderts’. In 799. Kunst und Kultur, ed. Stiegemann and Wemhoff, 3:183–96. Ganshof, François Louis. The Carolingians and the Frankish Monarchy: Studies in Carolingian History. Trans- lated by Janet Sondheimer. London: Longman, 1971. Gantner, Clemens, Rosamond McKitterick, and Sven Meeder, eds. The Resources of the Past in Early Medieval Europe. Cambridge: Cambridge University Press, 2015. Ganz, David. ‘Einhard’s Charlemagne: The Characterisation of Greatness’. In Charlemagne: Empire and Society, edited by Joanna Story, 38–51. Manchester: Manchester University Press, 2005. Garipzanov, Ildar. The Symbolic Language of Authority in the Carolingian World (c. 751–877). Brill’s Series on the Early Middle Ages. Leiden: Brill, 2008. Gauert, Adolf. ‘Zum Itinerar Karls des Großen’. In Persönlichkeit und Geschichte, edited by Helmut Beumann, 307–21. Vol. 1 of Karl der Große. Lebenswerk und Nachleben, edited by Wolfgang Braunfels. Düsseldorf: L. Schwann, 1965. Geertz, Clifford. ‘Centers, Kings, and Charisma: Reflections on the Symbolics of Power’. In Clifford Geertz, Local Knowledge: Further Essays in Interpretive Anthropology, 121–46. New York: Basic Books, 2000. ———. ‘Thick Description: Towards an Interpretive Theory of Culture’. In Clifford Geertz, The In- terpretation of Cultures: Selected Essays, 3–30. New York: Basic Books, 1973. Gelichi, Sauro. ‘Ravenna, ascesa e declino di una capitale’. In Sedes regiae (ann. 400–800), ed. Ripoll and Gurt, 109–34 Gentili, Gino Vinicio. La villa romana di Piazza Armerina. Palazzo Erculio. 3 vols. Osimo: Fondazione Don Carlo, 1999. Gerberding, Richard. The Rise of the Carolingians and the Liber Historiae Francorum. Oxford Historical Monographs. Oxford: Clarendon Press, 1987. Gerrard, James. The Ruin of Roman Britain: An Archaeological Perspective. Cambridge: Cambridge Univer- sity Press, 2013. Gibson, Sheila, Janet DeLaine, and Amanda Claridge. ‘The Triclinium of the Domus Flavia: A New Reconstruction’. Papers of the British School at Rome 62 (1994): 67–100. Gillett, Andrew. ‘Rome, Ravenna, and the Last Western Emperors’. PBSR 69 (2001): 131–67. ———. ‘Telling Off Justinian: Theudebert I, the Epistolae Austrasicae, and Communication Strategies in Sixth-Century Merovingian-Byzantine Relations’. EME 27, no. 2 (2019): 161–94. 381 Ghirardini, Gherardo. ‘Gli scavi del Palazzo di Teoderico a Ravenna’. Monumenti antichi 24 (1917): coll. 737–888. Goldberg, Eric. In the Manner of the Franks: Hunting, Kingship, and Masculinity in Early Medieval Europe. The Middle Ages. Philadelphia: University of Pennsylvania Press, 2020. –––—. ‘Louis the Pious and the Hunt’. Speculum 88, no. 3 (2013): 613–643. Grabar, André. Martyrium : recherches sur le culte des reliques et l'art chrétien antique. Vol. 1. Paris: Collège de France, 1946. Grewe, Holger. ‘Die Königspfalz zu Ingelheim am Rhein’. In 799. Kunst und Kultur, ed. Stiegemann and Wemhoff, 3:142–51. Grey, Cam. ‘Landowning and Labour in the Rural Economy’, in Companion to Ostrogothic Italy, ed. Ar- nold et al., 285–89. Guidobaldi, Federico. ‘Le domus tardoantiche di Roma come “sensori” delle trasformazioni culturali e sociali’. In The Transformations of Vrbs Roma in Late Antiquity, edited by William V. Harris, 53– 68. Journal of Roman Archaeology Supplements 33. Portsmouth, RI: Journal of Roman Ar- chaeology, 1999. Hales, Shelley. The Roman House and Social Identity. Cambridge: Cambridge University Press, 2003. Halfond, Gregory. Bishops and the Politics of Patronage in Merovingian Gaul. Ithaca, NY: Cornell University Press, 2019. Halsall, Guy. Barbarian Migrations and the Roman West, 376–568. Cambridge Medieval Textbooks. Cam- bridge: Cambridge University Press, 2007. ———. ‘The Decline and Fall of the Ancient Triumph’. In Der römische Triumph in Prinzipat und Spät- antike, edited by Fabian Goldbeck and Johannes Weinand, 555–68. Berlin: De Gruyter, 2017. ———. ‘Gender and the End of Empire’. Journal of Medieval and Early Modern Studies 34, no. 1 (2004): 17–39. ———. ‘The Preface to Book V of Gregory of Tours’ Histories: Its Form, Context, and Significance’. English Historical Review 122, no. 496 (2007): 297–317. ———. Settlement and Social Organisation: The Merovingian Region of Metz. Cambridge: Cambridge Univer- sity Press, 1995. Hamerow, Helen. Early Medieval Settlements: The Archaeology of Rural Communities in North-West Europe, 400–900. Oxford: Oxford University Press, 2003. Hampel, Andrea, and Heinz Schomann. Der Kaiserdom zu Frankfurt am Main: Ausgrabungen 1991–93. Beiträge zum Denkmalschutz in Frankfurt am Main 8. Nussloch: R. Angerer, 1994. Hauck, Karl. Tiergärten im Pfalzbereich’. In Deutsche Königspfalzen, 1, ed. Gauert, 32–50. 382 Heather, Peter. ‘A Tale of Two Cities: Rome and Ravenna under Gothic Rule’. In Ravenna: Its Role in Earlier Medieval Change and Exchange, edited by Judith Herrin and Janet Nelson, 15–38. London: Institute of Historical Research, 2016. Hen, Yitzhak. The Royal Patronage of Liturgy in Frankish Gaul to the Death of Charles the Bald (877). Henry Bradshaw Society. Subsidia 3. London: Boydell Press, 2001. ———. ‘The Uses of the Bible and the Perception of Kingship in Merovingian Gaul’. EME 7, no. 3 (1998): 277–90. Henning, Joachim, Michael McCormick, Lauro Olmo Enciso, Knut Rassmann and Eyub Fikrit Eyub. ‘Reccopolis Revealed: The First Geomagnetic Mapping of the Early Medieval Visigothic Royal Town’. Antiquity 93, no. 369 (2019): 735–51. Hidalgo Prieto, Rafael. Espacio público y espacio privado en el conjunto palatino de Cercadilla (Córdoba): el aula central y las termas. Cercadilla 1. Seville: Consejería de Cultura, 1996. Hirsch, Eric. ‘Introduction. Landscape: Between Place and Space’. In The Anthropology of Landscape, edited by Eric Hirsch and Michael O’Hanlon, 1–30. Oxford Studies in Social and Cultural Anthropology. New York: Oxford University Press, 1995. Hope-Taylor, Brian. Yeavering: An Anglo-British Centre of Early Northumbria. Department of