Understanding the British Iron Age
An agenda for action
This paper assesses the overall research priorities for British Iron Age archaeology in the coming decades. Discussion is organised around five general themes: within each of these, we set out particular topics on which research is essential or should be encouraged, based on current understandings of the period. Each section is followed by a summary of the key points and recommendations for specific changes in current archaeological practice. Given the new research opportunities brought about by developer-funding, the resultant framework will, we believe, be of particular value in supporting curatorial decisions, as well as helping to shape more detailed research agendas for the period formulated at regional level and articulating these with research issues at national level.
This is a Draft Report of a Working Party of members of the Iron Age Research Seminar. The Convenor was Colin Haselgrove (University of Durham); and the members were: Ian Armit (Queens University Belfast), Tim Champion (University of Southampton), John Creighton (University of Reading), Adam Gwilt (National Museum of Wales), JD Hill (British Museum), Fraser Hunter (National Museum of Scotland), Ann Woodward (Birmingham University Archaeological Field Unit).
Olivier 1996) in order to encourage the development of regional research strategies covering the whole of England. All sides of the profession see these as essential to provide local authority archaeologists with a framework for making judgements about the relative importance of threatened sites and to maximise the new research opportunities which PPG 16 has created. Since the 1996 initiative, archaeologists in England have been engaged in detailed resource assessments as a preliminary to formulating locally-relevant research frameworks (e.g. Glazebrook 1997) and there have been similar discussions in Scotland and Wales. Whilst significant progress has been made in defining future regional priorities, these discussions have also highlighted the critical issue that no single area can be considered entirely in isolation. The archaeology of each region contributes to a wider spatial and temporal picture, which in turn helps define what is unusual or exceptional about a given region or locality.
The need for national research frameworks to inform regional research strategies and vice-versa is endorsed by both English Heritage and Historic Scotland, who have issued general statements of their own (Barclay 1997; English Heritage 1998), although CADW has yet to produce a similar document for Wales. However, the wide remit of such papers makes them an inappropriate place to set out more detailed priorities for specific periods or themes. These can only emerge from archaeologists working in the particular subject area, but the foundation which exists for achieving this is very variable. In some fields such as Roman archaeology, there are period societies to sponsor such initiatives and recent statements on which to build (e.g. SPRS 1985). A new agenda for Roman Britain is in preparation, based on a day of presentations at the 1999 Roman Archaeology Conference (James and Millett forthcoming).
In contrast, discussion of Iron Age research priorities has generally been within the context of wider reviews, encompassing the whole of prehistory (e.g. Prehistoric Society 1988). For a detailed statement, we have to go back over 50 years (CBA 1948). A group of us working on the British Iron Age decided that the time had therefore come to undertake a detailed review of the subject and identify key research priorities for the 21st century as well as possible strategies for taking these forward at local, regional and national level. This initiative was endorsed at an Iron Age research seminar held in Cardiff in November 1997 and a year later an early draft of the present paper was discussed at a second research seminar in Sheffield to which both specialists and non-specialists were invited. The present text owes much to the detailed comments made by other participants during and after the Sheffield meeting.
A key aim of this document is to provide support for local curatorial decisions. Since PPG 16 was introduced most excavations on Iron Age sites have been as a result of development and this state of affairs is set to continue, yet the responsibility for recommending what is to be recorded or preserved generally resides with planning archaeologists who are not specialists in the period and have limited time to keep abreast of changing ideas in the many different fields with which they have to deal. In recent years, several volumes have appeared which challenge traditional views of Iron Age Britain (e.g. Hill 1995a; Champion and Collis 1996; Gwilt and Haselgrove 1997; Bevan 1999), laying stress on the complexity and diversity of the archaeological record created by its inhabitants. We cannot however expect developers to pay for recording threatened sites at the level of detail which these new ideas about the Iron Age demand unless local government archaeologists are made aware of the relevant issues and provided with justification in a form which they can use to inform the planning process. Of course, ensuring that rescue excavations contribute the maximum amount of new knowledge is only one aspect of research; giving new meaning to existing data and exploiting the mass of Iron Age material which already lies in museums and excavation stores are just as important.
