‘Good government and effective administration required cooperation. Accordingly, successful Romanisation of the native élite was essential to local administrative control through the urban network and, as such, no less important a means of controlling conquered territory than was military force’ (Hanson 1997: 78).

Our understanding of the relationships between colonisers and indigenous peoples, past and present, has changed dramatically over the last 20 years. This paper argues that although theories of resistance, identity, and hybridity now commonly inform interpretations of archaeological data, there are still areas deserving of closer scrutiny. My doctoral research (Harlow 2017; 2018) crosses chronological, geographical, and disciplinary boundaries. It focusses on practice and display, drawing on Crummy’s (1983) functional categories which are allied to praxis, and Eckardt (2014), Gardner (2007), and Revell’s (2016) work on identity and difference.

My work uses material culture to investigate how the indigenous people of northern East Anglia identified themselves within Late Iron Age society and in contrast to, or in alliance with, the Roman conquerors. The study spans the period 100 BC–AD 200 and the modern counties of Norfolk, Suffolk, and Cambridgeshire. Working with small finds provides opportunities to think about how changes in practice can occur with the same material culture and vice versa. New patterns of display and behaviour can be sought through conservatism, innovation, and hybridity in objects. Although material agency without context, and intentionality in resistance, have been questioned (Russell 2004; Ribeiro 2016), objects can express and create feelings of (in)equality.

One of the challenges, often debated at previous TRAC conferences and workshops, is to successfully integrate theoretical understanding with large material datasets. Contextual interpretations are difficult to apply to unstratified surface finds, which form the bulk of my data. Nonetheless, I believe there is significant research value in large-scale studies of metal-detector finds. What we do have is material culture, and lots of it: I have compiled a database of over 14,000 individual object records from the Portable Antiquities Scheme (PAS) and county Historic Environment Records (HERs) (Harlow 2018).

A discussion of boundaries necessitates talking about identities, but I would argue that the two are not identical. The idea of social boundaries encapsulates more nuanced divisions between people. I am interested in the social frontiers between coloniser and colonised, elites and non-elites, and those in between. The wide-ranging analysis in which I have been engaged demonstrates some of the possibilities and difficulties of integrating theory and data.

A brief outline of the historical and archaeological background to the Iceni territory and its distinctive regional characteristics sets the scene within which theoretical issues of cultural contact and change can be problematised. Then, I present four case studies, for which the underlying data is presented in Supplementary Table 1. Horse and chariot fittings are used to question social organisation, elite status, and the Icenian response to the invasion. I reconsider the role of brooches as symbols of group allegiance and resistance in light of intra-regional patterning. Themes of selective adoption and hybridity are explored through distributions of pipeclay figurines and seal boxes.

The Civitas Icenorum

Late Iron Age (100 BC–AD 60) ‘Icenia’ was atypical in comparison with other parts of southern Britain.1 The Iceni seem to have subscribed to a different way of life than their neighbours, following less hierarchical settlement and burial practices. Although it is unwise to interpret social differentiation directly from material remains, this may indicate a more egalitarian social structure. Most homesteads were unenclosed and there were few of the extensive nucleations usually described as oppida (Pitts 2010; Moore 2012; Fernández-Götz 2014). There is little archaeological suggestion of developed social hierarchies, and an absence of the ‘warrior chieftain’ funerary rites typified by Aylesford-Swarling or Welwyn in the southeast, or the extravagant ‘Arras-style’ chariot burials of Yorkshire (Clarke 1939; Hill 1995; 2007; Martin 1999; Davies 2011). Only a select few imports reached the region from the Mediterranean and Gaul. Instead, distinctive items of indigenous material culture, including jewellery and horse gear, are often found in hoards rather than graves. Production continued into the early Roman period, frequently developing hybrid forms which blended local and Roman manufacturing techniques and designs.2

I use the term civitas Icenorum to describe the territory populated by the people known to us as ‘Iceni’, effectively a Roman administrative unit with its capital at Venta Icenorum (modern Caistor St. Edmund, Norfolk).3 Recent numismatic studies (Chadburn 2006: 478; Talbot 2011; 2015; 2017; Leins 2012: 300–304) have determined that there were several loosely associated sub-groups within this area, which perhaps coalesced around a charismatic individual or leadership group during the turbulence of the invasion. This coalition may have been identified by Roman officials as ‘Iceni’ (Davies 1999: 41; Hutcheson 2007: 369). However, the latest synthesis found no evidence of ‘rigid centralised organisational structures’, observing that the connection between coinage and political power or kingship ‘does not appear to be appropriate in Late Iron Age East Anglia’ (Talbot 2015: 313–314; 2017).

Previous considerations of boundaries sought geographical borderlines between the Iceni and their neighbours (e.g. Clarke 1939; Allen 1970; Martin 1999). While there are some observable geographical confines to the circulation and deposition of Icenian coinage, my study of other portable artefacts (including horse harness, votive objects, and personal adornment) has revealed different distributions. Despite distinct groupings of material culture, there is no archaeologically clear pattern to be ascribed to a single unified polity or people. In light of this, discussing the Iceni (or the neighbouring Trinovantes in south Suffolk and Essex, or the Catuvellauni in Cambridgeshire) as discrete entities is problematic. The names are retained here as familiar epithets, but should be read as shorthand for the communities of these regions, without implications of ethnic identity or coherent leadership.