the Envi- ronment Archaeological Reports 7. London: HM Stationery Office, 1977. Hummer, Hans. Politics and Power in Early Medieval Europe: Alsace and the Frankish Realm, 600–1000, CSMLT, 4th Series, 65. Cambridge: Cambridge University Press, 2005. Humphrey, John. Roman Circuses: Arenas for Chariot Racing. Berkeley, CA: University of California Press, 1986. Ingold, Timothy. ‘The Temporality of the Landscape’. World Archaeology 25, no. 1 (1993): 152–74. Ioppolo, Giovanni, and Giuseppina Pisani Sartorio. La villa di Massenzio sulla Via Appia. Vol. 2. Mo- numenti Romani 9. Rome: Editore Colombo, 1999. Jacobsen, Werner. ‘Die Pfalzkonzeptionen Karls des Großen’. In Karl der Große als vielberufener Vorfahr: sein Bild in der Kunst der Fürsten, Kirschen und Städte, edited by Lieselotte Saurma-Jeltsch, 23–48. Schriften des historischen Museums Frankfurt am Main 19. Sigmaringen: Jan Thorbecke Ver- lag, 1994. Jaeschke, Verena. ‘Adapting to a New Concept of Sovereignty: Some Remarks on Tetrarchic Palace Architecture’. In Late Antique Palatine Architecture. Palaces and Palace Culture: Patterns of Transcul- turation, edited by Lynda Mulvin and Nigel Westbrook, 63–76. Architectural Crossroads. Stud- ies in the History of Architecture 5. Turnhout: Brepols, 2019. James, Edward. ‘The Militarisation of Roman Society, 400–700’. In Military Aspects of Scandinavian Society in a European Perspective, AD 1–1300, edited by Anne Nørgård Jorgensen and Birthe L. Clausen, 19–24. Copenhagen: National Museum, 1997. 383 James, Simon. ‘The Fabricae: State Arms Factories of the Later Roman Empire’. In Military Equipment and the Identity of Roman Soldiers. Proceedings of the Fourth Roman Military Equipment Conference, edited by J. C. Coulston, 257–331. BAR International Series 394. Oxford: British Archaeological Re- ports, 1988. Jarnut, Jörg. ‘Die frühmittelalterliche Jagd unter rechts- und sozialgeschichtlichen Aspekten’. In L’uomo di fronte al mondo animale nell’alto medioevo, 765–808. Settimane 31. Spoleto: CISAM, 1985. Johnson, Mark. The Roman Imperial Mausoleum in Late Antiquity. Cambridge: Cambridge University Press, 2009. ———. ‘Toward a History of Theoderic’s Building Program’. Dumbarton Oaks Papers 42 (1988): 73– 96. Jones, A. H. M. The Later Roman Empire: A Social, Economic, and Administrative Survey. 3 vols. Oxford: Blackwell, 1964. Jones, A H. M., J. R. Martindale, and J. Morris, eds. The Prosopography of the Later Roman Empire. 3 vols. Cambridge: Cambridge University Press, 1971–1992. Kähler, Heinz. Die Villa des Maxentius bei Piazza Armerina. Monumenta Artis Romanae 12. Berlin: Mann, 1973. Karkov, Catherine. The Art of Anglo-Saxon England. Boydell Studies in Medieval Art and Architecture. Woodbridge: Boydell Press, 2011. Kempf, Theodor. ‘Konstantinische Deckenmalerei aus dem Trierer Dom’. Trierer Zeitschrift 19 (1950): 45–51. ———. ‘Trierer Domgrabungen, 1943–54’. In Neue Ausgrabungen in Deutschland, ed. Krämer, 368–79. Kerr, Lorraine. ‘A Topography of Death: The Buildings of the Emperor Maxentius on the Via Appia, Rome’. TRAC 2001: Proceedings of the Eleventh Annual Theoretical Roman Archaeology Conference, Glasgow 2001, edited by Martin Carruthers, Carol van Driel-Murray, Andrew Gardner, Jason Lucas, Louise Revell, and Ellen Swift, 24–33. Oxford: Oxbow Books, 2002. Kinney, Dale. ‘The Evidence for the Dating of S. Lorenzo in Milan’. JSAH 31, no. 2 (1972): 92–107. Knapp, A. Bernard, and Wendy Ashmore. ‘Archaeological Landscapes: Constructed, Conceptualized, Ideational’. In Archaeologies of Landscape, ed. Wendy Ashmore and A. Bernard Knapp, 1–30. Malden, MA: Blackwell, 1999. Kolb, Frank. Herrscherideologie in der Spätantike. Studienbücher. Geschichte und Kultur der Alten Welt. Berlin: Akademie Verlag, 2001. Kondoleon, Christine. Antioch: The Lost Ancient City. Princeton: Princeton University Press, 2000. Koziol, Geoffrey. ‘The Dangers of Polemic: Is Ritual Still an Interesting Topic of Historical Study?’ EME 11, no. 4 (2002): 367–88. 384 Krautheimer, Richard. Corpus Basilicarum Christianarum Romae: The Early Christian Basilicas of Rome (IV– IX Century). Vol. 1. Monumenti di antichità cristiana, 2nd series, 2. Vatican City: Ponticio istituto di archeologia cristiana, 1937. ———. Three Christian Capitals: Topography and Politics. Una’s Lectures 4. Berkeley, CA: 1983. Krämer, Werner, ed. Neue Ausgrabungen in Deutschland. Berlin: Mann, 1958. Krencker, Daniel, Ludwig Hussong, and Heinz Cüppers, eds. Die Trierer Kaiserthermen. Trierer Grabun- gen und Forschungen 1. Augsburg: Filser, 1929. Kreusch, Felix. ‘Kirche, Atrium und Portikus der Aachener Pfalz’. In Karolingische Kunst, edited by Wolfgang Braunfels and Hermann Schnitzler, 463–533. Vol. 3 of Karl der Große. Lebenswerk und Nachleben, edited by Wolfgang Braunfels. Düsseldorf: L. Schwann, 1965. Kretzschmer, Fritz. ‘Die Heizung der Aula Palatina in Trier: Ein Versuch ihrer Deutung und der Auf- klärung ihrer Betriebsweise’. Germania 33 (1955): 200–10. Kulikowski, Michael. ‘The Notitia Dignitatum as a Historical Source’. Historia 49, no. 3 (2000): 358–77. Lavin, Irving. ‘The Ceiling Frescoes in Trier and Illusionism in Constantinian Painting’. Dumbarton Oaks Papers 21 (1967): 99–113. ———. ‘The House of the Lord: Aspects of the Role of Palace Triclinia in the Architecture of Late Antiquity and the Early Middle Ages’. The Art Bulletin 44, no. 1 (1962): 1–27. La Rocca, Cristina. ‘Mores tuos fabricae loquuntur. Building activity and the Rhetoric of Power in Ostro- gothic Italy’. Haskins Society Journal 26 (2015): 1–31. ———. ‘Una prudente maschera “antiqua”. La politica edilizia di Teoderico’. In Teoderico il grande e i Goti d’Italia. Atti del 13° congresso internazionale di studi sull’alto medioevo, 2:451–515. Spoleto: CI- SAM, 1992. La Rocca Hudson, Cristina. ‘“Dark Ages” a Verona. Edilizia privata, aree aperte e strutture pubbliche in una città dell’Italia settentrionale’. Archeologia Medievale 13 (1986): 31–78. L’Orange, Hans Peter. Art Forms and Civic Life in the Later Roman Empire. Princeton: Princeton Univer- sity Press, 1965. L’Orange, Hans Peter, and Ejnar Dyggve. ‘È un palazzo di M. Erculeo che gli scavi da Piazza Armerina portano alla luce?’