TRAC 1991:
THE PRE-INDUSTRIAL CITY IN ROMAN BRITAIN
By Simon Clarke

By detailed examination of the location of different types of building within a range of urban sites, the functions of Roman towns are considered. This paper rejects the suggestion that the form of the Roman town can be explained solely as the imposition of an alien culture. Indigenous social forces must be considered primarily responsible for the maintenance and adaptation of an institution which suroived for over three hundred years within the province. Various theoretical models of the pre-industrial city are considered. With these in mind possible explanations are sought for the variations in the form of RomanrrBritish urban topography. The differences in size and amenities offered between early civitas capitals and small towns are considered the product of radically different social conditions in each. This is quite contrary to the still frequently expressed opinion that small towns simply represent a later version of urbanism which had discarded as unnecessary the trappings of classical civilisation. The possibility that general principles exist which may extend throughout all Periods is suggested by parallels in the medieval period.

INTRODUCTION
Rarely have there been systematic attempts to consider in detail the functional implications of the urban topography of Roman sites. There is a belief that we already know all about the major Roman towns. Though rarely explicit there is an assumption that towns are a facade, an alien imposition, part of an ideology accepted by the local population during a vague process known as Romanisation (Reece 1985, 37). Such faith in diffusionism has largely disappeared from other periods in archaeology. Though there is no doubt that a significant flow of people from the continent did occur, the possibility that early towns served functional purposes for the indigenous population must be considered.

THE ORIGIN OF ROMANO-BRITISH URBANISM
In discussing the function of early public towns it is worth quickly reviewing the evidence for their origin. Two factors have traditionally been considered highly significant in the development of the early urban centres in Roman Britain. Both are essentially diffusionist explanations for the spread of towns: official encouragement, and the economic stimulation provided by the presence of the army.

The reference by Tacitus to ‘private encouragement and official assistance'(Agricola 21) regarding the foundation of public buildings has been seen as tantamount to a deliberate official policy of urbanisation (Frere 1987, 98-9; Wacher 1974). However no other references to an official urban policy are known to exist (Mann and Penman 1978, 61). Others have seen this passage as indicating the informal use of tax remissions and the use of the local nobility’s client system to raise funds and gather necessary resources (Blagg 1980; Millett 1990, 72). Millett has suggested that ‘private encouragement and official assistance’ are much more likely to have originated at the civitas rather than provincial level of government, with Agricola receiving undue credit for the actions of his political subordinates (ibid., 74).

Similarly the theory that skilled military personnel were used to carry out official policy has lost favour in recent years. Blagg notes that the military have a quite distinct style of stone masonry (Blagg 1980; 1984). Furthermore masons were far from always present in army units stationed in southern Britain. Cirencester’s fort apparently had none, its timber buildings (probably of two phases) being of post-in-trench and sill beam construction (Wacher and McWhirr 1982).

The theory that major early towns were stimulated by economic activity originally focused on a fort (Webster 1966) has also been questioned. On a general level for southern Britain it is now considered highly unlikely that camp followers were able to determine the location of political centres in a society in which local government continued to be in the hands of the traditional land owning elite (Jones 1987, 48-9). Certainly the traders and craftsmen known to have existed at fort vici are very unlikely to have possessed the resources to carry out a major public building programme. Although Corinium directly overlies at least one fort, it perhaps provides one of the best examples of the weakness of the military origin theory. On the creation of the public town neither buildings nor existing road system appear to have been incorporated into the town’s fabric. In fact the levelling of the fort and surrounding civilian settlement seems to have been so complete that the construction of the basilica, one of the first tasks undertaken, failed to make provision for possible subsidence over the fort’s defensive ditches (Wacher 1974, 298).

FIGURE 1. Models of pre-modern Urbanism

A more plausible suggestion is that a fort was located at Cirencester largely for political reasons, being just a short distance from the existing elite centre at Bagendon. It was this settlement which provided the origin of the Roman town. The creation of the public town at Cirencester signifies nothing more than a minor shift in the location of the region’s political centre, in order to take advantage of the new communications network focused on the fort (Jones 1987,48-9; Millett 1990, 74).