After the Claudian conquest, we are told, the Iceni were affiliated with the Roman Empire under a client king, Prasutagus (Tacitus, Annales 12.31; 14.31). The breakdown of this relationship and the ensuing Boudican ‘War of Independence’ (Cassius Dio, Historia Romana 62.1–12; Crummy et al. 2015) in AD 60/614 has been much discussed (e.g. Trow 2003; Collingridge 2005; Hingley and Unwin 2005; Gambash 2012). A narrative of widespread depopulation and abandonment in the aftermath, based on Tacitus (Annales 14.38), has gone relatively unchallenged, but is little supported by material evidence.

The Iceni have been constructed as an ‘Other’, not just as rebellious barbarians facing up to the Roman invaders, but also in opposition to inhabitants of the south-eastern kingdoms. They are perceived to have rejected the civic accoutrements of mosaics and inscriptions, and the architectural statements of large towns and villas, preferring traditional modes of status, display, and portable wealth (Fincham 2002: 83; Mattingly 2007: 111–113; Andrews 2012: 66). These absences may reflect significant pre-conquest differences, rather than an unwillingness to embrace a ‘Roman’ way of life (e.g. Atkinson 1932: 42; Millett 1990: 100–101). Their slow adoption of new material culture and practices has been ascribed to social inwardness and, more recently, in terms of resistance. Cultural conservatism can be part of a dynamic of resistance in response to Imperial oppression (Hill 2001). I define ‘resistance’ in Hingley’s (1997: 88) sense, in which material culture is used in subtle forms of opposition, or Scott’s (1990) ‘hidden transcripts’, which may have occurred as an alternative, or a counterpart, to outright armed rebellion.

The Material Elite Paradigm

An elite can be described as: ‘Individuals in a society with greater access to wealth and power than the majority of members’ (Wells 1999: 267). Elite status and identity in the Roman world have been presumed to be ‘self-evident, inevitable, and unproblematic’, as Revell (2016: 61) observes. Uncritically used, expressions like ‘native elites’ and ‘high-status’ fail to explore variability or discrepancy. Hoards and richly-furnished burials are often interpreted as deposits of individual wealth and status, resulting in the assumption that ‘material elites’ are readily visible in the archaeological record.

We need to be aware of how we use this terminology and apply it to archaeological groups or artefacts. We must also question the assumed membership of these elites, aristocracies, and status groups. I do not dispute the existence of social differentiation, but wish to raise the debate about how to recognise it archaeologically. Individuals could have held membership of various, sometimes contradictory, social groups, based on household and kinship connections, knowledge and skills, age and peer groupings, inherited class and power, as well as non-hereditary leadership (Tullett 2010: 65–67). An exclusive group based on access to specialised knowledge, such as a craft workshop or a religious cult, would leave very different material remains to one based on communal ownership of livestock, for example.

What other models might have existed in later prehistoric and early Romano-British societies? Hill (2011: 258) explored alternatives to stratified societies, such as segmentary and non-hierarchical groupings, which lack elites and hold resources communally. The idea of heterarchy, which allows for multiple and overlapping interest groups to exert influence, has been applied variously to complex societies, metalworking, and gender groups (Crumley 1987; 1995; Ehrenreich 1991; Ehrenreich et al. 1995; Levy 1999). Rathbone’s (2014) ‘anarchist archaeology’ and Revell’s (2016) search for non-elite provincial identities also offer interesting possibilities. Oosthuizen (2016) looks at the potential for collective land holding and ‘horizontal’ forms of governance. Even in ostensibly egalitarian societies, people hold strong opinions about values and (in)equality and monitor and assess each other accordingly (Wason 1994: 36–37).

Watts (2005: 2) makes a valuable point that increasing hierarchy may have come about as the result of interactions and trade during the first century BC. She also considers the likelihood of slave labour in certain industries in Roman Britain. Iron slave shackles have been discovered at Saham Toney in Norfolk (Bates et al. 2000: 230) (NHER 29429) and dredged from the River Wensum at Worthing (NHER 2984). Was there an ‘Icenian aristocracy’ (Gurney 1998: 27) which included Prasutagus and Boudica, or their equivalents? After the rebellion was subdued, were the remnants of this ‘elite’ enslaved or allowed to take up positions of power in the new administration? If material culture and craft skills were valued as sources of status, these may have provided a safe locus for the manifestation of rival identities. During the occupation, we can only imagine the dynamics at work between local power-brokers, resettled veterans, active military personnel, and the administrative class, as well as the average subsistence farmer. Nevertheless, this viewpoint still assumes that hierarchy was the norm. There are possibilities for a re-reading of the Boudican uprising, using concepts of consensus and collective action (DeMarrais 2016).