The paper focuses on five themes which we feel are of particular importance as Iron Age studies stand today. These are chronological frameworks; settlement patterns; material culture; regionality; and socio-economic changes during the period. For each of these strategic themes, we assess current knowledge and seek to identify specific research priorities; where appropriate refinements for collecting and analysing data are also suggested. We have concentrated on the building blocks and have not discussed problems with the models used in interpreting Iron Age societies, although a lack of concern with the 'big' issues of Iron Age social structure and political institutions has been noticeable in the last 15 years. So too is the lack of regional syntheses of the mass of new data which field archaeology is producing, which has become a major academic priority. A number of other topical areas of enquiry have received relatively little recent attention from Iron Age specialists; some like landscape are stressed below; others which we might expect to see debated more extensively in the future include ethnic identity (cf. James 1999) and gender. What potential coastal and maritime archaeology has for the Iron Age remains to be seen, but should not be ignored.
Our discussion covers England, Scotland and Wales, but not Ireland,
and is primarily directed at the period between the eighth century BC when
iron first came into common use in the British Isles (Needham
et al. 1998) and the first century AD, although many of the points
made here apply with equal force to Roman Iron Age communities outside
and indeed inside the Roman province, as well to the later Bronze Age.
The final section of the summarises the archaeological, financial and human
resources available for translating the research agenda into action and
suggests some ways of making good current deficiencies.
In many areas of Britain, we are faced with a profusion of sites and site types which we believe to be Iron Age, often on the basis of limited and potentially ill-founded parallels with sites in other regions, but where there is a total absence of any internal dating framework. A prime example is south-east Perthshire where intensive ground survey and aerial photography has produced a dense distribution of putative Iron Age sites, but where whole categories of monument (e.g. the interrupted ring ditches, thought to relate in some way to both ring-ditch houses and souterrains) are entirely undated. Other major classes of site, such as the Cornish rounds or the duns of Argyll, can be broadly defined as Iron Age settlement types, but lack sound internal chronologies.
Even in artefact "rich" areas like Wessex and south-east England, we often overlook how dependent the absolute dating is on a few key sequences and diagnostic artefact types. The existing, essentially ceramic-based, chronology relies heavily on the proposition that broadly similar regional assemblages were in use at the same time. The only thoroughly independently dated sequence is that from Danebury (Cunliffe 1995a), but this only begins in the early 5th century BC and the existence of one entire ceramic phase (cp 6) is still a subject of some contention. Supporting this framework are some 40 usefully stratified brooches and 60 radiocarbon dated assemblages from other sites. The numbers sound satisfactory enough, until the evidence is analysed region by region. Even if we are correct that similar style-zones were broadly contemporary most of the time, what of the exceptions where change in one region does not follow another? Identifying such disjunctions is important if we wish to analyse inter-regional relationships with any degree of subtlety, but must start from independently constructed local chronologies (e.g. Knight forthcoming) if we are to avoid circular arguments. The apparent persistence of handmade 'middle Iron Age' pottery traditions into the Roman period in parts of southern and eastern England, without an intervening 'late Iron Age' phase defined by wheelmade pottery, affords a good illustration of this point.
Past typological assumptions inescapably pervade modern interpretations of chronology and sequence with the result that schemes founded on different categories of artefact are often difficult to reconcile. The divergence between the brooch and pottery dating for the King Harry Lane cemetery has received particular attention (e.g. Mackreth 1994), since it directly affects the interpretation of the burial rituals, but many settlement assemblages display similar problems. Potentially crucial correlations with the European material also need to be addressed, as this is one of the major sources of the absolute dates used in Britain. The consequences of the earlier dating for continental La Tène D1-D2 are still being assimilated, but the implications for insular late Iron Age chronology are considerable (Haselgrove 1997, 56-58): evidently cremation was established in parts of southern England before the mid 1st century BC (Fitzpatrick 1997a), whilst the adoption of other Aylesford-Swarling traits was probably a long drawn out and selective process. An earlier dating for the inception of the East Yorkshire Arras burial tradition has also been gaining ground in recent years, although this is due more to a shift in intellectual fashion over how quickly typological innovations could travel than to new data (Collis 1994). Issues relating to the adoption of new artefact types and the spread of ideas among European communities cannot be adequately addressed without an independent dating framework.