. Symbolae Osloenses 29 (1952): 114–28. Longhi, Andrea. ‘Palaces and Palatine Chapels in 15th–Century Italian Dukedoms: Ideas and Experi- ences’. In A Renaissance Architecture of Power: Princely Palaces in the Italian Quattrocento, edited by Silvia Beltramo, Flavia Cantatore, and Marco Falin, 82–104. The Medieval Mediterranean 104. Leiden: Brill, 2015. Lefebvre, Henri. The Production of Space. Translated by Donald Nicholson-Smith. Oxford: Blackwell, 1991. 385 Le Jan, Régine. ‘Frankish Giving of Arms and Rituals of Power: Continuity and Change in the Caro- lingian Period’. In Rituals of Power: From Late Antiquity to the Early Middle Ages, edited by Frans Theuws and Janet Nelson, 281–309. The Transformation of the Roman World 8. Leiden: Brill, 2000. Ley, Judith. ‘Aquis palatium: Spätantiker Palast oder frühmittelalterliche Pfalz Architekturhistorische Überlegungen zur Ikonographie der Aachener Pfalz’. In The Emperor’s House, ed. Featherstone et al., 127–46. Leyser, Conrad. ‘Introduction: Making Early Medieval Societies’. In Making Early Medieval Societies: Conflict and Belonging in the Latin West, 300-1200, edited bz Kate Cooper and Conrad Leyser, 1– 15. Cambridge: Cambridge University Press, 2016. Lézine, Alexandre. Carthage-Utique. Études d’architecture et d’urbanisme. Paris : CNRS Editions, 1968. Liebeschuetz, J. H. W. G. ‘Ravenna to Aachen’. In Sedes regiae (ann. 400–800), ed. Ripoll and Gurt, 9– 30. Linn, Jason. ‘The Roman Grain Supply, 442–455’, Journal of Late Antiquity 5, no. 2 (2012): 298–321 Lobbedey, Uwe. ‘Carolingian Royal Palaces: The State of Research from an Architectural Historian’s Viewpoint’. In Court Culture in the Early Middle Ages: The Proceedings of the First Alcuin Conference, edited by Catherine Cubitt, 129–54. Studies in the Early Middle Ages 3. Turnhout: Brepols, 2003. Lusuardi Siena, Silvia. ‘L’origine dell’archetipo e il problema del Palatium: Una cronologia di VI se- colo?’. In La più antica veduta di Verona: L’Iconografia Rateriana. L’archetipo e l’immagine tramandata, edited by Antonella Arzone and Ettore Napione, 59–70. Verona: Comune di Verona, 2012. Luttwak, Edward. The Grand Strategy of the Roman Empire: From the First Century CE to the Third. Rev. ed. Baltimore: Johns Hopkins University Press, 2016. MacCormack, Sabine. Art and Ceremony in Late Antiquity. Transformation of the Classical Heritage 1. Berkeley: University of California Press, 1981. MacDonald, William. The Architecture of the Roman Empire. 2 vols. Revised edition. Yale Publications in the History of Art 17, 35. New Haven, CT: Yale University Press, 1982–86. Maioli, Maria Grazia. ‘Classe, podere Chiavichetta, zona portuale’. In Ravenna e il porto di Classe: venti anni di richerche archeologiche tra Ravenna e Classe, edited by Giovanna Bermond Monatanari, 65– 78. Imola: Santerno, 1983. Maioli, Maria Grazia, and Maria Luisa Stoppioni. Classe e Ravenna fra terra e mare – città, necropoli, monu- menti: un’avventura della archeologia, gli scavi nella zona archeologica di Classe. Ravenna: Sirri, 1987. Malmberg, Simon. ‘Triumphal Arches and Gates of Piety at Constantinople, Ravenna, and Rome’. In Using Images in Late Antiquity, edited by Stine Birk, Troels Myrup Kristensen, and Birte Poulsen, 150–89. Oxford: Oxbow, 2014. 386 Malnati, Luigi, Giuseppe Sassatelli, Andrea Augenti, and Maria Grazia Maioli. ‘Nuovi scavi archeolo- gici a Classe: campagne 2004–2005’. In Felix Ravenna. La croce, la spada, la vela: l’alto Adriatico fra V e VI secolo, edited by Andrea Augenti and Carlo Bertelli, 33–38. Milan: Skira, 2007. Mango, Cyril. The Brazen House: A Study of the Vestibule of the Imperial Palace at Constantinople. Arkælogisk- kunsthistoriske Meddelelser 4, no. 4. Copenhagen: Munksgaard, 1959. Mann, Michael. The Sources of Social Power. Vol. 1. 2nd ed. Cambridge: Cambridge University Press, 2010. Marlowe, Elizabeth. ‘Framing the Sun: The Arch of Constantine and the Roman Cityscape’. Art Bul- letin 88, no. 2 (2006): 223–42. ———. ‘Liberator urbis suae: Constantine and the Ghost of Maxentius’. In The Emperor and Rome, ed. Ewald and Noreña, 199–219 Martínez Jiménez, A. Javier. ‘Monte Barro: An Ostrogothic Fortified Site in the Alps’. Assemblage 11 (2011): 34–46. ———. ‘A Preliminary Study of the Aqueduct of Reccopolis’, Oxford Journal of Archaeology 34, no. 3 (2015): 301–20. Maskarinec, Maya. City of Saints: Rebuilding Rome in the Early Middle Ages. Philadelphia, PA: University of Pennsylvania Press, 2018. Matejčić, Ivan, and Pascale Chevalier. ‘The Episcopium of Poreč’. In Des domus ecclesiae aux palais épisco- paux. Actes du colloque tenu à Autun du 26 au 28 novembre 2009, edited by Sylvie Balcon-Berry, François Baratte, Jean-Pierre Caillet, and Dany Sandron, 151–60. Bibliothéque de l'antiquité tardive 23. Turnhout : Brepols, 2012. Mauss, Marcel. The Gift: The Form and Reason for Exchange in Archaic Societies. Translated by W. D. Halls. London: Routledge, 1990. Max-Planck-Institut für Geschichte, ed. Die deutschen Königspfalzen. Repertorium der Pfalzen, Königshöfe und übrigen Aufenthaltsorte der Könige im deutschen Reich des Mittelalters. 5 vols. Göttingen: Vandenhoeck & Ruprecht, 1986–2016. Mayer, Emanuel. ‘The Architecture of the Tetrarchy’. In A Companion to Roman Architecture, edited by Roger B. Ulrich and Caroline K. Quenemoen, 106–26. Oxford: Blackwell, 2014. ———. ‘Propaganda, Staged Applause, or Local Politics? Public Monuments from Augustus to Sep- timius Severus’. In The Emperor and Rome, ed. Ewald and Noreña, 111–34. ———. Rom ist dort, wo der Kaiser ist. Untersuchungen zu den Staatsdenkmälern des dezentralisierten Reiches von Diocletian bis zu Theodosius II. Römisch-Germanisches Zentralmuseum Forschungsinstitut für Vor- und Frühgeschichte. Monographien 53. Mainz: Verlag des RGZM, 2002. McClendon, Charles. The Origins of Medieval Architecture: Building in Europe, A.D 600–900. New Haven: Yale University Press, 2005. 