In the light of this the development of major towns should be studied with a view to understanding the needs of the indigenous society they served. This paper will consider the distribution of buildings and the accessibility of different parts of the town, using models of urban topography, some of which were developed originally to examine medieval towns.

PROBLEMS WITH EXISTING MODELS OF URBANISM
There are such great variations between the models concerned with urbanism that it must be wondered whether any useful generalisations can be transferred to the Roman period. A great many classifications of ancient and medieval urban settlement have been made. However a basic division can be made between those originating from economic stimuli and those which start as primarily political and administrative centres (Fig. 1).

Commercial Urbanism
Vance (1971) pictured the medieval city as primarily a commercial settlement. Authority was in the hands of the leading members of the various guilds, and as a result towns had not one but multiple foci, with practitioners of different crafts tending to live and work close to their own guildhall. The aristocracy, while it may have possessed town houses, was peripheral to urban life. Nevertheless society continued to be essentially feudal, guilds providing an alternative overiordship (ibid., 106), with strong ties being maintained across class divisions, hence the absence of zoning according to class. Because the society lacked a strong central authority no sense of planning is visible. The result is a mass of densely packed buildings served by a network of narrow irregular streets which were frequently encroached upon and built out over (ibid., 102).

A slightly different picture is proposed by Langton (1975), who described the pre-capitalist city based on observations of seventeenth century Newcastle. He recognised that a single mercantile quarter provided the focus to a town zoned by both class and occupation. The aristocracy was also present in the city, but as with Vance’s model was not of central importance occupying its own high class residential district away from the city’s centre. These observations are less acceptable as a model for urbanism in Roman Britain, however, for Newcastle was on the eve of the Industrial Revolution and is perhaps more aptly thought of as a hybrid premodern/ modern city.

Public Urbanism
Sjoberg’s (1960) pre-industrial city, like Finley’s (1975) ancient city, was dominated physically and symbolically by the edifices of administrative and religious institutions. Important commercial buildings were to be found in the central area, but were dwarfed by administrative and religious structures. These formed a single focus around which the residences of the aristocratic elite clustered. Sjoberg noted that most cities were multi-functional, usually fulfilling administrative, religious, cultural and economic roles. However the city was essentially geared to serving the needs of the landed elite, economic activity being of secondary importance. Finley stated that in defining the city, ‘the economy did not enter into consideration at all, apart from the requirement that the material goods indispensable for civilised amenities had to be available somehow’ (ibid., 124). Unlike many other writers Sjoberg saw the aristocracy as an urban class, in spite of their ownership of rural estates. To them the city was an essential tool of social control. It allowed the maintenance of communication within the elite class, provided physical safety, perpetuated their value system, confirming their ‘divine rights’ and provided consumer items which were the symbols of elite status. Outside the core area, isolated from it and each other by poor communications, and sometimes internal city walls, were further distinct districts. Each had a fairly homogeneous character, zoning being according to ethnic origin, family ties (including patronage) and occupation. A significant low status population (slaves and outcasts) was to be expected throughout the city and just outside the city walls, where rural poor were also to be found.

Gradual Development
Each of these theories has some basis in fact, but they are clearly incompatible. Collis (1984, 123) simply saw them as flawed by their ‘lack of historical depth’, leading him to construct his own ‘model of gradual development'(Fig. 2) to describe the evolution of a defended oppidum (ibid., fig. 8.15). This model has much to commend it, the settlement created closely resembling Silchester, which Collis takes as the Roman ideal (ibid., 124). However I do not believe that failure to consider the temporal dimension can completely account for the differences between the models of Vance and Sjoberg. A more reasonable suggestion is that they represent two distinct brands of urbanism.

Figure 2. Model of gradual urban deuelopment (after Collis).
Figure 3. Roman Cirencester: functions of buildings.

