A far-reaching material impact of Roman expansion in northwest Europe was the spread of standardised pottery vessels, supplementing a hodgepodge of handmade and wheelthrown wares from different local traditions. This phenomenon is routinely addressed in big picture scholarship focusing on a single category of pottery, the standardised fineware par excellence, terra sigillata (e.g. Millett 1990, 123–6; Woolf 1998, 185–205; Wallace-Hadrill 2008, 407–21; cf. Van Oyen 2016a). Despite the frequency of claims made about the social dispersal of terra sigillata, its early distribution was often concentrated around Roman military supply routes and newly founded urban centres. Many societies in the new province of Gallia Belgica (as well as large parts of southern Britain and adjacent areas) were more likely to encounter standardised vessels in the guise of so-called Gallo-Belgic wares. This reality is attested in the substantial quantities of Gallo-Belgic wares in indigenous settlements and cemeteries in northern Gaul and southern Britain, from ca. 25 BC to the late first century AD. The aim of this chapter is to explore the proliferation of Gallo-Belgic finewares as a phenomenon – their occurrences and combinations in different kinds of cultural context, their participation in social practices, and the degree of any relationship between the style of vessels and their itineraries (Hahn and Weiss 2013b). In other words, what did Gallo-Belgic wares do in the early Roman West?
‘Gallo-Belgic wares’ is an awkward umbrella term used to describe a broad repertoire of mostly fineware pottery produced in Gallia Belgica and adjoining regions, ca. 25 BC to AD 85. The most common Gallo-Belgic fabrics are orange-red terra rubra, fired in oxidising conditions approaching those of the production of terra sigillata, which it imitates in appearance, and black terra nigra, fired in a reducing atmosphere favoured in the production of pre-Roman black-gloss wares (Wightman 1985, 144) (see Fig. 4.1 for some common examples). Despite impressions of the apparent separation between ‘Romanising’ terra rubra and ‘native’ terra nigra, this division is largely arbitrary: there is great similarity and overlap in the shapes of vessels produced in red and black, their distribution and chronologies are largely complementary, and vessels in both fabrics are routinely deposited together in settlement and funerary contexts. While many Gallo-Belgic vessels may be seen to derive directly from terra sigillata forms, a roughly equivalent number of types have a more diagnostic north-western European genealogy, or indeed, are innovations of the period and the fusion of styles broadly associated with temperate and Mediterranean Europe.
Gallo-Belgic wares are typically viewed as ‘second-class’ finewares compared to sigillata, both in terms of their perceived importance to archaeologists and their perceived significance to ancient users – a view that is doubly problematic. Like sigillata, Gallo-Belgic wares were produced in a series of highly standardised shapes at multiple production centres and kilns, with manufacturing foci located near urban centres such as Reims, Tongeren, Trier, Metz, Cologne and Nijmegen (Deru 1996, 263–317). While long-held wisdom links the origins of Gallo-Belgic production with the requirements of military supply (e.g. Wightman 1985, 144), by the early-mid first century AD Gallo-Belgic wares merely supplemented more extensive supplies of terra sigillata for many military communities (Pitts 2014). In contrast, the impact of Gallo-Belgic wares was arguably greater among civilian communities in northern Gaul and southern Britain, for whom it frequently outstrips quantities of sigillata by a considerable margin. For example, while the ceramic assemblages from the Augustan-Neronian graves at the cemeteries of Feulen, Luxembourg (Schendzielorz 2006) and Wederath-Belginum, Rhineland-Pfalz (Haffner 1971; 1974; 1978; Cordie-Hackenberg and Haffner 1991; 1997) are composed of 40–60% Gallo-Belgic wares, the same assemblages feature less than 3% of sigillata. Towards the end of the first century, a tendency for increasing civilian supplies of terra sigillata to be viewed differently from Gallo-Belgic wares can be seen at cemeteries such as Septfontaines-Dëckt (Luxembourg), where sigillata is more frequently included separately on the pyre rather than in the grave (Polfer 2000).
Figure 4.1 A selection of Gallo-Belgic pottery types from Britain. Codes are from the Deru (1996) type-series.
The civilian predilection for Gallo-Belgic wares – whether by accident of supply, deliberate marketing, or cultural selection – ought to make them a priority for study owing to their potential to shed new light on material change among indigenous societies. Such potential is nevertheless constrained by the largely representational ways in which Gallo-Belgic wares have been treated so far, with some examples outlined in the next section. In the first place, their marginalisation from the study of terra sigillata reinforces a view of history in which the perceived products of the ‘core’ metropolitan civilisation are valued more than those of the ‘periphery’. Secondly, at an interpretive level, Gallo-Belgic wares are poorly served by prevailing thinking that has tended to view them as ‘proxies’, typically for certain kinds of economic activity (e.g. the imperial economy) or differentiated cultural practices (evidence for Romanisation).
