Wessex hillfort communities: consensus and controversy

The interpretation of hillforts has been fundamental in shaping the debate about how Iron Age communities were organized. Hill (Reference Hill, Moore and Armada2012) has recently caricatured the current state of this debate as between two polarized positions, the hierarchists and the levellers. The hierarchists regard hillforts as the elite residences of kings or chiefs and as the military strongholds and central places of social, economic, political and religious systems (Cunliffe Reference Cunliffe1984). On the other side are the levellers, who see hillfort societies as very communal in emphasis and ideology, lacking marked social distinctions (Collis Reference Collis and Guilbert1981; Sharples Reference Sharples1991; 2010; Hill Reference Hill, Champion and Collis1996). Although polarized, both of these positions are perfectly legitimate frameworks for understanding later prehistoric hillfort-using societies in Wessex and are dependent, in part, on whether one agrees that relatively complex prehistoric societies could only have operated as hierarchies or whether other social forms were possible. Both the hierarchist and leveller positions are allowed for because the nature of the archaeological evidence recovered from the excavation and survey of hillforts across Wessex is so highly variable and can be variously used to support or attack one side or the other. Despite these seemingly irreconcilable positions there are, however, a number of broad observations about hillforts which have received general consensus:

• • Many hillforts were constructed in the Early Iron Age, but only a few remained in use during the Middle Iron Age.

• • There is considerable variation in the occupation evidence at Early Iron Age hillforts and it is likely that they were created for a number of different purposes (e.g. enclosure for pastoral activities, defined space for ritual, storage for agricultural surplus, defence, territorial marker), many of which were not mutually exclusive.

• • Middle Iron Age hillforts exhibit evidence of intensive internal activity and were strongly defended, with multiple lines of ramparts and complex entranceways. This suggests a major social change, perhaps even the result of long, drawn-out conflict between different communities.

• • The evidence for contemporary settlement around the large, developed, Middle Iron Age hillforts such as Maiden Castle is minimal.

• • Most developed hillforts had been abandoned or gone out of use by the 1st century B.C.

However, these observations do not provide any explanation for why these conditions might occur. One explanation for the variability in settlement evidence and development is that some hillforts began at an early stage to perform certain functions that others did not. Collis (Reference Collis and Guilbert1981) has proposed the ‘crisis model’ to explain the mixed evidence of hillfort development. He argues that there are three phases of hillfort development, which he terms the pre-crisis, crisis and post-crisis situations. The size, siting, inhabitants and layout of a hillfort will therefore be dictated by the social structure and settlement pattern in the period immediately before it is established, and by the nature of the crisis. A hillfort's subsequent development will depend on the alternative ways a given community may react to the crisis, and what might happen after the crisis has passed. The variations are numerous and allow for multiple trajectories, which may help to account for variable distributions of wealth and settlement activity in Iron Age Wessex (ibid., figure 8).

Yet there are several problems with this model. Perhaps most notably, Collis assumes that hillfort ramparts were built primarily for defence and protection during a crisis. It has been argued convincingly elsewhere (Hingley Reference Hingley1984; Bowden and McOmish Reference Bowden and McOmish1987; Sharples Reference Sharples, Haselgrove and Pope2007) that the construction and maintenance of hillfort boundaries, rather than simply being for defence, may in fact have been an arena in which relations of dominance and subservience could be negotiated and reinforced. In these terms, the motivation for hillfort construction was more likely to have been a result of community prestige rather than a reaction to a crisis. The patterns of residence are also likely to have been much more complex than Collis accounts for, since there is the potential for variable groups of residents to live within the hillfort simultaneously: for instance, people from nearby settlements may have supplemented a permanent population at certain times of the year.

Though these models may give us an idea of how Iron Age hillfort communities might have been constituted, they help little in understanding the relationships and interactions between the different social institutions (e.g. household, neighbourhood, kinship, locality) that structure both the character of individuals and the communities they are part of (cf. Tullett Reference Tullett, Sterry, Tullett and Ray2010). This final point is crucial because it highlights that it is only through understanding the taxonomy of a hillfort community that it is possible to explain the nature and variability of hillfort settlement evidence and use.