Chronological issues are also treated differently across Britain. In
Scotland, it is common to view the Iron Age as part of a much longer period
of development than is traditional further south, with the 'long Iron Age'
being seen to continue until at least the Norse incursions. This trend
towards integrated study across inherited chronological boundaries is epitomised
by the founding of the First Millennia Studies Group as a forum for discussing
Scottish archaeology between 1000 BC-AD 1000. The interlude of interaction
with the Roman army is increasingly seen as insufficient reason to suspend
study after the 1st century AD, given the manifest continuity of Roman
and post-Roman Iron Age communities from their predecessors. Similar arguments
for continuity between certain late Bronze Age practices and the early
Iron Age in southern Britain (cf. Champion 1999)
have led many to argue that the first millennium BC, rather than the Iron
Age is a more appropriate unit of study. In Wales, however, the traditional
Iron Age and Roman divide has tended to remain in place, although the situation
in many areas between the 1st and 4th centuries AD was evidently not dissimilar
to that in southern Scotland.
An initial attempt at such an audit for Atlantic Scotland highlighted the pervasive influence of interpretations of the first few radiocarbon dates, made before many of the method's limitations were fully understood (Armit 1991). In the south, the longer chronologies advanced for the South-Western decorated wares and East Midlands Scored Ware rest on an equally small number of determinations (Cunliffe 1991), which might yet be open to re-interpretation. Such an audit can also serve to indicate the areas and site types in most urgent need of secure dating.
B2.2.1 Radiocarbon as routine
Whilst dating programmes must recognise the problems of the calibration curve between c. 800-400 cal BC, we should not over exaggerate them. Using a large enough number of AMS dates and multiple samples can overcome many of the difficulties. Where a suitable stratigraphic sequence exists, application of Bayesian modelling offers significant gains in precision (25-35%) for dating particular events or phases (Bayliss 1998). In many regions, using radiocarbon dating simply to differentiate earlier Iron Age sites or occupation phases from those of later Iron Age (or late Bronze Age) date would be a major advance.
B2.2.2 Single entity dating
B2.2.3 Organic residues
B2.2.4 Human remains
B2.2.5 Reporting of dates
B2.2.6 Targeted fieldwork
Excavation of limited sections of hillfort 'defences' is currently deeply
unfashionable. Yet, when a specific threat (rampart-loving rabbits) forced
this strategy to be adopted at the Brown Caterthun, the dating evidence
accrued transformed our understanding not only of the site, but, by implication,
of the Strathmore Iron Age (Dunwell and Strachan forthcoming).
Other similar programmes could produce enormous, yet cost-effective, advances
in our chronological understanding.
B2.3.1 The 'art' corpus
B2.3.2 'Classic' sites
TL dating is becoming increasingly more refined and Barnett (forthcoming) has shown its potential for dating first millennium BC ceramics in eastern England. Despite the wide error ranges, the use of multiple determinations from sites and single contexts offers considerable potential for close dating. It should be possible to build up a large enough corpus of dates to estimate the duration of more widespread fabric traditions by statistical means, since few Iron Age sites produce no pottery at all (cf. Willis 1999). It remains essential that wherever possible the dates are anchored in typology.
Dendrochronology has been used at wetland sites such as Fiskerton and
Goldcliff (Bell and Neumann 1997); at several Scottish
crannogs; and for dating the Hasholme boat (e.g. Hillam
1992). It is only a matter of time before Iron Age sites with extensive
artefact assemblages as well as suitable timbers for dating come to light.
Despite the obstacles encountered, the Somerset lake villages are obvious
candidates (Coles and Minnitt 1995). Wetland sites
also hold out the promise of combining radiocarbon and tree-ring dating,
producing further gains in precision, exemplified by the dating of the
waterlogged early Bronze Age timber circle uncovered recently at Holme-next-the
Sea in Norfolk (Bayliss et al 1999).
