”The Fall and Decline of the Roman Urban Mind”

The Urban Mind Cultural and Environmental Dynamics Edited by Paul J.J. Sinclair, Gullög Nordquist, Frands Herschend and Christian Isendahl African and Comparative Archaeology Department of Archaeology and Ancient History Uppsala University, Uppsala, Sweden 2010 Cover: NMH THC 9113 Artist: Cornelius Loos Panorama over the southern side of Istanbul facing north east. Produced in 1710. Pen and brush drawing with black ink, grey wash, water colour on paper. The illustration is composed of nine separate sheets joined together and glued on woven material. Original retouching glued along the whole length of the illustration. Dimensions (h x b) 28,7 x 316 cm Photograph © Erik Cornelius / Nationalmuseum English revised by Laura Wrang. References and technical coordination by Elisabet Green. Layout: Göran Wallby, Publishing and Graphic Services, Uppsala university. ISSN 1651-1255 ISBN 978-91-506-2175-4 Studies in Global Archaeology 15 Series editor: Paul J.J. Sinclair. Editors: Paul J.J. Sinclair, Gullög Nordquist, Frands Herschend and Christian Isendahl. Published and distributed by African and Comparative Archaeology, Department of Archaeology and Ancient History, Uppsala University, Box 626, S-751 26 Uppsala. Printed in Sweden by Edita Västra Aros AB, Västerås 2010 – a climate neutral company. 341 009 Trycksak Table of Contents Preface ....................................................................................................... 9 The Urban Mind: A Thematic Introduction Paul J.J. Sinclair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1. Climate Variability in the Eastern Mediterranean and the Middle East during the Holocene Martin Finné and Karin Holmgren . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2. Cultural Interaction and Cognitive Expressions in the Formation of Ancient Near Eastern Societies Kristina J. Hesse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3. Climate Change, Ecology and Early Sedentism in Interaction: Visible Traces of the Early Urban Mind in Continental and Northern Europe Julia Mattes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4. Cities and Urban Landscapes in the Ancient Near East and Egypt with Special Focus on the City of Babylon Olof Pedersén, Paul J.J. Sinclair, Irmgard Hein and Jakob Andersson .... 113 5. Social and Environmental Dynamics in Bronze and Iron Age Greece Erika Weiberg, Michael Lindblom, Birgitta Leppänen Sjöberg and Gullög Nordquist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 6. The Urban Mind is the Normalcy of Urbanity Svante Fischer and Frands Herschend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 7. The Role of Natural Phenomena in the Rise and Fall of Urban Areas in the Sistan Basin on the Iranian Plateau (Southern Delta) Behrooz Barjasteh Delforooz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 8. Concepts of the City-State in Ancient Greece Susanne Carlsson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 9. Long-term Resilience: The Reconstruction of the Ancient Greek Polis of Kos after Earthquakes in the Period c. 200 BCE to c. 200 CE Kerstin Höghammar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 10. The Fall and Decline of the Roman Urban Mind Svante Fischer, Hans Lejdegård and Helena Victor .......................... 277 11. Why Are the So-Called Dead Cities of Northern Syria Dead? Witold Witakowski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 12. Lost in the City: An Essay on Christian Attitudes towards Urbanism in Late Antiquity Mats Eskhult . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 13. Constantinople in the Transition from Late Antiquity to the Middle Ages Ewa Balicka-Witakowska . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 14. The Urban Anthropocene: Lessons for Sustainability from the Environmental History of Constantinople John Ljungkvist, Stephan Barthel, Göran Finnveden and Sverker Sörlin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367 15. Innovative Memory and Resilient Cities: Echoes from Ancient Constantinople Stephan Barthel, Sverker Sörlin and John Ljungkvist . . . . . . . . . . . . . . . . . . . . . . . . . . 391 16. What’s in a Name? Mistra – The Town. Gullög Nordquist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407 17. The Linguistic Landscape of Istanbul in the Seventeenth Century Éva Á. Csató, Bernt Brendemoen, Lars Johanson, Claudia Römer and Heidi Stein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415 18. Multilingualism and Language Contact in Urban Centres along the Silk Road during the First Millennium AD Christiane Schaefer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441 19. Is There an “Urban Mind” in Balochi Literature? Carina Jahani . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457 20. ‘ James His Towne’ and Village Nations: Cognitive Urbanism in Early Colonial America Neil Price . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471 21. Early Urbanism in Scandinavia Charlotta Hillerdal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 22. Greening the Ancient City: The Agro-Urban Landscapes of the Pre-Hispanic Maya Christian Isendahl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527 23. Southeast Asian Urban Minds: An Example From Laos Anna Karlström . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553 24. Conceptualising the Urban Mind in Pre-European Southern Africa: Rethinking Mapungubwe and Great Zimbabwe Munyaradzi Manyanga, Innocent Pikirayi and Shadreck Chirikure . . . . . 573 25. Towards an Archaeology of the Future: the Urban Mind, Energy Regimes and Long-term Settlement System Dynamics on the Zimbabwe Plateau Paul J.J. Sinclair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591 Participants ........................................................................................... 