A good example of the partiality of representational thinking concerns the interpretation of Gallo-Belgic wares at the site of Sheepen, Camulodunum, the oppidum and royal centre of the Catuvellaunian dynasty, and major stronghold of resistance to the Claudian conquest of southern Britain in AD 43. Writing on the extensive 1930s excavations at the site, Hawkes and Hull (1947, 49–50) surmised the following about the impact of (mainly Gallo-Belgic) pottery imports on the pre-conquest community:
The sheer material superiority of the imported pottery must have made for greater refinement, and in addition, there is its clear-cut specialization of form. This is of course no mere typological parade: it implies specialization of use, connected mainly with food and drink. The wine imported in the amphorae we have found may have been primarily for the rich who would drink it from the appropriate Arretine cups, but the wider popularity of Gallo-Belgic cups, less costly but still specialized for wine as no doubt were beakers for beer, proves a more general diffusion of civilized manners. Such a diffusion is still better attested by the abundance of platters, off which men had never been accustomed to eat in Britain before.
Despite recognising the social roles of Gallo-Belgic wares in eating and drinking, for Hawkes and Hull the appearance of this pottery reified Romanisation as a civilising process – the diffusion of manners from centre to periphery. Cups and platters were something new, and are therefore assumed to have participated in practices diffused from the distant continent – in this example Gallo-Belgic wares stood for the trope of the diffusion of civilised manners.
By the time the next major excavations at Camulodunum were published in the 1980s, the social and intellectual climate of British archaeology had changed. Despite the rigour of their excavations, the diffusionist perspective of Hawkes and Hull had been discredited, and economic perspectives came to take precedence over cultural concerns. Niblett’s (1985, 25) interpretation of the site, which essentially considers the same artefactual evidence as Hawkes and Hull, reaches a very different conclusion:
One of the most interesting aspects of the contents of the pits, however, is the light they throw on the life of the inhabitants of Sheepen. Samian, fine glass, Gallo-Belgic wares, and amphoras were abundant… These argue for a fairly high standard of living, certainly much higher than that apparent from the small huts found along the line of the filled-in Sheepen dyke in the 1930s. It was argued then that the inhabitants of Sheepen were downtrodden and enslaved native corvées forced to work supplying the new colonia (Tacitus, Annales, xiv, 31; Hawkes and Hull 1947, 38). Although such an element presumably existed in the Sheepen population, the impression given by the 1970 excavation is rather one of independent craftsmen successfully exploiting the opportunities provided by the new market.
Where the craftsmen originate is a matter for speculation but it is interesting to note that fragments of 135 amphoras from the 1970 excavations indicate the use of Italian and provincial wine, fish sauce, and nearly 1400 litres of olive oil. The large number of such vessels compared with what is normally found on native sites gives a picture of a cuisine that was much more Roman than British and may well imply the presence of immigrants from abroad.
While Niblett’s account maintains the logic that new pottery vessels represented the spread of civilised Roman eating habits, here the presence of Gallo-Belgic wares stood alongside other imported wares to represent entrepreneurial immigrants, a far cry from the uncouth ‘natives’ identified by Hawkes and Hull. But as with that of Hawkes and Hull (1947), this interpretation seems to be more reliant on the interpretive packaging of seemingly like artefact classes (the imports) together, rather than scrutinising them more closely as independent evidence.
More recent studies that compared contemporary patterns of pottery supply and deposition at the Sheepen site at Camulodunum with those of the fortress and colonia (ca. AD 44–61) highlight the fragility of both the 1940s and 1980s interpretations of Gallo-Belgic wares (Bidwell 1999, 488–93; Pitts and Perring 2006; Pitts 2010a; 2010b; Perring and Pitts 2013, 232–8). These studies show that after conquest, the pre-existing communities at Camulodunum and its environs continued to use ‘suites’ of Gallo-Belgic wares in traditional forms of social display evident in funerary contexts and feasting assemblages, often alongside increasingly prevalent terra sigillata vessels. In contrast, the veteran colony established at Colchester in AD 49 (Colonia Claudia Victricensis) eschewed most Gallo-Belgic vessels completely, receiving only a trickle compared to its huge supply of sigillata. Being primarily interested in what artefacts might reveal about social and cultural identities, I interpreted these patterns in terms of cultural differences between the local community and the incoming colonists. While this effectively re-stated the position of Hawkes and Hull (1947), it emphasised the agency of Sheepen’s community to distance themselves from the Roman colonists through their selection and use of Gallo-Belgic wares, rather than assuming that all imported wares represented a linear diffusion of civilised values from the continent. Gallo-Belgic wares had now come to represent indigenous cultural practices.
