With contributions by Rosemary Cramp, Nick Higham, Philip Holdsworth, Mark Leah, Sue Stallibrass and John Trippier
The early medieval period is perhaps the most challenging of all those eras defined by archaeologists, other than those at the very beginning of human activity in Britain. Whilst there is a veneer of historical information concerning the North West for this time, it can frequently create confusion, rather than illuminate the population of the region, since it tends to be of a very general nature, and was clearly written for purposes other than objective description. This, when coupled with the difficulties of identifying a cultural assemblage, or assemblages, which can instantly signify a site of this period when under excavation, means that these vital 700 years are even more under-represented in the archaeological record of the region than they are in the country as a whole. It was, however, clearly a time of immense change, when the structure of Roman Britain dissolved and society slowly moved towards that recognisable in the medieval period.
The last 30 years or so has seen a growing body of information concerning the period, which has moved any discussion away from the fragmentary framework created by the scant documentary sources towards sites that can be proved conclusively to have been occupied at some point between the 5th and 11th centuries AD. This has largely, although not exclusively, been associated with the growth of radiocarbon dating as a routine technique. One of the clear conclusions to be drawn from a review of the evidence for the period, however, is that frequently the evidence has come from multi-period sites, or even those where excavators have expected to find a site of another period altogether.
The challenges of site recognition are therefore considerable. In urban areas, activity may have continued on from Roman occupation, or, alternatively, there may be evidence for such activity, either proto-urban or of an entirely different nature, pre-dating a medieval town or village. It is also clear that there is considerable evidence to be recognised in the dispersed settlement of rural areas, both in the lowland and the upland parts of the region, if methodologies likely to identify such activity are adopted. The potential to identify early medieval activity is tremendous, both from commercially-driven projects, and also in those areas where development is minimal, such as in the Cumbrian fells. However, it depends on an awareness of this potential on the part of the excavator, and the adoption of very careful methodologies to maximise information retrieval, particularly in Roman urban or military sites, where early medieval remains can easily be mistaken for residual material or are indistinct in comparison with the large stone buildings preceding them.
The early medieval period nationally is generally divided into three broad phases, although the terminology used for these varies between different regions, since it is based largely on an assumption of population movement. Thus, in the south of England, the later two phases tend to be referred to as the Middle and Late Saxon periods, since it is assumed that an ethnically Saxon population had a significant influence there. It is arguable that such terminology would not be inappropriate in Cheshire. In the rest of the region, however, the Anglian kingdom of Northumbria held sway in the 7th to 9th centuries, and thus reference to ‘Saxon’ here is at best generic. Whereas in southern England the 10th and 11th centuries saw distinctive traits, such as the development of urban settlements, in the north perhaps the significant event of that time was the growing influence, culturally and politically, of people of Scandinavian origin. The lack of any common terminology that is applicable to all regions in Britain has led to considerable confusion, with a multiplicity of terms being used, often very loosely, for both the sub-phases and the material culture belonging to them.
A lack of clearly datable finds assemblages, and similarly the lack of publication of those few that we do possess, has meant that precise and detailed chronologies for material culture have not been developed. Indeed, apart from some culturally diagnostic metalwork, such as weaponry and personalia, there is a complete lack of material that can be readily associated with the period. Such diagnostic material tends to be found only in ‘extraordinary’ circumstances, with weaponry almost exclusively from graves or hoards, and personalia from metal detector finds, and to date almost exclusively from potential monastic sites, such as those at Dacre, Carlisle and Workington. Only a very few pieces have clearly come from other types of settlement. Pottery is almost non-existent, although handmade ceramics have been recognised as being made at Fremington (C), which, if found without any other evidence, would almost certainly be identified as of Bronze Age date. This clearly has potential repercussions for existing material in museum collections and other archives, which need to be examined as a matter of urgency.
For many years, whilst the importance was recognised of the ‘dark earth’ deposits overlying Roman towns in the south-east of England, horizons, whether dark or otherwise, above Roman sites in the North were given little or no priority, since they contained only Roman pottery and were declared ‘residual’. This southernocentric view has meant that these layers have tended to be removed as a block, sometimes even by machine, and excavation techniques were not formulated to attempt to identify any detailed phasing within them, nor to record in detail the material from them. In analytical programmes, the material from them, in towns such as Carlisle, was largely ignored because of this assumed residuality, and thus, even in published reports, this crucial period of change has largely been ignored, or is grossly under-represented. Similarly, rural sites containing only a few sherds of Roman pottery have traditionally been dated to that period, particularly in excavations from before the advent of radiocarbon dating. Recently, the adoption of such absolute dating as a relatively routine technique in rural excavations has demonstrated a far more complex picture.
In recent years the vital role played by absolute dating of material in the growth in our understanding of the period is self-evident. The number of sites dated in such a manner is not as yet sufficient, however, to allow this closely dated sequence to be applied to any cultural assemblages. In addition, the abundance of waterlogged deposits of the early Roman period in places such as Carlisle and Ribchester does not extend into the later Roman levels, and thus the only waterlogged material from the early medieval period recognised as yet comes from deep negative features, such as wells and a very few pits. To date, dendrochronological analysis has only been possible on three sites in Carlisle, and even there, it is not clear that the dates produced can firmly locate the feature chronologically, since some of the samples showed signs of reuse. It is therefore vital that every opportunity should be taken for radiocarbon or other scientific dating, to try to create a firmer foundation for the results of excavations in both urban and rural contexts.
The early medieval period has, perhaps, the greatest level of potential population movement of any time in the region’s past, and certainly the cultural influences from those centuries laid the foundations for the differences in some customs and in the regional dialects still spoken today. The extent to which cultural and linguistic influences equate with the physical movement of people continues to be debated (Fellows-Jenson 1985; Higham 1986; Hadley 2002; Abrams & Parsons 2004; Griffiths 2004), and the lack of skeletal material from the period hampers any scientific examination of this contentious subject. From the historical sources, however, it is clear that in the North West, certainly more than in any other region of England, there is potential influence from an enormous range of ethnic and cultural groupings. Within this region, influence might be expected from:
the native Romano-British population;
individuals or even groups of other ethnic populations within the Roman Empire who continued to live in the region after the ending of Roman governance, and their descendants;
peoples from Ireland and modern-day Scotland;
Anglians who first settled in the North East, presumably at least in part inter-marrying with the local population;
the Welsh (and presumably descendants of the inter-married Welsh and Irish populations);
peoples of Scandinavian origin, perhaps largely, but not exclusively, from present-day Norway, although inter-married with peoples from the Northern and Western Isles, as well as mainland Scotland, Ireland, and the Isle of Man;
Anglo-Saxons from the kingdoms of southern and central England; Britons from Strathclyde, and those descended from Irish settlers in western Scotland, and potentially also some influence from the Picts;
and, at the very end of the period, Normans, at least at the level of lordship.
The reason for these wide cultural influences is quite simple. The borders between England and Wales, and England and Scotland, only became firmly established during the medieval period. The coalescence of smaller polities to the south into the nascent England, to the north into the nascent Scotland, and to a much more limited extent, to the west of the region, particularly from the 10th century onwards, however, meant that the North West came under political pressure at intervals throughout the latter half of the 1st millennium AD, as it was competed over by more powerful political entities. The North West is unique in England in that it has two national boundaries and a lengthy seaboard dividing it from other nations. The early medieval period is unique in that it was the time when these nations began to form, whilst other conscious or unconscious bids for permanent political influence, such as that from Scandinavia, failed. This crucial element in the formation of nations has never been examined archaeologically, to date having been studied only through historical sources and a consideration of place-names.
For part of the period, at any rate, Cheshire looked to the south and the growing kingdom of England, whilst in the northern part of the region the preoccupations of the population seem to have been with the political developments in the North East, to the north in the growing kingdom of Scots, and to the west, across the Irish Sea. Over and over again, it is clear that there are culturally distinct traits in Cheshire, which differ from those evident in Cumbria, perhaps seen most vividly in the differing styles of the ecclesiastical stone sculpture of the 8th to 10th centuries, and the clear linguistic boundary in the vicinity of the river Ribble. In every case, the differences are clear between these two counties, but the lack of information from Lancashire hampers an understanding of the process of change. It is thus crucial to ascertain whether the apparent ‘black hole’ in this county is a result of a genuine lack of activity, or a lack of antiquarian or more recent research, or whether the methodologies adopted more generally across the region are not suitable to the specific geological and environmental conditions pertaining there.
There are several important initiatives that are crucial to forward the study of the region in the early medieval period. Such work is needed to create a more coherent context for all the early medieval research themes than exists at the present.
4.1 Establish an agreed common terminology across the region for the types and phases of activity found in the early medieval period.
4.2 Establish methodologies for the excavation of sites where there is potential for early medieval activity, to maximise the chances of its identification and recording, including regular metal detector surveys to aid artefact recovery.
4.3 Undertake radiocarbon dating as a matter of routine on any site with the potential to produce early medieval activity.
4.4 Prioritise dating by other methods, particularly dendrochronology, where appropriate.
4.5 Ensure that any palaeoenvironmental sampling and analysis should not be focused only on the prehistoric levels, but should include, and date, later material.