617 10. The Fall and Decline of the Roman Urban Mind Svante Fischer, Hans Lejdegård and Helena Victor Contact details Dr Svante Fischer Department of Archaeology and Ancient History Uppsala University Box 626 751 26 Uppsala, Sweden [email protected] Dr Hans Lejdegård Department of Archaeology and Ancient History Uppsala University Box 626 751 26 Uppsala, Sweden [email protected] Dr Helena Victor Kalmar Läns Museum Box 104 391 21 Kalmar, Sweden [email protected] ABSTRACT This chapter discusses the 5th-century west Roman imperial residences of Rome and its substitutes Arles and Ravenna, as understood within the framework of an imperialist ideology of urbanism, the “Roman urban mind”. During the late Roman Empire, the city of Rome was the central focus of the old Roman infrastructure. Ideally, the highest echelons of the imperial administration also ought to be located in Rome. There was an underlying idea that the purpose of the Roman Empire was to sustain the city of Rome – the capital of the world – and its ever-growing population. In this paper the authors argue that in spite of the fascination with Rome as the caput mundi, urban sustainability and resilience were problematic matters within the West Roman Empire. The imperial state apparatus proved incapable of resolving these issues in the face of barbarian attacks and internal strife. This spelled the end for the Roman urban mind. Introduction This chapter discusses the 5th-century West Roman imperial residences of Rome and its substitutes Arles and Ravenna, as understood within the framework of an imperialist ideology of urbanism, the “Roman urban mind”. Paraphrasing and reversing the order of Edward Gibbon’s classic title, The Decline and Fall of the 277 Roman Empire1 the paper seeks to offer a new, widened frame of interpretation. It attempts to take into account different perspectives on how urban societies are incapable of change and transformation not only vertically in terms of order and hierarchy, but also horizontally in terms of mutuality between city and hinterland and the overall support system in terms of infrastructure. In addition, the perspective of the barbarian outsider, who prefers to act outside of established networks and mutuality, must be taken into account. In this chapter we argue that, in spite of the fascination with Rome as the caput mundi ‘the capital of the world’, urban sustainability and resilience were problematic matters within the West Roman Empire. The imperial state apparatus proved incapable of resolving these issues in the face of barbarian attacks and internal strife. When an urban population is faced with warfare, an irregular climate and food shortages, the urban leadership has to act. Yet the ensuing consequences of these actions may be quite far-reaching and go well beyond the intended results. During the apex of the Roman Empire, the city of Rome was the central focus of the Roman infrastructure. Ideally, the highest echelons of the imperial administration also ought to be located in Rome. This was no longer the case in the late 5th century, however. Roman political action appears to have lagged behind and followed the law of least possible resistance, relying on the recruitment of barbarian warriors who were paid in gold coinage. The reason for the western Roman military establishment’s increased recruitment of barbarian troops lay at least partially in the two major defeats suffered by western usurpers against Theodosius I in the late fourth century (Maximus in 388 and Eugenius in 394). In 395 Theodosius died, leaving Stilicho in charge of the depleted western armies. Before Stilicho could rebuild the military strength of the West he was faced with severe challenges in the form of invasions of Italy by Alaric, the fall of the Rhine-frontier in 406 and the usurpation of Constantine III in 407. The only course open to him was to recruit barbarian troops on a scale hitherto unseen. For the rest of the western empire’s existence barbarian troops would form the core of the army. It was very dificult for the inancially strained state apparatus to pursue, at times, very ambitious but short-lived enterprises against barbarian attacks and food shortages with other alternatives once the irst initiatives failed. Cases in point are the unsuccessful military campaigns of the emperors Majorian and Anthemius against the Vandals in North Africa and the Visigoths in Gaul and Spain in 458–461 and 467–471. Both Majorian and Anthemius were murdered by disgruntled barbarian oficers. Anthemius’ death coincided with a lengthy siege of Rome and a subsequent plundering in 472, as well as a widespread famine followed by an epidemic.2 Many unexpected reactions followed instead, notably internal strife amongst the military leadership within the imperial administration, which beneited the barbarian aggressors. This led to the collapse of the West Roman Empire as a political entity which, in turn, caused substantial de-urbanisation. The late Roman Empire of the 5th century may be described accordingly: two emperors rule one half of the empire each by means of a hierarchical state apparatus. Each emperor rules his half from a palace. This was usually located within an urban environment, normally Rome and Constantinople (Fig. 1). These two cities were to become dominant in imperial ideology during the 5th century, 1 Gibbon 1897. 2 Stathakopolous 2004. 278 Fig. 1. Map of 4th-century Rome since the 4th-century western capital of Trier was no longer used as an imperial residence after 391 and fell prey to a barbarian plundering in 406. In times of crisis, the western imperial residence could only be relocated from Rome or Milan to the safe outpost of Ravenna or to Arles, the residence in southern Gaul. And the crises came all the more frequently in the 5th century. Arles was to assume the position as seat of praetorian prefect of Gaul and thus as the imperial residential city of Gaul after the fall of Trier in 406. In Italy, the previous alternative capital of Milan was permanently replaced by Ravenna in 402 (see Fig. 2). Ravenna then also periodically came to replace Rome, especially during the reign of Valentinian III (425–455). Why did these shifts take place? In regard to the 5th- and 6th-century decline of the city of Rome, the fall of the West Roman Empire, and the establishment of Germanic successor kingdoms on its former territory, there are two main theories: gradual transformation and rapid collapse. Peter Brown formulated an inluential theory on a slow transformation with important regional differences.3 Walter Goffart has also argued for a slow reshaping of Roman society rather than an abrupt break.4 Lately, however, the traditional theory of catastrophe has become en vogue in Late Antique research. Bryan Ward-Perkins and Peter Heather describe the same chain of events as the collapse of a civilization with tangible repercussions.5 In particular, WardPerkins points to the fall in overseas commerce, notably the import of pottery to Italy. Sustaining the Roman Urban Mind During the Late Roman Empire, the city of Rome was the central focus of the old Roman infrastructure. Ideally, the highest echelons of the imperial administration also ought to be located in Rome. There was an underlying idea that 3 Brown 1971. 4 Goffart 1980. 5 Ward-Perkins 2005; Heather 2005. 