Since PPG16, the number of Iron Age settlements and other sites investigated archaeologically in one form or other (whether by fieldwalking, geophysical survey, evaluation or small-scale excavation) has increased dramatically, although differential pressures of development have ensured that this new knowledge is unevenly spread across the country. The approach has been largely site-specific, however, with less concern for wider landscape contexts and settlement patterns. With some exceptions (e.g. Ferrell 1997), there has been little attempt to interpret overall patterns of settlement. Fields have not proved a popular target for study (exceptions include Drewett 1982; Pryor 1996), while compared to earlier prehistory, the symbolic and social aspects of Iron Age landscapes have received more limited attention, although there are exceptions (e.g. Bradley et al 1994; Bevan 1997; Taylor 1997).
One lasting impact of the excavations in the 1970s and early 1980s was the revolution in our knowledge of Iron Age farming. Key studies of animal and plant remains from Wessex and the Upper Thames Valley were in the forefront of developing these areas of bioarchaeology on a global level (e.g. Grant 1984; Jones 1984, 1985; Maltby 1981; 1985; Wilson 1978), for the first time providing a detailed picture of Iron Age herd composition, age profiles and culling strategies; crop types and husbandry regimes; and coppicing patterns. Significant contributions also came from pollen and soil studies (e.g. Wilson 1983; Robinson 1984; Bell 1996, Tipping 1997), exposing the major intensification of agriculture in the later Iron Age. This work also helped raise issues of wider impact, animal bone studies leading, for example, to the investigation of ritual deposition and the symbolic organisation of settlement space, while palaeo-botanical analysis (e.g. Jones 1984; 1985; Van der Veen 1992) provided new ways to consider the large scale organisation of agricultural systems - and, by extension, the political articulation of social groups - as well as ideas about how specific farming regimes and innovations related to the strategies which family groups employed in competitive social systems. Recent research on Iron Age agriculture is summarised in various papers (e.g. Jones 1996; Maltby 1996; Van der Veen and O'Connor 1998), but has yet to be adequately linked with work on social and symbolic aspects of the period.
In the last fifteen years, our approach to interpreting settlement layouts and artefact deposition has significantly altered. In the 1970s, techniques of spatial analysis to define functional activity areas were prevalent, although rarely explicitly employed (e.g. Clarke 1972; Fasham 1985, 127-134). Subsequently, certain workers began to question the very nature of deposition in the Iron Age and to suggest that the layout of settlements was integral to the values and dynamics of the groups who lived in and through them (e.g. Barrett 1989). Work in Atlantic Scotland showed how the syntax of settlement space provided evidence for changes in social organisation (Foster 1989) and has since been extended to consider the monumental (display) and ritual (symbolic) aspects of the Iron Age house (e.g. Armit 1989; Hingley 1992; 1995; Sharples and Parker Pearson 1997).
Equally radical changes have taken place in our understanding of southern British sites. Pit and ditch deposits are no longer seen as random dumps of rubbish (e.g. Cunliffe 1992; Hill 1995a, Chs.8-9; Parker-Pearson 1996). Often, contents were deliberately placed, with different artefact types occupying distinct areas of a site, reflecting ideological and cosmological referencing. The orientation and layout of houses is now seen as heavily influenced by cosmology (e.g. Parker-Pearson 1996; Fitzpatrick 1997b; Oswald 1997), while the practical functions of the banks and ditches that surrounded Iron Age settlements and hillforts have been questioned (e.g. Bowden and McOmish 1987; Sharples 1991a; Hill 1995b). The impact of these studies in transforming our perspective of Iron Age societies cannot be underestimated. Nonetheless, this work has relied heavily on data from Wessex and Atlantic Scotland, and we need to consider the nature of depositional practices and houses in other regions and through time. Equally, now that ritual and symbolic interpretations are in vogue, it is important to recognise that not everything on an Iron Age site is explicable as 'ritual'. The identification of such aspects always needs to be argued through in detail.
Future developments include the idea that settlement and landscape space
can be viewed in relation to the way the human body moves within that space.
This archaeology of 'inhabitation' has been explored in an innovative presentation
of the structural and artefactual data from Alcock's excavations at Cadbury
Castle, Somerset (Barrett et al. forthcoming). A
logical development from these ideas is an analytical approach based on
the human senses of sight, hearing, taste, smell and touch like that used
on the large settlement excavation at Crick, Northamptonshire (Hughes and
Woodward in preparation). Similarly, Giles (in preparation) has used the
abundant evidence from East Yorkshire (e.g. Dent 1982;
1991a), to extend the human body approach beyond the occupation areas
to the fields and burial places, opening up possibilities for deciphering
the 'inhabitation' of whole landscapes.