While I stand by my previous interpretations of the pottery from Colchester, I concede that the conclusions I drew are partial in certain key respects. This is in large part because I was interested in what the pottery might reveal about the elaboration of group identities, perhaps at the expense of a more thorough understanding of pottery as material culture. But this problem is by no means confined to the recent spate of identity studies in Roman archaeology. Indeed, commenting on the same patterns, Fulford (2009, 255) preferred to place the distribution of Gallo-Belgic wares in a wider economic context – as ‘a competitive, entrepreneurial mode of distribution and consumption’ running ‘alongside a system driven and paid for by the state’, a valid – if similarly partial – interpretation using representational logic. In this case, Gallo-Belgic wares became proxies for non-state economic networks.
So far, this introduction to the archaeology of Gallo-Belgic wares in British contexts has emphasised the pitfalls of representational modes of interpretation. The Gallo-Belgic assemblage from just one site has variously stood for the diffusion of civilised manners (Hawkes and Hull), entrepreneurial immigrants (Niblett), indigenous cultural practices (Pitts and Perring), non-state modes of distribution (Fulford), as well as offerings by religious pilgrims (Willis 2007, 121–2). It is clear that the overwhelming urge to ‘people the past’ and to use objects to stand for social categories (or more recently, social practices and identities) has obscured more nuanced ways of interpreting the evidence. As I demonstrate below, taking a less overtly representational approach to the same class of artefacts not only reveals richer patterning, but also provides a valuable new perspective on the same historical scenario at early Roman Colchester.
In contrast to the approaches discussed above, a rather different perspective on Gallo-Belgic wares is provided by the prehistorian Chris Gosden in his article ‘What do objects want?’ (Gosden 2005, 205–7). The starting point for Gosden is not to pursue abstract notions of the Roman economy or identity formation, and nor is it to uncover the definitive meaning of Gallo-Belgic pottery. Instead, Gosden focuses on the material properties of Gallo-Belgic ware, most notably its eclectic stylistic genealogy (drawing from traditions across the Mediterranean and Gaul), its genealogical impacts in influencing local pottery (in Britain), and its effect on human users. In terms of archaeological material culture, ‘genealogy’ refers to specific material traits in an object or group of objects that can be seen to derive directly from similar traits in an identified group of chronologically preceding objects. For Gosden (2005, 206–7), Gallo-Belgic ware is not simply a passive indicator of culture or society, but rather a historical agent in its own right:
The ability of different materials, shapes and finishes to impose themselves on people shows the true promiscuity of pottery which acted as a nexus for influences coming from many parts of the object world…
Gosden’s discussion of Gallo-Belgic pottery is regrettably brief and does not provide archaeological examples to develop his approach further. Nevertheless, it does provide elements of a blueprint for the non-representational analysis of Gallo-Belgic wares – a focus on pottery’s material properties, notions of stylistic genealogy and lineage in standardised pottery repertoires, and the capacity of mobile objects to influence the adoption of similar stylistic traits in more local repertoires (cf. Biddulph 2013 on memes in the reproduction and imitation of terra sigillata). With these broad concepts in mind, the following sections aim to explore what Gallo-Belgic wares did in early Roman northwest Europe, starting with regional variations in the make-up of assemblages in terms of vessel shape, and considering the localised impacts of vessel genealogy.
The rest of this chapter is based on the analysis of Gallo-Belgic assemblages drawn from across the breadth of northwest Europe, from the interior of Britain to the Rhine and Moselle. This is essentially the same dataset as that of a recent study that contextualised the early fineware assemblages of the first Roman cities in Britain (Pitts 2014), in which most continental data was drawn from Deru’s (1996) corpus (see Fig. 4.2). To ensure a more holistic comparison of assemblages across the wider region, the pottery vessels are re-classified here in terms of the Deru (1996) type-series as opposed to the Anglo-centric Camulodunum series (Hawkes and Hull 1947), which enables the comparison of extra sites, such as Tongeren and Liberchies. This area corresponds with the core region in which Gallo-Belgic wares circulated before ca. AD 70, and as such provides vital information on Gallo-Belgic wares as a broader canon of style (cf. Gosden 2005). Only by comparing a representative sample of locations in which Gallo-Belgic wares circulated is it possible to properly appreciate its local and global impact.
To gain an impression of the pan-regional and local trajectories of Gallo-Belgic wares, I use the concept of object-scapes to refer to repertoires of objects at hand in given localities in particular historical moments (Versluys, this volume; Pitts and Versluys forthcoming). In analytical terms, the concept can help to characterise recurrent combinations and selections of objects en masse at a variety of scales, from micro/local to macro/global. Analogous to the notion of ‘relational constellations’ (Van Oyen 2016b), thinking in terms of object-scapes can help place the relationality of material culture at the centre of analysis (e.g. through object-object relations at different analytical scales), without recourse to representational logic.