Creating a taxonomy of communities using grid–group

A great deal of research has been undertaken over the last 50 years by social anthropologists wanting to better understand the relationship between social organization and culture. In an influential paper published in the early 1970s, anthropologist Mary Douglas devised grid–group analysis to classify and explain different ‘cultures’ within any community (Douglas Reference Douglas1970). The analysis is represented pictorially on a matrix with individuals or communities plotted along horizontal and vertical axes into four opposed quadrants. The horizontal axis measures what Douglas (Reference Douglas2006) calls ‘group’, by which she means social norms, while the vertical axis measures what she calls ‘grid’, or social regulation or control. For Douglas (ibid.), every community is composed of individuals or groups who can be placed in each quadrant, although crucially their position is not fixed and they can move across the matrix according to circumstance. People in these groups define themselves by their relationships with each other and these relationships are critical and competitive (Douglas Reference Douglas2005). In other words, grid–group defines every community as consisting of four different types of social organization that are in constant fluctuation.

Mary Douglas's idea of communities divided by ‘grid and group’ has been largely ignored by archaeologists. However, they have recently been pioneered in an Iron Age context by Niall Sharples (Reference Sharples2010), who has used grid–group as a means of classifying very different Iron Age communities across Britain and as a method of explaining relationships within a particular society, that of the Iron Age academic community (Sharples Reference Sharples, Moore and Armada2012). The great explanatory power of grid–group to archaeologists is that since the four types of social organization within a community are not static, but in fact compete with each other (Douglas Reference Douglas2006), the model allows us to both classify communities and also trace changes over time. In order to create a taxonomy of both hillfort and football spectator communities, therefore, grid–group will be employed here.

A taxonomy of spectator identities of modern football clubs

To help shape our understanding of the constitution of a hillfort community we will start by analysing the contemporary spectator identities of modern professional football clubs. Clearly there are many dangers in using such an anthropological parallel of communities, particularly given the differences in both physical environment and chronology (Wylie Reference Wylie and Schiffer1985; 1988). However, used carefully, an ethnography of football spectator communities can, at least, provide some ideas to help unlock the possible organization of a hillfort community. After all, communities are not just collective identities, but the convergence of shared experiences, activities, histories and places (Cohen Reference Cohen and Cohen1982; Amit Reference Amit and Amit2002). The perception of a shared hillfort or football spectator community, for instance, is not just imagined but experienced through embodied and intimate relations with other people and places. In this sense all communities are constructed from embodied, sensual and emotionally charged affiliations, which is a means for people to apprehend and locate themselves in the world.

In the early part of the 20th century, before British football clubs became global brands, the cultural relationship of people with football was generally focused on the local team. Critcher (Reference Critcher, Clarke, Critcher and Johnson1979) has stated that these traditional fans viewed themselves as club ‘members’, even considering themselves to hold some representative status for the club. Since the 1960s, the old working-class supporters have increasingly been squeezed out as a result of the game's controlling forces’ pursuit of wealthier, middle-class audiences (Taylor Reference Taylor and Cohen1971). The culmination of this process has been a dissolving of local team identification, replaced by the mass consumption of televised, market-driven sport (Giulianotti Reference Giulianotti2002, 28).

Using grid–group, Giulianotti (ibid.) has analysed the identities of modern spectators and their relation to football clubs. He has argued that there are four ideal-type categories which can be used to classify spectators: supporter, follower, fan and flâneur (an ideological ‘stroller’). These he places into four quadrants, which map the relationships of spectator communities with specific clubs (figure 2). The traditional/consumer horizontal axis evaluates the extent of an individual's cultural relationship with the club, whereas the hot/cool vertical axis measures the level to which an individual's personhood is based upon their association with the club (hot emphasizing intense identification and cool the reverse) (ibid., 31).