It must not be assumed that published sites have been dealt with and 'finished'. Reports are increasingly selective in both content and approach. Vast banks of uninterrogated data lie in museum archives, which need to be continually researched and reassessed in the light of developing theory and new methodologies. Museum collections also contain substantial repositories of unpublished finds and records relating to many key first millennium BC sites, such Beckford, Ham Hill, Moel-y-Gaer, Mucking and Traprain Law. It would also be beneficial to re-evaluate many sites that were published between 1935 and 1970. This could be achieved by the selective reworking of artefacts and records, linked to new excavations to retrieve specific data such as environmental remains and samples for dating (as was done in the 1980s at Maiden Castle; Sharples 1991b). Key sites which might profit from fresh work include All Cannings Cross, Little Woodbury, Merthyr Mawr, and West Harling, as well as hillforts such as Bredon Hill and Traprain Law (where a small-scale programme has recently started; Armit et al 1999).
If their potential for interpreting life in the Iron Age in new and exciting ways is to be realised, sites excavated ahead of development need to be investigated and analysed according to some stringent and novel guidelines, developed in partnership with curators and contractors. Two main areas of innovation are required: first, in relation to sampling fractions as specified in project briefs; and second, regarding the analysis and publication of finds assemblages. In general, sample fractions need to be rather larger and far more flexible than at present. Site briefs should contain suggested sampling fractions for the different types of context expected, ranging for instance from total excavation of graves, preserved floor levels, structural elements of buildings and key placed deposits to lesser fractions for gullies and ditches; the figure of 2% adopted for many field evaluations has no reasoned basis (Williams 1999) and is patently inadequate. A suggested minimum for hut gullies and enclosure ditches is 20% evenly spaced around the circuit and this should always include the terminals (Hill forthcoming). Key intersections must be excavated to facilitate site phasing, but as many other segments as possible should be excavated to recover a sufficient number of single-phase context assemblages.
Bulk sampling for botanical remains and sieving for animal bone and artefacts should be routine requirements in briefs for potential Iron Age sites, supported by scientific techniques such as phosphate analysis, magnetic susceptibility and soil analysis. While the quantities of finds are generally going to look small compared to later periods, maximising their retrieval is essential to define the regionally-specific practices around which Iron Age social relations were evidently articulated. It is also imperative to look beyond visible settlement boundaries (Haselgrove 1999b). At many 'enclosed' sites, pits, structures and even burial areas have been found outside the ditches, while external patterns of deposition (ritual and otherwise) frequently seem to differ from those in the interior. In some cases, ditched enclosures can be shown to form one element of a larger inhabited area, while numerous others constitute phases in a much longer sequence of open and enclosed settlement. There are also indications that the extensive 'open' settlements found in certain areas of eastern England are actually a product of frequent locational shifts by relatively small communities (e.g. Willis 1997; Hill 1999), although some may be genuine nucleated settlements. Non-invasive survey techniques clearly have an important role here, both for planning excavation strategies and for providing information about the extent of the inhabited area (e.g. Clogg and Ferrell 1991; Bewley 1994; Biggins et al. 1997).
The second key area of innovation concerns the analysis and publication of excavated finds. Discussions of assemblages needs to be more contextual and less compartmentalised into specialist categories than at present (D1.1 below). As well as obvious topics like exchange and interaction with other groups, headings for exploring the data more holistically might include bodily adornment; identity; food and feasting; rubbish deposition; deliberate discard; exchange between groups; and decoration and colour; all of which contribute to our understanding of social and ritual practices. Minimum levels of quantitative and contextual data are essential to allow easier comparisons of finds assemblages and for others to consider issues of structured deposition and spatial organisation. Above all, the basic finds catalogues need to be properly cross-referenced to information about their phase and context. Such changes, which will affect recording systems as well as post-excavation procedures, have significant financial implications, but these will be far outweighed by the quality of information gained.
Many briefs and specifications do not currently include explicit project
research designs or statements of intent. These are essential for all sites.