To get an impression of what Gallo-Belgic ware object-scapes looked like in early Roman northwest Europe, Fig. 4.3 considers them in terms of the proportions of six major fineware shapes (beakers, butt-beakers and their derivatives, bowls, cups, dishes and platters) at over 20 archaeological sites. It is worth bearing in mind at the outset that this broad-brush analysis simplifies complex chronological and stylistic variation, and glosses over vital information on depositional context. Nevertheless, it gives a basic picture of the geography of Gallo-Belgic wares that may be further contextualised.
The most obvious pattern in Fig. 4.3 is the apparent similarity of assemblages from civilian sites in Britain (upper-tier) and Gallia Belgica (middle-tier). The majority of these assemblages are dominated by platters (e.g. A5, Fig. 4.1) and butt-beakers (e.g. P21, Fig. 4.1). Looking closer, the similarities in supply may be broken down into regional groupings (i.e. southern Britain, eastern Britain, western Belgica, and eastern Belgica), as well as along axes of connectivity (the Somme–Amiens and southern Britain, and the Moselle–Rhine–Thames – eastern Belgica and eastern Britain). At this point, there are a number of routine presumptions that might be invoked in interpreting such patterns. In economic terms, the supply of what might be described as the primary market (Belgica) looks very similar to that of the secondary market (Britain), suggesting blanket supply or even uniformity of demand. Likewise, the smaller regional variations may be explained by the peculiarities of local production and minor variations in regional shipments.
In contrast to the patterns just discussed, the Gallo-Belgic ware assemblages from this seemingly eclectic group of sites in the lower-tier of Fig. 4.3 could not be more different. This creates problems for a model of blanket supply that might be proposed from looking at the civilian centres alone. The sites in the lower-tier received very different combinations of Gallo-Belgic vessels in quantitative terms. Unlike those from civilian centres across the Channel and to the south, the assemblages from Tongeren and Liberchies are more heavily dominated by butt-beakers, whereas those from Nijmegen (the Hatert cemetery) and Cologne feature an emphasis on different beakers and bowls, each case suggesting a local predilection for particular vessels. In contrast, the Gallo-Belgic assemblages from known military (Exeter), colonial (Colonia Victricensis) and suspected colonial (London) centres in Britain share a similar dominance of dishes, and a virtual absence of butt-beakers. But the exception that really proves the rule is the idiosyncratic assemblage from Usk, which eschews both military and civilian templates of supply for Britain, displaying closest affinity to Cologne (in this sample). This last pattern is surely the result of the historically attested movement of Legion XX from Neuss (a military base just north of Cologne) to Usk, and adds a new dimension to the interpretation of the big picture, with object-scapes from the Rhine being transplanted to south Wales.
Figure 4.2 Map showing the location of selected cities and sites. Drawn by A. Montesanti.
So far, a description of Gallo-Belgic ware object-scapes in terms of vessel shape in Fig. 4.3 can be summarised as follows: rough homogeneity across the majority of civilian Belgica and southern Britain, with more distinctive combinations of vessels in northeast Belgica and in association with military foci in Britain and the Rhine. What can this crude overview of proportions of basic shapes contribute in historical terms? Returning to the example of Colchester and its various interpretations, the wider context provided by Fig. 4.3 suggests that the contrasting supply and deposition of Gallo-Belgic wares at Sheepen (Camulodunum) compared with Colonia Claudia Victricensis were not simply a ‘native’ reaction to Roman colonial domination. Instead, the assemblage at Sheepen fits a typical profile of civilian object-scapes on either side of the Channel, whereas the material from Colonia Victricensis closely follows the pattern of object-scapes from military sites in Britain (Exeter) and those of likely official origin (London) (Perring 2011; Pitts 2014; contra Wallace 2014).
However, the main headline from the general picture is that consistently different yet replicated configurations of Gallo-Belgic ware object-scapes mattered, possibly contributing to what Greg Woolf described as ‘styles of consumption’ (Woolf 1998, 176). This is acutely illustrated by the transplanting of combinations of Gallo-Belgic vessels from the Rhine to Usk in south Wales – probably carried by soldiers over great distance rather than maintained in later provisions – and more generally in the distinctive regional character of assemblages around northeast Belgica, the Rhine, and for military and colonial communities in Britain.