387 McCormick, Michael. Eternal Victory: Triumphal Rulership in Late Antiquity, Byzantium, and the Early Me- dieval West. Past and Present Publications. Cambridge: Cambridge University Press, 1990. McEvoy, Meaghan. Child Emperor Rule in the Late Roman West, AD 367–455. Oxford Classical Mono- graphs. Oxford: Oxford University Press, 2013. McKitterick, Rosamond. Charlemagne: The Formation of a Euroepan Identity. Cambridge: Cambridge Uni- versity Press, 2008. McLynn, Neil. Ambrose of Milan: Church and Court in a Christian Capital. Transformation of the Classical Heritage 22. Berkeley, CA: University of California Press, 1994. ———. ‘The Transformation of Imperial Churchgoing in the Fourth Century’. In Approaching Late Antiquity: The Transformation from Early to Late Empire, edited by Simon Swain and Mark Ed- wards, 235–70. Oxford: Oxford University Press, 2004. Meier, Mischa. Geschichte der Völkerwanderung: Europa, Asien und Afrika vom 3. bis zum 8. Jahrhundert n. Chr. Historische Bibliothek der Gerda Henkel Stiftung. Munich: C. H. Beck, 2019. Menéndez Pidal, José. ‘La basílica de Santianes de Pravia, Oviedo’. In Actas del Simposio para el Estudio de los Códices del ‘Comentario al Apocalipsis’ de Beato de Liébana, 2:281–97. Madrid: Joyas Bibliográfi- cas, 1980. Metlich, Michael Andreas. The Coinage of Ostrogothic Italy. London: Spink, 2004. Metz, Wolfgang. Das karolingische Reichsgut: eine verfassungs- und verwaltungsgeschichtliche Untersuchung. Berlin: De Gruyter, 1960. Millar, Fergus. ‘Emperors, Frontiers and Foreign Relations, 31 B.C. to A.D. 378’. Britannia 13 (1982): 1–24. Miller, Maureen. The Bishop’s Palace: Architecture and Authority in Medieval Italy. Conjunctions of Religion and Power in the Medieval Past. Ithaca, NY: Cornell University Press, 2000. Minoprio, Anthony. ‘A Restoration of the Basilica of Constantine, Rome’. Papers of the British School at Rome 12 (1932): 1–25. Mirabella Roberti, Mario. Milano romana. Milan: Rusconi, 1984. Mitchell, W. J. T. ‘Imperial Landscape’. In Landscape and Power, edited by W. J. T. Mitchell, 5–34. 2nd ed. Chicago: University of Chicago Press, 2002. Mitteis, Heinrich. Der Staat des hohen Mittelalters: Grundlinien einer vergleichenden Verfassungsgeschichte des Lehnzeitalters. 4th ed. Weimar: Hermann Böhlaus Nachfolger, 1953. Moorhead, John. Theoderic in Italy. Oxford: Clarendon Press, 1992. Morsel, Joseph. ‘Jagd und Raum. Überlegungen über den sozialen Sinn der Jagdpraxis am Beispiel des spätmittelalterlichen Franken’. In Jagd und höfische Kultur im Mittelalter, edited by Werner 388 Rösener, 255–88. Veröffentlichungen des Max-Planck-Instituts für Geschichte 135. Göttin- gen: Vandenhoeck & Ruprecht, 1997. Mulvin, Lynda, and Nigel Westbrook, eds. Late Antique Palatine Architecture. Palaces and Palace Culture: Patterns of Transculturation. Architectural Crossroads. Studies in the History of Architecture 5. Turnhout: Brepols, 2019. Müller, Harald, Judith Ley, Franck Pohle and Andreas Schaub. ‘Pfalz und vicus Aachen in karolingi- scher Zeit’. In Aachen: von den Anfängen bis zur Gegenwart, edited by Thomas Kraus, 2:1–408. Veröffentlichen des Stadtarchivs Aachen 14. Aachen: Mayersche Buchhandlung, 2013. Napione, Ettore. ‘Breve storia di un’Iconografia perduta’. In La più antica veduta di Verona: L’Iconografia Rateriana. L’archetipo e l’immagine tramandata, edited by Antonella Arzone and Ettore Napione, 25–32. Verona: Comune di Verona, 2012. Nelson, Janet. ‘Aachen as a Place of Power’. In Topographies of Power in the Early Middle Ages, edited by Mayke de Jong, Frans Theuws, and Carine van Rhijn, 217–41. Transformation of the Roman World 6. Leiden: Brill, 2001. ———. ‘Charlemagne and Ravenna’. In Ravenna: Its Role in Earlier Medieval Change and Exchange, edited by Judith Herrin and Janet Nelson, 239–52. London: University of London School of Ad- vanced Study, 2016. ———. ‘The Lord’s Anointed and the People’s Choice: Carolingian Royal Ritual’, in Cannadine and Price, eds., Rituals of Royalty, 137–80. ———. ‘Queens as Jezebels: Brunhild and Balthild in Merovingian History’. In Medieval Women: Essays dedicated and Presented to Professor Rosalind M. T. Hill, edited by Derek Baker, 31–77. Studies in Church History. Subsidia 1. Oxford: Blackwell, 1978. Reprinted in Janet Nelson, Politics and Ritual in Early Medieval Europe, 1–48. London: Hambledon, 1986. ———. ‘The Setting of the Gift in the Reign of Charlemagne’, in The Languages of Gift in the Early Middle Ages, edited by Wendy Davies and Paul Fouracre, 116–148. Cambridge: Cambridge University Press, 2010. ———. ‘Was Charlemagne’s Court a Courtly Society?’ In Court Culture in the Early Middle Ages: The Proceedings of the First Alcuin Conference, edited by Catherine Cubitt, 39–57. Studies in the Early Middle Ages 3. Turnhout: Brepols, 2003. Nibby, Antonio. Del circo volgarmente detto di Caracalla. Rome: Filippo e Nicola de Romanis, 1825. Noble, Thomas F. X. ‘Theodoric and the Papacy’. In Teoderico il grande e i Goti d’Italia. Atti del 13° con- gresso internazionale di studi sull’alto medioevo, 1:404–16. Spoleto: CISAM, 1992. O’Donnell, James. Cassiodorus. Berkeley, CA: University of California Press, 1979. Olmo Enciso, Lauro, ‘Fuentes escritas y primeras investigaciones sobre Recópolis’, in Recópolis y la ciudad, ed. Olmo Enciso, 22–39. 