4.6 Attempt to link the fragmentary artefact sequences to some kind of chronologically robust framework.
4.7 Undertake focused research on issues of borders and cultural identity, including cross-border research initiatives with Wales and Scotland, and also the Isle of Man and perhaps Ireland.
4.8 Link place-name studies with wider landscape research in an attempt to place the known sites into some sort of landscape context.
The end of Roman governance in Britain has been viewed as one of the key dates in British history, yet the archaeological record is increasingly demonstrating that this was a continuing and quite long drawn-out process, rather than a single catastrophic event. Changes can be identified in the archaeological record from at least the mid 4th century, if not earlier, which form a continuum into the 5th century, and arguably even later. To date, this sequence of change is only clearly visible in the North West at Birdoswald on Hadrian’s Wall (Wilmott 1997), but ever since the evidence from that site was made public, the same types of activity have been recognised elsewhere, in particular in the fort at Carlisle (Zant forthcoming). Similarly, the apparent collapse of urban life in Chester and Carlisle, and of the proto-urban activities recognised in extramural settlements such as at Lancaster (Howard-Davis et al forthcoming), in some areas seems to begin in the mid 4th century. Yet evidence from Scotch Street (McCarthy 1993) and Blackfriars Street (McCarthy 1990) in Carlisle would suggest that to the residents of those areas, at least, a ‘Roman’ way of life was continuing well beyond the early 5th century. If Bede’s famous comment is to be believed, of course, it implies that a functioning Roman water system still existed in Carlisle in the late 7th century (Webb 1998). There are hints in places such as Lancaster that populations survived longer in fort sites than they did in the adjacent open, extramural settlements. This has never been truly tested, however, yet is crucial to an understanding of the administrative, as well as societal, changes occurring at the end of Roman governance in Britain.
The dating of such sub-Roman activity is tenuous depending largely on coinage sealed beneath it, such as at Birdoswald (Wilmott 1997) and Scotch Street, Carlisle (McCarthy 1993), since the one thing that does seem secure is that the early 5th century saw the collapse of the monetary economy. It is therefore crucial that every effort is made to provide secure dating for this type of activity, if possible by absolute means. In this context, it is vital that any organic matter, such as animal bones or charcoal, that is clearly not residual, should be dated. In addition, in such circumstances, any buried soils take on an added significance, since these might not only provide material for dating, but could also contain evidence of the environment, such as pollen, which will indicate whether urban conditions prevailed, or whether vegetation associated with abandonment becomes visible in the record.
However much Roman towns may have survived in some form into the 5th century and beyond, they clearly changed, and one of the most obvious challenges for students of the early medieval period is to develop methodologies that allow these changes to be defined and analysed. Where dark earths are recognised, or indeed any other apparently amorphous build-up of soil, it is again vital that geoarchaeological samples should be taken routinely. These should investigate soil formation, for instance whether the material formed as a result of agricultural or horticultural mixing, whether there is evidence of nutrient enrichment, or importation of material from elsewhere, or in situ burning, or the clearance of materials from areas which have been burned. It is only by the adoption of such techniques that questions can be addressed as to whether such build-ups demonstrate abandoned, waste land, or land that was still being used actively, as gardens, or even for wider, agricultural purposes, or whether there had been deliberate slighting of standing buildings for some reason, such as clearly happened to the Principia at Carlisle (Zant forthcoming). In terms of the artefacts from such deposits, rather than assuming, as was common in the past, that such material is residual, the artefacts and ecofacts need to be checked for evidence of abrasion, which would demonstrate some longevity in a deposit which was being worked, or evidence of reworking themselves. For instance, Roman pottery is often found, having been worked into gaming counters or spindle whorls. There has been a general assumption in the past that this occurred immediately after the end of its use as a ceramic artefact, but this need not be the case. Even where there are clearly elements of residuality, such as the presence of abraded pottery, this need not be the case for the whole assemblage; it is perfectly possible for residual pottery to be found alongside animal bones that are considerably later. Indeed, if, as is commonly assumed, the use of ceramic containers for cooking and eating became rare in the early medieval period, such a scenario would be highly likely, since even when using other materials for storage and cooking, people still needed to eat, even when not using ceramic plates! Bulk sediment sampling for plant remains and fish bones will have the added advantage of ensuring full recovery of small artefacts or artefactual fragments and of industrial residues (such as hammer scale or droplets from other types of metalworking or glass making).
In other parts of the country, such as the South West, there is evidence that towns, and also forts, became increasingly depopulated in the 4th and 5th centuries (Wacher 1995, 342; 375), and this seems to go hand-in-hand with a re-occupation of earlier defended sites, such as the hillfort at South Cadbury (Alcock 1972). There are several place-names throughout the region with elements that might echo this, such as caer and possibly also burh/g, although in Cheshire, at any rate, this latter should not be confused with the burhs established by Aethelflaed, Lady of the Mercians, and Edward the Elder, in the early 10th century (Thacker 1987). In this context, the emergence of Caerluel out of Luguvalium is of considerable interest. Such a population shift can only be speculation at present, but again work to prove or disprove such an hypothesis is of immense importance in the effort to create any sort of proper framework for the region before the growth of the Anglo-Saxon kingdoms in the 7th century.
The few historical sources which seem to reflect events immediately beyond the traditional end of Roman governance indicate that there were numerous war bands in the region at that time (Miller 1975). The excavations at Birdoswald (Wilmott 1997) suggest that the limitanei, the garrison troops manning the forts of northern Britain, were not removed at the same moment, or even subsequently to the withdrawal of the field army, which seems to have occurred in the later 4th century, and by AD 406 at the latest (Johnson 1980, 104). Instead, as the imperial hold over Britain slackened, such garrisons would have found themselves increasingly relying on self-provision, rather than supplies from a central command, perhaps in return for protection in increasingly unstable political circumstances. As the army pay chests became more irregular in their arrival, until the moment they ceased altogether, each garrison commander is likely to have found himself in a position of increasing autonomy. It is thus easy to contemplate such an individual moving seamlessly from being an officer of the Roman army to the leader of what was, in essence, a war band, and the garrison beneath him changing from a unit in the Roman army into such a war band (Shotter 2004). This begs the question, of course, of whether the many small kingdoms that seem to have existed in the North in the sub-Roman period were based on a disintegrating Roman administrative system, or whether they reflect much older Iron Age tribal divisions, or, indeed, whether there was any difference between these two systems. The absence of large multivallate hillforts in the North West, and the generally small sizes of the univallate hillforts that have been identified, might support the argument that power in the pre-Roman Iron Age in this region was more locally based than elsewhere. If the hypothesis of population shift away from towns and forts can be sustained, then this apparently decentralised pre-Roman political structure would have been ideally suited to the needs of the post-Roman period.
Given the paucity of any material culture that can be clearly associated with the 5th and 6th centuries in the North West, the period before Anglo-Saxon influence is often dismissed with little or no comment, as researchers have tended to dwell on positive evidence, perhaps rightly believing that anything else is mere speculation. Absence of evidence, however, is very different from evidence of absence, and there is an urgent need to explore exactly what was happening culturally at this time. The people living in the North West in these centuries have been homogenised, or even effectively airbrushed out, despite the clear indications of a visible social disparity both at the end of the Roman period and also within the Kingdom of Northumbria. The fundamental issue that even in the 11th century there were indications of a recognisably ‘British’ culture in the region has been largely ignored. When raised, the presumption has been that this resulted from the influence of the kingdom of Strathclyde in the 10th and 11th centuries (Jackson 1953, 218-22), rather than the survival of an indigenous way of life. Yet Saint Cuthbert is recorded as having met a reeve called Waga, a British name, in Carlisle in AD 685 (Colgrave 1940), and the Vita Wilfridi (Colgrave 1985) describes how Wilfrid, during the dedication of his new church at Ripon (AD 671-678), read out a list of consecrated places that were clearly west of the Pennines which the British clergy had deserted when fleeing from a Northumbrian army. This seems to refer to the Northumbrian annexation of the northern parts of the region, and indicates vividly a surviving culture, at least until that moment, that was clearly Christian. In addition, a legal text, the Northleoda Laga, that seems to have been compiled by Archbishop Wulfstan of York in the mid 9th century, incorporated clauses on wergild for ‘Welshmen’ from the law codes of Ine of Wessex (AD 688-726). Patrick Wormald has noted (1999, 394) that ‘the author looked up Ine’s law on “Welsh” status for guidance in a northern society with a still significant British ingredient’. The flowering of Northumbrian stone sculpture, in particular, and certainly when coupled with the influence of the writings of the Venerable Bede (Colgrave & Mynors 1969) and the number of Old English place-name roots, makes it all too easy to over-estimate the effects of Northumbrian rule; surviving British cultural remains should therefore not only be anticipated, but methodologies should be developed to seek them actively.
4.9 Re-examine old excavation archives from key Roman sites, for instance Carlisle, to look again at evidence for latest activity.
4.10 Ensure that methodologies for the excavation of late Roman deposits are appropriate for recognising, characterising and dating later material above the more easily identifiable Roman features.
4.11 Ensure suitable sampling strategies for dating, geoarchaeological, and palaeoenvironmental analysis of deposits overlying Roman stratigraphy, such as testing for soil formation, agricultural or horticultural mixing, and nutrient enrichment.
