Internet Archaeol. 17. Willis. Samian and residuality

5.7 Samian and residuality

5.7.1 Residual samian

The presence of residual material culture among recovered site assemblages presents a range of problems. Residual items are pieces of reworked rubbish, including sherds from vessels broken previously, which come to be incorporated into later deposits. Pronounced levels of residuality have been observed at a variety of sites; for instance, at the forum-basilica site, Silchester (Timby 2000, 180-2, figs 101d and 102) and at the rural site at Birdlip Quarry, Gloucestershire (Mudd et al. 1999, 233-4), where, in both cases, samian has been used as a guide to establish the degree of residuality. At Dover, stratified deposits associated with the 'Classis Britannica II fort' covering the period c. AD 130/140-270 contained sherds from 34 South Gaulish La Graufesenque vessels which one might expect to be passing out of use c. AD 100-120, if not before (Philp 1981). These items came from specific spatial areas and were particularly associated with ground make-up deposits of Period I, c. AD 130/140-155. In those areas where pieces of La Graufesenque samian occurred in the earliest deposits there appears to be a tendency for further examples to occur through the sequence, presumably being derived from these earlier deposits; when absent from the earliest deposits they appear to be absent from the rest of the sequence. Hence these items would appear to represent genuinely residual pieces, not survivals in use, but relating to earlier occupation from (presumably) the near vicinity.

5.7.2 Residuality: site and context specific

Residual material is sometimes difficult to recognise, and its presence may distort site dating and archaeological understanding. There has been, perhaps, little advance in comprehending and dealing with this domain with regard to Roman pottery since the paper published by Evans and Millett (1992). One aspect of that study was to use the incidence of samian within site sequences as an index of residuality. These authors concluded that levels of residuality were site and context specific, dependent on the highly variable character and history of deposit formation processes at sites. This is a key point. Levels of residual material will alter from context to context and hence this aspect is best considered on a case by case manner. Residual items may be detectable through qualitative and quantitative indicators, such as, for instance: typologically earlier sherds occurring in a stratigraphically later context or with later pottery; abraded sherds; small sherds of low average weight, and so forth.

5.7.3 Samian: a good indicator of residuality?

Since samian types have a wide distribution, were standardised, are identifiable and are associated with comparatively tight date-ranges, samian may be thought a particularly useful indicator of residuality. However, there is an attendant problem with the potential use of samian in this manner since this distinctive ware seems to have been treated differently from other contemporary pottery, and in particular was curated over long periods (cf. Section 5.8). Samian vessels appear in some quantity among deposits which post-date the normal range of the pottery type; some degree of survival into later periods is to be expected and should be considered normal (cf. Millett 1987a). With samian, though, this aspect appears to be exaggerated with a comparatively high proportion of samian vessels surviving into later periods and presumably still in use (cf. Section 5.8). It therefore becomes complicated to distinguish, among site groups, the 'longer surviving examples only subsequently lost/discarded to appear in later horizons' and truly residual sherds. Thus it may be that in fact samian is not a subtle indicator of residuality, and other pottery types might be more reliable.