279 Fig. 2. City Plan of Ravenna the purpose of the Roman Empire was to sustain the city of Rome – the capital of the world – and its ever-growing population.6 What was wrong with Rome, then? Urban areas were in general incapable of supporting their own populations. The residential cities were supplied with fresh water and comestibles from the outside. The population of Rome received state subsidies in terms of cheap foodstuffs, the annona. Oil and grain were transported from Egypt and North Africa via the ports of Portus and Ostia some 20 km downstream the Tiber River from Rome.7 In the late 4th century, the imperial administration had divided Italy into two administrative units, Italia annonaria and Italia suburbicaria, that is, the part of Italy that was providing the annona, and the part of Italy that was within reach of the city. This had little effect on sustaining the city of Rome, however. It is apparent that the 5th-century urban centres were vulnerable to sudden changes in the equilibrium of military security, fresh water supply, grain import, and the real monetary value of commodities. In order to maintain the status quo, the urban population of the imperial residential cities had to rely on a functioning military protection while being supported by the emperor by means of fresh water via aqueducts and heavily subsidized grain imports. On occasion, the imperial bureaucracies failed to provide all of this to the urban population. Life in 5th-century Rome proved to be dificult in comparison to life in the expanding Constantinople. During the 4th century and most of the 5th century the Roman infrastructure guaranteed the city of Rome both security and a relatively stable supply of neces- 6 Sirks 1991. 7 Sirks 1991. 280 sary foodstuffs. 8 With the fall of the West Empire in 476 and the Ostrogothic conquest of Italy in the 490s, there were only subtle changes to this order, and they were more in terms of scale.9 But the Byzantine conquest (534–554) and the Lombard invasion (568–571) caused signiicant changes. Breaks in settlement patterns and demographic reproduction along the main roads and junctions became apparent. It is clear that suburban settlements along the main entry routes into Rome, as well as larger towns and villages in the peripheral intersections, declined at a rapid pace at this time.10 People did not simply move to different locations; faced with changing conditions which had an adverse effect on their ability to sustain themselves, they disappeared completely. The Roman infrastructure remained to a large degree intact. But the network that enabled favourable conditions at nodal settlements within the infrastructure had become disadvantageous. Roads were open, but the opportunities to use them were limited and entailed considerable risks. A still functioning urban network was abandoned in a time of crisis. The traditional use of the network, for which it was built, disappeared and was replaced by one in which those who had previously beneited from it came to suffer. This is a reversed infrastructural order which beneited a few, but worked to the detriment of the population at large. The impressive infrastructure of roads and aqueducts has been seen as a necessary prerequisite for urban growth within the Roman Empire and especially for the viability of the city of Rome itself. 11 Cartographic material indicates that similar ideas were current among Romans. There was a clear emphasis on the layout of roads rather than an accurate graphical representation of the areas through which the roads passed.12 However, as will be shown below, this infrastructure could be equally instrumental in providing the means for depopulation and de-urbanisation. The main Roman roads in Italy were well built with the central part made of basalt stone.13 They therefore required relatively little maintenance. During the lengthy periods when the safety of the infrastructure was guaranteed by the military strength of the Roman Empire, road intersections generally attracted trade and settlement, that is mutuality. In the Roman Empire irst roads, then aqueducts and fortiications spurred urban growth during half a millennium. But what happened during periods of economic decline and deurbanisation? Without military security, the land-based network ceased to be a resource for mutuality and for the reproduction of the hierarchical order and became a broad, horizontal security risk within the immediate reach of small units of highly mobile cavalry units engaging in asymmetrical warfare.14 The likelihood of being exposed to violence and terror is greater the closer one is to an intersection in the network, as it is most likely that this is where the asymmetrical intrusion will take place. The city of Rome underwent depopulation during the 5th century. This depopulation process then spread from the city of Rome to the entire Italian countryside during the 6th century. Current population estimates of the city of Rome around 400 suggest approximately 800000 inhabitants. Some 300000 urban 8 Sirks 1991; cf. Symmachus Rel. 18, 35, 37. 9 Moorhead 1992. 10 Ward-Perkins 1984; 2005, 139–141. 11 Ward-Perkins 1984; Laurence 1999. 12 Salway 2005 13 Laurence 1999. 14 Mack 1975. 281 denizens disappeared during the crisis of 408–419 when Italy was invaded by the renegade Visigoth army unit under Alaric. By the 6th century the population of Rome had fallen to a mere 80000.15 This made things easier for the new Ostrogothic rulers of Italy in the late 5th and early 6th centuries. With a city population that had suffered a 90% decrease since 400, the Ostrogoths no longer needed to use the ports to import the annona from abroad. Instead, it was gathered within Italy and transported by means of the regional road network. “The Noble Order of Cities” An Apology for the Roman Urban Mind How is one to appreciate the Roman ideological perception of urbanism? Civilization, according to the Romans, was based on the city.16 Only barbarians lived in non-urban societies. It is furthermore important to understand that the Roman imperialist conception of the state and its urban civilisation was one of inertia, very slow to change, largely reactionary, and eager to put the blame on irrelevant factors. There were ixed notions that were supposed to be followed slavishly and little room for any major reform or radical critique. This was especially true in regard to urbanism as understood by the imperialist ideology. While working as a tutor for young Gallo-Roman nobles in Burdigala (present-day Bordeaux, France) c. 334–364, the poet Ausonius ranked the city of Arelate (present-day Arles, France) tenth among the top 20 Roman cities within the following order in his poem Ordo Urbium Nobilium (The Noble Order of Cities), see Table I: Table 1.Twenty Roman Cities Ranked According to the Ordo Urbium Nobilium Latin name Roma Constantinopolis Carthago Antiochia Alexandria Augusta Treverorum Mediolanum Capua Aquileia Arelate Hispalis Corduba Tarraco Baraco Athenae Catania Syracusa Tolosa Narbona Burdigala 15 Durliat 1990. 