4.12 Ensure that opportunities are taken to examine the veracity of the commonly-held view of a population shift away from towns into defended places at the end of the Roman period.
4.13 Ensure appropriate methodologies are adopted on all sites with a potential for early medieval remains, to identify and characterise activity leaving scant and possibly confusing cultural indicators.
4.14 Target sites that may contribute to an understanding of social disparity in the early medieval period.
4.15 Develop methodologies to identify British cultural remains in the post-Roman milieu.
An understanding of settlement patterns for the early medieval period throughout the North West is in its infancy, and knowledge of associated land-use is even more fragmentary. Yet again, however, the historical sources hint at the probability of recognisable landholdings and ways of life beginning to come into existence, at least in the later part of the period, that would develop clarity in the 13th and 14th centuries. This can be seen in the emergence of a series of sub-regional names, often first recorded in the Domesday Survey (Faull & Stinson 1986; Morgan 1978), such as Amounderness, and in the 12th century in Cumbria, such as Copeland. The predominating pattern throughout the region seems to have been that of dispersed settlement, any evidence of nucleation, particularly in the form of urbanism, fading with the collapse of the Roman monetary economy, and not developing again until the 10th century in the south of the region, and after the Norman takeover in the north.
There is currently very little known about land-use, landscape and climate change from the 5th to the 11th centuries, on both a regional and national level. Coastal studies (eg Tooley 1978) have demonstrated that the period was one of some volatility in terms of sea levels, although some changes may have been localised, and a trend towards greater stability can be seen through time. Fluvial activity has also been recognised, which may at least in part be linked to changing sea levels, such as the evidence from excavations at Damside Street, Lancaster, where early Roman deposits were covered by more than 1m of increasingly sterile silts, suggestive of a riverine incursion, which were in turn cut by medieval features (LUAU 1991). The peat resource has been demonstrated to retain late Holocene deposits that span the early medieval period in much of the region, but little systematic work has been undertaken to produce a comprehensive picture, nor has dating tended to concentrate on the upper parts of the peat sequence. At present, such work as has been carried out is biased towards the north of the region. Palaeoenvironmental studies there have demonstrated a complex sequence of both woodland regeneration and the creation and maintenance of clearances, but there is currently little synthesis of the data. Recent studies of sediments and peat deposits in the Howgills and the Forest of Bowland have indicated several periods of accelerated soil erosion on upland slopes during the Holocene, including a period possibly starting during the 10th century AD (Harvey & Chiverrell 2004), although other studies have dated such activity to later in the medieval period (eg Wild et al 2001), indicating, perhaps, a complex pattern of localised events. Many more areas have suitable sediments for studies of soil erosion, river sediment transportation and deposition, or for vegetation studies, all of which contribute to an understanding of changes in vegetation cover and erosion/sedimentation in response to climate change and to human interference.
On a regional scale the excavated evidence for land-use is at present, sparse, although much of the region is rich in features that may have originated from the early medieval period onwards. These consist primarily of civil parish boundaries, many of which still survive as banks, hedges, or ditches, and which usually began as township and ecclesiastical boundaries (Lewis & Pepler 2002a; 2002b). Many, of course, will have originated in the medieval period or later, but some may go back to the pre-Conquest period. Indeed, it is possible that some major features might even reflect pre-Roman territorial divisions which were given renewed importance by the fragmentation of society following the collapse of Roman rule. The likelihood of an early medieval origin is enhanced where a boundary originally marked the limits of one of the large, multi-township parishes that were characteristic of the region in the past. These features may sometimes seal or contain valuable palaeoenvironmental evidence, such as buried soils, pollen and macrofossils (Gifford & Partners 2002). Excavated data relating to early medieval settlements and landscape boundaries, however, remain rare and, if theories are to be tested, there is a need for more information from excavations. Synthesis of such evidence will also need to include a much wider dataset, including the pattern, as far as it is understood, of routes, particularly Roman roads, appropriately drawn information from medieval landholdings, and place-names, to develop the picture as far as possible. There is at present little evidence for agricultural regimes and economies, or even for everyday diet and subsistence, and it is clear that any fieldwork project should develop methodologies to maximise retrieval of material which will contribute to an understanding of these crucial topics. Any early medieval site is likely to have had evidence of diet and subsistence practices deposited during occupation, and charred, mineralised or waterlogged plant remains should survive on many sites. As for all periods, animal bones are only likely to survive in sediments with neutral or alkaline pH, and this is most likely to occur in urban conditions or on base-rich bedrock such as limestone. Shells of marine shellfish, however, are likely to survive relatively well.
Despite the poor quality of survival of many animal bone assemblages, considerable potential remains. For instance, there is a large assemblage of butchered animal bones from Castle Street, Carlisle (Rackham et al 1991), although it is dated through association only, giving a possible range for the activity, either in the 7th to 10th centuries or the 12th to 13th centuries. An estimated 2,700,000 fragments of cattle bones from specifically selected parts of the carcass (deriving from approximately 1500 animals) were recovered, which had been butchered in a very distinctive manner, presumably representing an extremely organised and specifically-targeted activity. Such a degree of organisation is remarkable for any period, and is particularly important if it does relate to the early medieval period, since so little is known about the continued occupation of urban sites or agricultural practices. Likewise, the date of the deposits at the base of a well at Castle Street (McCarthy 1991) is controversial and needs to be resolved. There, the presence of grain beetles may imply relatively intense activities of a type normally to be found in an urban environment, such as bulk grain storage and/or importation of grain supplies from Europe, and could potentially make a revolutionary contribution to the understanding of the period as a whole.
One context where bone survival is good is within caves, and most especially those of south Cumbria and Lancashire. While there is little evidence for the use of caves in the region during the early medieval period, this area of research has, like many others, been presumed to be an unprofitable avenue of study because of negative evidence. Many caves have evidence of activity throughout the prehistoric and Romano-British periods, and then it is presumed to have ceased. Recent dating of worked antler and a cattle metatarsal from the Doghole in south Cumbria (H O’Regan pers comm) clearly confirms the excavators’ belief (Benson & Bland 1963) that activity in the cave continued from the Romano-British into the early medieval period. The practices of deposition within caves are still poorly understood, but it must be borne in mind that such activities may have continued beyond the end of the Roman period.
Many of the problems regarding settlement studies of the period stem from the perceived lack of chronologically diagnostic artefacts, coupled with a timber building tradition, at least in the southern part of the region. This has rendered much of the evidence for human occupation difficult to detect, particularly in these lowland parts of the region, or to be seemingly invisible. Indeed, the ephemeral nature of even quite substantial buildings (eg McCarthy 2002, 136) might suggest that many more structures from the period may have been encountered during the course of antiquarian and archaeological excavations, but remained unrecognised. The distinctive sunken-featured buildings known as grubenhauser are only known for certain from one site (Oliver et al 1996). In addition, few of the rectangular stone buildings identified in the uplands, or the series of enclosures in the Eden valley, have been dated in any meaningful way.
This situation has to some extent fuelled the traditional view that the ending of Roman Britain led to the collapse of an agricultural economy that was primarily supplying the Roman army. Thus enclosures, houses and field systems, that have been dated largely through morphology to the Romano-British period, have at times been presumed to be abandoned at the end of Roman governance, and there has been almost no work to address the resulting hiatus in a detectable settlement pattern before visible dispersed farmsteads apparently emerged in the medieval period. The physical evidence of past land-use, clearance and division is visible in both the uplands and lowlands, but these features are not normally ascribed an early medieval date. This is despite the fact that there are few dates for cairnfields, stone banks or field ditches from the entire region and that, on one occasion at least, an early medieval date has been recovered from a clearance cairn (Quartermaine & Leech forthcoming). Similarly, the system of transhumance in the uplands, particularly of Cumbria, can be distinguished in the medieval period, and has linguistic links with dialect derived from Scandinavian roots, but there has been no coherent work to date this site type on anything other than morphological grounds or from ceramics or documentary evidence; both of these are, by their very nature, going to provide a medieval or later date. It is therefore possible that this pattern of landscape exploitation originated in the early medieval period. Apart from a small number of well-studied sites, data on the nature, preservation, and extent of archaeological deposits at many important early medieval centres (usually those identified on the basis of place-name or documentary evidence) are often entirely lacking. There is evidence for possible continuity of occupation in some of the Roman proto-urban centres and military stations, but the exact date and nature of this activity is unclear.
The dearth of excavated sites from the sub-Roman period onwards has on occasion led to the development of models of settlement relying on place-name evidence and historical data. This can lead to the unsafe presumption that place-names reflect the ethnicity of either their original inhabitants or those in the surrounding area, rather than a much more subtle and complex pattern involving not just new settlement, but the renaming of existing entities to suit changing ownership or even taste. The small number of known sites is also exacerbated by the lack of a morphological typology for sites of this period. Many rural settlements are known from aerial photography and field survey, represented by earthworks in the uplands and cropmarks in the lowlands, but these sites have often been interpreted as dating to other periods, although many may well have been occupied in the early medieval period, and some may have had continuous occupation from the Romano-British period, or even before. The marginal nature of so much land in the North West means that, in the uplands at any rate, settlements will have existed at the interface between land suitable for cultivation and the ‘outfields’ suitable for grazing for many thousands of years, as they do to this day. This makes the challenge of identification greater, since these confined areas are likely to have a palimpsest of occupation, with only limited lateral movement. In effect, the only obvious sites in the landscape in such a pattern are the most recent, and those that have failed, for whatever reason. Fieldwalking, evaluation and excavation of sites may fail to recover chronologically diagnostic artefacts due to an apparent low level of usage of inorganic materials during the 1st millennia BC and AD. Ultimately, only excavation coupled with absolute dating techniques will address some of these questions, but such projects will need to be developed with care, to maximise data retrieval.