389 ———., ed. Recópolis y la ciudad en la época visigoda. Zona Arqueológica 9. Alcalá de Henares: Museo Arqueológico Regional: 2008. Olmo Enciso, Lauro, and Manuel Castro Priego. ‘Coins, Cities, and Archaeological Contexts in the Centre of the Iberian Peninsula Between the 6th and 8th Centuries AD: Reccopolis and Toledo’. In Numismatica e archeologia. Monete, stragrafie e contesti: Dati a confronto, edited by Giacomo Pardini, Nicola Parise, and Flavia Marani, 557–72. Rome: Quasar, 2018. Olmo Enciso, Lauro, Manuel Castro Priego, Amaya Gómez de la Torre-Verdejo, and Álvaro Sanz Paratcha. ‘Recópolis y su justificación científica: la secuencia estratigráfica’, in Recópolis y la ciudad, ed. Olmo Enciso, 64–75. Oosthuizen, Susan. The Anglo-Saxon Fenland. Oxford: Windgather, 2017. Ousterhout, Robert. Eastern Medieval Architecture: The Building Traditions of Byzantium and Neighboring Lands. Onassis Series in Hellenic Culture. Oxford: Oxford University Press, 2019. Patroni, Giovanni. ‘Avanzi di edificio sovrapposto al pavimento romano scoperto sotto il Corso Vit- torio Emanuele. Altri frammenti architettonici’ Notizie degli scavi di antichità, 1924, 265–68. Pattitucci Uggeri, Stella. ‘La politica navale di Teoderico: Riflessi topografici nel Ravennate’. In Teode- rico il Grande e i Goti d’Italia. Atti del 13° congresso internazionale di studi sull’alto medioevo, 2:785–86. Spoleto: CISAM, 1992. Patzold, Steffen. ‘Der König als Alleinherrscher? Ein Versuch über die Möglichkeit der Monarchie im Frühmittelalter’. In Monarchische Herrschaft im Altertum, edited by Stefan Rebenich, 605–633. Schriften des Historischen Kolleges 94. Berlin: De Gruyter, 2017. Peduto, Paolo. ‘Arechi II e Salerno’. In Tra i Langobardi del Sud: Arechi II e il Ducato di Benevento, edited by Marcello Rotili, 355–74. Padua: Il Poligrafo, 2017. ———. ‘Consuetudine ed evoluzione dell’antico nelle costruzioni di Arechi II’. In Salerno. Una sede ducale della Langobardia meridionale, edited by Paolo Peduto, Rosa Fiorillo, and Angela Corolla, 1–19. Studi e Ricerche di Archeologia e Storia dell’Arte 16. Spoleto: CISAM, 2013. ———. ‘Paolo Diacono e la Cappella Palatina di Salerno’. In Paolo Diacono e il Friuli altomedievale (secc. VI-X). Atti del XIV Congresso internazionale di studi sull’Alto Medioevo, 2:655–70. Spoleto: CISAM, 2001. Peduto, Paolo, Matilde Romito, Maria Galante, Daniela Mauro, and Irma Pastore. ‘Un accesso alla storia di Salerno: stratigrafie e materiali dell’area palaziale longobarda’. Rassegna Storica Salerni- tana 5, no. 2 (1988): 9–64. Périn, Patrick. ‘Paris mérovingien, sedes regia’. Klio 71, no. 2 (1989): 487–502. Perrugot, Didier. ‘Le palais mérovingien de Malay (Yonne): histoire et archéologie’. In Palais royaux et princiers au Moyen Âge : actes du colloque international tenu au Mans les 6–7 et 8 october 1994, edited by Annie Renoux, 147–56. Le Mans: Publications de l'Université du Maine, 1996. 390 Petolettii, Marco. ‘L’Iconografia Rateriana: Le didascale e i versi celebrativi’. In La piü antica veduta di Verona: L’Iconografia Rateriana. L’archetipo e l’immagine tramandata, edited by Antonella Arzone and Ettore Napione, 33–46. Verona: Comune di Verona, 2012. Peyer, Hans Conrad. ‘Das Reisekönigtum des Mittelalters’. In Vierteljahrschrift für Sozial- und Wirtschafts- geschichte 51 (1964): 1–21. Pflug, Jens. ‘Die Bauliche Entwicklung der Domus Augustana im Kontext des sudöstlichen Palatin bis in severische Zeit’. In Palast und Stadt im severischen Rom, ed. Sojc et al., 181–212. Picard, Jean-Charles. ‘Le quadriportique de Saint-Laurent de Milan’. MEFRA 85, no. 2 (1973): 691– 712. Pisani Sartorio, Giuseppina, and Raissa Calza. La villa di Massenzio sulla Via Appia. Vol. 1. Monumenti Romani 6. Rome: Istituto di studi romani, 1976. Pitz Martina, and Roland Puhl. 'Trisgodros = Triguères (Loiret) ? Pour une nouvelle localisation d'une villa publica énigmatique mentionnée dans une charte de Pépin le Bref'. Nouvelle revue d'ono- mastique 49-50 (2008): 55-81. Pohl, Walter. ‘Staat und Herrschaft im Frühmittelalter: Überlegungen zum Forschungsstand’. In Staat im frühen Mittelalter, ed. Airlie et al., 10–38. Polci, Barbara. ‘Some Aspects of the Transformation of the Roman Domus between Late Antiquity and the Early Middle Ages’. In Theory and Practice in Late Antique Archaeology, edited by Luke Lavan and William Bowden, 89–105. Late Antique Archaeology 1. Leiden: Brill, 2003. Pryor, John, and Elizabeth Jeffreys. The Age of the ΔΡΟΜΩΝ: The Byzantine Navy, ca. 500–1204. The Medieval Mediterranean 62. Leiden: Brill, 2006. Raddatz, Klaus. ‘Studien zu Reccopolis I. Die archäologischen Befunde’. Madrider Mitteilungen 5 (1964): 213–23. Rasch, Jürgen. Das Maxentius-Mausoleum an der Via Appia in Rom. Mainz: Philipp von Zabern, 1984. Rauch, Christian. Die Ausgrabungen in der Königspfalz Ingelheim 1909–1914. Edited by Hans Jörg Jacobi. Studien zur Königspfalz Ingelheim 1. Mainz: Verlag des Römisch-Germanischen Zentralmu- seums, 1976. Raw, Barbara. ‘Royal Power and Royal Symbols in Beowulf’. In The Age of Sutton Hoo: the Seventh Century in North-Western Europe, edited by Martin Carver, 167–74. Woodbridge: Boydell, 1992. Rembold, Ingrid. Conquest and Christianization: Saxony and the Carolingian World, 772–888. CSMLT, 4th Series, 108. Cambridge: Cambridge University Press, 2017. Reusch, Wilhelm. ‘Die Aula Palatina in Trier. Ergebnisse der Untersuchungen in den Jahren 1951– 1953’. Germania 33 (1955): 180–99. ———. ‘Die Außengalerien der sog. Basilika in Trier’. Trierer Zeitschrift (1949): 170–87. 