16 Revell 2009, e.g. 54. 282 Current name Rome, Italy Istanbul, Turkey Carthage, Tunisia Antioch, Turkey Alexandria, Egypt Trier, Germany Milan, Italy Capua, Italy Aquileia, Italy Arles, France Sevilla, Spain Cordoba, Spain Tarragona, Spain Braga, Portugal Athens, Greece Catania, Italy Syracuse, Italy Toulouse, France Narbonne, France Bordeaux, France Sack 410, 455, 472 439 484 406, 419 409, 488 488 411, 471 409–411 409–411 409–411 409–411 455 455 418 418 418 This urban ranking is clearly subjective. It is meant as an expression of local chauvinism and a didactic piece for young Gallo-Roman aristocrats, with the learned addition of towns credited with a great past such as Capua and Athens. Meanwhile, important 4th-century cities that struck gold coinage, such as Heraclea, Nicomedia, and Thessalonica, are missing. Note that Ausonius further claimed that Arles was the Rome of Gaul: ”Pande, duplex Arelate, tuos blanda hospita portus, Gallula Roma Arelas, quam Narbo Martius et quam accolit Alpinis opulenta Vienna colonis, praecipitis Rhodani sic intercisa luentis, ut mediam facias navali ponte plateam, per quem Romani conmercia suscipis orbis nec cohibes populosque alios et moenia ditas, Gallia quis fruitur gremioque Aquitania lato.”17 “Open your gates, double Arles, friendly hostess, Arles – little Rome of the Gauls, neighbour of Narbonne and Vienne, which have enriched the settlers of the Alps. The rapid current of the Rhône divides you in such a way that a boat-bridge forms a place in your middle. This river brings you all the produce of the Roman world, you do not steal it, and you enrich the other peoples and cities within Gaul and Aquitaine.” It is quite debatable, however, whether the adjective “duplex” in conjunction with the place name Arles was there to emphasize ideological grandeur or simply to emphasize the fact that the city was divided by the Rhône.18 It is rather evident that Arles never quite measured up to the two sisters, Rome and Constantinople, nor was it ever allowed to take on an ideological identity of its own, like Trier did in the 4th century when Ausonius composed his greatest poem Mosella in praise of the most distant imperial capital. Arles remained, at best, a provincial copy of Rome. Eleven of the twenty cities listed by Ausonius were under barbarian control by the 5th century and most cities had been sacked by barbarians or renegade army units. These cities were thus already part of the general urban decline beyond the reach of the imperial administration. Another city not mentioned by Ausonius is Ravenna, which effectively came to replace Rome as the most favoured place of residence in the irst half of the 5th century. The ideological reason for Honorius to transfer his residence from Rome to Ravenna was debatable, but it was clearly a wise decision given the fall of Trier in 406, Milan in 408, Rome in 410, and Arles in 411.19 But despite the fact that Ravenna was, by and large, the exclusive imperial residence in the west during 402–440, and the seat of the praetorian prefect of Italy, this did not confer it a new status of ideological power at the expense of Rome – on the contrary. Rome was the burial ground of the Theodosian dynasty ever since Theodosius I had been buried there after his death in Milan in 395. The destroyed Church of San Giovanni Battista in Ravenna has been attributed to the empress Galla Placidia as a sort of dynastic church with a strong expression of the Christian faith of the Theodosian dynasty, but religious piety is not tantamount to ideological orthodoxy. Instead, the idea that Ravenna somehow stood for the supreme ideological power in Italy and that this was tied to religion should rather be attributed to the 6th century and 17 Ausonius, Ordo Urbium Nobilium, 73–80. 18 Constans 1921. 19 Lejdegård 2002. 283 the Arian Ostrogothic king Theoderic and later the Byzantine emperor Justinian with the foundation of the Exarchate of Ravenna. Rome and Its Hinterland Very few emperors were able to reside in Rome, although this was a key point in the imperial ideology that constituted the Roman urban mind. This had to do with the inconvenient location of Rome. Rome lies in the province of Lazio on seven hills on the Tiber River, a body of water that allows for boat trafic some 100 km upstream from the seaports of Ostia and Portus, the distance to Rome from the ports being some 20 km. As the Tiber is relatively narrow and shallow, it is prone to looding with ensuing complications for the grain stores along the city shores. If the ports were shut down or the Tiber was blocked, the road network remained the only alternative. The water supply of Rome, known as the cura aquarum, was maintained by a water system covering over 500 km. Eleven aqueducts and a further eight channels delivered water to the fountains and baths of Rome. Even if the main part of the water supply system was constructed between 312 BC and AD 109, the aqueducts were maintained and repaired by the government well into the mid-6th century.20 Ward-Perkins argues that the aqueducts fulilled an ideological rather than a sustaining role.21 In support of this view, he argues from a number of cases during the many 6th-century wars when aqueducts do not appear to have had a major impact on the sustainability of the urban population. However, the water supply that was able to provide 800000 people with water in the city of Rome during the late 4th century does not stand in reasonable proportion to that of a population that had shrunk by 90% a century later. The imperial capital of Rome was both a liability and an asset. The main defence of the city was its outer wall. First begun in the 270s, it measured 6 metres at its lowest point and was 3.8 metres wide. It stretched some 19 km with 29 gates and 381 towers, and enclosed some 2500 hectares and 46000 building blocks known as insulae, which at the population peak in the mid-4th century may have housed as many as one million people.22 The wall underwent signiicant changes during the irst half of the 5th century, especially after the 442 earthquake. Despite this massive defensive structure, Rome was plundered in 410, 455 and 472. This was mainly due to lack of food. When communications between Rome, and Ostia and Portus, the two main ports of Rome, were cut by a besieging army the problem of feeding the huge population became insurmountable. Later in the 6th century, when the population of Rome had shrunk signiicantly; the city walls of Rome were simply so vast that they could not be defended by the military resources available to the city authorities. The emperor was no longer safe there. So what were the alternatives to Rome? Arles and Its Hinterland Arles is located in the Rhône River delta in the south of France (Fig. 3). This 20 Lancon 2000; Bruun 1991. 