CORINIUM: A PUBLIC TOWN HISTORIANS OF TECHNOLOGY
Cirencester (Fig. 3) like most major Roman towns had a relatively low density of buildings set within a highly regular grid of streets. These were generally between six and nine metres in width but could be as much as twelve metres across in central areas of the town (McWhirr 1981, 30). Far from lacking a single strong focus, the town was dominated by a large basilica and forum, with a large temple precinct probably occupying the insula immediately to the south-east and public baths possibly to the west (Wacher 1974, 300). The theatre next to the city’s north gate, the amphitheatre just outside the west gate and a number of possible temple sites, at locations scattered throughout the city, detract somewhat from the impression of centralisation of public buildings. Only very limited commercial and industrial activity is suggested from existing excavations. These are confined to a limited area adjacent to the central public buildings with a single light industrial building located just outside the west gate. The rest of the city displays no sense of zoning by building type, houses of the highest quality being scattered fairly evenly amongst domestic premises of more modest character.

In further contrast to Vance’s model of medieval urbanism the aristocracy seems to have been central to the life of Roman cities. Corinium, as has been argued above, served as a replacement for the elite native settlement at Bagendon. It may perhaps even have been named after Corio, the last pre-Roman ruler of the Dobunni (Wacher 1974, 293). Cirencester’s service community was probably consistent with serving the needs of the resident population without providing a central place facility to the surrounding countryside.

COMMERCIAL URBANISM IN THE ROMAN PERIOD
The nature of the urban development at Cirencester outlined above fits quite well with Sjoberg’s view of the pre-industrial city, but Vance’s concept of commercial urbanism, in which the landed elite are peripheral, is not completely alien to Roman Britain. Roman Britain appears to have had no equivalent to the medieval guilds and very little evidence exists as to the Roman attitude to land holding in urban centres. However small towns appear to have a great deal in common with Vance’s medieval model. Some at least lacked much cohesion, being strung out along a great length of highway. The Roman settlement at Bourton-on-the-Water seems to have consisted of a number of occupation areas both on and some distance away from the Fosse Way. Small towns nearly always lacked either the will or the means to establish a regular street network. Where the building land provided by the frontage of a highway was not adequate small tracks often seem to have been laid to serve individual buildings as the need arose. A reasonably full street pattern of this type is known at Wycomb. Only two branch roads are known from the patchy investigation of Bourton Bridge settlement (Fig. 4). Even from the very short lengths of these side roads so far recognised it is clear that encroachment by buildings was a regular occurrence.

Figure 4. Bourton Bridge settlement.
Figure 5. Thiessen polygons applied to larger Romano-British walled towns (after Hodder 1972).

So far relatively minor commercial settlements have been discussed. A site like Kenchester however, with the resources and the will to build a defensive circuit, is more interesting. This suggests that the wealth generated from commercial activity was enough to support a large population, with fairly high material standards, but did no~ lead to the urban planning and strong central architectural focus which characterised the public towns. Even in the ancient period it is clear that such sites were confused with public towns. However Strabo (3.4.13) emphasises that they represent something quite different when he protested against the term city being applied to the large villages (komai) of the Spanish peninsula (Finley 1975, 124).

The Temporal Development of Classical Urbanism?
It might be argued that the differences between the public towns, which were mostly early foundations, and the small towns, which often did not reach their maximum extent until the later Roman period, were due to changes in the nature of urbanism. It has been said (e.g. Reece 1980) that classical urbanism was in decline by the late period, the population of civitas capitals having failed to grow, their public buildings falling into decay or at least not being added to by new foundations. The larger walled small towns which grew spontaneously on the fringes of the early civitas territories (Hodder 1972) may have performed essentially the same role as an early civitas capital (Fig. 5). Their ‘natural’ growth, without the public trimmings, was sustained by the rising power of the money economy. The term ‘false start’ has been applied to the public towns Gones 1987), their genus being seen as an artificial creation of a foreign ideology which was ultimately unsustainable by the native population.