Aerial photography has always played an important role in the recognition of rural settlement, particularly where arable agriculture predominates. Rural settlement sites will always be under-represented in areas of permanent pasture, where cropmarks are less likely to form, and the North West has large tracts of land which have been used for arable in the past, but are now down to permanent pasture. In such locations early medieval sites are most likely to be discovered during development projects, often while in the process of investigating sites of other periods. New aerial photography should, however, be used intensively in those areas of potential, particularly targeting places that have not been flown to any great extent in the past. Attention is increasingly being given to the significance of littoral parts of the region throughout the early medieval period, both in terms of the potential for headland settlements, and the possibility of landfalls in inlets and estuaries, and clearly such areas would benefit from targeted programmes of flying.
Recent excavations have shown that the period is not entirely aceramic (eg Oliver et al 1996) but artefacts cannot be relied upon to date excavated features or sites. Some forms of locally made pottery are difficult to identify and are not currently chronologically diagnostic; scientific dating techniques are thus required. Absolute dating techniques that are most relevant to this period are radiocarbon dating and archaeomagnetic dating of materials heated in situ (hearths, ovens, kilns etc); Optically Stimulated Luminescence (OSL) may be applicable to sediments buried by subsequent deposition, eg beneath banks. Sites for which some artefactual dating evidence exists (eg Roman or medieval pottery) should be included in any re-assessment of the chronology of rural settlement, and should also be subject to absolute dating techniques. Particular attention must be given to post-Roman and pre-medieval stratigraphy, even when seemingly sterile.
A site forming an important focus in the early medieval period, if archaeological deposits can be shown to survive, will have great potential to address issues concerned with the organisation, chronology, function and economy of the period, and thus an holistic approach is vital. Clearly, the challenge is to recognise such sites. Rosemary Cramp (1994) has pointed to the conjunction of Roman fort, or, indeed, earlier fortified site, early place-name and later medieval estate centre on the same parcel of land, which can be identified again and again, such as in the vicinity of Brougham and Penrith in Cumbria, of course at Chester, Lancaster and Carlisle, and perhaps also in the juxtaposition of the fort at Ravenglass and Muncaster Castle, also in Cumbria. It is quite clear that, to some extent at least, the same strategic parameters were relevant in the medieval period as were current in the previous millennium, and thus centres of power that proved their strategic importance in both the Roman and medieval periods are good candidates for significant early medieval activity. This may be particularly important on key routes, such as that along the Mersey Valley, following the Lune in Lancashire and southern Cumbria, and down the Eden in Cumbria. In addition, locales where there is a conjunction of a medieval estate centre and a church producing early medieval sculpture are also clearly significant (Cramp 1983). Indeed, the presence of a church with evidence of activity before the Norman Conquest can be used as a clear indicator of settlement, for whilst the surrounding settlement pattern is likely to be dispersed, nevertheless the church will have acted as a focus.
A further dataset that may in the long term aid the discovery of early medieval settlement sites is the information being accumulated by the Portable Antiquities Scheme (PAS). Typologically distinctive high-status metal artefacts may be indicative of specific activities or forms of deposition (such as a grave or ecclesiastical site), but a distribution of even a single artefact may indicate activity or the location of settlement. Access to the finds data is currently problematic, and there is a time delay for the dissemination of information to the region’s HERs. A periodic regional synthesis of reported early medieval finds would greatly aid this form of research. Similarly, metal detectorist rallies can target localised areas, as can individuals over a longer period, the value of such an approach being obvious, providing sufficient levels of recording are undertaken. Encouragement of detailed metal detector surveys, particularly following a regular grid and with precise positions being marked (for instance with GPS techniques) would also greatly enhance the quality of the data.
In contrast to almost every other region of England, the re-establishment of towns in the North West is largely a medieval phenomenon. Indeed, it would seem that any move away from a dispersed settlement pattern towards any sort of nucleation, be it rural or proto-urban, does not seem to have been an early occurrence. The exception would seem to be in Cheshire, the only part of the region to lie beyond the borders of Northumbria, and where the expansion of the nascent kingdom of England in the 10th and 11th centuries created an economic climate not seen further north until after the Norman Conquest. In Cheshire, Chester quickly became a centre of production from the 10th century onwards, and accrued most of the attributes of urbanism in the next 200 years (Carrington 1994). Chester’s renaissance appears to have followed the establishment of a Mercian burh there in AD 907. It is thus tempting to equate the growth of proto-urbanism elsewhere with this phenomenon, to the extent that Higham (2004a, 140-1) has suggested that Penwortham, on the opposite side of the Ribble from Preston, and the only site north of the Mersey to be listed with burgages in Domesday Book (Morgan 1978), may have been such a creation, although in the reign of Edward the Elder (AD 900-24) rather than his predecessors. No coherent study of these sites has been undertaken, however, and hardly any excavation has taken place outside of Chester.
Elsewhere, there are slight indications that some sites were favoured for the exchange of goods, for instance, Meols on the Wirral coast (Griffiths 2004), and there is a recurrent suggestion that markets were developing in association with church sites, hardly surprising, perhaps, as these would be the focus of the local population at specific times of the year. Some at least of these foci seem then to have acted as a catalyst for nucleation. It seems clear that the church played an important role in the development of nucleated settlement, but it is to date uncharacterised in the region, and there is almost no firm evidence to support the contention.
4.16 Integrate palaeoecological and archaeological data to investigate how people’s use of the landscape altered it and to try to distinguish changes relating to climate from changes related to human activities.
4.17 Date late Holocene peat sequences by radiocarbon assay, particularly in the south of the region, which currently lacks any well dated analysed sequences.
4.18 Target river basins and valley bottoms to investigate erosion sequences potentially associated with the intake of fields, woodland clearance or soil disturbance. This could be undertaken in conjunction with investigations of peat deposits for climatic indicators and changes in settlement and land-use.
4.19 Re-examine and date deposits of animal bones and environmental material from Castle Street, Carlisle, to resolve at least some of the uncertain date of existing material excavated from urban contexts.
4.20 Analyse any animal bones, and charred, mineralised or waterlogged plant remains, identified from any site of the period as a matter of priority.
4.21 Record, wherever possible, sections through extant civil parish boundaries threatened by development. Opportunities are most likely to arise through linear development projects, such as pipelines and road schemes, as well as housing developments. Work should concentrate not just on recording sections but also on assessing the potential for the preservation of environmental data (buried soils, pollen, macrofossils etc).
4.22 Create a large-scale map for each district or county of the various boundary types as an essential curatorial tool.
4.23 Ensure the upper stratigraphy of Roman period sites and the lower stratigraphy of medieval sites is subject to detailed scrutiny for what may be ephemeral traces of early medieval activity.
4.24 Ensure method statements and research designs for the excavation of rural settlement in the region build in costings for absolute dating, in most cases for a series of dates, particularly targeting short-lifespan material, such as twiggy charcoal, animal bones and charred cereal grains, as well as archaeomagnetic dating of materials heated in situ (hearths, ovens, kilns, etc).
4.25 Examine, wherever possible, presumed centres of early medieval activity, particularly the nature of surviving archaeological deposits, coupled with a detailed programme of dating. These might include settlements with churches producing early sculpture, or medieval estate centres re-occupying Roman sites.
4.26 Monitor the distribution pattern of material reported to the Portable Antiquities Scheme throughout the region, to highlight ‘hot spots’ of early medieval activity, and individual sites with archaeological potential.
4.27 Undertake systematic metal detector surveys in such ‘hot spots’ to attempt to define and characterise early medieval activity.
4.28 Sample suitable deposits throughout urban stratigraphic sequences systematically, for scientific analyses, coupled with absolute dating techniques, to refine both the chronology and an understanding of the activity taking place.
4.29 Prioritise significant sites for publication.
In many ways the pattern of Christian life, in the form of stone sculpture, certainly from the 8th century onwards, is more visible than any other type of activity in the early medieval period in the North West. Despite this, a true understanding of what the presence of these sculptural fragments actually mean, and what this tells us about the North West at this time, is in its infancy. Much of the sculpture came to light in the 19th century, during the massive ‘campaign’ to refurbish the dilapidated parish churches of the country, which in many cases in the North West led to an almost total rebuilding. In most of these cases, the early material was found built into medieval or later masonry, or the circumstances of the discovery were not recorded; in no case was formal archaeological work undertaken to explore the context of the material. Since then, conservation has been a key theme for ecclesiastical sites, which presents its own challenge for furthering understanding of their developmental sequence, since opportunities for anything other than very small-scale work are extremely limited. Such opportunities as do present themselves should be taken, however, particularly in association with refurbishment of the church fabric itself, or in any development that affects key elements of the churchyard, such as well-established boundaries, paths, trees, and so on.