391 Reuter, Timothy. ‘Assembly Politics in Western Europe from the Eighth Century to the Twelfth’. In The Medieval World, edited by Peter Linehan and Janet Nelson, 432–50. London: Routledge, 2001. Reprinted in Reuter, Medieval Politics and Modern Mentalities, 193–216. ———. Medieval Politics and Modern Mentalities. Edited by Janet Nelson. Cambridge: Cambridge Uni- versity Press, 2006. ———. ‘Plunder and Tribute in the Carolingian Empire’, Transactions of the Royal Historical Society, 35 (1985): 75–94. Reprinted in Reuter, Medieval Politics and Modern Mentalities, 231–50. ———. ‘Regemque quem in Francia pene perditit, in patria magnifice recepit: Ottonian Ruler Representation in Synchronic and Diachronic Comparison’. In Herrschaftsrepräsentationen im ottonischen Sachsen, edited by Gerd Althoff and Ernst Schubert, 363–80. Vorträge und Forschungen 46. Sigmarin- gen: Jan Thorbecke, 1998. Reprinted in Reuter, Medieval Politics and Modern Mentalities, 127–46. ———. ‘“You Can’t Take It With You”: Testaments, Hoards, and Moveable Wealth in Europe, 600– 1100’. In Treasure in the Early Medieval West, ed. Elizabeth Tyler, 11–24. York: York Medieval Press, 2000. Reydellet, Marc. ‘La regalità teodericiana’. In Dall’età bizantina all’età ottoniana, bk. 2, Ecclesiologia, cultura e arte edited by Antonio Carile, 9–30. Vol. 2 of Storia di Ravenna. Ravenna: Marsilio, 1992. Reynolds, Susan. ‘The Historiography of the Medieval State’. In Companion to Historiography, edited by Michael Bentley, 117–38. London: Routledge, 1997. Richmond, Ian. The City Wall of Imperial Rome: An Account of its Architectural Development from Aurelian to Narses. Oxford: Clarendon Press, 1930. Rheinisches Landesmuseum Trier. Trier: Kaiserresidenz und Bischofssitz Die Stadt in spätantiker und früh- christlicher Zeit. Mainz: Philipp von Zabern, 1984. Ripoll, Gisela. ‘Sedes regiae en la Hispania de la Antigüedad Tardía’, in Sedes regiae (ann. 400–800), ed. Ripoll and Gurt, 385–90. Ripoll, Gisela, and Josep Gurt, eds. Sedes regiae (ann. 400–800). Barcelona: Rreial Acadèmia de Bones Lletres, 2000. Rizzardi, Clementina. ‘Note sull’antico episcopio di Ravenna: formazione e sviluppo’. In Actes du XIe congrès international d'archéologie chrétienne. Lyon, Vienne, Grenoble, Genève, Aoste (21-28 septembre 1986), edited by Noël Duval, 2:711–31. Rome: École Française de Rome, 1989. Roach, Levi. ‘Hosting the King: Hospitality and the Royal iter in Tenth-Century England’. Journal of Medieval History 37 (2011): 34–46. Röder, Josef. ‘Neue Ausgrabungen in Andernach’. Germania 39 (1961): 208–13. 392 Roberts, Michael. ‘The Description of Landscape in the Poetry of Venantius Fortunatus: The Mo- selle Poems’. Traditio 49 (1994): 1–22. Roffia, Elisabetta. ‘Le fortificazioni di Sirmione: nuove ricerche’. In Le fortificazioni del Garda, ed. Brogi- olo, 21–38. Rollason, David. The Power of Place: Rulers and their Palaces, Landscapes, Cities, and Holy Places. Princeton: Princeton University Press, 2016. Rouche, Michel. ‘L’héritage de la voirie antique dans la Gaule du Haut Moyen Âge’. In L’homme et la route en Europe occidentale au Moyen Âge et aux temps modernes, edited by Charles Higounet, 13– 32. Flaran 2. Toulouse: Presses universitaires du Midi, 1982. Roueché, Charlotte. ‘Acclamations in the Later Roman Empire: New Evidence from Aphrodisias’. JRS 74 (1984): 181–99. Rotili, Marcello. ‘Arechi II e Benevento’. In Tra i Langobardi del Sud: Arechi II e il Ducato di Benevento, edited by Marcello Rotili, 181–226. Padua: Il Poligrafo, 2017. Rubner, Heinrich. ‘Vom römischen Saltus zum fränkischen Forst’. Historisches Jahrbuch 83 (1964): 271– 77. Rusconi, Antonino. ‘La chiesa di S. Sofia di Benevento’. Corso di cultura sull’arte ravennate e bizantina 14 (1967): 339–59. Saal, Eveline. ‘Forschungen zum antiken Rheinhafen in Andernach’. In Der Rhein als europäische Ver- kehrsachse: Die Römerzeit, edited by Heike Kennecke, 63–77. Bonn: Vor- und Frühgeschichtliche Archäologie, Rheinische Friedrich-Willhelms-Universität, 2014. Safran, Linda. ‘Points of View: The Theodosian Obelisk Base in Context’. Greek, Roman, and Byzantine Studies 34, no. 4 (1993): 409–35. Said, Edward. Orientalism. New York: Pantheon, 1978. Sage, Walter. ‘Die Ausgrabungen in der Pfalz zu Ingelheim am Rhein 1960–1970’. Francia 4 (1976): 141–60. Samson, Ross. ‘Carolingian Palaces and the Poverty of Ideology’. In Meaningful Architecture: Social Inter- pretations of Buildings, edited by Martin Locock, 99–131. Worldwide Archaeology Series 9. Al- dershot: Avebury, 1994. ———. ‘The Merovingian Nobleman’s Home: Castle or Villa?’. Journal of Medieval History 13, no. 4 (1987): 287–315. ———. ‘The Residences of the Potentiores in Gaul and Germania in the Fifth to Mid-Ninth Century’. PhD Dissertation, Glasgow, University of Glasgow, 1991. 393 Santangeli Valenzani, Riccardo. ‘Public and Private Building Activity in Late Antique Rome’. In Tech- nology in Transition, A.D. 300–650, edited by Luke Lavan, Enrico Zanini, and Alexander Sar- antis, 435–49. Late Antique Archaeology 4. Leiden: Brill, 2007. ———. ‘Residential Building in Early Medieval Rome’. In Early Medieval Rome and the Christian West: Essays in Honour of Donald Bullough, edited by Julia Smith, 101–12. The Medieval Mediter- ranean 28 (Leiden: Brill, 2000). Sarti, Laury. ‘Die spätantike Militärpräsenz und die Entstehung einer militarisierten „Grenzgesell- schaft“ in der nordwesteuropäischen limes-Region’. In Militärische Migration vom Altertum bis zur Gegenwart, edited by Christoph Rass, 43–56. Studien zur Historischen Migrationsforschung 30. Paderborn: Ferdinand Schöningh, 2016. Schama, Simon. Landscape and Memory. New York: Alfred A. Knopf, 1995. Schaub, Andreas, and Tanja Kohlberger-Schaub. ‘On the Origins of the Great Carolingian Place of Power: Recent Excavations at Aachen Cathedral’. In Churches and Social Power in Early Medieval Europe: Integrating Archaeological and Historical Approaches, edited by José C. Sánchez-Pardo and Michael G. Shapland, 474–94. Studies in the Early Middle Ages 42. Turnhout: Brepols, 2015. Schieffer, Rudolf. ‘Von Ort zu Ort. Aufgaben und Ergebnisse der Erforschung ambulanter Herr- schaftspraxis’. In Orte der Herrschaft, ed. Ehlers, 11–24. Schönberger, Hans. ‘The Roman Frontier in Germany: An Archaeological Survey’. JRS 59, no. 1–2 (1969): 144–97. Schulte, Aloys. ‘Anläufe zu einer festeren Residenz der deutschen Könige im Hochmittelalter’. Histo- risches Jahrbuch 55 (1935): 131–42. ———. ‘Pavia und Regensburg’. Historisches Jahrbuch 52 (1932): 465–76. Schlunk, Helmut. ‘Las pinturas de Santullano. Avance al estudio de la pintura mural asturiana de los siglos IX y X’. Archivo Español de Arqueología 25 (1952): 15–37. Schlunk, Helmut, and Magín Berenguer. La pintura mural asturiana de los siglos IX y X. Translated by María de los Angeles Vázquez de Parga. Madrid: L. Sanchez Cuesta, 1957. Schmidt, Walter. ‘Die deutsche Hauptstadt-Diskussion in der Revolution von 1848/49’. Sitzungsberichte der Leibniz-Sozietät 6 (1995): 5–30. Schutte, Sven. ‚Der Aachener Thron’. In Krönungen: Könige in Aachen – Geschichte und Mythos, edited by Mario Kramp, 213–22. Mainz: Phillipp von Zabern, 2000. Schwartz, Barry. ‘The Social Psychology of the Gift’. American Journal of Sociology 73, no. 1 (1967): 1– 11. Segal, Zef. ‘The Two-Edged Sword: Capital Cities and the Limits to State Centralization in Mid Nine- teenth-Century Germany’. Journal of Historical Geography 60 (2018): 52–63. 