However the image of public towns as failed social institutions by the fourth century is denied by the distribution of villas. These late Roman expressions of landed wealth generally occur in distinct clusters centred on the public towns, but are noticeably absent around most larger walled small towns. Some of the so-called small towns enclose areas not much less than that of the smaller public towns, and with their extensive extramural settlements may have had populations somewhat larger. But similarities in the size of the settlements were not matched by similarities in form. The larger small towns were very often heavily involved in industrial activity. The largest, Water Newton, was a major centre of the pottery industry. Worcester, another large walled site, was heavily involved with iron production. In contrast, even in the late Empire public towns were not important production or redistribution centres. Even their own needs had to be met by importation from areas some distance away; north Wiltshire and Oxfordshire supply Cirencester’s pottery, for example. The absence of a regular street grid already commented upon should also not simply be dismissed as a degeneration of urban style in the later Empire. Drinkwater has noted that the insula system was most inflexible. Once laid down the network was not easily added to or altered, even to accommodate the largest and most important subsequent building programmes (Drinkwater 1985, 53). This inflexibility was apparently acceptable in public towns throughout their history in spite of the inefficient use it made of space, but was not considered appropriate in small towns.

Figure 6. Late Saxon Winchester.

These differences argue very strongly for public towns continuing to have very different functions, in relation to the surrounding countryside, to even the very largest small towns.

MEDIEVAL PUBLIC TOWNS
In spite of frequent references to the dissimilarity between classical and medieval urbanism (e.g. Reece 1985), commercial urbanism was not absent from Roman Britain. Nor does public urbanism appear to have been confined to the Roman period. Saxon Winchester (Fig. 6) would appear to fit the public town model much more closely than Vance’s model of a medieval city supported by commercial activity. Between the seventh and ninth century the city’s population seems to have comprised four main elements. These were a royal residence, a cathedral church and its attendant community, an unknown but probably small number private estates belonging to nobles and a small service population (Biddle 1976, 119). The latter was numerically small and dependent on the other three elements in the town and certainly did not perform a central place function as a market for the surrounding countryside (Hodges 1982, 42). What is particularly startling however was the towns provision around AD 900 with a highly regular street grid planned and executed as a single feature (Biddle 1976, 119). While the town had probably started to take on certain trade and production roles from the late ninth century, the impetus for this administrative and cultural capital in Wessex clearly came from society’s traditional elite: king, nobles and church. Hodges sees this as the result of the kings of Wessex looking to southern Europe where the classical street grids still persisted as a fossil from the Roman Empire (Hodges 1988, 5), but this is surely a weak argument for so strong a similarity. For such a massive expenditure of labour as the creation of a planned town, there must have been sound practical rather than purely sentimental reasons.

THE SEPARATION OF PUBLIC AND COMMERCIAL URBANISM
Sjoberg has characterised public towns as populated by a significant merchant class, as well as an aristocratic elite, and lists a number of reasons why the two elements of urban society should be located together (Sjoberg 1960). Prime amongst these was the elite’s need to maintain a firm grip over all potential sources of social power. Elite control of the means of production is a central theme in Marxist social theory and is seen by Hodges as the driving force behind early medieval urbanism (Hodges 1982; 1988). However there are sound reasons why these two important elements of urban life should repel as well as attract each other. Sjoberg himself noted that the religious-philosophical value systems promoted by most landed elites in pre-industrial societies depreciates non-intellectual activity as degrading, while glorifying leisure, literary pursuits and conspicuous consumption. These values had the effect of maintaining the status of the elite over the common classes. A secondary consequence of this mentality however is that initiative and enterprise is stifled, as there is little incentive to increase productivity (Sjoberg 1960, 186). This may help to explain why so little industrial activity is situated in the public towns. The presence of the elite while providing a natural market, also provided a powerful disincentive to the initiation of commercial activity. The lack of evidence for industrial activity at Cirencester is almost complete. This is largely explained by the site’s complete unsuitability as a redistribution centre. Although the city is at the hub of a road network, bulky items like pottery become prohibitively expensive unless water transport is possible. Cirencester’s lack of a navigable river therefore effectively prevented its development as a commercial centre. However, that such a site should be chosen as a capital is telling in itself. There may be a similar division between planned towns and organic development at small towns in Gaul. Drinkwater noted that sites like Arion St Ambroix, while not possessing a classical planned layout, displayed a higher level of commercial and industrial activity than many civitas capitals (Drinkwater 1985, 54).