The paradigm of the material elite is based largely on the expectation that past societies operated hierarchically through material wealth, projecting the values of late Western capitalism backwards in time. Social status can nevertheless be predicated upon many other factors, including ownership of livestock, magico-ritual capital, prowess in battle or hunting, and links with the ancestors or land. Many of these forms would leave little archaeological trace, but it is important to consider their existence (Hill 2011: 243). I am not implying that such assumptions are made by all Late Iron Age and Roman archaeologists, rather that they are rarely overtly problematised in the same way that gender bias, for instance, is now commonly questioned.

The idea of inherited status is also seldom interrogated. Caesar (Bellum Gallicum 5.14) states the Britons practiced polygamy and had customs which determined the place of children within this marriage network, although there is nothing to corroborate this. Tacitus (Annales 14.29) introduces other elements of social hierarchy, such as the role of priests or ‘Druids’. These texts were intended to emphasise the ‘transgressive customs’ of the ‘barbarian’ Britons (Stewart 1995: 3). Hoards of torcs, coins, and horse harness are a ‘culturally distinct phenomenon’ of the region (Chadburn 2006: 174). Based on their frequency, it is conceivable that the Iceni held their wealth in common, rather than individually, in the form of precious metalwork or livestock. This social organisation would have an important impact on inheritance of land, power, and wealth, and on our ability to recover these structures archaeologically.

The material elite paradigm has seeped into our very understanding of artefacts and practices such as hoarding. The language speaks of ‘treasuries’ (e.g. Stead 1991: 463), objects are ‘in circulation’ or ‘cached’ for safe-keeping. There is also a tendency to refer to Roman ‘industrial production’ (e.g. Allason-Jones 1989: 68), and Gaulish and Romano-British ‘factories’ (Green 1976: 15, 20; Watts 2005: 3), perhaps a relic of the progressive understanding of Roman Imperialism (see papers in Webster and Cooper 1996 for critique). Mackreth (2011: 68, 130) goes further, referring to the ‘marketing area’, factory owner, and mass production of certain types of brooch. In fact, the values placed upon precious metalwork, even coinage, in this period, are poorly understood. Talbot (2015; 2017) found conflicting evidence for the use of Icenian coins as part of a ‘market economy’. Walton (2011: 290–291) describes the patchy monetisation of Roman Britain, especially in the early centuries of occupation. During the early colonial period in North America, the Algonquians were eager to acquire refined copper at the expense of gold and silver. This seemed naive to the European colonists, but ‘within the indigenous world view it made complete sense’ (Farley 2012: 152–156): copper was a powerful, spiritual, and political substance, particularly valued for its red colour. This historical example should remind us to avoid retro-fitting materialist and capitalist doctrines onto prehistory.

The division of labour which supports specialised craft practices is sometimes considered evidence for social differentiation, although this has been disputed (e.g. Ehrenreich 1991). In this model, supposed ‘elites’ exercise symbolic and economic control over the skills of craftspeople and the negotiation and exchange of prestige items. Commissioning high-quality artefacts is seen as a political strategy to distinguish the (material) ‘elite’ from other social strata (Wason 1994: 107–108). Certainly, there may have been a relationship between social stratification and craft specialisation, but this seems rather too analogous with a medieval European system of patronage.

In the production of Middle Iron Age cauldrons, there is little sign of the dedicated elite-sponsored workshops once thought to have been a major stimulus for production of ‘prestige metalwork’ (Baldwin and Joy 2017: 103–105). The frequency of metalworking debris on settlements confirms that small-scale, local production with idiosyncratic styles and methods continued into the Late Iron Age. Increasing standardisation is considered to exemplify the metalwork of the early Roman occupation, but this is superficial. Many artefacts created in the first two centuries AD were individually finished, even when manufactured in relatively large amounts (Jackson 2010: 26). Therefore, I would argue that a hierarchical, competitive social milieu need not exist in order for highly-skilled craft production. The trappings used for harnessing horses to carts and chariots form one such category of elaborate decorative metalwork. In the next section, I present my interpretation of these artefacts and suggest that the symbolic and social value of horses may represent a more communal system of wealth.

Horse Harness

In Icenia, ‘the Kingdom of the Horse’ (Davies 2009: 109), alternative expressions of status may have been displayed through livestock and highly-ornamented harness and chariot fittings. The density of terrets (rein rings from vehicles) is especially high in Norfolk, whilst low in Cambridgeshire, perhaps due to the inaccessibility of the Fenland by road (Figures 1, 2). Horses are one of the most frequent motifs on Icenian coins, but are rarely represented in other pre-conquest material culture. They may therefore have played an emblematic role within appropriate contexts, subject to certain prohibitions (Creighton 1995: 289; Davies 2014: 31). A horse would certainly qualify as a ‘prestige’ belonging: rare, expensive, even dangerous. It would embody an expenditure of effort in terms of breeding, training, and maintenance. It might hold symbolic capital in a range of guises, from religious to practical. The number of horses owned may have conferred status on a person or group and this could have been further enhanced if the horses and vehicles were bedecked with brightly-enamelled metalwork.

Figure 1 

Distribution of Late Iron Age and early Roman terret rings (Source: Author).

Figure 2 

Density of all terret rings (Source: Author).