394 Settia, Aldo. ‘Le fortificazioni dei Goti in Italia’. Teoderico il grande e i Goti d’Italia. Atti del 13° congresso internazionale di studi sull’alto medioevo, 1:101–32. Spoleto: CISAM, 1992. Shils, Edward. ‘Center and Periphery’. In Edward Shils, Center and Periphery: Essays in Macrosociology, 3– 16. Selected Papers of Edward Shils 2. Chicago: University of Chicago Press, 1975. Sfameni, Carla. Ville residenziali nell’Italia tardoantica. Munera 25. Bari: Edipuglia, 2006. Small, Alistair, and Robert Buck. The Excavations of San Giovanni di Ruoti. 2 vols. Phoenix Supplementa 33–34. Toronto: University of Toronto Press, 1994–1997. Smith, Adam. The Political Landscape: Constellations of Authortiy in Early Complex Polities. Berkeley: Univer- sity of California Press, 2004. Smith, Julia. Europe after Rome: A New Cultural History, 500–1000. Oxford: Oxford University Press, 2002. Soja, Edward. Postmodern Geographies: The Reassertion of Space in Critical Social Theory. London: Verso, 1989. Sojc, Natascha, Aloys Winterling, and Ulrike Wulf-Rheidt, eds. Palast und Stadt im severischen Rom. Stuttgart: Franz Steiner, 2013. Sonne, Wolfgang. Representing the State: Capital City Planning in the Early Twentieth Century. Munich: Pres- tel, 2004. Squatriti, Paolo. Landscape and Change in Early Medieval Italy: Chestnuts, Economy, and Culture. Cambridge: Cambridge University Press, 2013. Stafford, Pauline. ‘Queens and Treasure in the Early Middle Ages’. In Treasure in the Early Medieval West, edited by Elizabeth Tyler, 61–82. York: York Medieval Press, 2000. Stamm, Otto. ‘Zur karolingischen Königspfalz in Frankfurt am Main’, Germania 33 (1955): 391–401. Stathakopoulos, Dionysios. Famine and Pestilence in the Late Roman and Early Byzantine Empire: a Systematic Survey of Subsistence Crises and Epidemics. Birmingham Byzantine and Ottoman Monographs 9. Farnham: Ashgate, 2004. Steinby, Eva Margareta, ed. Lexicon Topographicum Urbis Romae. 6 vols. Rome: Quasar, 1993–2000. Stiegemann, Christoph and Matthias Wemhoff, eds. 799. Kunst und Kultur der Karolingerzeit. Karl der Große und Papst Leo III. in Paderborn. Vol. 3. Mainz: Phillipp von Zabern, 1999. Stocking, Rachel. Bishops, Councils, and Consensus in the Visigothic Kingdom, 589–633. Ann Arbor, MI: University of Michigan Press, 2000. Stoclet, Alain. Autour de Fulrad de Saint-Denis (v. 710–784). Hautes Études Médiévales et Modernes 72. Geneva: Librarie Droz, 1993. 395 Stoner, Heidi. ‘Heaven and Hall: Space and Place in Anglo-Saxon England’. In Place and Space in the Medieval World, edited by Meg Boulton, Heidi Stoner, and Jane Hawkes, 159–77. Routledge Research in Art History. New York: Routledge, 2018. Strayer, Joseph. On the Medieval Origins of the Modern State. Princeton: Princeton University Press, 1970. Stumberg Edmunds, Martha Mel. Piety and Politics: Imaging Divine Kingship in Louis XIV’s Chapel at Ver- sailles. Studies in Seventeenth- and Eighteenth-Century Art and Culture. Newark, DE: Univer- sity of Delaware Press, 2002. Swoboda, Karl. ‘Palazzi antichi e medievali’. Bollettino del Centro di studi per la storia dell’architettura 11 (1957): 3–32. ———. The Problem of the Iconography of Late Antique and Early Mediaeval Palaces’. JSAH 20, no. 2 (1961): 78–89. ———. Römische und romanische Paläste: eine architekturgeschichtliche Untersuchung. Vienna: Anton Schroll, 1919. Syvitski, James, Albert J. Kettner, Anna Correggiari, and Bruce W. Nelson. ‘Distributary Channels and Their Impact on Sediment Dispersal’ Marine Geology 222–23 (2005): 75–94. Sykes, Naomi. ‘Deer, Land, Knives and Halls: Social Change in Early Medieval England’. Antiquaries Journal 90 (2010): 175–93. Talbot, Richard. Rome’s World: The Peutinger Map Reconsidered. Cambridge: Cambridge University Press, 2010. Theuws, Frans. ‘Maastricht as a Centre of Power in the Early Middle Ages’. In Topographies of Power in the Early Middle Ages, edited by Mayke de Jong, Frans Theuws, and Carine van Rhijn, 155– 216. Transformation of the Roman World 6. Leiden: Brill, 2001. Tuan, Yi-Fu. Space and Place: The Perspective of Experience. Minneapolis, MN: University of Minnesota Press, 1977. Ulbert, Thilo, ed. Ad pirum (Hrušica). Spätrömische Passbefestigung in den Julischen Alpen: der deutsche Beitrag zu den slowenisch-deutschen Grabungen, 1971–1973. Münchner Beiträge zur Vor- und Frühge- schichte 31. Munich: Beck, 1981. Urbano, Arthur. ‘Donation, Dedication, and Damnatio Memoriae: The Catholic Reconciliation of Ra- venna and the Church of Sant’Apollinare Nuovo’. Journal of Early Christian Studies 13, no. 1 (2005): 71–110. Velázquez, Isabel, and Gisela Ripoll. ‘Toletvm, la construcción de una vrbs regia’. In Sedes regiae (ann. 400– 800), ed. Ripoll and Gurt, 521–78. Veyne, Paul. Bread and Circuses: Historical Sociology and Political Pluralism. Translated by Brian Pearce. London: Penguin, 1990. 396 von Massow, Wilhelm. ‘Der Circus des römischen Trier’. Trierer Zeitschrift 18 (1949): 161–67. von Petrikovits, Harald. ‘Fortifications in the North-Western Roman Empire from the Third to the Fifth Centuries A.D.’