21 Ward-Perkins 1984, 123–124. 22 Durliat 1990; Fields 2008. 284 Fig. 3. The Rhône River Delta region faces the Mediterranean with an area of intermittent brackish-water lagoons. It is therefore not a good arable area, although it served for the grazing of livestock. In the 2nd century, the export of wine from the Rhône valley was quite important for Arles.23 By the fourth and ifth centuries Arles was important both for the commercial distribution of goods and as a major centre of the annona, the state-controlled distribution system devised to supply the imperial court, army and the city of Rome.24 It is safe to assume that trade volumes continued to expand throughout the 3rd century and the irst half of the 4th century. Strategically located on the shores connecting the Iberian and Italian peninsulas right on the major body of water from inland Gaul, it became a vital military strongpoint after the fall of Trier, the invasion of Spain and Gaul by Vandals and Visigoths. Fifth-century Arles is interesting in that it has many features that initially qualiied it to be a successful imperial residence. But the three irst usurpers of the 5th century residing there must be regarded as failures. This had repercussions for the city. The town underwent no less than seven different sieges 411–508. After having taken control over much of Gaul and Spain after the fall of Trier in 406, Constantine III was trapped in Arles during a siege in 411. The city fell, and Constantine III was captured there and later executed, his and his son Julian’s decapitated heads reaching Ravenna on September 18th 411. Later the usurper Jovinus surrendered to Honorius’ forces further south at Narbonne in 413, his and his brothers Sebastianus’ and Sallustius’ decapitated heads eventually reaching Ravenna on August 30th 413. From his base in Ravenna, Honorius appears to have been able to re-establish a real military presence in Arles, minting in the city in 418. Later Avitus led to Arles after his failure to hold power in Rome during the famine of 455 caused by the Vandal sack of the city. Once secure in Arles, he raised an army to go back into Italy but was defeated on October 17 at Placentia. He was allowed to resign and was ordained a bishop, soon after which he died. His death was attributed to his successor, Majorian, who ran the West Empire from Arles 458–461 after having irst established military security in Italy before his campaigns in Spain and Gaul that were to cost him his life (Plate 1). Gold coinage was later struck in Arles by the two last western emperors Julius Nepos and Romulus Augustus, and this coinage reached Italy. After the fall of Romulus Augustus in 476, Arles became less important. Yet in 508 the Ostrogothic king 23 Constans 1921. 24 Loseby 1996, 46–47. 285 Plate 1. Reverse image of a solidus struck for Majorian in Arles, 458. Photo courtesy of the Royal Coin Cabinet, Stockholm. of Italy, Theoderic, wrote to the habitants of Arles urging a reconstruction of the ramparts.25 Whether for sentimental or strategic reasons, Theoderic was still prepared to invest in the upkeep of the fortiied city that was a key bridgehead between Spain, Gaul and Italy. Ravenna and Its Hinterland Ravenna lies at the southern end of the Po River delta along the Adriatic shores of north-east Italy. It is speciically located in a moist, brackish-water area, which is a perfect breeding ground for mosquitoes. Indeed, malaria is endemic to the Veneto region and the Po Valley.26 Malaria infection is most frequent in the late summer and early fall. Ravenna was always a relatively minor settlement on a few sandy islands along a long, coastal, sand dune facing the Adriatic to the east and inland marshes and lagoons to the south, west and north. The harbour was enlarged during Augustus, who also had grand plans for connecting Ravenna directly to the Po River by means of the Fossa Augusta, a vast canal project, but the main harbour would remain further south at Classe. Ravenna replaced Milan, as the latter city had inadequate defences to withstand a siege. Besieging Ravenna was a completely different matter. It was fortiied by its surrounding biotope, hostile to human intervention. In contrast to Milan and Rome, Ravenna was also easily accessible from the Adriatic Sea and thus close to Constantinople. Milan remained an important garrison and imperial mint in close contact with Ravenna, but the imperial court usually refrained from staying in Milan after Alaric besieged the emperor Honorius there in 401. The link to Rome was another issue. Mobile military units could cut off communications between Ravenna and Rome across the Italian peninsula without exposing themselves to more than minimum danger. A key explanation of the decline of the city of Rome, the isolation of Ravenna, and the inversed function of the network in the 6th century is that a hierarchical order was needed to sustain the old Roman network. A disciplined army under competent leadership with an ability to act for a prolonged period of time within rather than outside the network was needed in order to guarantee its security. But this could not be organized from Ravenna as time progressed. 25 Constans 1921, 224–225; Cassiodorius Var VIII. 26 Stathakopoulos 2004. 286 Table II The Late 4th- and Early 5th-Century Imperial Residences in Comparison City Rome Constantinople Trier Population (est) 500000 350000 100000 Arles 75000 Ravenna 50000 Hippodrome Theatre Arena 270000 50000 50000 30000 20000 20000 18000 10000 20000 Imperial residence 440–455, 462–465, 468–474 324– 293–311, 367–391, 407–408 409–411, 455–456, 458–461 402–440, 467–468 474–476 Mints (oficinae) 5 4–5 3 3 1 One may contrast the three imperial residences in the 5th-century West Empire by irst evaluating them according to a number of criteria in terms of infrastructure and public accommodation, and then contrasting their collective timeline against the successful eastern imperial residence of Constantinople and the lost 4th-century western capital of Trier. In Table II, the various aspects of urbanism and imperial administration are plain to see. Firstly, Rome, Trier and Arles all have large arenas, theatres, and hippodromes to accommodate the inhabitants of their respective hinterlands, who entered the cities to partake in spectacles and games on given holidays. Rome could theoretically have itted all the inhabitants of Constantinople into its three main public arenas. By contrast, the still smaller and younger city of