Figure 2 The four spectator categories. Each category constitutes a stable social configuration that is associated with distinctive values (based on Giulianotti Reference Giulianotti2002, figure 1).

Giulianotti (ibid., 33) defines the traditional/hot spectator as a supporter of a football club who is deeply and emotionally invested in the club. He argues that ‘for this community showing support for the club in its multifarious forms is obligatory, since the individual has a relationship with the club that resembles those with close friends and family’ (ibid.). As a result he claims that to switch allegiance to a rival is inconceivable because ‘traditional supporters are culturally contracted to their clubs’.

In this sense, Giulianotti argues that for traditional supporters the club is representative of the surrounding community. This link between club and community is emphasized and communicated through embodied processes – club crests and names are tattooed on bodies, club colours are worn during matches, and the whole supporter community resonates and moves in union through the chanting of various football songs (ibid., 33–34). Within the supporter community there is variation in the status of members. For instance, Giulianotti observes that distinction is acquired by supporters who continue to attend matches when their clubs are unsuccessful, while the ‘embodiment of key values also accords status, such as perceived “dedication” to the club or vocal appreciation of the team's playing style’ (ibid., 34).

The traditional/cool spectators are described by Giulianotti as ‘followers of clubs, but also of players and managers’ (ibid.). In this sense, the follower understands the specific sense of community identity associated with particular clubs because they closely follow the developments and changes amongst football clubs, managers and players in which they have an interest, both favourable and unfavourable. However, as Giulianotti (ibid., 35) highlights, the follower does not possess the close spatial association of the supporter with a club and its surrounding communities. Affiliations to other clubs are complex, often reflecting ideological and historical attractions or links to the favoured club. For instance, Giulianotti observes that groups of followers may possess intricate webs of friendship with the traditional supporters at other clubs, such as in Italy, where networks of friendship across supporter groups are frequently manifested: for instance many Sampdoria supporters are also Parma followers.

The hot/consumer spectator is defined as ‘a modern fan of a football club or its specific players, particularly its celebrities’ (ibid., 36). Thus the fan possesses or develops a form of intimate association with the club, or more frequently its particular celebrity players, but this relationship is more detached and less active than that evidenced by supporters. Giulianotti (ibid., 36–37) proposes that the scale of the fan's identification with the club is defined through ‘the consumption of club-related products’ such as merchandise or pay-per-view subscriptions to club matches. Therefore he identifies fans as typically strong in affection for the club and its players, but likely to be physically distant from the club's stadium.

The cool/consumer spectator, on the other hand, is defined by Giulianotti (ibid., 38–39) as a football flâneur, or ideological traveller, who develops a ‘detached relationship to football clubs, even favoured ones’. He argues that flâneurs desire senses of pleasure and excitement, which means they are likely to shift their allegiances and associations with teams or players until they find such sensation. In this sense, the clubs are chosen to reflect the identity and sense of personhood of the flâneur. Supporters and flâneurs are entirely opposed to one another, but they are also dependent on each other: local supporters, for instance, may grudgingly come to realize that the club must attract the custom of flâneurs in order to both maintain its prestige amongst other clubs and strive for more successes with which to satisfy both the supporters and the flâneurs (ibid., 42).

Toward a taxonomy of Iron Age hillfort communities

All of this is a long way from Iron Age Wessex, but we can use a similar model to map changes experienced by Iron Age people in their relationships with particular communities, specifically those brought together at hillforts. Rather than individual football spectators, the appropriate scale used here to study Iron Age hillfort communities will be the household. While it is now generally accepted that the later prehistoric household was unlikely to have been an independent productive unit (Moore Reference Moore2007; Sharples Reference Sharples, Haselgrove and Pope2007; Davis Reference Davis, Moore and Armada2012; Hill Reference Hill, Moore and Armada2012) it was still the key focus of social relationships. As Hill (Reference Hill, Moore and Armada2012, 250) has argued, these relationships could often be messy and entangled, simultaneously operating in local or wider social networks. Therefore it was the ways that these webs of social relationships were recognized, contradicted or challenged that textured the taxonomies of larger communal entities.