Late Roman histories record that Christianity became the official and sole religion of the Roman Empire in the reign of Constantine in the early 4th century (MacMullen 1987), and yet there is almost no physical evidence for its practice in the North West (Thomas 1971). Superficially, the lead salt pans found at Shavington (Ch), which record an episcopus named Viventius (Penney & Shotter 2002), appear to be the best evidence in a late Roman or sub-Roman context, but at that time an episcopus could have held a secular adminstrative post, and there is no independent proof of a Christian context. The name Viventius, however, seems to have Christian affinities, and thus it is at least possible that this may represent a rare survival of an administrative structure that otherwise is known only from the few documentary sources for the period. This demonstrates the essentially serendipitous ways in which the scattered information that informs the period can come to light, and highlights the need for seizing every opportunity to explore new evidence to the full.
Since Kenneth Cameron’s work in the 1960s, it has been recognised that the place-name element ‘Eccles’derives from the Welsh eglēs ‘a church’, itself a loan from Latin ecclesia. The element is particularly concentrated in English place-names in the historic counties of Lancashire and the West Riding of Yorkshire. In Lancashire, there is a coincidence between Eccles place-names and the structure of hundreds or wapentakes (Barrow 1969), as they are understood to have existed by the high Middle Ages. This may reinforce the case for their earlier role as churches serving an extensive territory of the type represented by the Northumbrian shire identified by Jolliffe (1926). If these place-names, as is generally believed, do offer the opportunity to localise post-Roman but pre-English Christian cult sites, then their research should be given a high priority and a systematic study take place of one at least of these concentrations.
Some sites, such as Eccles in Salford (GM) and Eccleston to the west of St Helens (M), have been heavily developed and thus may be more challenging for archaeological survey or excavation, other than desktop research. In contrast, other potential sites, such as Eccleshill in Blackburn and Little Eccleston and Great Eccleston on the south side of the Wyre (L), remain largely undeveloped and might benefit from both desktop and ground survey to attempt to locate the historic cores. A series of burials was actually excavated at Eaglesfield (C) in the 19th century, of which one contained recognisably Scandinavian artefacts, but the remainder were oriented east to west, and apparently without grave goods (Cowen 1967). These burials are therefore not closely dated but may represent an early Christian cemetery. The potential for further work at this site remains, to place these early finds in context and to establish the longevity of burial activity.
Certain churches within the region have unusual dedications, which, if not clearly belonging to post-medieval re-dedications, may imply a medieval tradition of an early site. For instance, few medieval churches are dedicated to St Martin, yet there are three in Cumbria, and others have associations with the Celtic church, principally through references to Ninian and Kentigern (or Mungo). At one such site, Ninekirks near Brougham (C), such a tradition is backed up by both the physical presence of an enclosure, seen on aerial photographs (Loveluck 2002), and metalwork dated to the 8th century (Bailey 1977a), but it has never been subject to formal excavation. Excavations close to sites such as Brampton Old Church (C), dedicated to St Martin, have concentrated on the earlier Roman activity in the vicinity, rather than the church itself. Other dedications have clear Irish associations, such as those to St Bega/St Bridget on the west coast of Cumbria, and there has been a long tradition, dating to at least the 16th century (Fowler 1903), of linking dedications to St Cuthbert with places where the saint’s body rested during the travels (and travails) of his community in the 9th and 10th centuries. The way in which such traditions can be tested to ascertain whether they are based on fact needs careful thought, however, given that the majority of such sites are either living churches or have a long tradition of burial in and around them. In the same vein is the tradition of holy wells, many of which also have unusual dedications, and some of which may theoretically at least have origins in the pagan past. Such sites are fraught with difficulty for the archaeologist, though, and so far it has not been possible to prove their longevity. In an area with few documentary records until the relatively recent past, this makes judgement difficult as to whether these are genuinely ancient, or the creation of later traditions.
The recovery of early medieval metalwork through metal detecting has the potential to locate previously unknown churches and also aid in their classification, even if the metalwork itself is not directly Christian in its iconography. Indeed, at those rare sites recorded in documentary sources, such as Dacre (C), most of the metalwork recovered through excavation would not have been out of place on a high-status secular site (Newman & Leech forthcoming). At Dacre, and elsewhere in the country, the recovery of items such as styli and book clasps has provided unambiguous evidence of literacy which in a pre-Conquest environment is almost certain to mean a connection with a religious community.
The origins of churches in the North West, as in most regions, are shrouded in mystery, although a foundation as a minster or fully functioning monastery for some at least is probable (Morris 1989) and many of the churches at the heart of the larger medieval parishes are likely to have originated as pre-Conquest minsters and/or important estate centres. It is at least possible that Anglo-Saxon kings saw the placing of monasteries within the ruins of Roman sites as a means of revival (on the Continent and in Kent these tended to be specifically female houses; R Cramp pers comm), and whether this is true or not, the association between early churches and Roman centres, as at Chester and Carlisle, is not surprising (Matthews & Wilshaw 1995; McCarthy 2002). It would be tempting in this way to see the placing of medieval parish churches in the ruins of Roman forts, a relatively common occurrence in the northern part of the region, as evidence of some sort of continuity, but at Brampton Old Church (C) at least, there is no evidence from excavations that the Roman site was occupied beyond the mid 2nd century at the latest (Simpson & Richmond 1936). There is also evidence that, in parts of Northumbria, headland sites have significance, as at Whitby, Hartlepool, and Monkwearmouth (Cramp 1976). The recent work at St Michael’s, Workington (C), as well as at Heysham (L) (Potter & Andrews 1994), has indicated that this may be a more general trend in Northumbria, and to this end, study of sites such as Moresby (C) would be of interest.
Whilst modern evidence suggests a less precise boundary between monasticism in the 7th to 9th centuries and its collapse under the onslaught of Viking raiders in the 10th (Morris 1989), nevertheless, the little evidence recovered to date does suggest an increasing ‘secularisation’ of such sites from the 9th century onwards, although without any sharp dislocation in activity (see, for instance, Newman & Leech forthcoming). The development of so-called church markets, whilst probable, has not been proven beyond doubt, and deserves attention. Given the complete lack of detailed documentary sources, careful fieldwork is the only way in which evidence will be gathered. This creeping secularisation is one route by which monastic sites could have metamorphosed into nascent parish churches, although to what extent the pattern was infilled by new foundations by secular lords in the 10th to 12th centuries is a matter for debate. The presence of early sculpture is the most sure indicator of a church founded in the early medieval period, particularly those few examples with Northumbrian art forms and also Scandinavian motifs, indicating some form of continuity into the 10th century and beyond. Whilst the corpus is clearly incomplete, and new sculpture continues to be identified, the pattern is sufficiently understood that some analysis of the monuments themselves could now be attempted. For instance, there are clear differences between the material found in Cumbria and north Lancashire, in other words, the Kingdom of Northumbria, and that in Cheshire (Mercia). Simple division between clear grave markers and more complex monuments would help to refine our understanding of the purpose of the sculpture, particularly for the 10th and 11th centuries, when the number of identified grave markers rises. This would allow a focus on the unusual monuments, such as the Dacre Stone, with its unique theological programme of iconography (Bailey 1977b), and in-depth study of whether this could represent some teaching or even missionary activity at that site (R Cramp pers comm). In the same vein, understanding the meaning of the linking of Christian and Scandinavian subjects on the Heysham hogback, or at Gosforth, would be useful. Such sculptural material is often associated with centres that later developed into towns, as at Penrith, Kendal, Carlisle and Workington (all Cumbria), but there are significant exceptions, although these tend to be major medieval church sites. Some consideration of the factors at work here would be useful, and a regional landscape study of where the major medieval churches developed would be welcome, although this should, perhaps, be linked into a wider consideration of churches as indicators of settlement.
Little is known of the structure of early medieval churches since, with the possible exception of a very small group in eastern Cumbria, and St Patrick’s Chapel at Heysham (L), no firmly dated structure pre-dating the 12th century remains (Taylor & Taylor 1965). Unsurprisingly, these remains are all of stone bonded with mortar, but this does not mean that all, or even most, early churches were built in this way. Indeed, it is likely that dry stone construction as well as wholly timber buildings were the norm, and in areas such as the Solway, the tradition of clay buildings may be very ancient, although none earlier than the medieval period have been proven to survive (Messenger 2000; Jennings 2003, 33-45). The distinctive group of apparently early churches in east Cumbria, when linked to others sharing the same long segmented structure, but with no clearly early masonry, such as at Brough and Orton (C), are clearly worthy of further study, particularly when evidence of cobbles is visible beneath the ashlar masonry, perhaps indicating an earlier structure used in the foundations of the later. In this case, it would also be beneficial to expand such a study beyond modern administrative boundaries, as there are commonalities with the adjacent part of Yorkshire, in particular.