Constantinople had no large theatres or arenas and only a relatively small hippodrome, which was a very complicated imperial propaganda piece that required substantial military security measures. Following the precedent set by the Circus Maximus and the imperial palace in Rome, the hippodrome was in direct contact with the imperial palace via the VIP section known as the kathisma (Plate 2). Indeed, this hippodrome was the scene of two serious rebellions in 505 and 531, respectively. No comparable buildings have yet been identiied in Ravenna, although at least a minor hippodrome is likely. Rome and Arles were old-fashioned cities tied to a functioning infrastructure whereas Trier was a vast outcrop at the very end of the infrastructure. By contrast, Constantinople was a successful meta-city above the network (albeit with some serious internal security problems) while Ravenna was too insigniicant to play any major role for its hinterland at all. Roman Gold Coinage as Analysed by the LEO Project To pay for all supplies and military security, the imperial state apparatus had to use real money – hard currency meant only for the state and its functionaries. 287 Plate 2. The Theodosian imperial family in the kathisma surrounded by their Germanic bodyguard. From the Theodosian obelisque celebrating the defeat of Magnus Maximus in 388 in the Constantinople hippodrome. Photo by Anneli Sundkvist 2008. Roman gold coinage, solidi, was only struck in the imperial residences and usually in the vicinity of the emperor. Roman gold coinage is thus very apt research material as its relative frequency is highly indicative of the allotted importance of a given imperial city in a speciic period. The research of the LEO project is focused around two databases, BLEO and CLEO. BLEO (Baltic/European Liber Excelsis Obryzacusque) is constructed at a micro level and currently consists of approximately 7300 individual gold coins from the period AD 249–565. CLEO (Continental Liber Excelsis Obryzacusque) is constructed on a macro level and consists of 180 gold hoards in Europe, the Middle East, and North Africa with approximately 22000 gold coins.27 The ultimate goal of the LEO project is to relate the two databases to climate data and to correlate this with historical sources.28 It will thus establish a more coherent timeline for coinage, climate data and historical sources, visualizing a multi-dimensional frame of reference where urban sustainability and resilience can be assessed from a new perspective. One way of establishing the comparative role of urban centres in the Roman Empire is to look at the aftermath of the coinage reforms of Diocletian in 294 and Constantine in 326 and Valentinian I in 367. By means of these reforms, 13 Roman cities were allotted the right to strike coinage for the empire within 10 given dioceses out of a total of 13, see Table III. 27 Fischer et al. 2011. 28 Stathakopoulos 2004; Fischer et al. 2009. 288 Table III. Cities with Imperial Gold Mints after the 326 Reform City Trier Lyon Arles Aquileia Rome Siscia Thessalonica Heraclea Constantinople Cyzicus Antioch Alexandria Diocese Galliae Galliae Viennensis Italia annonaria Italia suburbicara Pannoniae Moesiae Thracia Thracia Pontica Oriens Aegyptus In regard to Table III one may irst note that the three dioceses of Britanniae, Hispaniae, and Africa were not given mints even though urban centres such as Sevilla, Carthage and London clearly could have supported these. Secondly, when one looks at the number of imperial mints located in each city, it is abundantly clear that Rome and Constantinople retained an absolute hegemony. They struck gold, and lots of it. Arles rarely struck gold but had three mints for the substantial amounts of bronze coinage that were needed in the hinterland economy of southern Gaul. After 367, the imperial mints became more mobile and tied to the imperial persona. Following the deaths of two senior emperors, Valentinian I in 375 and Valens in 378, the imperial administration became unstable. A lengthy interim period with a number of itinerant rival emperors followed, so that state inances in the western part of the Empire became sporadic, see Table IV. Table IV. Sporadic Gold Mints of the Late 4th Century AD City London Sirmium Lyon Lyon Lyon Issuer Magnus Maximus Theodosius I Magnus Maximus Valentinian II Eugenius Chronology 383 383–388 385–388 388–392 392–394 Table IV shows how rival emperors had to move the imperial mints during their internal wars that further weakened the empire. This coinage was struck in alternate cities when the imperial courts had to inance military expeditions away from the normal residences. In 394, Theodosius I emerged as the sole victor, settling down momentarily in Milan. With the death of Theodosius the next year, these alternative mints ceased to be of any major importance save for the usurpation of Constantine III in 407–411 when the latter reactivated the dormant imperial mints of Trier and Lyon. Instead, the number of imperial mints became even smaller, yet the output was to become even more disproportionate, see Table V. 289 Table V. Mid- to Late 5th-Century Cities with Imperial Gold Mints City Arles Ravenna Milan Rome Thessalonica Constantinople Diocese Galliae Italia annonaria Italia annonaria Italia suburbicara Moesiae Thracia Table V is quite revealing, showing how Constantinople monopolized state inances in the East Empire while the West Empire was undergoing a comparative downsizing in the number of imperial mints but without the corresponding growth and afluence demonstrated by Constantinople. Gold coinage began to be issued in Ravenna in 402 with the transfer of the Milan mint to Ravenna under Honorius. Ravenna was soon given its own acronym on coinage, RV. It was to be frequently imitated. The Ravenna mint does not identify separate oficinae within its coinage although it struck gold and silver to begin with, suggesting that separate oficinae must have been active from early on. The Ravenna mint was not in constant operation, although its peak is clearly measurable against Milan in particular. Its hiatuses are quite easily measured, though, in particular in relation to the mints of Rome in 450–455 and Arles in 455–561. After 450, Ravenna only struck gold coinage, just as Milan had before. A larger overview of all the gold coins recorded in CLEO with certain mint marks identifying the city in question shows the ascendancy of Constantinople in an even more obvious fashion, see Table VI. Table VI. Gold Coinage Recorded in CLEO with Certain Mint Marks, AD 317–565. City Constantinople Milan Rome Ravenna Trier Thessalonica Antioch Lyon Arles Sirmium Total: Gold Coins 3300 804 683 425 360 94 92 89 71 70 5988 