In an Iron Age sense, then, the four spectator categories examined here should be regarded as households that are distinguished according to their different identities, their relationships with specific places, and their distinctive participation within the community. Accordingly, these four household categories are differentially motivated to associate and affiliate with larger groups or communities (figure 3). One of the physical manifestations of such communities is the construction and maintenance of hillforts.

Figure 3 The four categories of a hillfort community.

I argue that there are four ideal-type groupings with which we can categorize Iron Age households in relation to their association with hillforts that closely resemble those of modern football supporters. The four household categories are based upon two binary oppositions: hot–cool and introvert–extrovert. This results in four quadrants into which households can be categorized: introverted/hot, introverted/cool, extroverted/hot, extroverted/cool. The four quadrants therefore represent ideal-type groupings, but crucially there is scope for movement between groups. This allows us to assess the cultural differences and changes over time of particular groups of Iron Age households and their relationships with specific hillfort communities.

The introvert/extrovert horizontal axis measures a household's motivation and ability to create networks of social relationships. Introverted households will tend to socialize with their immediate neighbours and kin, while extroverted households will have far-reaching networks of interaction that are not geographically restricted. The hot/cool vertical axis represents the extent to which the hillfort community is central to a household's personhood. Hot forms of loyalty reflect powerful types of identification and solidarity with the hillfort community, and cool forms the reverse.

Each of the four groupings therefore evinces particular kinds of identification with a specific place – a hillfort – and a specific motivation for that relationship. Each grouping also displays a specific kind of spatial relationship with the hillfort. Variation and difference amongst the constituents is given through their manifestations of thick or thin solidarity, which may reflect various status graduations.

Applying the model: Maiden Castle and other Wessex hillfort communities in the first millennium B.C.

The results of various landscape studies in Wessex suggest that in certain areas the Early Iron Age landscape consisted of closely spaced univallate hillforts, around three kilometres to 10 kilometres apart (Palmer Reference Palmer1984; Shennan Reference Shennan1985; Barrett, Bradley and Green Reference Barrett, Bradley and Green1991; Woodward Reference Woodward and Sharples1991; Gingell Reference Gingell1992; Bradley, Entwistle and Raymond 1994; Fulford et al. Reference Fulford, Powell, Entwistle and Raymond2006). In Dorset, there are more than 30 such sites ranging in size from less than two hectares to more than six hectares. Maiden Castle is one of the largest and most extensively excavated (Wheeler Reference Wheeler1943; Sharples Reference Sharples1991). Originally constructed around 500 B.C., a single timber and earth rampart and ditch defined an area of 6.4 hectares (figure 4, Phase 5). Access into the enclosure was provided by both an entrance on its north-west side and a peculiar ‘double’ entrance in the centre of eastern side. Multiple entrances into Early Iron Age hillforts are commonplace (Cunliffe Reference Cunliffe1984; Cunliffe and Poole Reference Cunliffe and Poole2000) and may have been created because the hillforts sat astride a number of previously separate Bronze Age territorial units (Sharples Reference Sharples1991). At Maiden Castle, one linear boundary ran east–west along the ridge of the hill and would have passed between the two gateways if extended to the entrance (figure 4, Phase 4). A possible interpretation is that as previously separate farming communities came together to live in the fort, the multiple entrances were a result of a desire to provide separate access to the different land units (Sharples Reference Sharples1991, 72).

Figure 4 Bronze Age to Middle Iron Age phases (4–6) of occupation at Maiden Castle (based on Sharples Reference Sharples1991, figure 33).

Unfortunately, little is known of the Early Iron Age occupation at Maiden Castle because in the excavated areas earlier settlement appears to have been largely destroyed by extremely intensive Middle Iron Age activity (Sharples Reference Sharples1991, 258). However, excavations at the neighbouring hillforts of Poundbury and Chalbury, which also date to the Early Iron Age (Richardson Reference Richardson1939; Whitley Reference Whitley1943), may provide useful parallels (figure 5).