It is recognised that church archaeology has been an understudied discipline in most parts of the region. The opportunity for intrusive archaeological work within or around churches is limited, but not altogether out of the question. As well as groundworks for maintenance, the current emphasis on providing toilet and kitchen extensions has provided several opportunities to examine the depth and nature of the archaeological deposits at these sites, usually through evaluation trenching, and a valuable body of data has already been collected in Cheshire. It is likely that similar opportunities will continue to arise in the future and it is therefore important that the curatorial framework is in place so that opportunities are not missed. An understanding of the development of ecclesiastical structures in the early medieval period should not, of course, concentrate solely on the church itself, since a monastic site would lie at the centre of a much wider estate, but study of the wider landscape is again in its infancy. The detached elements of parishes, seen particularly in medieval Cumbria and Cheshire, may be faint echoes of such as, for instance, at St Bees. Similarly, the process through which parishes developed, widely believed to have fully formed by the 13th century (Higham 2004a), is neither well understood nor dated. Whilst place-names can aid in the identification of ancient institutions, the boundaries of many parishes were marked physically, often by earthworks, and every effort should be made to establish a firm date for such features through excavation and radiocarbon dating.
With notable exceptions, the region is almost entirely lacking in burial evidence from the early medieval period. Beyond those associated with a church, the few known instances often involve the secondary use of prehistoric burial mounds for inhumation, examples ranging from the cemetery at Southworth Hall, Warrington (Freke & Thacker 1987), to fragmentary single inhumations at Withington (Ch) (Wilson 1981), and perhaps also at Hardendale (C) (Williams & Howard-Davis 2004). The suspicion is that this represents a much more widespread phenomenon than is currently recognised and an aspect of a tradition well-known from elsewhere in the country (Williams 1997). Many excavated unaccompanied human remains in barrows have been presumed to be prehistoric, although in many cases this has not been confirmed by scientific dating techniques. Only a small proportion of the region’s barrows have been excavated to modern standards, but there is a great deal of material in local museums from the many less rigorous excavations carried out in the past. Only a few individuals are to be expected from any single site, but this is in no wise a valid reason for giving the material a low priority. The putative ‘battle cemetery’ at Heronbridge (D Mason pers comm) is not typical, although the ongoing study should throw fascinating new light on this period.
The period is characterised by migration and movement, but the extent and details of such movement and settlement are not known. Whilst place-name evidence and distinctive artefact styles are signifiers to some degree, they do not necessarily indicate people of a distinct ethnic origin. Stable isotope analysis cannot identify ethnicity, but may help to identify where an individual spent their childhood, and hence the area of their origin. Soil conditions in the North West are generally inimical to the preservation of bone, and thus few burials with bone surviving can be expected. The small number of burials dating from this period makes it imperative that all available material should be analysed for stable isotopes as a matter of routine.
The recently excavated cemetery at Cumwhitton (C) has demonstrated that completely unforeseen sites can be identified fortuitously, in this case by metal detectorists (Brennand 2006). There, the finds reported to the Portable Antiquities Scheme were so distinctive that a burial could be anticipated, and funding for formal excavation was secured. In other cases, evidence of potential burials may be more subtle. A survey of other metalwork in the region, both recently discovered and in museum collections, is urgently needed to establish the potential for other such sites in the region, where further research would be beneficial. A detailed survey of those churches and churchyards where potential or actual Viking-Age burials have been identified would also be useful.
4.30 Archaeological curators need to formalise and agree appropriate protocols with the Diocesan authorities which will enable all intrusive work on and around these sites to be monitored archaeologically, and for evaluation or excavation to be carried out where appropriate.
4.31 Target appropriate sites with Eccles place-names with systematic survey of surrounding settlement.
4.32 Target churches on headlands for study as possible early sites, particularly when they are associated with Roman material or early medieval stone sculpture.
4.33 Examine all types of evidence for church markets.
4.34 Analyse differences between stone sculpture at a sub-regional level.
4.35 Prioritise the study of long segmental churches in both east Cumbria and adjacent parts of Yorkshire.
4.36 Prioritise any possibly early medieval burials or funerary material for full scientific analysis including absolute dating, stable isotopes and DNA investigations.
4.37 The presence of stray finds recovered by metal detectorists needs to be highlighted, which may indicate burials of this period being damaged by agricultural practice, and they should be prioritised for investigation.
4.38 Previously excavated skeletal and artefactual material (particularly pottery) needs to be re-examined, and, where appropriate, scientifically dated, to establish whether a phase in the development of burial practice in the region has, to date, largely escaped notice.
4.39 Maximise opportunities to excavate and date parish boundaries.
4.40 Key sites for publication need to be prioritised.
The 4th and 5th centuries undoubtedly witnessed widespread changes in both the scale, form and means of production of most types of material, as well as in the routes of trade and exchange. Models of Romano-British agricultural systems have stressed the primary importance of supplying the Roman military, with the personal and social aspects of Romano-British communities being a secondary consideration. The decline of the Roman army did not necessarily indicate that the means of agricultural production or that for other goods collapsed entirely within rural communities. Major sources of archaeological information do, however, cease to be found on Roman military and proto-urban sites. The decline in pottery use may be a factor of production and supply, but equally one of desirability, where organic containers of leather and wood were preferred to ceramic ones. It must also be noted that some ‘Roman’ artefacts apparently enjoyed a continuation of use in the sub-Roman period (Cool 2000). It does seem logical, however, to suggest that the scale of production was generally reduced, from one of mass consumption, to a more personal and immediate supply. This certainly seems to be indicated by the evidence from Birch Heath, Tarporley (Ch) (Fairburn 2003). In most instances, though, sites of this date are often indistinguishable from those of other periods, without scientific dating.
The region is commonly perceived to be materially impoverished in the early medieval period, with an almost total absence of pottery throughout the area. Other classes of artefact are usually considered to be almost as rare, except for metalwork from sites with an ecclesiastical, if not monastic, connection. Work in the 1990s has begun to alter this perception, with the recovery of hand-made pottery from an excavation in Cumbria (Oliver et al 1996), and also the identification of its means of production, in a simple clamp kiln. This has, however led to the possibility that at least some early medieval material has been misidentified as of Bronze Age date in the past, given its crude nature. An increasing body of metalwork has also been recovered, more recently recorded through the Portable Antiquities Scheme. The distribution, quantity, and meaning of this material, however, remains poorly studied, but it has the potential to identify unknown sites. The definition of early medieval traditions of hand-made pottery would aid the identification of sites of this date in the field, and an examination of museum collections for misidentified material would be beneficial.
Evidence of copper mining and smelting during this period is sparse on a national scale, and it has sometimes been considered that the raw materials were imported from the Continent (Ponting forthcoming). The evidence for metal working and smithing is confined to iron working, no ore extraction sites being currently known in the region, although they should be anticipated in areas such as Alderley Edge (Ch), the Pennine fringes, particularly the Alston area (C), and the Lake District. If, however, evidence for either Romano-British or medieval mining is identified then post-Roman activity must also be anticipated, since post-Roman societies are equally likely to have needed and exploited the natural resources of the region even though the evidence may not be so easily identifiable.
On a national scale there is an increasing incidence of mixed alloys of non-ferrous metals during the Romano-British period (Bayley 1998, 167), which appears to have continued into the early medieval period. This mixing is thought to be a result of the recycling of existing metalwork. From about the 8th century onwards, however, specific alloys (especially brass) reappear, indicating a new input of ‘pure’ raw materials. Other classes of material would also seem to have been recycled, particularly glass, since cullet is a relatively common find on any site of the period (for instance, at Dacre; Howard-Davis forthcoming). Apart from noting its existence, however, little formal work has been undertaken to forward an understanding of the levels and complexity of the industrial process. In terms of data collection, there is a need for every excavation strategy on any site, whether rural or urban, where early medieval activity might be present, to allow for extensive sampling, both of specific material and bulk samples, not just for palaeoenvironmental material but for small artefacts and industrial residues. Such material, including that not necessarily easily identifiable without sieving, could include hammerscale and droplets of metal, indicating metal working or glass production that has otherwise left no understandable trace in the archaeological record.
A synthesis of metal artefacts from this period within the North West may provide a foundation for the study of both imported and locally manufactured items, the distribution of which may provide an indication of manufacture or trading centres. The contribution of the Portable Antiquities Scheme needs to be acknowledged in the development of regional models for this period. Significant material continues to be reported, particularly in the south of the region, and it seems that a similar pattern is emerging as the scheme establishes itself in the north. The vast majority of these items are dated only by typological analogies, rather than by stratigraphy or associated absolute dates, these analogies for the most part referring to other parts of the country. A detailed classification of the material from the region might help to establish a more localised chronological framework, both for metalwork and other classes of material. Studies of early medieval material culture need to ascertain not just what cultural affiliations the styles indicate, but where and how such items were made. Metallurgical studies of metalwork, whether from metal detecting or from stratigraphically excavated deposits, are also required to investigate this, and to test the contention that there was a general continuation of practice from the Romano-British period until this was replaced by an influx of new raw materials during the latter part of the early medieval period.