The pre-conquest Icenian tradition of producing horse harness and chariot fittings continued into the early Romano-British period. Highly-decorative terrets flourished in the mid-first century AD. Certain finds are restricted to the north of the region, including ‘platform-decorated’ terrets, adorned with geometric polychrome enamelling (Hutcheson 2004: 30) (Figures 3, 4). These items would have been very visible. It could be argued that the increased prominence of terret rings served to demarcate the importance of horses and vehicles during the client-kingdom. Enamelled terrets were also deposited in hoards, suggesting a votive role. Hybrid forms are recorded which blend formal characteristics and decorative schemes of indigenous and Roman origin.5 This is a small part of the evidence which contradicts the Classical accounts of the merciless depopulation of the region in the Boudican aftermath.

Figure 3 

Distribution of platform-decorated terrets (Source: Author).

Figure 4 

Matching platform-decorated terrets, Saham Toney and Carleton Rode, Norfolk (NWHCM Accessions: 1847.66.3 and 2006.349, courtesy Norfolk Museums Service).

What was the status of metalworkers in Late Iron Age and Romano-British societies? Were they part of a craft elite based on expert, perhaps arcane, knowledge? Did the expansion of material culture in the first century AD (Hill 1995: 66) follow from a democratisation of the knowledge of production? The concept of the ‘itinerant metalworker’ is frequently cited (Johns 1996: 20; Bayley et al. 2001: 97–100; Mackreth 2011: 242), but there is little support for this other than occasional hoards of scrap metal, unfinished items, and tools, which may have had non-functional associations. The evidence for established workshops is similarly limited (Bayley and Butcher 2004: 35–40). The individuality of finished artefacts might imply that pieces were made to a considered aesthetic as part of a negotiated process between maker and recipient. Manufacturers may have catered to local tastes and styles, or simply provisioned the nearest communities. We need to develop a better understanding of the processes of commissioning, using, and depositing artefacts and how they travelled from their places of origin.

We might ask what the evidence for itinerant bronzesmiths would look like. Did workshops exist in the vicinity of areas of deposition? A consideration of a specific group of objects can illuminate some of the issues raised above. Enamelled flat-ring terrets are typically recovered in west Norfolk and northeast Suffolk (Figure 5), including a matched set of eight from the Westhall hoard. They date to between the late first century BC and mid-first century AD. The distributions of such highly-decorative harness in north Suffolk and Norfolk were previously used to attribute Icenian origins (Martin 1978: 137; Gregory and Martin 1985: 35). The terrets from Weybread (SHER WYB003), Westhall (SHER WHL007), and Rushmere (SHER RMA004) in northeast Suffolk are so similar that they are probably from the same production centre. Such a workshop may have existed at Waldringfield in southeast Suffolk (SHER WLD001), considered Trinovantian territory (Martin et al. 1986; Rigby 2013). Westhall-type terrets have been found as much as 60 miles away from Waldringfield, although remaining tightly within northern East Anglia (Figure 6). Are we seeing here the passage of one of these travelling craftspeople? It is also notable that the distribution of Westhall terrets differs quite clearly from that of platform-decorated types, which may have had a manufacturing base near Saham Toney, hinting at intra-regional preferences.

Figure 5 

Westhall-type terret, Sporle with Palgrave, Norfolk (PAS: NMS-6599B4. Reproduced under CC-BY-SA 4.0 licence).

Figure 6 

Distribution of Westhall-type terrets (Source: Author).

Both the quantity and quality of horse and chariot fittings in the north of the region suggest that this area was an important focus of Late Iron Age activity and one of the ‘wealthier’ sub-regions. It may also show a shift in the Icenian powerbase, from the era of the Snettisham torcs in northwest Norfolk in the last centuries BC, to the more central Breckland area by the time of the Boudican uprising. Later first and second century AD versions, such as dropped-bar and skirted terrets, are less conspicuously decorated and may be continental imports. These are not commonly found in hoards. Terrets still fulfilled their functional role in transport, but modes of display and disposal altered (see Lewis 2015: 220–222). A variety of possible social boundaries suggest themselves here. Stratification or difference may have occurred between the makers of fine artefacts, those who knew the techniques of breeding and training horses, or the people with the power or means to commission these things. What is clear is that artefact distributions do not correlate with boundaries based upon ‘tribal’ identifiers known from Classical texts. This is reinforced by the following discussion of cultural resistance.

Resistance is Useless!

Artefacts mediate in power relationships: they have potency over people and give people authority. In this way, objects interact with power at the boundaries between colonised and coloniser. The agency and potential of the masses have often been overlooked in discussions of Roman colonisation. The over-concentration on elites leaves the majority of the population ‘neutralized’, ‘disempowered’, and ‘unprivileged’ (James 2001: 199). Cultural resistance can be a constructive practice within dominated and marginalised groups, unwilling or unable to risk more direct action against oppression (e.g. Hingley 1997; Wells 1999; Webster 2003). It is therefore possible to speak of the power of the dispossessed, which may take a variety of forms: ‘subordinate groups in the province resisted the power and ideological projections of the Roman rulers in everyday practices using everyday things’ (Hill 2001: 14). Prior to the rebellion and for several generations beyond, items of portable wealth and display may have been used to resist incorporation into the Roman world (Davis and Gwilt 2008: 146).