In cases where large zones of potential first millennium BC landscape are
due to be developed, detailed and probably multi-authored research designs
must be commissioned. Recent examples of such an approach are the brief
for Crick - prepared by Northamptonshire Heritage (the curator) in consultation
with a group of academic consultants, and developed by BUFAU (the contractor)
- and the research design for the new Heathrow Terminal 5 (Andrewes
and Barrett 1997).
Most Iron Age settlements were farmsteads, most Iron Age people were farmers, and farming formed the basis of Iron Age societies. Although archaeobotanical and archaeozoological studies are offering more sophisticated elucidation of Iron Age agricultural regimes and their variation in space and time (e.g. Jones 1996; Hambleton 1999), this work is only loosely articulated with research on other aspects of material culture and society. A more inclusive approach is required, which would transcend the normal separate reports on the animal and plant remains. One answer is to develop an agrarian sociology for the Iron Age. Animals provided meat, but just as important were their roles as draught and pack animals, and as symbolic wealth (cattle and horse); as raw materials for clothing (leather, skins, wool, sinews); as providers of dairy produce and manure (cattle and sheep); and as aids to herding and hunting (dogs). Plants provided staple foods, but also fodder; fuel for the ever-essential fires and ovens; and bedding and roofing material (heather, straw, peat, turf). Managed woodland provided resources for building, fencing and hurdling, wheels and carts. Fields and boundaries were developed in abundance; pasture was maintained within fields and as controlled grazing; both animals and plant products moved around the landscape; and seasonal exploitation has been demonstrated. Such themes need to be welded together to produce a new picture of farming life in the Iron Age. How was the daily work schedule arranged, and how did it vary with the seasonal cycle? How much time was devoted to various crafts and processing activities, and how was labour apportioned by age, class and gender?
We have already stressed the need to think (and excavate) beyond the site. A dynamic sociological approach to Iron Age farming is dependent on detailed understanding of how the landscape around settlements was used. One aspect is mapping and excavating field systems and other boundaries, as well as neighbouring habitation sites. But recent work is recognising an increasing range of 'non-settlement' components of the Iron Age landscape. These include sites with clearly identifiable functions, such as salt production, quarrying, iron smelting and shrines. Others were probably temporary camps and shielings connected with seasonal movements of people and livestock over potentially long distances, like the buildings recently excavated at Goldcliff and other sites on the Severn Levels (Bell and Neumann 1997). Yet others comprise isolated wells, shallow pits, or other features with few finds, often ill-defined and usually over-looked, but important evidence for landscape utilisation. These uses of the near and faraway landscapes were all part of the way in which Iron Age peoples perceived and understood their world.
Symbolic aspects of the landscape need to be considered as well. As several recent studies have shown, the patterns of structured deposition seen on settlements extended across the landscape (e.g. Fitzpatrick 1984; Bradley 1990; Hunter 1997). The placing of weapons and other metalwork in rivers and bogs; the concealment of torcs, harness and coins (singly or in hoards) in dry parts of the landscape; and ritual deposits of pottery and bone in wells or ditches were all important aspects of Iron Age behaviour. It is also clear that there are more Iron Age burials outside settlements than hitherto appreciated, making radiocarbon dating of unaccompanied inhumations essential (B2.2.4 above). Other work has emphasised the importance of earlier monuments as ritual foci and for the laying out of agricultural landscapes (e.g. Hingley 1999; Gillings and Pollard 1999). Greater recognition and more careful study of all these 'off site' activities in their immediate and wider landscape contexts is of vital importance, as is the integration of this evidence with environmental data in order to understand fully how specific settlements 'worked' in their social landscapes.