Figure 5 Simplified comparative plans of Early Iron Age occupation at Chalbury, Poundbury and Maiden Castle (based on Sharples Reference Sharples1991, figure 199).

Poundbury is located around three kilometres north of Maiden Castle. A V-shaped ditch and accompanying bank, very similar to the rampart at Maiden Castle, enclosed 5.5 hectares, although excavation in the 1930s revealed very little evidence for occupation (Richardson Reference Richardson1939). More recent aerial photography of parchmarks in the summer of 1976 has, however, suggested the presence of a cluster of 12 houses in the south-east corner, though the rest of the hillfort was apparently empty (Green Reference Green1987). The implication is that settlement within the enclosure was not intensive, and any permanent resident population was small. In contrast to Maiden Castle and Poundbury, which are situated on the chalk, Chalbury is located on an outcrop of oolitic limestone on the south side of the Dorset Ridgeway about five kilometres south-east of Maiden Castle. Enclosing 3.4 hectares, the rampart was constructed of limestone slabs quarried from the enclosure ditch and was rebuilt at least once (Whitley Reference Whitley1943, 101). Unlike Poundbury, the interior of Chalbury was densely occupied, with over 40 houses dispersed across the hilltop (ibid., 98). It is uncertain which of these hillforts provides the closest parallel for occupation at Early Iron Age Maiden Castle, although the intensity of activity in the Middle and Late Iron Age suggests that it was likely to have been densely occupied, like Chalbury.

Despite the similar size of their ramparts and enclosed areas, the variable evidence for occupation at these three neighbouring hillforts suggests that they may have been created for different purposes that reflected both the needs and the nature of the hillfort communities that built and used them. The initial creation of the boundary was an important act to establish both social cohesion and a sense of ‘placedness’ – an attachment to both community and place. This was presumably undertaken by the households who lived in the range of smaller enclosed and unenclosed settlements surrounding the hillforts (for a detailed landscape survey of the local region see Woodward (Reference Woodward and Sharples1991)).

However, after the boundaries had been set up, it seems that the hillforts’ distinctive trajectories of use differed. At Chalbury, and possibly Maiden Castle, the large number of houses surely suggests that a large proportion of the hillfort community was constituted by permanent hillfort residents who had moved into the newly created enclosure. Through the physical act of dwelling within the enclosure, these permanent resident households would have developed long-term personal and biographical connections with the hillfort and the hillfort community. By contrast, at Poundbury there are far fewer houses and it follows that permanent residents must have represented a much smaller component of the hillfort community. Perhaps from an early stage the majority of households that comprised the Poundbury community wanted to maintain some sense of social and agricultural independence by remaining within their settlements amongst their fields, and only came together at certain times of the year for social gatherings, rituals or defence. It seems logical to assume that as a larger proportion of the community was constituted by regular or rare hillfort visitors it must also have been structured in a different way to that at Chalbury, since the itinerancy of these households meant that they did not possess the same strong spatial embedding as permanent residents (figure 6). The implication is that after establishing the hillfort community through the setting up of the boundary they may have visited for ritual or social gatherings, but if the hillfort community failed to deliver these, or a rival provided a more attractive proposition, then these households may have been tempted to drift away.

Figure 6 Hypothetical taxonomy of the Early Iron Age hillfort communities at Chalbury and Poundbury.