As a theory, cultural resistance has been questioned. Sauer (2004: 119) accuses it of being ‘a misguided research strategy’ and an ‘anti-colonialist mission’ seeking proof in the archaeological record of ‘otherwise unrecorded widespread resistance’. He challenges the idea that the ‘subdued population of the provinces’ was constantly struggling for independence and asks why covert resistance rarely spilled over into open rebellion in the Roman Empire. Of course, in the case of the Iceni, it did. Carr (2006: 15–16) puts forward the idea of ‘persistence’ and ‘multi-faceted local struggles’ in an attempt to move away from the polarisation of oppressive Romans vs. rebellious Natives. Resistance is also problematic from a practical perspective, given what we know of intra-regional diversity and the relatively slow spread of change and communication. In this section, I discuss three categories of objects and how they each tell a different story of Icenian resistance.

Brooches: Badges of Belonging?

Regional variation and change in personal belongings have been interpreted as indicators of identity and resistance. Were specific brooches worn by the Iceni as expressions of their group identity and opposition to Roman rule (Mackreth 2011: 145; Davies 2014: 30)? Mackreth (2009: 139) identified the Rearhook brooch (a Colchester Derivative with a rear-facing hook which secures the spring; Figure 7) as a distinctively Icenian type.6 He dated them tightly to the invasion in AD 43, and drew a marked endpoint for manufacture in AD 60/61, which he assigned to the ‘savage repression of the Iceni after the defeat of Boudicca’ (Mackreth 2009: 144). According to Mackreth’s dating, the Rearhook subsequently ceased to be made or distributed, replaced by hinged brooches in the later first century AD: ‘This Icenian brooch type had become a souvenir or a bleak survival’ (Mackreth 2009: 144, 146).

Figure 7 

Colchester Derivative Rearhook brooch, Pulham Market, Norfolk (PAS: SF-C65E1A. Reproduced under CC-BY-SA 4.0 licence).

Mackreth (2009: 147) went on to propose that members of the disaffected Iceni were removed in ‘a kind of enforced diaspora’, developing their distinctive brooch type in supervised exile. An alternative narrative has metalworkers fleeing as economic migrants ‘from a devastated and impoverished land to a more congenial area’ (Mackreth 2011: 234). This is based on a distinct sub-group of Rearhooks clustered around Cirencester, far outside their usual distribution. If, according to Mackreth, the Rearhook was the mark of the Iceni, then the Harlow brooch (a Colchester Derivative with a double pierced lug arrangement; Figure 8) was commonly worn by their neighbours the Catuvellauni and Trinovantes (Mackreth 2011: 50, 64–65).

Figure 8 

Colchester Derivative Harlow brooch, South Elmham St Mary, Suffolk (PAS: SF5751. Reproduced under CC-BY-SA 4.0 licence).

Figures 9, 10, 11, 12 show the distributions and concentrations of Rearhook and Harlow brooches (see Table 1). At first glance, the distribution maps look similar, but the densities tell a different story. Major concentrations of Rearhook brooches occur around the central belt of pre-conquest settlement. Harlow brooches have a more southerly distribution, in line with Mackreth’s observations, with a strong focus around Coddenham. The parishes around Withersfield show concentrations of Harlow brooches and a paucity of Rearhook brooches, again fitting the ‘tribal’ ascriptions. However, Harlow brooches are also found in west Suffolk around Mildenhall and west Norfolk near Fincham, usually considered within Icenian territory. Neither type is common in northeast Norfolk or in Cambridgeshire, perhaps reflecting recovery bias and geographical constraints, such as wetlands.

Figure 9 

Distribution of Colchester Derivative Rearhook brooches (Source: Author).

Figure 10 

Distribution of Colchester Derivative Harlow brooches (Source: Author).

Figure 11 

Density of Colchester Derivative Rearhook brooches (Source: Author).

Figure 12 

Density of Colchester Derivative Harlow brooches (Source: Author).

Table 1

Rearhook and Harlow brooches by county.

County Harlow Rearhook

Cambridgeshire 74 38
Norfolk 297 684
Suffolk 839 671
Total 1210 1393

The small towns of Coddenham, Wenhaston, and Hacheston in Suffolk are hotspots for both brooch types, areas recognised as zones of interaction by Martin (1999: 86, 90). This Gipping Valley area shows ‘no clear tribal affiliation’ and ‘may reflect greater contact and cultural mixing’ (Moore et al. 1988: 13). These settlements certainly do not represent a boundary or a dividing point, rather a nexus in which different groups met and merged. I would argue that many people could have owned and worn both types of brooch. It must be noted that brooch spring fixings are not especially visible, even to the wearer, so would have been limited as a means of distinguishing one group from another, except in close proximity (Cool 2016: 415). Regardless of the impact of the Boudican revolt, the Rearhook spring fixing was a particularly ineffectual technique of holding the pin which soon went out of use.