In order to achieve a more mature understanding of Iron Age landscapes,
the results of new excavations of settlements and other loci of activity
- sampled and contextually analysed along the lines recommended above -
and data obtained by reworking old assemblages and archives, need to be
combined with the strategic study of fields and boundaries in compact landscape
blocks. Such study areas could usefully include the environs of major hillforts
such as St. Catherine's Hill or Traprain Law, to set beside the results
already obtained from Wessex sites like Danebury and Cadbury Castle; an
upland farming landscape, perhaps in Northumbria or Wales; a zone on the
fringe of Salisbury Plain; or a river valley and its neighbouring hills
- a promising candidate might be the Bredon Hill forts and the gravel terrace
sites like Kemerton and Beckford below. It is here that evaluations have
become the most important research tool in contemporary archaeology. On
their own, the small areas exposed and finds assemblages may seem inconsequential,
but imaginatively combined they provide a sample which can be used to plot
settlement patterns, human presence and social change across whole landscapes
(e.g. Yates 1999). Compared to the more intensive
excavation strategies recommended above, such projects would be relatively
inexpensive, but the research rewards would be far-reaching indeed. On
a more general level, basic mapping and analysis of the aerial photographic
archive, similar to that undertaken by Stoertz (1997)
in East Yorkshire or by Whimster (1989, figs 22-28)
in the Welsh Marches, remains a pressing need in many areas.
Most of the points made regarding the development and implementation of project briefs for settlement investigations and the need for holistic contextual analysis and publication (C2.1 above), apply equally forcefully to mortuary evidence, whether recovered as 'formal' cemeteries, or in other types of context. Where burial grounds are discovered, it is important to look beyond the actual graves, for instance for associated structures or pyre sites which may provide insights into earlier stages of the funerary rituals, as at Westhampnett (e.g. Fitzpatrick 1997a). As this and other recent studies (e.g. Pearce 1997; Parker Pearson 1999) have shown, detailed contextual analysis of body treatment and spatial patterning can reveal a great deal about regional age, sex and status differences, as well as local descent systems and kinship organisation, and the wider cosmological and symbolic principles underlying the relationship between the dead and the living in different parts of Britain.
There is evident need for research into the location of Iron Age burials and how these relate to other components of the settlement pattern. Outside East Yorkshire, where the Arras rite barrows produce readily identifiable cropmarks, Iron Age cemeteries and burial places are difficult to detect and are mostly chance discoveries. In south-east England, as well as in other areas like Cornwall or East Lothian where the numbers of known Iron Age burials have gradually mounted over the years, GIS analysis of their landscape setting might well prove worthwhile. Any patterns which this revealed could be used to predict the zones where Iron Age burials are most likely, so that evaluations could be carried out ahead of development. Other contexts which merit particular attention include settlement boundaries and the vicinities of settlement enclosures; the landscapes within and around earlier Neolithic and Bronze Age monuments, many of which produced evidence of re-use as burial places in the Iron Age (e.g. Warrilow et al. 1986; Murphy 1992), and caves (Savile and Hallén 1994).
The potential of DNA analysis for Iron Age populations also needs to
be explored. DNA studies ought, for example, eventually to be able to determine
if the individuals buried within sites like Danebury were closely related
to one another, or genetically heterogeneous - as might be expected if
it was outsiders who received such treatment. Extracting DNA from British
and continental burials could provide genetic data relevant to current
arguments over the ethnic identity of insular Iron Age peoples (e.g. James
1999), and might help to resolve long standing controversies such as
whether or not the Arras burials of East Yorkshire, or the Aylesford burials
of south-east England, included a significant immigrant element.
While basic data gathering is by its nature detailed and specialist,
the interpretations which can be built on it should be both exciting and
contribute to general understanding of the period. We must not forget that
in terms of public perception, it is often objects rather than sites which
bring the past to life - and therefore justify the archaeological pursuit.
This section will first consider data collection and other basic issues
of artefact research, before moving on to examine some key aspects of the
study of artefact life cycles and priorities for individual material types.
Attempts to define standards for post-excavation analysis, while worthy, are often flawed by a lack of pragmatism and flexibility in their application. However, there is a need at least to recognise minimum standards of recording and publication where specialist groups have defined them, as for example with pottery (PCRG 1995) and the special case of coins (Harrison 1992). For most other materials, there are no accepted standards. This is not a particular problem, provided efforts are made to publicise and expound best practice, as defined by the relevant specialists (D1.2 below).