Despite the substantial number of sites and quantity of material associated with Roman and medieval salt production now known from Cheshire, there is no firm evidence for production in the early medieval period. Penney and Shotter have speculated (2002, 59), however, that the finding of a bronze pennanular brooch close to the findspot of one of the lead brine pans from Shavington (Ch) suggests that these brine springs may have been the focus for continuing activity beyond the formal Roman period. Given the evidence for a continuing industry to the south at Droitwich (Worcestershire), any further work at Shavington or in centres of salt-production in both the Roman and medieval periods should specifically address this issue. The identification of seaweed at Fremington (C) might also suggest that such material was being transported from the coast in the north of the region (Oliver et al 1996), and highlights the importance of targeted palaeoenvironmental studies as part of any excavation campaign. Other organic material can also point towards industries that are to date little understood. For instance, the recovery of the extensive assemblage of cattle bones from Castle Street, Carlisle (Rackham et al 1991), which had been butchered in a very distinctive manner, must represent an organised and targeted activity, such as, perhaps, tanning. If the deposit dates from the early medieval period, which whilst not proven is a definite possibility, then the material could revolutionise our understanding of activity in former Roman centres at this time. Similarly, the radiocarbon dating to this period of what would appear to be evidence of hemp retting at Glasson Moss (C) (Cox et al 2000) demonstrates the importance of work on the wetlands and also highlights the need for close dating of any features identified, as one of the few ways that industries of the period will be recognised in the archaeological record.
4.41 Raise awareness in those responsible for the excavation of sites that ceramics initially classed as prehistoric might represent examples of a regional tradition of early medieval hand-made pottery.
4.42 Examine any relatively undiagnostic ceramic material currently held in museums and classed as prehistoric, to establish whether any of this could actually be early medieval in date.
4.43 Recognise the continued use of ‘Roman’ artefacts into the 5th century as a topic worthy of further research, and ensure all late Roman assemblages are scrutinised for such.
4.44 Consider controlled metal detecting on archaeological excavations to aid artefact recovery for early medieval sites and those of other periods.
4.45 Enact a full sampling and dating strategy for production residues and deposits on all urban and rural sites, to detect evidence of hammerscale, metalworking or glass production.
4.46 Establish a regular period-based review or publication scheme for artefacts recorded by the PAS, including metallurgical studies, as appropriate.
4.47 Undertake metallurgical studies of early medieval metalwork as a matter of routine.
4.48 Prioritise the absolute dating of key assemblages, such as the animal bone from Castle Street, Carlisle.
4.49 Ensure methodologies for any excavations of late Roman or medieval sites are sufficiently robust to allow for the possibility of early medieval industrial/craft production, including salt.
The western seaboard of the North West formed a major routeway in the early medieval period, with traffic apparently moving between the Continent, the south-west of mainland Britain, Dublin, the Isle of Man, Western Scotland, the Northern Isles, and Scandinavia. Indeed, in many ways the evidence for intercommunication in the so-called Irish Sea Province is more vivid in the early medieval period, particularly in the 10th and 11th centuries, than at almost any other time. Firm evidence to date is, however, very limited in the region, in contrast to that for both the seaboards to the west and south, in Wales, and the north, along the Galloway coast.
In particular, the absence of imported Mediterranean wares of the 5th and 6th centuries is striking, apart from a small quantity from Chester (Carrington 1994), and others recently recognised from Carlisle (Swan et al forthcoming). The exception to this general absence of evidence is that material indicating the presence of Scandinavians, particularly the Hiberno-Norse from Ireland (Griffiths 2004; Higham 2004b), since both place-name and burial evidence exist throughout the region.
Models indicating that a higher relative sea level existed during the Roman period, particularly in the Lancashire/Morecambe Bay area (Tooley 1980), have come under recent scrutiny, and it has been suggested that such a change may actually have occurred in the later Roman or post-Roman periods (Zong & Tooley 1996; Wells 2003). The implication for this is far-reaching in relation to the navigability of rivers and the location of coastal settlement and activity during the early medieval period, although it must be borne in mind that such change could be relatively localised. Whilst much of the coastal part of the region is low-lying, headlands exist in North Lancashire and northern Cumbria, which, on analogy with the North East, could have significance both as trading points and as a focus for early Christianity. Obvious examples of this are at Heysham (L) (Potter & Andrews 1994) and Workington (C) (P Flynn pers comm). The use of GIS mapping to predict changes to the coastline, and from this to highlight areas where settlement might have been likely, is an important tool for the future.
Few early medieval coastal settlements are known between Meols on the Wirral and Whithorn in Galloway, although promontories and headlands, and around natural inlets, harbours and estuaries, are all potential locations for settlements and trading centres. The potential improved navigability of major rivers during the period also increases the likelihood of trading contacts and settlement on river courses further inland. There is no immediate reason to assume that the formal end of payments to the Roman army in the north signalled an end to all coastal transport and trade, and the water routes to former Roman proto-urban sites may well have been navigable for small craft. In this respect, the estuaries of the Esk, Irt and Mite (C), as well as those of the Ribble and Lune (L), may well be significant, since all have evidence of Roman military and civilian settlement close by, as well as early medieval activity in the form of stone sculpture. In addition, those coastal or riverine areas where any late Roman or post-Roman coinage has been found, not just hoards, should be studied.
The artefacts from Meols attest to trading contacts from far afield continuing beyond the end of the Roman period in the North West. Goods and commodities traded and exchanged would undoubtedly have varied widely, including organic and perishable goods, and potentially people themselves. There is, however, evidence of antiquarian selection of material from Meols, given the relative lack of ceramic material within the surviving assemblage (Griffiths 1992). Meols, though, offers a type-site from which, perhaps, predictive modelling of the coastline could be developed, which would aid an understanding of not just the Wirral but the North West more generally.
Those materials most likely to be available to archaeologists are metalwork, glass and amber, and possibly ceramics, although foodstuffs such as grain and animal bone may be recovered from contexts with preferential preservation. Inadvertent contaminants, such as insects, may also be an indicator of the movement of goods. For instance, Kenward (in press) has suggested that the incidence of grain beetles is a good indicator of urban activities, given the bulk storage of foodstuffs needed for a large, relatively static population. In general, these insects are almost absent from sites of the early medieval period, although frequent in both Roman and medieval towns. In Carlisle, such beetles have been found in deposits at the base of a well at Castle Street (McCarthy 1991), the dating of which is controversial.
The huge assemblage of cattle bone from Castle Street (ibid; Rackham et al 1991) is also important, comprising, as it does, material from selected parts of the carcass, which had been butchered in a very distinctive manner. This presumably represents an extremely organised and specifically-targeted activity, as well as suggesting importation from elsewhere, although this may simply have been the surrounding rural hinterland. It is, however, dated only by association, either to the 7th to 10th centuries or the 12th to 13th centuries. Such a degree of organisation is remarkable for any period, and is particularly important if it does relate to the early medieval period.
Some evidence of trading patterns, or lack of movement, may have wider connotations. For instance, the lack of Chester ware from north of the Mersey may indicate that its distribution was limited to the old kingdom of Mercia, which, at the time of the production of this pottery, was rapidly becoming part of the expanding English kingdom. During the course of the 10th century, however, England would have incorporated the area that became Lancashire, and thus a purely political reason for its absence is harder to sustain.
In the absence of urban centres, and their substantial cultural assemblages, until late in the period, and then only in the south of the region, such as at Chester (Carrington 1994), methods of distribution are little understood. Meols stands out as an example of a coastal or beach market, but elsewhere there is slight evidence that church sites might have acted in many parts of the North West as a focus for the exchange of goods, a trend reflected in the tradition of markets being sited at the gates of churches in the medieval period. This can, perhaps, be seen in changes that can be inferred from the artefact assemblages from monastic sites such as Dacre, as well as others in the North East, where there seems to have been a move to a more secular lifestyle, although the religious focus survived (R Cramp pers comm).
It is clear that the monetary economy collapsed at the end of Roman rule in Britain, and presumably a system of barter became prevalent. Whilst coinage was reintroduced in the 8th century, and, indeed, there are indications of a mint in Chester from the late 9th century (Dolley 1970), no physical trace of this, or indeed of other coin production or formal exchange, exists in the region. It is likely that, in addition to barter as the principal method of exchange, redistribution of surplus crops and livestock formed a key element of tribute, or taxation, in the region, an echo of which is seen, perhaps, in the taxation of Cornage, still evident in the 12th century in Cumbria. This was a payment of cattle, a largely royal due paid throughout the old counties of Cumberland and Westmorland, as well as parts of the North East, and represents payments outside the monetary system, which was at this point still not well established in the area (Kapelle 1979).
4.50 Take into account in all coastal survey or research, any former sea level rises and the possible locations of settlement on promontories, river inlets and on water courses, including former channels now in-filled.
4.51 Undertake detailed landscape surveys of headland sites.
4.52 Treat any early medieval coin and artefact finds, not just hoards, in coastal and river-edge areas as potential sites, and monitor such areas closely.
4.53 Develop a GIS to aid understanding of historic changes in sea level, and their effect on possible coastal settlement.
4.54 Develop a programme of sampling and dating of suitable samples, such as those from Castle Street, Carlisle, to enlarge the corpus of insects from the region, and our understanding of the conditions in which they thrived.
The continued occupation of at least some Roman military sites beyond the formal end of Roman governance, such as Birdoswald (Wilmott 1997) and Carlisle (Zant forthcoming), has been demonstrated, but as yet there has been little work at other forts in the region specifically targeted towards the later sequence. There is also as yet little evidence to place such activity in the context of the wider population of the North West. An Iron Age date has traditionally been attributed to upland defended sites or ‘hillforts’, although there is a strengthening argument to place the origins of some of these monuments within earlier periods of prehistory (Edmonds 2004, 176), and a Bronze Age origin has been proven at some sites (Ellis 1993). It is ironic therefore that the only radiocarbon assay from a Cumbrian upland defended site dates to the 6th or 7th century cal AD (LUAU 1999). The position of this dated sample, in the base of a rock-cut ditch, suggests either construction or an extensive overhaul at this time. Data from other similar sites in Cumbria and Lancashire are currently lacking but the potential for early medieval occupation would seem to be extremely high at many sites. There is now, however, increasing evidence for the actual construction and occupation of defended sites in the centuries following the Roman period. Recent work at Heronbridge suggests that an embanked enclosure was constructed to the south of Chester before the 8th century AD (D Mason pers comm). Similarly, evidence from the fortified site at Buckton Castle, Tameside (GM), appears to be pre-Norman in origin (M Nevell pers comm).