Mackreth (2009: 139) names the Rearhook as an Icenian brooch, but my study shows a much more complicated picture. There is an inherent danger in mapping group identities indiscriminately onto material culture and this is certainly not my intention. Brooches are, of course, not directly comparable to people. Certain types and styles clearly had restricted distributions, but I am wary of the idea that brooches were worn as badges of ethnic or social identity. If this is the case, then neither can brooches effectively be used to show cultural resistance. The regional patterning is more likely to reflect local workshops and fashions (Davies 2014: 30–31; Eckardt 2014: 218).

Brooch moulds can highlight indigenous manufacturing and distribution processes. They are extremely rare: only four metal (as opposed to clay) examples are known, all from Norfolk.7 Rearhook moulds have been discovered at Felmingham (NHER 34732) in northeast Norfolk, and at Brancaster (NHER 1003) on the North Norfolk coast. At Old Buckenham (NHER 30864) to the southwest of Wymondham, an unprecedented collection of bi-valve mould fragments, some still containing unfinished brooches, was found. These were for producing Harlow brooches with their characteristic double pierced lug (Bayley et al. 2001).

Evidence for a bronze-casting workshop at Venta Icenorum includes clay investment moulds and both triangular Iron Age and hemispherical Roman crucibles (NHER 9816). Investigations revealed a layer of ash containing metalworking debris; hearths, a kiln, and a furnace were found nearby. Failed castings provided evidence for production of razors, brooches, bracelets, and pins. Some burnt mould fragments preserved the shapes of the objects within, including the only known manufacturing evidence for Mackreth’s Harlow Type 1.a1 brooch (Tylecote 1969; Mackreth 2011: 50).8 Harlow brooches flourished from c. AD 40–70, although they do run into the later first century AD, putting these finds either in the client-kingdom or immediate post-revolt period. This presents a new set of questions: were Harlow brooches being made here during the client-kingdom, before the inception of the civitas capital as a structured settlement? Was the workshop active into the second century AD, producing a late version of the Harlow decades after the revolt had been quelled? If so, who was manufacturing them? Why were so few deposited in the hinterland of the workshop?

Three observations can be made here. Firstly, no substantial Late Iron Age settlement is known in the vicinity of the civitas capital (contraDavies 2011: 103–104). The formal layout of the street grid and main buildings has been convincingly re-dated no earlier than AD 120 (Bowden 2013). Therefore, we would not expect many mid-first century AD brooches in and around the capital, and indeed this is shown by the density maps. Secondly, the presence of Iron Age crucibles alongside Roman versions, and the use of investment moulding, presents the intriguing possibility that Icenian bronzesmiths were undergoing an experimental phase in which both indigenous and introduced technologies worked in parallel. This strategic approach of selective adaptation refutes some of the accusations of deliberate rejection or backwardness attributed to the Iceni during this period. Finally, if the Harlow brooch is characteristic of the Trinovantes (Mackreth 2011: 50), its manufacture at Venta Icenorum and Old Buckenham in the heart of Icenian territory, is unexplained.

The idea that people expressed resistance to colonisation through subtle signals encoded in everyday speech, dress, and actions (Scott 1990) is attractive, but does not seem to be supported by the evidence from brooches. Opposition can also be demonstrated through refusal and avoidance, or the appropriation and subversion of external ideas, including religious behaviour and literacy. Using an object-based approach to these concepts, I will now consider the Icenian reputation for cultural conservatism.

Pipeclay Figurines

Theories of resistance, creolisation, and hybridisation still focus on Rome as the key originator of ‘non-indigenous cultural forms’ during times of change. This fails to acknowledge the importance of pre-existing exchange networks and relationships with other neighbouring provinces, such as Gaul (Revell 2016: 47). Resistance, adaptation, and acceptance can occur simultaneously (Carr 2006: 110). One class of imports may demonstrate this process of selectivity and receptivity to non-Roman origins. Pipeclay figurines were produced in large quantities in Gaul and the Rhineland during the first and second centuries AD (Jenkins 1977). There is currently no evidence that they were manufactured in Britannia (Fittock 2016: 1). The range of figurine types found in Britain is restricted compared to that from Central and Eastern Gaul. This suggests that only certain forms were demanded for import, or offered for export.

In my study area, there is a distinction between anthropomorphic and zoomorphic figurines. Norfolk has almost exclusively female deities, such as Venus and Dea Nutrix, with one rare Mercury figure from the civitas capital (Figure 13). The occurrence of child burials with primarily zoomorphic figurines in south-western Cambridgeshire allies this zone with Essex and Hertfordshire, rather than Norfolk and northern Suffolk. The location of pipeclay figurines across the three counties is notably different compared with that of other small finds (Figure 14). They are barely represented in areas with high densities of Late Iron Age finds and settlements. Instead, they are found at sites with post-conquest foundations, small towns, and temples.

Figure 13 

Two joining pieces of a pipeclay figurine of Mercury, Venta Icenorum, Norfolk (CRT SF5139 and SF5140, courtesy Ian Jackson, Caistor Roman Project).

Figure 14 

Distribution of pipeclay figurines (Source: Author).