Percentage 55.4 % 13 % 11 % 7% 6% 1.6 % 1.5 % 1.4 % 1.1 % 1.1 % 100 % Furthermore, Table VI shows that the city of Rome was in a poor inancial state that progressively grew worse during the 5th century. More than half of all gold coinage with certain mint marks was struck in Constantinople and dates to the mid-5th century and the reign of Theodosius II (408–450) and Leo I (457–474). Much of this coinage must be regarded as subsidies directed to aid the inept western emperors. The entire Italian output amounts to only 31% of the total. Note also that the two alternative residences of Arles and Ravenna produced the equivalent of two thirds of the total output from Rome, but only during relatively short periods. The study of Roman gold coinage of the LEO project shows that the imperial administration did not have the inancial means to support the costs for sustaining the city of Rome, and that the attempts to run the administration 290 from the alternative cities were doomed to fail, given the constant threat of barbarian incursions and the inancial dependency on Constantinople. The 6th-Century Decline of the Roman Urban Mind After the deposition of the last West Roman emperor in 476, the supply system to Rome continued to function more or less as usual under the Ostrogothic king Theoderic. Subsistence demands had diminished as the urban population had decreased, yet a precondition for the continuity of the Italian urban system was the relatively strong and legitimate central power that Theoderic had established and which was able to maintain and control the network that sustained urbanism. But during the Byzantine conquest of Italy 534–554, this precondition was notably absent. Neither the Byzantines nor the Ostrogoths after the death of Theoderic could maintain a central authority. It was by no means obvious to Italian urban magistrates who represented legitimate power. In 536 the city of Rome accepted the troops of the Byzantine general Belisarius. Naples, on the other hand, refused to do so, causing Belisarius to lay siege, storm and sack the city. Furthermore, the small military forces available to the combatants prevented both sides from establishing any sort of long-term and viable hegemony. It was not until after 562 that the Byzantine armies inally succeeded in defeating the Ostrogoths after Narses in 553 had secured Via Flaminia by conquering the strategic town of Spoleto. Before that, the control of Italy and Rome had changed hands several times. The Roman urban mind that had survived the fall of the West Empire a hundred years earlier disappeared with the Byzantine conquest. The economy was shattered and there was a signiicant depopulation. The Lombard Invasion and the Dissolution of the Roman Urban Mind In 568, the Lombards invaded and gained control of substantial parts of Italy. They consciously avoided attacking large urban centres, successfully negotiating surrenders on a local level. Coniscated lands were divided between the leading Lombard families. This is mirrored in the Lombard cemetery of Castel Trosino where as many as 90% of all buried men in the period 570–610 have swords in their graves.29 It is possible to trace the creation of a new network of genealogically related power spheres diffused throughout Italy in the shape of Lombard landownership in the old Roman network. The Lombards, however, lacked resources to control all of Italy and soon lost hierarchical cohesion. A periodically strong kingdom was established in the north with the town of Pavia as its capital.30 Meanwhile a number of smaller principalities were established to the south.31 The result of the Lombard invasion was the creation of a number of smaller enclaves, power spheres above and outside yet within the old Roman network. These developed along road intersections which had previously been of lesser 29 Åberg 1923; Lindqvist 1926; Paroli & Ricci 2008. 30 Paulus Diaconus 27–28. 31 Christou 1991, 177. 291 importance, such as Monza and Pavia to the north of Rome, and Spoleto, Fermano and Benevento along Via Flaminia and Via Appia further south. During the Byzantine conquest and the Lombard invasion, military tactics were dictated by the old Roman network. The antagonists strove to avoid fortiied intersections in the network by staying off the roads as much as possible before amassing suficient strength to attack more vulnerable intersections. This type of asymmetrical warfare required a high level of mobility. The solution was a reliance on small cavalry units.32 One may note a failed Frankish attempt to invade northern Italy with infantry in 539–540. The Frankish force was severely weakened by starvation, disease and illness and soon disappeared from the historical sources.33 Belisarius intentionally recruited his cavalry units from distant places and thus often employed Alans and Huns34, a practice also used by the Lombards. Being short of manpower, the Lombards invited other groups of Indo-Iranian and Slavic origin to settle in Italy. Late Antique and early medieval cavalry units did not as a rule use shod horses. Since hoofs are sensitive35 and since the surface of Roman roads were of stone, cavalry units would avoid the road network as much as possible. Instead they used the surrounding hinterland where suburban green areas and abandoned cultivated areas provided cover prior to assaults on settlements and intersections. Thus, the mobile military units could cut off communications between Ravenna and Rome across the Italian peninsula without exposing themselves to more than minimum danger. An explanation for the decline of the city of Rome and the inversed function of the road network in the 6th century is that a hierarchical order was needed to sustain the old Roman network. A disciplined army under competent leadership with an ability to act for a prolonged period of time within rather than outside the network was needed in order to guarantee its security. The long wars did not lead to any deinite results except the demographic collapse of the Italian population. For want of real results all combatants allowed their small cavalry units to roam the hinterlands of enemy strongholds. After 533 there was no legitimate central power in Italy. The last emperor, Romulus Augustus, had been deposed in a coup d’état two generations earlier in 476 by the Germanic oficer Odoacer. The Ostrogoths, Byzantines, and Lombards could not claim to represent a legitimate power that could guarantee sustainability and mutual low within the old Roman network. Nor could any of these groups be considered to be native to Italy. A trait shared by all three groups is that they lacked a number of common interests or mutuality with the Italian population, especially