Around 450 B.C., Maiden Castle was significantly enlarged from 6.4 hectares to 19 hectares (figure 4, Phase 6). This coincided with the abandonment of all of the smaller settlements and enclosures in its vicinity, and also some of the hillforts (Woodward Reference Woodward and Sharples1991) – presumably the households were voluntarily moving, or being compelled to move, into Maiden Castle. Sharples (Reference Sharples1991, 84) argues that this suggests that Maiden Castle and its occupants had become the most important community in south Dorset, with power manifested in the control of large areas of agricultural land and resources. It seems that Poundbury was also entirely abandoned at this time (Richardson Reference Richardson1939); the community presumably moved wholesale into Maiden Castle. There is little sign of an armed struggle between the hillfort communities – there is no evidence of burning or slighting of defences at Poundbury, for instance – so how should we understand this change? One possibility is that the large number of regular or rare visitors which constituted the greatest proportion of the Poundbury community may have been easily enticed or persuaded to transfer their allegiances since they lacked the strong emplaced connections of the permanent residents. These households could then have been absorbed into the Maiden Castle community until the number of households comprising the Poundbury community was so small that it was untenable to maintain the hillfort's defences or continue the practical exploitation of the surrounding agricultural land, and they too threw in their lot with Maiden Castle.

The situation at Chalbury was somewhat different. Here there was clearly intensive occupation, suggesting that the hillfort community was constituted from a very early stage by a large proportion of permanent residents with a strong topophiliac relationship with the hillfort. Such spatial embedding meant it was likely to have been much harder to simply seduce and absorb this community. This may be reflected in the slightly longer chronological occupation of Chalbury, which was not abandoned until the early Middle Iron Age (c.400 B.C.; Whitley Reference Whitley1943). This suggests a more protracted and difficult process of amalgamation, probably requiring complex negotiations, agreements, threats and even possibly violence.

It is interesting to note that the abandonment of Chalbury appears to broadly coincide with the early Middle Iron Age refurbishment of the eastern entrance at Maiden Castle. The original eastern entrance rampart was fronted by a vertical wall, supported by large timbers, but the rebuild was furnished with a carefully constructed limestone wall. The blocks which formed this impressive limestone facade were sourced from the southern side of the Dorset Ridgeway, presumably in the vicinity of Chalbury (Sharples Reference Sharples1991, 76). The sourcing of limestone slabs from the locality of Chalbury is unlikely to have been arbitrary and it is interesting to consider that they may even have been brought from the hillfort itself. The final-phase rampart at Chalbury appears to have been deliberately slighted with large limestone slabs that constituted the boundary forming a distinct destruction layer within the enclosure ditch (Whitley Reference Whitley1943, figure 2). This act of boundary destruction, the very physical manifestation of social cohesion, could be interpreted as an analogy for the symbolic death of the resident hillfort community. In that light, the subsequent incorporation of limestone slabs into the rampart at Maiden Castle, some possibly from Chalbury itself, could be understood as a merging both of monuments and of communities – a literal physical embedding of Chalbury into the fabric of Maiden Castle.

However, as Sharples (Reference Sharples1991, 260) has highlighted, the agricultural land around Chalbury was probably too far away to have been farmed directly by the permanent residents of Maiden Castle, and this would have required many households to remain living in the surrounding landscape. Part of the Chalbury community must, then, have been forced to move into the smaller outlying settlements, such as Quarry Loddon (Bailey and Flatters Reference Bailey and Flatters1972), which see continual occupation into the Late Iron Age. These households would have represented a sizeable population of seasonal and regular visitors whose emplaced connections with the hillfort community were reinforced through annual rampart construction events rather than permanent residence. Others, whose agricultural land was closer to Maiden Castle, must have moved into the hillfort. This would surely have included the previous occupants of Poundbury. It is interesting to note that there was a reorganization of settlement within Maiden Castle during the Middle Iron Age with the construction of regimented rows of similarly sized houses. This may have been a strategy to break down the extended kinship ties that existed between the previously independent households that had moved into the hillfort and to strengthen the importance of the larger Maiden Castle community (Sharples Reference Sharples1991, 262).