The arrival of pipeclay figurines, representing alien deities, sacrificial practices, and mythological allusions must have provoked mixed reactions in the local inhabitants: from mystified or intrigued to disinterested. Their early adoption in some areas, especially in funerary contexts, is contrasted with later deposits in other parts of the region, suggesting a lag between their manufacture, subsequent use, and later discard. These Gaulish imports, along with amphorae, and other Mediterranean artefacts, may have been rejected by those in the Iceni heartland, where practices of devotion and disposal of the dead followed older customs into the early Roman period.

This is in marked contrast to second century AD brooches, for example, which remained popular on most settlements with continuity from the Late Iron Age into the early Roman period. Brooches were less selectively deposited, perhaps highlighting a difference between everyday ornament and religious beliefs or funerary rites. The discriminating uptake of pipeclay figurines implies that communities had an element of choice or negotiation over which deities they chose to worship or how they incorporated new ideas and objects into their daily lives. This counters both the stereotype of Icenian cultural retardation and the enforcement of Roman power structures in the civitas. Local evidence of manufacturing is the key difference between these two categories of finds, which raises issues of demand, distribution, and disparate markets. Hybrid artefacts can show how new objects and ideas were appropriated into indigenous manufacture and usage.

Seal Boxes

Literacy is considered part of the domination and power of conquest (Eckardt 2014: 177–207). The practices of reading, writing, and speaking in a new language would have been part of the internalisation of power structures and the rules governing social order. These repetitive physical actions presented opportunities for people to identify and belong, or to resist and stand apart. Seal boxes can be interpreted as part of this language of literacy and power. They date broadly to the second and third centuries AD. They have no known Late Iron Age predecessors, suggesting they were part of a new lexicon of material culture.

Seal boxes are small copper alloy containers, often with decorative lids. A retaining knot was sealed within by impressing an intaglio ring into beeswax, which served to guarantee identity and prevent tampering (Bagnall Smith 1999: 40). These objects are usually considered to seal writing tablets, although Andrews (2012; 2013) provides strong evidence that they could equally have fastened small pouches of valuables in transit. Bagnall Smith (1999) relates them to the ‘Roman’ practice of making religious vows at temples, based on a large group from the rural shrine at Great Walsingham, Norfolk.

A marked bias towards military sites and urban centres including civitas capitals (Eckardt 2014: 184–186), suggests administrative control over the conquered areas. Andrews (2012: 63) hypothesises that the high numbers of seal boxes from the Icknield Way area in Norfolk may reflect post-Boudican confiscation and reallocation of lands. He states that ‘almost complete native depopulation’ was followed by ‘repopulation by officially sanctioned incomers, such as auxiliary and legionary veterans’. Alternatively, I found that seal boxes were widespread at many rural as well as urban settlements (Figure 15). There is little evidence for the claimed discontinuity and repopulation. Few settlements reveal signs of deliberate destruction or a hiatus in activity in the post-revolt period. The distribution of finds appears very similar on both sides of the Boudican revolt. Without the Classical sources, it would be very difficult to tell the Late Iron Age and early Romano-British periods apart, in terms of material culture and settlement patterns.

Figure 15 

Distribution of seal boxes (Source: Author).

One interesting, possibly hybrid, find from the civitas capital is an enamelled leaf-shaped seal box lid which bears a stylised phallus and crescent motif (Andrews type P4D14) (CRT SF4483; Figure 16). The phallus is not in itself a rare design on seal boxes, but they are always separately cast and riveted attachments (Andrews 2012: 13). Many of the designs on seal box lids may have been intended to avert the ‘evil eye’, either from the private contents of a letter, or to protect valuables from theft. Andrews (2012: 102–103) considers both lunula and phallus powerfully apotropaic.

Figure 16 

Seal box lid with enamelled crescent and phallus, Venta Icenorum, Norfolk (CRT SF4483, courtesy Jenny Press, Caistor Roman Project).

Phallic symbolism is seemingly absent in Iron Age Britain (Plouviez 2005: 161), but had a long tradition in Roman culture, where it was not directly sexual, instead inviting good fortune and averting evil. Small apotropaic phallic pendants were worn by Roman boys (Johns 1996: 108), as well as soldiers, although whether these customs applied to Britain is debateable.9 Swift’s (2017: 168–173) research has shown that rings with phallic motifs were often diminutive and could have been worn by very young children. Plouviez (2005: 161–162) studied phallic amulets from Suffolk, noting a connection between phallic objects and minor towns, with examples from Icklingham, Coddenham, and Pakenham. Small towns may have played a role in the introduction and distribution of new material culture and practices in the civitas.

The crescent is a motif frequently found on Icenian coins and other metalwork and may represent lunar symbolism, associated with the feminine and protection (Creighton 1995; Johns 1996: 143; Davies 2014). The combination of symbols on this artefact is unusual as it hints at a convergence, or even parity, between male and female elements, as well as blending indigenous and incoming protective symbolism. Its discovery at the civitas capital may signify the spread of literacy and the language of conquest into the Icenian territory. The Iceni, like other Britons, were powerless to resist literacy as both a useful skill and a way to display authority. Perhaps the mingling of the iconography, in this instance, was a way of putting their own ‘seal’ upon it.