For communities in Gallia Belgica, particularly those close to major Gallo-Belgic ware production, for example, at Reims and Trier, it may be safe to assume that the combinations of different vessel shapes effectively represented local demand. However, the apparent similarity in configurations of Gallo-Belgic object-scapes in southern Britain is less easy to explain. One option is that the British communities effectively saw themselves as part of the same cultural milieu as those in Belgica, as attested by various strands of evidence – written sources, funerary practice, coinage, tribal names etc. It is possible that the similarity of Gallo-Belgic ware object-scapes might represent degrees of (fossilised) shared cultural practice, involving the use of finewares in the spheres of domestic consumption and funerary practice. At face value, the case for the patterns belying some form of cultural choice is also reinforced by the massive impact of Gallo-Belgic style on local pottery production, a point not lost on Gosden (2005).
Another way of explaining the apparent homogeneity in Gallo-Belgic ware object-scapes in civilian Britain and Belgica might be to cast such patterns in terms of the over-supply of the civilian market in Belgica, and the relative disinterest of military communities in Britain in Gallo-Belgic wares (in contrast to terra sigillata). This sort of economic explanation is frequently encountered in Roman pottery studies and shifts the agency and choice to a different social group – the military, who are often implicitly seen as exerting more agency than indigenous communities due to their representational connections (real and imagined) with Rome. Leaving aside these overtones of centre-periphery logic, it is possible that the combinations of Gallo-Belgic vessels on civilian sites in southern Britain did have limited cultural significance, as surplus (and therefore much cheaper) products that would have been effectively dumped on uncaring recipients. In this scenario, combinations of vessels that might have had particular cultural significance in Belgica would have been completely lost on British users, who may have instead associated Gallo-Belgic wares with the general changes attendant on Roman annexation (in line with Niblett 1985).
To address these competing hypotheses to explain similar configurations of Gallo-Belgic ware object-scapes in civilian Britain and Belgica, it is necessary to take a closer look at the assemblages in question, returning to the theme of what the objects did. As has been pointed out, in order to visualise patterns across such a wide area, the data in Fig. 4.3 are summarised in a way that masks a great deal of the variation in the style and appearance of the pottery. Taking up some of the themes discussed in Gosden’s (2005) consideration of Gallo-Belgic wares, the following sections return to the topic of object genealogies, focusing on the itineraries of styles of vessels of broadly Gallic genealogy (butt-beakers and their derivatives), as well as those deriving from Mediterranean inspiration (especially platters).
Figure 4.3 The relative proportions of Gallo-Belgic tablewares at selected sites in NW Europe. Top tier: sites in Britannia with pre-conquest origins; middle tier: civilian centres in Gallia Belgica; bottom tier: military/colonial sites in Britain and sites in NE Belgica.
A strong pattern in Fig. 4.3 is the overwhelming civilian popularity of butt-beakers and related forms like the grätenbecher and girth-beaker (P21, P23 and P29 respectively in Fig. 4.1). While I have characterised the butt-beaker as a Gallic design, its genealogy of influences is complex. Hawkes and Hull (1947, 237) describe the general vessel form as a ‘late La Tène and early Provincial adaptation of pre-Imperial Mediterranean prototypes.’ One source of likely genealogical inspiration for this design are the tall jar vessels of the La Tène D1a (later second century BC) that occur in cemeteries such as Wederath (e.g. Haffner 1971, grave 300). However, other influences are possible. It is notable that some of the earliest dated occurrences of butt-beakers, such as in the princely graves at Goeblange-Nospelt (Luxembourg, ca. 30–15 BC) (Metzler and Gaing 2009) are contextually associated with Italian thin-walled ware ACO beakers, a vessel which bears a striking typological resemblance to some early butt-beakers in terms of shape, size and decoration. While the ACO beaker, a.k.a. Conspectus R12, was produced in sigillata and non-slipped forms in northern Italy as well as enjoying short-lived production at Lyon (Ettlinger et al. 1990, 182), it had limited distribution and longevity in northwest Europe, and is more typically associated with early military sites. There are also no long-lasting direct equivalents of the butt-beaker in the repertoires of either Italian or Gallic sigillata. Therefore, while Gallic and Italian influence may have been factors in the initial genesis of the butt-beaker, the phenomenon it became was an unquestionable innovation of northwest Europe.
Figure 4.4 The relative proportions of the most common butt-beaker (and related beaker) types at civilian sites in Britannia and Gallia Belgica, using the Deru (1996) type-series.
What did the butt-beaker and its related forms do in early Roman Europe? Fig. 4.4 breaks down the Gallo-Belgic butt-beaker category according to its most common circulating types, according to Deru’s type-series (1996). It is worth pointing that Fig. 4.4 concentrates almost exclusively on civilian sites, owing to the relative scarcity of butt-beakers in military contexts in this sample – a significant finding in its own right. The patterning within Fig. 4.4, however, hints at a much more complex story. While the supply of British sites in the upper-tier appears largely homogenous (as with Gallo-Belgic vessels in general from Fig. 4.3), Gallia Belgica is characterised by pronounced regional differences in combinations of butt-beaker types.
To attempt to make sense of Fig. 4.4, I begin by focusing on the itineraries of individual butt-beaker types. It is clear that while the butt-beaker was in some sense a universal form, styles and types of butt-beaker varied widely across the Roman northwest. While some forms, such as the P6/7 (Cam 112), P23 (Cam 114) and P29 (Cam 82/4) occurred in modest quantities at most sites, others have more pronounced regional distributions. For example, the P1 butt-beaker completely dominates assemblages in the territory of the Treveri (Metz, Dalheim and Trier), but does not appear at any of the other centres in particular quantity (apart from Reims, located immediately to the west). The same appears to hold for the P10/11 and P13/17/18 with Tongeren and northern Belgica, and the P20/21 with southern Britain and Amiens.
Why did some butt-beakers dominate particular object-scapes but fail to travel longer-distances, while other Gallo-Belgic vessels achieved much more widespread distribution? Why did other butt-beakers achieve more universal distributions in the civilian sphere without dominating in any single region? Standard representational logic might account for these patterns in terms of market economics or the expression of regional identities. However, as this book stresses, viewing artefacts only in terms of their capacity to stand for social categories or economic processes is intrinsically partial. Instead, focusing on the stylistic and typological properties of the vessels in question reveals more promising lines of inquiry.
An important observation seems to be that the butt-beakers with more pronounced regional distributions also look very different from one another. There is no obvious direct genealogical connection between the P1, P10/13 and P21 types (Fig. 4.5). At the same time, all of these regionally distinct types share clearer similarities with the seemingly universal P6/7 (i.e. in terms of characteristics such as pronounced rim, bands of decoration and vessel profile). In other words, the regionally-specific butt-beaker types appear to be localised innovations that only make sense through their collective reference to a shared universal style maintained by circulating objects with which they share a mutual dependency (cf. Collins, this volume, on material culture as circulating references). While further quantitative research is needed to verify this interpretation, it provides a more holistic explanation than simply considering the patterns in terms of local identities and niche markets, which do not do justice to the global picture. Indeed, it is difficult to envisage individual butt-beaker types as isolated local developments without the existence of a shared cultural milieu.
Closer inspection of the British material in Fig. 4.4 emphatically refutes the idea that Britain was an uncaring ‘secondary market’ for Gallo-Belgic wares. The most notable feature of the British Gallo-Belgic butt-beaker assemblage is the dominance of type P21 (Cam 113), which is only paralleled in quantity in this sample at Amiens (crucially in a different fabric). Both Hawkes and Hull (1947, 238) and Niblett (1985, 23) firmly believed that the Cam 113 beakers found at Camulodunum were not imports, but instead products of local manufacture close to the site. This case is supported by the sheer quantity of vessels at the Sheepen site – over 2,750 individuals, if the figures of Hawkes and Hull are to be believed, the most common fineware vessel at the site by some margin. Although this view has been questioned by Timby (2000, 205), citing the lack of identified kilns in Britain and the similarities with the assemblage at Amiens, the virtual absence of Cam 113 vessels across Gallia Belgica in fabrics and quantities akin to those routinely found in Britain is inexplicable if they were of Gallic manufacture. The Cam 113/P21, being typologically similar to the P6/7 (Cam 112), must be interpreted as an innovation that took place within Britain that directly referenced a larger repertoire of circulating objects.
Figure 4.5 Common Gallo-Belgic butt-beaker types from northwest Europe.
Returning to the bigger picture, the situation at Camulodunum and in southern Britain begins to make more sense. The Gallo-Belgic butt-beaker was an enigmatic design borne of the unique fusion of styles rooted in potting traditions from northern Italy to southern Britain. Not only did it dominate fineware assemblages at civilian centres spanning a wide area that crossed multiple tribal and indeed provincial boundaries, it also spawned major local innovations that often corresponded to the territories of powerful political entities, such as the Treveri (P1), the Tungri/Batavi (P11 and P13), and Catuvellauni (P21). The impact of this new style of object was such that while the same communities would happily rely on imports of other standardised Gallo-Belgic wares, they were compelled to create their own distinctive forms of butt-beakers – even if the technical expertise to do so was lacking amongst local potters. Both Hawkes and Hull (1947) and Timby (2000) agree that the high quality of Cam 113 (P21) vessels at Camulodunum was such that they must have been produced by migrant potters from the continent. If this was the case, it underscores the importance of the butt-beaker within a shared northwest European cultural milieu that was independent from military and colonial communities associated with the Roman state.
The social distribution and subsequent local innovation in butt-beaker design is highly reminiscent of the recent discussion by Van Oyen (2016a; 2016b) of the ‘rootedness’ of Trier Rhenish wares (third century AD). Through their colour, shapes and inscriptions on the vessels, Rhenish wares were created as a local category distinct from the more homogenous and universal terra sigillata, despite sharing many of the same technical characteristics in production. A definition of rootedness as denoting ‘local entanglement as explicit and fundamental to the definition of a thing and its possibilities’ (Van Oyen 2016a, 112) fits both the innovative off-shoots of Gallo-Belgic butt-beakers and Rhenish ware motto beakers very well – accepting that both phenomena likewise depended on the existence of a more homogenous, universal canon of style. It is perhaps no coincidence that both phenomena took place in the same region, and involving beakers ostensibly for alcohol consumption. Indeed, despite the gap of over a century, the general designs of both kinds of vessel are similar and a genealogical connection seems likely. The connection, however tentative, speaks volumes on the long-term historical impact of distinctive objects and styles (cf. Gosden 2006).
The previous section, addressing the trajectories and local innovations of a vessel with broadly ‘Gallic’ genealogy, stressed the likely existence of a shared cultural milieu underpinned by a universal canon of styles of evocative objects (cf. Poblome, this volume, on koine and pottery productions in the Roman East). In contrast, this section examines Gallo-Belgic vessels with a more obvious Mediterranean genealogy. The other most numerous class of Gallo-Belgic wares favoured at civilian sites in southern Britain and Gallia Belgica was the platter – of which some of the more common forms are illustrated in Fig. 4.1 (A5, A17, and A39). Unlike the butt-beaker, most of the platter forms in the Gallo-Belgic repertoire can be seen as imitations or derivations of equivalent forms in Italian-style and south Gaulish terra sigillata. Whereas the butt-beaker was evidently more ‘rooted’ in northwest European styles of pottery production as well as spheres of practice such as communal drinking and funerary display, the platter is often taken to represent the arrival of new cuisine and new individualised habits of consumption (Hawkes and Hull 1947, 49–50; Cool 2006, 164–8).
Figure 4.6 The relative proportions of the most common platter types at sites in Britannia and Gallia Belgica, using the Deru (1996) type-series.
Fig. 4.6 outlines the proportions of the most common circulating Gallo-Belgic platter types at civilian centres in southern Britain and Gallia Belgica. In contrast to the eclectic butt-beaker itineraries in Fig. 4.4, the picture is overwhelmingly uniform. Most platter types achieved wide circulation in the sample of sites considered, and most centres received similar quantities of each type. While closer inspection reveals many of the same regional groupings apparent from Fig. 4.3, such as the similarity of Amiens and central southern Britain, and parts of south-east Belgica with south-east Britain, these are not very pronounced, and are likely to be affected by chronological variations in the floruit of the different sites and their pottery supplies.
Given the striking variation in the global object-scape of Gallo-Belgic butt-beakers (Fig. 4.4), the uniformity among local configurations of Gallo-Belgic platters (Fig. 4.6) is initially puzzling. The absence of equivalently rooted local innovations suggests that the platter, as a general stylistic object category, lacked the same special qualities of the butt-beaker. This is not to downplay the importance of the platter – indeed, it rivalled the butt-beaker in terms of raw numbers of vessels, and was subject to greater stylistic variation with over 50 different types in the Deru (1996) type-series. Unlike the butt-beaker, which was effectively a design unique to the Gallo-Belgic repertoire, the platter closely referenced vessels from a firmly established global object-scape in terra sigillata. The same can be said for other popular Gallo-Belgic forms, notably cups, which likewise have a broadly universal distribution of types across northwest Europe. The uniform and universal itineraries of Gallo-Belgic vessels with Mediterranean genealogy must have mattered – not due to some oversight in production which required them to be dumped wholesale on an unsuspecting ‘secondary market’ – but rather for their capacity to evoke and complement competing suites of terra sigillata.
A combination of rooted (butt-beakers) and universal (platters and cups) styles of vessel ensured that the Gallo-Belgic repertoire was ultimately successful in both cultural and economic terms. In the first place, this was achieved by appealing to the traditional aspects of local practice, involving large and elaborately decorated drinking vessels. Likewise, the standardised vessels of Mediterranean genealogy in the Gallo-Belgic repertoire must have helped bridge the huge cultural gulf apparent with the more exotic and expensive styles of consumption and object-scapes associated with Roman military and colonial communities. These universal shapes would go on to form a major element of provincial pottery assemblages well into the second century AD. This is not, however, either the end of the story or a simple case of the old wine of Romanisation in new skins. Despite the popularity of Mediterranean-derived Gallo-Belgic platter and cup forms in the mid-first century AD, both lacked the same local ‘rootedness’ of butt-beakers. In the longer-term, platters and cups were gradually eclipsed by dishes and beakers in northwest European fineware object-scapes, yet the rooted butt-beaker and its derivatives arguably had much longer lineages. The story of changing pottery assemblages in the Roman northwest seems less one of progressive Romanisation, but rather an eclectic fusion of influences that gradually gave way to regional styles as the Roman world ceased expanding.