The development of large Middle Iron Age developed hillfort communities can in some ways be likened to the establishment of modern professional football clubs. Teams such as Manchester United have progressively become dependent upon the custom of fans and flâneurs to maintain their status, although they retain strong local support. The large communities of developed hillforts such as Maiden Castle may have been similarly structured. Such a community was potentially unstable because it would have consisted of a large number of households whose allegiances could change rapidly as they did not have a historical or emplaced connection with the hillfort. The active participation in constructing Maiden Castle's ramparts, combined with compelling people to reside permanently or seasonally within the boundaries, can be seen as an attempt to create such a place relation between hillfort and community. Within such a community there may have been distinctions of status. The original Maiden Castle residents or their descendants may have claimed greater status over regular and rare visitors, but recognized that the construction and maintenance of the large hillfort and the working of its agricultural land were not possible without their support. In this sense, the necessity to create a sense of community overrode the desire to establish distinction.

Around 100 B.C. there was a reappearance of settlements in the immediate vicinity of Maiden Castle when occupation within the hillfort was at its densest. The greatest volume of material culture at Maiden Castle also belongs to this final period of occupation. This increase in commodities arriving at Maiden Castle coincides with a major rationalization of settlement inside the hillfort and abandonment of any great interest in the defences (Sharples Reference Sharples1991). The provenance of material also appears to change over time. In the early phases of the Middle Iron Age, pottery appears to be primarily produced at a number of local sources and the forms are a standardized range of shouldered jars. By the second century B.C. more distant industries became much more important until around 100 B.C., when Poole Harbour pots made up 95 per cent of assemblage represented by a more varied range of vessels from large storage jars to small bowls. These later bowls and jars are often elaborately decorated, perhaps to define distinct territories conforming to tribal groupings. After 50 B.C. there was a gradual breakdown of occupation of the hillfort and the importance of the ramparts diminished. There was a breakdown in the structure of the occupation within the hillfort, while there was an increase in the number and size of surrounding settlements and the development of field systems. Sharples (Reference Sharples1991) has suggested that this indicates the breakdown of communal ownership and a new emphasis on individual status.

How should we understand the demise of Maiden Castle? One possibility is that as the dependent territory and population grew, it became increasingly more difficult to integrate and coordinate this large community through monumental hillfort construction. The exchange of material culture, particularly metalwork, became much more important as a means of creating and maintaining relationships. Yet the acquisition of this material may have undermined the hillfort system. Sharples (Reference Sharples2010) has highlighted that although pottery and stone were embedded within local gift-exchange networks, this was not the case for metals. Indeed, metalwork appears to have been produced by specialist artisans (Iron Age flâneurs) located in areas on the periphery of the chalklands. Sharples (ibid.) has argued that the primary production of metalwork was removed from the process of gift-giving as hillfort communities like Maiden Castle now had to barter with artisans for products to use in gift exchange between and within other hillfort communities. This meant that the act of producing and distributing material was separated from the role that material culture played in creating the links between hillfort communities. These changes increased the importance of certain individuals who were closely involved in the exchange processes and were able to manipulate exchange relationships. This increased their cultural capital within hillfort societies. Ultimately this was expressed by the gradual decline of some developed hillforts like Maiden Castle, the catastrophic destruction of others like Danebury, and the re-emergence of individual settlements (ibid.).

The developed hillfort model was therefore ultimately unstable. Hillforts were dependent on the growth and maintenance of a large territory and a dependent population, yet the links that held such a community together were messy and tangled, and cross-cut households and larger groups. In the long run, households without an emplaced connection with the hillfort may have become easily disenfranchised from the community, seeking new partnerships and associations elsewhere, which would send the hillfort community into gradual or sometimes rapid contraction. The emergence of tribal identities centred on enclosed and territorial oppida in the 1st century B.C. may have been the result of the undermining of such hillfort communities by households and individuals creating new forms of political and social organization. These emergent kingdoms of Late Iron Age southern and eastern England could be caricatured as modern fashionable football clubs like Manchester City or Paris St Germain, whose existences are sustained by immensely wealthy and charismatic individuals who attract followers and flâneurs through the acquisition of star players (high-status individuals) with transferable allegiances.