Conclusion: Is Resistance Useless?

My archaeological approach with its focus on portable artefacts has the advantage of looking at object patterning across a wide area, not simply the keyhole view of excavation. These predominantly surface finds can add a different perspective to the narratives derived from written sources or known settlements, towns, and forts. Studying a 300 year period also affords the potential to take the long view of cultural contact and change. An object-centred method allows the intricate processes by which new ideas and things arrive, take hold, and become commonplace to be explored.

Novel technologies, fashions, and social structures can all leave their marks on the archaeological record. The charges of cultural backwardness or rebellious difference are often levelled at the Iceni as a foil to other southern British peoples as early adopters or ‘neophiles’. There has been a tendency to fall back on the paradigm of the material elite and a teleological view in which neophilia is seen as positive and evolutionary, whereas neophobia is a trait which stifles growth and creativity. Carr (2006: 15–16) questions whether the persistence of earlier practices and artefacts could be evidence of poverty: an inability to acquire new things, rather than deliberate resistance. This does not seem likely given the abundance of finds from my study area.

‘Tribal elites’ are frequently used to explain the development of emergent aristocracies in the Late Iron Age, which maintained power through the control of rare and imported products (Millett 1990: 38, 68). My study of the widespread, high-quality, horse harness presents an interpretation against hierarchical social organisation. However, if we are to discuss status and elites we must be much more explicit about the potential alternatives and the limits of our knowledge. The ostentatious decoration of horses and vehicles in the client-kingdom signalled their importance and relevance to earlier, more communal forms of wealth and status. This may indeed demonstrate an overt, non-violent, form of cultural resistance, and also challenges the idea of a hiatus in the post-revolt period. Production and deposition of horse gear and brooches suggest strong regional connections, which cross social and geographical boundaries. This raises the need for further investigation of the idea of transient metalworkers and the processes of manufacture and trade in this period. My discussion of the two brooch variants proposes that intra-regional patterning may derive from workshops or fashions rather than resistance through the badging of identities. This breaks the straightforward link between material culture and ethnic identity and contests outmoded concepts of ‘tribal’ boundaries.

The construction of the Iceni as an ‘Other’ does not stand up to much scrutiny. Instead, the case studies of pipeclay figurines and seal boxes challenge the stereotype of the Iceni as culturally conservative and backwards-looking. These objects can inform our understanding of the processes of selection and strategic adoption of new techniques, practices, and material culture. Both were sometimes deposited in ritual contexts, whether at temples or in burials. The avoidance of certain figurine types may relate to the persistence of local burial customs, or the strength of earlier cosmologies and religious preoccupations. This could be seen as resistance, but it is not a blanket rejection of all imports or ideas. In the case of seal boxes, the clearly advantageous practices of literacy and mark-making were adopted no more slowly in Icenia than other parts of Britain. The evidence here is for a lack of resistance, twinned with hybridity, suggesting that people were actively negotiating how these objects fitted into indigenous symbolism and lifestyles.

There are undoubtedly links between objects, social boundaries, and power. Manifest in these are people’s varied responses to change, from discrimination and resistance to acceptance and desirability. Imported goods were treated differently for a considerable period after the conquest and locally made versions were frequently hybridised. This is also evident in religious behaviour which transferred from natural places in the landscape to more formalised temples and shrines. The end of the Icenian client-kingdom brought new deities and changing patterns of ritual practice, as well as increased literacy. In the mid-Roman period, stylistic, material, and formal changes in material culture show a shift from diversity and fragmentation towards greater similarity and integration. By the third century AD, the bloody uprising a distant memory, another phase begins. Innovation and radical change slow and many items of display discussed above go out of use, for currently unknown reasons. The practice of hoarding carries through into the Late Roman period, albeit with different artefacts, such as pewter plate and Christian symbolism (Moore et al. 1988: 73, 82).

At the risk of describing ‘Romanisation by any other name’, I would argue that these developments are related to the increased connectivity which occurs at times of contact, with its attendant access to new resources, styles, and concepts. This does not presuppose Roman cultural superiority, but rather demonstrates the willingness of the Iceni to engage with and appropriate these novelties. The processes of adjustment, acceptance, appropriation, and adaptation happened over generations.

So, was resistance beneficial to the inhabitants of the Iceni territory? Their confrontations with Imperial might were ultimately unsuccessful. Yet, there is no real show of advanced militarisation of the polity, unlike on the northern frontier. Perhaps the evidence of post-revolt continuity presented here implies that the remaining insurgents simply returned to their rural lifestyles. Icenian resistance, both physical and symbolic, may have resulted in a subsequent century or two of peace, in which complex displays of identity and affiliation became less crucial as their ‘political and economic advantages’ waned (Wells 1999: 197–198). Resistance was not a reaction to a unilateral process of ‘Romanisation’, or a two-way dialogue between elites, but a multivocal discourse which reflected the claims and identities of all sides in the battle for Britannia.

Additional File

The additional file for this article can be found as follows:

Supplementary Table 1

Resistance is Useless Dataset: Object data tables used to generate maps discussed in this article (MS Excel). DOI: