Internet Archaeol. 17. Willis. Examining the distribution of samian across townscapes

10.3 Samian and intra-site analysis: examining the distribution of samian across townscapes

It should not be surprising that different areas of sites produce differing finds assemblages. This aspect reflects the rich diversity in the archaeological record for the period, which presents us with considerable opportunities to explore the cultural life and organisation of settlements. Variations in finds assemblages from area to area mean that one should be cautious before generalising, especially if this is on the basis of a single excavated area, as Taylor has pointed out in the case of Carlisle (cf. Taylor 1990, 199). Some 'case studies' highlight significant aspects of the relationship between samian and spatial factors.

Canterbury and Silchester
Not infrequently the differences in finds between various locations across a site can be considerable. It was observed in Section 7.3.7 that two groups from different interventions at Canterbury had remarkably low proportions of decorated samian ware (Table 35), unlikely to be representative of the levels for Canterbury generally. The explanation for these conspicuously low proportions would seem to lie in their location which, at this time, was on the margins of the centre and associated with agricultural activity and evidently low-status occupation. In the case of the defences sites at Silchester, Fulford suggested a similar explanation for the fall in the presence of imported pottery, including samian, following the Conquest period.

'Imported pottery is considerably less common after the Flavian period in the groups published here. It has to be remembered that all the present groups come from the edge of the Roman settlement. Lack of imports may be a reflection of comparative poverty and/or the non-domestic nature of activities carried out in these areas. Such an interpretation would suggest some shift in the location of different types of activities and residence between the pre-conquest and Claudio-Neronian period and the Flavian and later periods. The highest figures for imports occurred in the cultivated-soil build up ([Group] 2.7) beneath the 1974 rampart section ...' (Fulford 1984, 123).
Lincoln
In a pathfinding study, Darling (1998) has examined the nature of samian assemblages from various excavations conducted in Lincoln. She demonstrated that there were both consistencies and differences between Lincoln sites at various levels (as one might expect). A number of problems impacting on this type of approach were noted by Darling (1998), such as variation in sample sizes from different excavation areas, redeposition of deposits in the mid-late Roman period, and residuality. The nature of the archaeology at Lincoln, particularly site-formation processes, made the latter two factors particularly challenging. Darling's findings, however, present a nuanced picture of samian consumption in various areas of the city. One aspect identified was that, 'The percentages of decorated wares vary considerably between sites' (1998, 172); in the case of South Gaulish ware, approximately contemporary with the fortress era at Lincoln, decorated samian was found to be more common in the extra-mural area at Wigford than in the area of the fortress (Darling 1998, 171 and fig. 3). The latter finding is consistent with the pattern noted elsewhere (Section 7.3) for extra-mural settlements outside military installations to have markedly higher proportions of decorated samian bowls than areas within the fortresses/forts. (Darling notes, though, that some deposits from the extra-mural suburb at Wigford might be ground make-up dumps from other areas of Lincoln (1998, 173)). Later in the 2nd century, during the era of the colonia, the balance alters, with proportionally more decorated ware being consumed in the city area as opposed to the Wigford suburb (1998, 173-6). Darling argues that this is likely to be an indicator of the relative high status of the city area during the 2nd century, when Wigford is mainly characterised by utilitarian strip buildings fronting on to Ermine Street.
London
Numerous excavations have, of course, been conducted in London, across the geography of the Roman city. A wealth of finds data has been forthcoming. Studies of samian from London have an excellent pedigree. Much valuable synthetic work has been undertaken (e.g. Davies 1993), most recently by Gwladys Monteil who is, among other aspects, researching spatial aspects of samian incidence. Variations in the character of samian groups from various locations in London have been discussed as part of the present project (Willis 1998a, 89; Section 7.2.4).

Pottery finds from a series of buildings dating to the early Roman period at Leadenhall Court, by the centre of the city, were examined by Groves using EVE in an attempt to discern the status and functions of the buildings (1993). In contrast to the procedure at Shepton Mallet, the finds were considered at the level of the building, rather than of individual rooms (due to the modest size of the samples). The survey showed that among the 13 groups associated with the use of the buildings, samian generally formed between 16-26% of the pottery by EVE; two groups registered lower frequencies at 7%, and one higher at 30%, though in all three of these cases outside the normal range, small sample size is thought to have produced distorted figures (Groves 1993, 130-2). The general range of between 16-26% of the groups comprising samian is, in the round, similar, if not somewhat higher, than the levels of samian registered by EVE method elsewhere (e.g. Table 28). This is higher than the average percentage for samian present in groups of the Flavian period from London (by EVE), which according to Groves is 13% (1993, 134). The groups, overall, showed a considerable degree of ceramic homogeneity, with deviations relating to small sample sizes (1993, 133). Groves interpreted these as domestic assemblages and suggested potential status differences between the assemblages based on the samian and other pottery evidence, which presumably reflected status differences between households (1993, 134; cf. Davies 1993, 143). The evidence of the samian, together with other indicators, points to a residential zone.
Heterogeneity in samian distribution
Sites have their own ceramic identities (cf. Going 1987; Monaghan 1998). Often, too, their samian assemblages have threads of distinctiveness resulting from the nature of supply, and consumption trends. Yet the composition of samian assemblages from interventions at larger sites, when compared, typically varies from one excavation to another in terms of the samian manufacturers represented. Dickinson, for instance, notes in the case of the Causeway Lane and Shires sites, Leicester, that, '... while there are some similarities between the two assemblages, they do not have many potters in common, and even fewer dies ...' (Dickinson 1999d, 104). This is a reminder of the fact that samian vessels from specific workshops were widely diffused once they had left the workshop. Even groups associated with likely pottery shops trading samian show only modest numbers of 'batches' from particular workshops (cf. Dickinson and Hartley 2000). Indeed, lists of samian potter's stamps from excavations are an index of the heterogeneity of samian product consumption, in terms of who made this pottery. Of course, this testimony to the nature of samian distribution is seen at all types of site. Hence we should not expect lists of samian producers represented among samian assemblages to be similar even when these come from adjacent areas of forts and towns.