. JRS 61 (1971): 178–218. Vuolo, Antonio. ‘Agiografia beneventana’. In Longobardia e longobardi nell’Italia meridionale: le istituzioni ecclesiastiche. Atti del 2° Convegno internazionale di studi promosso dal Centro di Cultura dell’Università Cattolica del Sacro Cuore Benevento, 29–31 maggio 1992, edited by Giancarlo Andenna and Giorgo Picasso, 199–238. Bibliotheca Eurdita Studi e Documenti di Storia e Filologia 11. Milan: Vita e Pensiero, 1996. Waitz, Georg. Deutsche Verfassungsgeschichte. 8 vols. Kiel: Schwers’sche Buchhandlung and Ernst Homann, 1844–1878. Wagner, Heinrich. ‘Zur Topographie von Königsgut und Pfalz Salz’. In Pfalzen – Reichsgut – Königshöfe, edited by Lutz Fenske, 149–83. Deutsche Königspfalzen 4. Göttingen: Vandenhoeck & Ruprecht, 1996. Walker, Rose. Art in Spain and Portugal from the Romans to the Early Middle Ages: Routes and Myths. Kala- mazoo, MI: Medieval Institute Publications, 2016. Ward, Elizabeth. ‘Caesar’s Wife: The Career of the Empress Judith’. In Charlemagne’s Heir: New Perspec- tives on the Reign of Louis the Pious (814–840), edited by Roger Collins and Peter Godman, 205– 27. Oxford: Clarendon Press, 1990. Ward-Perkins, Bryan. From Classical Antiquity: Urban Public Building in Northern and Central Italy, AD 300–850. Oxford Historical Monographs. Oxford: Oxford University Press, 1984. Weber, Max. Economy and Society: A New Translation. Translated by Keith Tribe. Cambridge, MA: Har- vard University Press, 2019. Weise, Georg. Zwei fra ̈nkische Königspfalzen: Bericht u ̈ber die an den Pfalzen zu Quierzy und Samoussy vorgenom- menen Grabungen. Tübingen: Fischer, 1923. Werner, Mattias. Der Lütticher Raum in frühkarolingischer Zeit. Untersuchungen zur Geschichte einer karolinischer Stammlandschaft. Veröffentlichungen des Max-Planck-Instituts für Geschichte 62. Göttingen: Vandenhoeck & Ruprecht, 1979. Werther, Lukas. ‘Der Königsgutkomplex Salz und das Neustädter Becken – ein frühmittelalterlicher Zentralraum im Wandel der Zeit’. In Zentrale Orte und Zentrale Räume des Frühmittelalters in Süd- deutschland, edited by Peter Ettel and Lukas Werther, 89–112. RGZM-Tagungen 18. Mainz: Verlag des Römisch-Germanischen Zentralmuseums, 2013. Werther, Lukas, Janet Nelson, Franz Herzig, Johannes Schmidt, Stefanie Berg, Peter Ettel, Sven Lin- zen, and Christoph Zielhofer. ‘792 or 793? Charlemagne’s Canal Project: Craft, Nature and Memory’. EME 28, no. 3 (2020): 444–65. Whittaker, C. R. Rome and its Frontiers: The Dynamics of Empire. London: Routledge, 2004. 397 Weizman, Eyal. Hollow Land: Israel’s Architecture of Occupation 2nd ed. London: Verso, 2017. Wickham, Chris. Early Medieval Italy: Central Power and Local Society, 400–1000. London: MacMillan, 1981 (repr. Ann Arbor, MI: University of Michigan Press, 1989). ———. Framing the Early Middle Ages: Europe and the Mediterranean, 400–800. Oxford: Oxford Univer- sity Press, 2005. Widdowson, Marc. ‘Merovingian Partitions: A “Genealogical Charter”?’. EME 17 (2009): 1–22. Wijnendaele, Jeroen. ‘Generalissimos and Warlords in the Late Roman West’. In War, Warlords and Interstate Relations in the Ancient Mediterranean, edited by Toni Ñaco del Hoyo and Fernando Lopez-Sanchez, 427–51. Impact of Empire 28. Leiden: Brill, 2017. ———. The Last of the Romans: Bonifatius, Warlord and comes Africae. London: Bloomsbury Aca- demic, 2014. Wilkes, John Joseph. Diocletian’s Palace, Split: Residence of a Retired Roman Emperor. Sheffield: University of Sheffield, 1986. Wilson, Roger J. A. Piazza Armerina. Archaeological Sites. London: Granada, 1983. Wood, Ian. ‘Giovardi, MS Verolensis 1, Arichis and Mercurius’. In Zwischen Niederschrift und Wieder- schrift: frühmittelalterliche Hagiographie und Historiographie im Spannungsfeld von Kompendienüberlieferung und Editionstechnik, edited by Richard Corradini, Maximilian Diesenberger, Meta Niederkorn- Bruck, 197–210. Forschungen zur Geschichte des Mittelalters 18. Vienna: Verlag der Öster- reichischen Akademie der Wissenschaften, 2010. ———. The Merovingian Kingdoms, 450–751. London: Longman, 1994. ———. The Modern Origins of the Early Middle Ages. Oxford: Oxford University Press, 2013. ———. ‘Theoderic’s Monuments in Ravenna’. In The Ostrogoths from the Migration Period to the Sixth Century: An Ethnographic Perspective, edited by Sam Barnish and Federico Marazzi, 249–77. Stud- ies in Historical Archaeoethnology. Woodbridge: Boydell, 2007. Wood, Jamie. ‘Defending Byzantine Spain: Frontiers and Diplomacy’. EME 18 (2010): 292–319. Wortley, John. ‘The Trier Ivory Reconsidered’. Greek, Roman, and Byzantine Studies 21, no. 4 (1980): 381–94. Wulf-Rheidt, Ulrike. ‘Die Bedeutung der severischen Paläste für spätere Residenzbauten’. In Palast und Stadt im severischen Rom, ed. Sojc et al., 287–306. ———. ‘Die schwierige Frage der Nutzung des römischen Kaiserpalastes auf dem Palatin in Rom in der Spätantike’. Antiquité Tardive 25 (2017): 127–48. 398 Wyss, Michäel. ‘Die Klosterpfalz Saint-Denis im Licht der neuen Ausgrabungen’. In Splendor palatii. Neue Forschungen zu Paderborn und anderen Pfalzen der Karolingerzeit, edited by Lutz Fenske, 175– 92. Deutsche Königspfalzen 5. Göttingen: Vandenhoeck & Ruprecht, 2001. ———. ‘Saint-Denis’, in 799. Kunst und Kultur, ed. Stiegemann and Wemhoff, 3:138–41. Zangara, Vincenza. ‘Una predicazione alla presenza dei principi: la Chiesa di Ravenna nella prima metà del sec. V’. Antiquité Tardive 8 (2001): 265–304. Zanker, Paul. The Power of Images in the Age of Augustus. Translated by Alan Shapiro. Ann Arbor, MI: University of Michigan Press, 1988. Ziemssen, Hauke. ‘Die Kaiserresidenz Rom in der Zeit der Tetrarchie’. In Rom und Mailand in der Spätantike. Repräsentionen städtischer Räume in Literatur, Architektur und Kunst, edited by Therese Fuhrer, 87–110. Topoi. Berlin Studies of the Ancient World 4. Berlin: De Gruyter, 2011. Zotz, Thomas. ‘Pfalzen zur Karolingerzeit. Neue Aspekte aus historischer Sicht’, in Splendor palatii. Neue Forschungen zu Paderborn und anderen Pfalzen der Karolingerzeit, edited by Lutz Fenske, 13–23. Deutsche Königspfalzen 5. Göttingen: Vandenhoeck & Ruprecht, 2001. 399