Over Britain as a whole most Roman period industrial activity can be seen to have been situated either in the countryside or in small towns, the largest of which seem to have developed at the political periphery. The development of these small towns away from the political centre might represent displacement of merchant and artisan classes to locations with more favourable social conditions.

GLEVUM: COMMERCIAL URBANISM WITH PUBUC TOWN TRAPPINGS? –
Of all the major towns which developed in the early Roman period, the colonia at Gloucester provides evidence for the best objections to the theory for public towns as set out above. Figure 7 is based on the extensive excavations published by Hurst (1972; 1974; 1975; 1988). Glevum developed on the site of a legionary fortress and its attendant civilian settlement. Both would appear to have made significant contributions to the material development of the colonia. The fortress street grid was inherited almost intact, as in all probability were at least some of the barrack blocks (ibid., 56). This is in stark contrast to Cirencester where, as already noted, levelling seems to have been total. The tilery and much of the civilian settlement outside Glevum’s north gate also experienced a high level of continuity. Therefore the suggestion that the city developed from the commercial activity stimulated by the fortress seems not unreasonable in this case. Furthermore the city was established in an area marginal to the pre-existing Dobunni tribal grouping and for this reason it must to some extent have been a conscious decision to create a new centre of urbanism. By this I mean not simply the material trappings of a monumental town but also its social fabric, as no naturally urban, landed elite previously existed in the area of the territorium.

Certainly there is little evidence for a landed elite in the area immediately surrounding the town, villas being relatively sparse and small in scale by comparison with the nearby Cotswolds. However, that some sort of elite was extracting surplus in the form of rent or some other unearned income seems almost certain to be due to the colonia’s lack of a visible means of support. In spite of Glevum’s promising location as a commercial centre, at the lowest bridging point of a major river system, manufacturing on the site actually declined after the military’s withdrawal.

In many ways the commercial elements of the city remained separate from the monumental town in the same way that the military and civilian settlements had been. Although a few manufacturing premises were present within the walled area the main concentration of shops seems to have been outside the north gate, where a building interpreted as a market hall has also been noted. Wealthy houses are very definitely concentrated within the walled town along with a cluster of monumental buildings. In many ways Glevum was two communities, a walled settlement of the landed elite dominated by public buildings, and a commercial settlement, without a street grid and consisting mainly of strip buildings fronting the main road. This situation has some similarities with Langton’s model in which elite residential and commercial districts were quite separate.

CONCLUSION
In conclusion, public cities as exemplified by Corinium cannot be explained purely as a facade of classical culture which was dropped in favour of ‘natural’ town development. Rather the form of the classical city should be considered highly functional. An emphasis on public buildings and a strong sense of planning would appear to be diagnostic features for towns of all periods which had their origins as political, administrative, cultural, educational and religious central places. In their earliest stages it is doubtful that such centres can truthfully be described as urban. Winchester was not urban until the late ninth century at least (Biddle 1976, 120), while few would consider the oppida of southern England to have been. Nevertheless they perfonned the role of urbs as understood by Roman writers and should be thought of as one of two quite distinct origins of urbanism. The other was commercial urbanism in which the town, lacking the central authority of the aristocracy, developed organically without any sense of planning and often without a single strong central focus to the settlement. While several writers, Sjoberg in particular, have noted that there are many reasons why these two elements of urbanism should come together to form a single settlement, forces also seem to exist which repel merchant and aristocratic classes from one another. In the Roman period the two types of urbanism seem to have remained particularly rigidly separate. This has given rise both to towns with segregation of the commercial from administrative elements, such as Glevum, and to towns in which the landed elite and their administrative infrastucture are largely or wholly absent, such as Kenchester.