Rather than casting consumption patterns of Gallo-Belgic wares as (variously) local reactions to colonial domination, expressions of local identity, or the products of non-state controlled entrepreneurs, I have attempted to sketch a different account of Gallo-Belgic pottery by investigating aspects of their configurations in local object-scapes, genealogical origins, and stylistic innovations. While this attempt at an object-centred discussion by no means avoids the representational logic that has dogged artefactual interpretation in Roman archaeology, I hope that it has opened up some fresh avenues for approaching pottery and artefacts in the Roman northwest. At the very least, there is good reason to think that paying closer attention to what objects did can result in better interpretations that put archaeological data first instead of subordinating them to prevailing assumptions from the outset. If anything, this is the major lesson from the countless failed attempts to find direct archaeological correlates of historical phenomena such as named ethnic groups (cf. Collins, this volume). Rather than the complexity of the archaeological record making such tasks impossible or undesirable, it follows that objects are better understood independently from historical phenomena first, from their production and genealogies to their itinerant pathways and relational associations within broader object-scapes. Indeed, continuing to subordinate the study of material culture to research that treats artefacts representationally risks a kind of methodological circularity in which the framing of questions constrains rather than nurtures the possibilities for new findings.
Taking a more object-centred approach to the study of Gallo-Belgic wares in particular highlights the flaws of previous interpretations that attempted to cast them as a unified category, either with a single universal meaning (old-fashioned blanket Romanisation), or indeed meanings that are contingent entirely upon local context (to caricature nativist and post-colonial approaches). A major implication of asking what standardised objects did entails looking at particular types across a sample of the full extent of their distribution, including multiple modern nations, ancient tribal territories, and provincial divides. Such an approach has significant implications for new interpretations at both global and local levels. This example underlines the inherent flaws of the prevailing tendency in (Roman) archaeology to think in terms of ‘containers’ (Versluys 2015, 143), from individual provinces to regions demarked by modern national boundaries, and even single sites. A connected empire mattered not just for the machinations of a centralised taxation system or the long-distance exchange of commodities like olive oil, but at a more fundamental level of shared cultural imaginations that are most evident in a richly varied global object-scape. To do proper justice to this mélange of innovating styles and objects at a historical level means engaging more closely with ideas of globalisation – both in terms of big-picture cross-cultural analogy (Hodos et al. 2017) and methodological perspectives for coming to grips with objects in motion (Versluys 2014; Pitts and Versluys 2015a, cf. Foster 2006).
Lastly, I return to address a common interpretive challenge in the study of Roman artefacts, especially pottery – the apparent paradox that a distribution pattern must either be explained in terms of local (cultural) choice or the manifestation of a bigger (economic) market systems (see also Poblome, this volume). This is an especially familiar debate in the study of commodities exchanged over long-distances in amphorae, as well as mass-produced finewares like terra sigillata. As the examples in this chapter demonstrate, supply versus choice is an entirely false dichotomy. Indeed, Roman archaeology is poorly served by the arbitrary separation of the study of cultural and economic phenomena, as if one does not matter at all to the other. Mass-produced and standardised objects in motion require both cultural demand and economic imperative to achieve pan-regional distribution – the two are intrinsically connected, as an object-centric approach makes abundantly clear. In the case of Gallo-Belgic wares, neither the replication of configurations of vessels from the Rhineland at Usk (south Wales) nor the general popularity of Gallo-Belgic forms in Britain’s first civilian communities (for example) would have been possible without underlying cultural demand and economic logic. Likewise, the relative failure of Gallo-Belgic wares (especially butt-beakers) to ‘compete’ with terra sigillata in many Roman military and colonial communities can be attributed to an intersection of economic and cultural rationale. In such contexts, the ‘rootedness’ of butt-beakers hand-in-hand with perceptions of inferior pottery favoured by recently conquered provincials may well have made Gallo-Belgic wares less ‘profitable’.
This chapter represents a preliminary attempt to think through the core issues for a larger project, ‘Mass consumption in the early Roman northwest’, supported by study leave from the College of Humanities, University of Exeter (2015–6). I thank Xavier Deru and Rien Polak for their helpful suggestions on earlier presented versions of this research at the AHRC-funded ‘Big Data on the Roman table’ network in Leicester (2015) and Exeter (2016). I am greatly indebted to the comments from Astrid Van Oyen, which have significantly helped to clarify and sharpen the argument. Any errors or omissions remain my own.
* Department of Classics and Ancient History, University of Exeter.
