The large sample of sites examined in Section 4.4 show a continuum of coin and artefact loss that subsumes excavated and metal-detected sites. To a large extent this supports Richards' (1999a) contention that the only thing that is special about a 'productive site' is that it was discovered though a combination of ploughzone agriculture and metal-detecting. On the other hand there is clearly a category of site that is found predominantly in eastern England, from Yorkshire to Kent with some extension around the south coast, and that relates to a peculiar phenomenon of these regions during the 8th and 9th centuries AD. These sites are not wics, although they have much in common with wic assemblages.
Nearly all of the sites considered above are established on sites that are newly occupied in the Middle Saxon period. While there may be Early Saxon activity in the parish there is rarely any direct continuity of settlement. In a few cases – Burgh Castle, Caistor St Edmund, Reculver, Richborough – the sites lie within or adjacent to Roman walled forts, but in each of these cases a minster church appears to have been established within the enclosed area. In two or three other cases – Hartlepool, Whitby Abbey, Whithorn – there is an association with a very early monastic site.
Activity continues at some of the sites considered into the Late Saxon period, indicating that the market was so well established as to survive the disruption caused by Viking raiding, or at least to return to the same site subsequently. However, it is often difficult to demonstrate direct continuity as the spatial resolution of finds is insufficient to determine if they have exactly at the same focus. Where detailed recording and excavation has allowed the observation of movement through time – as at Cottam A – it is clear that there is a disjunction of activity that parallels the shift in trading activity seen at a higher level at Hamwic, and between York's Fishergate and Coppergate sites. The majority of sites have finds only of the Middle Saxon period, or where they have finds from both periods the Middle Saxon finds are in a significant majority. The sites at Meols and Torksey stand out as unusual in having as many Late as Middle Saxon finds, and in both cases it is significant that the late finds include a very strong Scandinavian component. These later finds may represent the same sort of trading activities as witnessed during the Middle Saxon period, and may indicate a reversion to the same rural market network, away from an urban environment. Perhaps the Scandinavian traders preferred the less regulated markets available at such rural centres, rather than the regulated urban trade.
Nonetheless, it is clear that the productive site is largely a Middle Saxon phenomenon, which was generally brought to an end by settlement nucleation and overtaken by the more urbanised and industrialised economy of the 10th century onwards. Rural-based production was replaced by industrialised town-based crafts, and with the break-up of the great estates following Viking settlement there were fewer major landowners in eastern England. Was the decline of productive sites therefore due to the disappearance of their sponsors?
Some of the sites examined in Section 4.4 have assemblages dominated by coinage, but at the majority coins make up only 10-15% of the assemblage, at some it is 20-30%, and in a few it can be up to 50%. In most cases dress accessories make up the bulk of the remaining finds, with varying proportions of pins, strap-ends and hooked tags, according to regional trends in dress fashion. Equestrian equipment, while important in the national PAS fingerprint, is not generally found on these sites, mainly because it is a Late Saxon and Anglo-Scandinavian artefact.
While only a few sites seem genuinely to have assemblages dominated by coin finds, the majority of sites where metalwork as well as coin finds have been recorded seem to have assemblages dominated by copper alloy artefacts. It is difficult to escape the conclusion that while coins were lost in the process of the small-scale economic transactions that were taking place on these sites, the dress accessories represent the durable – and easily lost – items that were among the items being exchanged.
Most of the sites considered are at key nodes in communications routes, sometimes linking different systems. In some cases this appears to be waterborne, and the sites may have been coastal, such as Bamburgh Castle, Bawsey, Burgh Castle, Flixborough, Meols, Reculver, Richborough, Sandtun, and Shalfleet. Other sites within the sample are at the confluence of a river system, or a key point along it, such as Barham, Burnham, Burrow Hill, Cliffe, Ixworth, Ryther, Tilbury, and Torksey. In other cases they are positioned to exploit land routes, on major prehistoric trackways or Roman roads, and often near crossroads or junctions of different routes, such as Barham, Congham, Hollingbourne, Melton Ross, Nettleton, 'Near Royston', and South Newbald. Dyer suggests that the location of such markets in remote locations away from towns or other settlements explains why they did not survive the concentration of trade and industry in urban environments in the 10th century: 'Their location sometimes at road junctions or river crossings supports the suggestion that they were periodic markets or fairs, which failed to establish themselves as centres of exchange in the long term' (Dyer 2003, 91).
The association of many 'productive sites' with minster churches is striking but not exclusive. There is evidence for trading activity at the known major monastic sites within the sample, including Hartlepool, Whitby Abbey and Whithorn, and also some possible early minsters or monasteries, such as Beverley, Burgh Castle, Burrow Hill, Caistor St Edmund, Cliffe, Coddenham, Flixborough, Reculver and Richborough. Further sites in the sample are also associated with churches, although there is not always direct evidence that these had minster status, as in the case of Barham, Barton Bendish, Bawsey, West Walton and Wormegay. In many cases it is clear that the church provided a focus for regular markets and may well have used the market to convert the surplus from its own estates into cash. Elsewhere there is no direct evidence for ecclesiastical involvement, although there may be a secular estate centre. The sites considered are often, like Burnham and Middle Harling, at the centre of a multiple estate. But where there has been intensive archaeological activity in one area it is clear that within a large estate there may have been more than one focus where coins and portable artefacts were exchanged, as at East and West Rudham, or Cottam A, Cottam B, and Cowlam for example.
Finally, it should also be remembered that there are some sites of this period from which very few metal artefacts have been recovered, despite excavation. At North Elmham in Norfolk excavations have revealed an episcopal palace complex (Wade-Martins 1980). A Middle Saxon cathedral community abandoned the site in the early 9th century, possibly as a result of Viking raids, but sometime after 917 the site was cleared and levelled for a new church and ancillary buildings. Yet despite being in East Anglia, very little metalwork was recovered from the backfilled ditches and wells, and only two coins. In addition, there are excavated sites outside the 'productive site zone' which confirm that exchange was operating differently in these regions. At Pennylands in Buckinghamshire, Middle Saxon enclosures, pits and buildings have been excavated that are comparable to those found at Cottam B. The site is not too far inland to have participated in trade as it is not far from the River Thames, and there are fragments of lava quernstone and sherds of Ipswich Ware, but no coins and only a single pin is recorded (Williams 1993). To gain a full understanding of the 'productive sites' one needs to adopt a holistic perspective, as sites with few coins or metalwork can tell us just as much about trade and exchange as those that have profuse finds.
© Internet Archaeology/Author(s)
URL: http://intarch.ac.uk/journal/issue25/2/4.5.html
Last updated: Tues Apr 21 2009