Figure 7. Roman Gloucester.

Bibliography

Biddle, M. 1976 ‘Towns’ in D. Wilson (ed.), 99-150.

Blagg, T. F. C. 1980 ‘Civilian and military architecture in the province of Britain’ World Archaeology 12 (1), 27-42.

Blagg, T. F. C. 1984 ‘An examination of the connections between military and civilian architecture in Roman Britain’ in T. F. C. Blagg and A. C. King (eds), 249-64.

Blagg, T. F. C. and A. C. King (eds) 1984 Military and Civilian in Roman Britain. Oxford; British Archaeological Reports (British Series 136).

Clarke, D. L. 1972 (ed.) Models in Archaeology. London.

Collis, J. 1984 Oppida. Sheffield; Sheffield University, Department of Prehistory.

Drinkwater, J. R. 1985 ‘Urbanisation in the three Gauls: some observations’in F. Grew and B. Hobley (eds), 49-55.

Finley, M. 1975 The Ancient Economy. London.

Frere, S. S. 1987 Britannia 3rd edition. London.

Grew, F. and B. Hobley (eds) 1985 Roman Urban Topography in Britain and Western Empire. Council for British Arachaeology Research Report 59. London.

Hodder, I. 1972 ‘Locational models and the study of Romano-British settlement’ in D. L. Clarke (ed.), 887-909.

Hodges, R. 1982 Dark Age Economics. London; Duckworth.

Hodges, R. 1988 ‘The rebirth of towns in the Middle Ages’ in R. Hodges and B. Hobley (eds).

Hodges, R. and B. Hobley (eds) 1988 The Rebirth of Towns in the West AD 700-1050. Council for British Archaeology Research Report 68. London.

Hurst, H. 1972 ‘Excavations at Gloucester, 1968-1971, first interim report’Antiquities Journal 52, 24-69.

Hurst, H .1974 ‘Excavations at Gloucester, 1971-1973, second interim report’ Antiquities Journal 54, 8-52.

Hurst, H. 1975 ‘Excavations at Gloucester, third interim report: Kingsholm 1966-1975’ Antiquities Journal 55, 267-294.

Hurst, H. 1988 ‘Gloucester’ in G. Webster (ed.), 48-73.

Jones, R. F. J. 1987 ‘A false start? The Roman urbanism of western Europe’ World Archaeology 19 (I), 47-57.

Langton, J. 1975 ‘Residential patterns in pre-industrial cities. Some case studies in seventeenth century Britain’ Transactions of the Institute of British Geographers 65, 1-27.

Mann, J. C. and R. C. Penman 1978 (eds) Literary Sources for Roman Britain. London.

McWhirr, A. 1981 Roman Gloucestershire. Gloucester.

McWhirr, A. 1988 ‘Cirencester’ in G. Webster (ed.), 74-90.

Millett, M. 1990 The Romanisation of Roman Britain. An essay in archaeological interpretation. London.

Reece, R. 1980 ‘Town and country: The end of Roman Britain’ World Archaeology 12 (I). 77-92.

Reece, R. 1985 ‘Roman towns and their plans’ in F. Grew and B. Hobley (eds),37-40.

Sjoberg, G. 1960 The Pre-Industrial City. New York.

Vance, J. E. 1971 ‘Land assignment in the pre-capitalist, capitalist and postcapitalist city’ Economic Geography 47, 101-120.

Wacher,J. 1966 (ed.) The Civitas of Roman Britain. Leicester.

Wacher,J. 1974 The Towns of Roman Britain. London.

Wacher, J. and A. McWhirr 1982 Early Roman Occupation at Cirencester. Cirencester.

Webster, G. 1966 ‘Fort and town in early Roman Britain’ in Wacher (ed.) .

Webster, G. (ed.) 1988 Fortress into City. The Consolidation of Roman Britain, First Century AD. Batsford; London.

Wilson, D. 1976 (ed.) The Archaeology of Anglo-Saxon England. London.