In the later part of the period, it is clear from documentary records that sites in Cheshire were being occupied and defended as part of a concerted programme of acculturation, accompanying the development and expansion of England in the 10th century (Thacker 1987). To date, there has been little archaeological investigation of this activity, apart from in Chester, where clearly urban attributes began to develop from this period (Carrington 1994). If this policy was extended northwards in later reigns, then the Domesday reference to six burgesses at Penwortham (Farrer & Brownbill 1911, 57) may be significant, perhaps representing a forward position in the reign of Edward the Elder (Higham 2004a), as may the morphology of Penrith (Winchester no date). The fact that the river Eamont seems to have been a frontier between the Scots and English in the reign of Athelstan (Earle & Plummer 1892) suggests that this, too, could have acted as a forward, defended base. In each case, however, firm archaeological evidence is lacking.
It is unlikely that any political boundary, however important, will have been marked in a formal manner, as for instance was the border between Mercia and Wales by Offa’s Dyke,or the northern frontier of the Roman Empire, defined by Hadrian’s Wall. Indeed, the manner in which Hadrian’s Wall ceased to be an effective boundary is of considerable interest, since it is clear that, by the 7th century, it had no political relevance, since from the reign of Oswald (AD 633-42), the kingdom of Northumbria was said to have held sway over the Picts and Scots (Hunter-Blair 1976). Certainly, it was seen by Bede as an object of historical curiosity only (Colgrave & Mynors 1969).
It is likely that, to a considerable extent, boundaries were defined in natural terms, and therefore the rivers which flow east to west in the region are important. The extent of the polities of the region has long been a source of considerable discussion. Both the Mersey and Ribble have been suggested as the southern boundaries of Northumbria, and it is, of course, perfectly possible that each formed the border at different times. In this context the dating of the Nico ditch, Greater Manchester, is of importance (Higham 1993), although it is possible that it had no significance in terms of a political boundary. Clearly, the cultural differences between Cumbria and north Lancashire in the north and Cheshire in the south, as reflected by the stone sculpture of the period and linguistic differences, indicate that the border between the two Anglo-Saxon kingdoms lay somewhere in Lancashire, at the point at which the differing traits were developing, presumably in the 8th to 9th centuries. In the 10th century, it seems likely that, at least for a short time, the Eamont/Eden corridor near Penrith formed the border between the expanding English kingdom and the lands to the north, since Athelstan met his northerly neighbours there (Earle & Plummer 1892), such meetings frequently occurring at such borders. As late as the 12th century the bishop of Glasgow was claiming influence over the land as far south as the Rey Cross on Stainmore (Collingwood 1927, 3), lending support to the importance of this river system at some point in the development of the two major nations in Britain.
Despite the region having clearly been in a state of political flux throughout much of the early medieval period, few, if any, battle sites have been identified with any degree of confidence. Bromborough remains a strong candidate for the Battle of Brunanburh in AD 937 (O’Hanlon 1986), although the candidates for the site of this battle range as far afield as the Humber (Kirby 1975) and Burnswark in Scotland (Anderson 1922). Several other possible battle sites are listed in the various SMRs and HERs, but the provenances are loose and all are based on tradition, although place-name evidence perhaps points more strongly than most to the Battle of Arthuret (Armterid; AD 573) having been fought in the vicinity of Arthuret or Netherby in northern Cumbria (McCarthy 2002). The recent excavation of a mass grave at Heronbridge (Ch), some five miles to the south of Chester, produced skeletons that have been dated to cal AD 430-640 and cal AD 530-660 (D Mason pers comm). This has indicated the potential to link historical references to physical evidence, as such a large number of bodies buried at once must surely reflect a catastrophic event. Given the apparent injuries to many of the bodies, one candidate at least could be the Battle of Chester of AD 615/6.
4.55 Ensure excavation methodologies are in place to recognise and characterise post-Roman activity in Roman forts.
4.56 Ensure all opportunities are taken to date material from both the defences and interiors of hillforts, and from potential burh sites in Cheshire.
4.57 Obtain appropriate samples from the Nico ditch to provide radiocarbon dating.
4.58 Examine the evidence for Arthuret being the site of a significant battle.
The early medieval period in the North West, as in much of the rest of England, is largely archaeologically invisible, and is perhaps the least understood period, in purely archaeological terms, of any in the region. Indeed, until recently, without the few documentary references, the ‘Dark Ages’ would have been utterly black. Absence of evidence, though, is very far from evidence of absence, nor does it imply an impoverished society, fallen on hard times following the collapse of the ‘sophisticated’ Roman lifestyle. With the growth of absolute dating techniques, the development of methodologies which encourage the recognition and recording of all archaeological features, and perhaps also a growing interest in the period, sites are beginning to be recognised and information about these centuries is accumulating rapidly.
In the nostalgic literature coming out of Wales from the 9th century, the north of what is now England was an heroic place in the 5th and 6th centuries, with a sophisticated society moving forward from its Roman antecedents. Similarly, in literature and in history, the floruit of the Anglian Kingdom of Northumbria was seen as a golden age, when art and architecture, literature and learning flourished, and some of the most important legal and political institutions of the period developed. As this kingdom declined, so the might of the Kingdom of Mercia grew, and their kings, in particular Offa, assumed the mantle of Bretwalda or over-king, from Northumbria. The North West, then, was inhabited by societies that were seen by their contemporaries as important, and were looked back on with nostalgia: yet this apparently vibrant society is hardly glimpsed in the archaeological record. The region was, however, only a part of these wider political entities, and it would be limiting in the extreme to see the North West in isolation, without reference to the North East and Yorkshire, and the Midlands. Thus, cross-regional comparisons are essential to further an understanding of the development, not just of the North West region, but of the nations of the island. For the early medieval period is the time when the nations we know today as the United Kingdom really formed, and whilst the political boundaries, both ecclesiastical and secular, only become clear in medieval documentation, there is little doubt that some, at any rate, are ancient. This is reflected in the archaeological record in the differences between Cumbria and Cheshire, one firmly in the northern and Northumbrian aegis, whilst the other looked to the south and Mercia. Lancashire is therefore crucial to an understanding of the ebb and flow of these polities, and yet it is precisely there that the record is lacking, or the hardest to interpret.
The position of the North West, with national boundaries to both the north and west, makes ethnicity a live issue, particularly in the early medieval period, where there is both historical, linguistic, and archaeological evidence for different cultural influences. A strong element of this reflects the position of the region on a seaway from the Mediterranean through to Scandinavia, which brought material from both north and south, as well as from the west across the Irish Sea. In many ways, whilst the record is certainly biased at present, those parts of the coast that do not show evidence of such contacts are equally interesting as those that do.
There is clearly much work needed to create a coherent framework for the period. In particular, it is increasingly clear that information exists, whether in the Portable Antiquities Scheme records, within museum collections, or in unpublished archives, that would benefit from re-evaluation, and bringing into the public domain. Targeted projects will begin to shed light on certain types of site, particularly if they are coupled with absolute dating. It is now clear that no potential site of the period should be excavated without a programme of absolute dating, as it is only through this that assumptions can be tested. However, it is equally clear that as many of the key pieces of evidence will be gained in a serendipitous fashion, during the excavation of sites of other periods, and therefore the ability to recognise such material is equally crucial.
Fig 4.1 Pottery kiln under excavation at Fremington, Cumbria (OA North).
Fig 4.2 Fragment of decorated stone displaying ‘Pictish style’ zoomorphic carving recovered from Scordale, Cumbria (Richard Newman).
Fig 4.3 Principia building within the Roman fort at Carlisle at the end of the 4th century (OA North).
Fig 4.4 Possible early medieval building under excavation at Shap, Cumbria (Cumbria County Council).
Fig 4.5 Part of a probable early medieval bow-sided building at Irby, Wirral (National Museums Liverpool).
Fig 4.6 Cross Hill, Thingwall, Wirral (National Museums Merseyside).
Fig 4.7 8th – 9th-century pectoral cross from Faddiley, Cheshire (PAS).
Fig 4.8 Cross shaft at Halton, Lancashire (Lancashire County Council).
Fig 4.9 Potential Viking grave group from Meols, Wirral (National Museums Liverpool).
Fig 4.10 Late 8th – 9th century Irish hanging bowl mount from Arnside, Cumbria (PAS).
Fig 4.11 Pins and ‘purse clasp’ from Dacre, Cumbria, dating from the 9th to 11th centuries (OA North).
Fig 4.12 Early medieval silver sword pommel from Mouldsworth, Cheshire (PAS).
Fig 4.13 Thingmount at Little Langdale, Cumbria, (William Rollinson/LDNPA).