regarding sustainability in terms of agriculture and manufacture. Conclusion The city of Rome was the ball and chain of the western imperial administration. It could not be sustained in the face of the onslaught of rapid change and had a very limited capacity for resilience as the administration was unable to reform or 32 33 34 35 292 Mack 1975; Elton 1996, 43–51, 59. Procopios 6, XXV; Stathakopolous 2004, 275–276. Procopios 5. Sundkvist 2003; Fischer 2005, 92–93. to inance an indeinite status quo. When the population of Rome had dwindled into that of a regular Roman town, there were no longer any great transports by sea. The old road network, however, could not be used to sustain Rome having become a liability in a new age of asymmetrical warfare. Arles and Ravenna, the two western alternatives to Rome, were advantageous as they could be supplied from the sea, the cheapest form of transport. Located in lagoons, they were easy to defend as any siege would require huge resources beyond the capacity of barbarian invaders. The two alternative residences had relatively small populations while still being able to provide suficient ideological legitimacy. Rome on the other hand was dificult to defend, had initially a huge population and was dificult to supply. But a puppet imperial administration, pampered with gold coinage, soldiers and food by the increasingly resilient and dominant city of Constantinople from overseas while hiding out in Ravenna, was no solution to the Roman problem. In the end, Constantinople dispatched an Ostrogothic general, King Theoderic, to take over the remains of urbanism in Italy. This he did. But he also set up his own rule where Constantinople wielded little inluence. A generation later, Constantinople dispatched its own army against Rome and laid to waste much of the Roman urban mind for good. Arles and Ravenna soon fell back into relative obscurity in the early medieval world, very much like Trier had done a century earlier. The Dark Ages had arrived in Western Europe. This spelled the end for the Roman urban mind. References Abbreviations of Classical sources Mosella Ausonius, Mozella. The works of Ausonius, R.P.H. Green (ed.), 1991. Oxford: Clarendon Press. Ordo Urbium Nobilium Ausonius, Ordo Urbium Nobilium. The works of Ausonius, R.P.H. Green (ed.), 1991. Oxford: Clarendon Press. Cassiodorius Cassiodorius Variae, S.J. Barnish (transl., ed.), 1992. Liverpool: Liverpool University Press. Paulus Diaconus Paulus Diaconus, Historia Langobardorum, Helge Weimarck, (transl.), 1971. Langobardernas Historia. Uddevalla: Bohuslänningen. Procopios Procopios, History of the Wars, H.B. Dewing (transl.), 1952. Cambridge, Mass.: Harvard University Press, Heineman. Symmachus Symmachus, Relationes. In Prefect and emperor: the relationes of Symmachus A.D. 384, R.H. Barrow (transl. ), 1973. Oxford: Clarendon Press. Modern literature Åberg, Nils 1923. Die Goten und Langobarden in Italien. Uppsala: Almqvist & Wicksell. Brown, P. 1971. The World of Late Antiquity AD 150–750. London: Thames and Hudson. Bruun, C. 1991. The Water supply of Ancient Rome. A Study of Roman Imperial Administration.. Helsinki: Societas Scientiarum Fennica. 293 Christou, K. P. 1991. Byzans und die Langobarden. Von der Ansiedelung in Pannonien bis zur endgültigen Anerkennung (500–680). Athens: Historical Publications St. D. Basilopoulos. Constans, L.A. 1921. Arles antique. Paris: E. de Boccard. Durliat, J. 1990. De la ville antique à la ville byzantine : le problème de subsistence. Coll. De l’école francaise à Rome 136. Rome: Ècole francaise de Rome. Elton, H. 1996. Warfare in Roman Europe AD 350–425. Oxford: Oxford University Press. Fields, N. 2008. The Walls of Rome. London: Osprey. Fischer, S. 2005.Roman Imperialism and Runic Literacy. The Westernization of Northern Europe 150–800 AD. AUN 33. Uppsala: Department of Archaeology and Ancient History. Fischer, S., H. Victor and N.L. Wicker (eds) 2009. Climate and Coinage. Political Conlict and Social Crisis.2nd International LEO Conference (LEO II) November 5th–6th 2009. Uppsala. Fischer, S., F. López Sánchez and H. Victor 2011. A Preliminary Result of the LEO-Project: The 5th Century Solidus Hoard in Stora Brunneby, Öland, Sweden. Fornvännen (forthcoming). Gibbon, E. 1897. The Decline and Fall of the Roman Empire, J.B. Bury (ed.). London. Goffart, W. 1980. Barbarians and Roman, AD 418–584. The Techniques of Accommodation. Princeton: Princeton University Press. Heather, P. 2005. The Fall of the Roman Empire: A New History of Rome and the Barbarians. Oxford: Oxford University Press. Lançon, B. 2000. Rome in Late Antiquity. Everyday Life and Urban Change, AD 312–609. Edinburgh: Edinburgh University Press. Laurence, R. 1999. The roads of Roman Italy: mobility and cultural change. London: Routledge. Lejdegård, H. 2002. Honorius and the City of Rome. Unpublished dissertation, Uppsala University. Lindqvist, S. 1926. Vendelkulturens ålder och ursprung. KVHAA Handlingar 36:1. Stockholm: Kungl. Vitterhets-, Historie- och antikvitetsakademien. Loseby, S.T. 1996. “Arles in Late Antiquity: Gallula Roma Arles and Urbs Genesii“, in Towns in transition. Urban Evolution in Late Antiquity and the early Middle Ages, N Christie & S.T. Loseby (eds.), 45–70. Aldershot: Ashgate. Mack, A. J.R. 1975. „Why Big Nations Lose Small Wars: The Politics of Asymmetric Conlict”, World Politics, Vol. 27, No. 2 (January 1975), 175–200. Moorhead, J. 1992. Theoderic in Italy. Oxford: Clarendon Press. Paroli, L. and M. Ricci, 2008. “The necropolises in Castel Trosino and Nocera Umbra in Italy”. In Rome and the Barbarians. The Birth of a New World, Jean-Jacques Aillagon (ed.), 476–485. Milan: Skira. Revell, L. 2009. Roman imperialism and local identities, Cambridge: Cambridge University Press. Sirks, Boudewijn. 1991. Food for Rome. The Legal Structure of the Transportation and Processing of Supplies for the Imperial Distribution in Rome and Constantinople. Studia Amstelodamensia ad epigraphicam, ius antiquum et papyrologicam pertinentia 31. Amsterdam: Gieben. Stathakopolous, D. Ch. 2004. Famine and Pestilence in the Late Roman and Early Byzantine Empire. A systematic Survey of subsistence, crises and epidemics. (Birmingham Byzantine and Ottoman Monographies vol 9), Aldershot: Ashgate. Sundkvist, A. 2003. Hästarnas land: aristokratisk hästhållning och ridkonst i Svealands yngre järnålder. OPIA 28. Uppsala: Department of Archaeology and Ancient History. Ward-Perkins, B. 1984. From Classical Antiquity to the Middle Ages. Urban Public Building in northern and central Italy AD 300–850. Oxford: Oxford University Press. Ward-Perkins, B. 2005. The Fall of Rome and the End of Civilization. Oxford: Oxford University Press. 294