A key issue is the ability to quantify (C2.1 above). Examples from pottery studies (e.g. Evans 1995) show the great value of simple descriptive statistics for making comparisons and revealing patterns. This does however require data to be easily available from published sources. As microfiche falls from favour there has been a tendency to relegate detailed listings to archive, imposing severe limitations on the subsequent use of the data. If print or fiche is truly unfeasible, then developing technologies such as the Web (e.g. Sharples 1998, 10), CD-ROM, or the electronic repository of the Archaeology Data Service should be used. Paper archives are still required, but electronic archiving provides the potential for access which current archiving policies deny. Improvements in the archiving of the actual material (and the accompanying records) in museums and stores are also essential if future researchers are to make the most of its potential; this clearly has cost implications both for the preparing of archives for deposition and for their long-term curation.
Another role of material culture studies is to aid site interpretation. The structure of the report is crucial to this. At present, too many reports leave integration of site and artefact data to the user. The ground-breaking report on the Colchester finds showed that a function-based structure was both possible and valuable (Crummy 1983). Whether this is done within or across materials depends on the assemblage involved, as certain categories remain resistant to easy functional classification. By involving function, a report is immediately made an active part of site interpretation. It also forces a more holistic view of finds, rather than one compartmentalised between specialisms. The results of using multivariate correspondence analysis to compare Roman assemblages from York (Cool et al 1995) show a promising way forward.
Not all finds arise from excavation. A key consideration in rescue terms
is keeping a closer watch on river widening or dredging schemes. Old finds
from the Thames, Trent, Witham and Tay show the lurking potential of votive
deposits (Fitzpatrick 1984), while work in France
illustrates the rewards awaiting focussed research (Bonnamour
and Dumont 1996). Similarly, although strategies to deal with the explosion
of information produced by metal-detecting are emerging - at least in England
and Wales - there remains a need for a more proactive role in encouraging
finds reporting in many areas. Only with coinage has substantial use been
made of the resulting data (e.g. Haselgrove 1987;
1996). Yet these datasets often represent considerable quantities of
largely untapped information, which call into question existing excavation-based
pictures of relative Iron Age wealth and poverty. Systematically recorded,
metal detector finds have the potential dramatically to change understandings
of the period. At present, however, the majority of finds which are reported
remain unpublished. Here, the onus is surely on national and local museums
to publish at least a basic record of recent finds. If an excavation backlog
is unacceptable, so too is a backlog of finds. Outlets such as Discovery
and Excavation in Scotland are invaluable in this respect, as will be the
online publication scheme for the Portable Antiquities Recording Scheme
in England and Wales. In areas where fieldwalking for Iron Age sites is
largely futile, as in much of northern and western Britain, metal detecting
can also be a valuable prospection tool, as well as opening avenues to
non-normal sites and ritual deposits, which are neither found nor prioritised
by conventional approaches. The Salisbury hoard provides a good example,
while also highlighting the problems resulting from thoughtless use of
metal detectors (Stead 1998).
The much derided corpus continues to serve an invaluable role as a vehicle of record and a springboard for further research. Many important categories of artefact, however, have not been studied for decades. The standard work on swords (Piggott 1950) is fifty years old, despite many new finds, while for pins we have to go back to Dunning (1934). Even more recent catalogues such as those for northern decorative metalwork (MacGregor 1976); glass beads (Guido 1978); horse bits (Palk 1984); strap unions (Taylor and Brailsford 1985); and pre-La Tène D brooches (Hull and Hawkes 1987), while still invaluable, require updating due the mass of recent excavation and metal detector finds. Rather than waiting on the vagaries of individual researchers, new databases need to be systematically instigated, perhaps most realistically on a regional basis, and then regularly updated. A possible model is provided by numismatics, which publishes regular round-ups of recent coin finds (e.g. Bateson and Holmes 1997). Specific institutions might perhaps take responsibility for maintaining information on key aspects of material culture, as SMRs do for sites, although this will certainly require additional funding.
There is a pressing need for the relevant specialists to publish articles defining recording standards and research agendas for particular types of material (cf. Bishop 1988 on Roman harness fittings). This would encourage wider application of best practice, and help raise awareness of artefactual work, but will require funding, as many of the individuals concerned are freelance. Syntheses are equally essential. Without a comparative framework, artefacts from individual excavations rarely make independent sense. While they may answer questions of chronology and site function, to understand their use and significance requires a wider view. Such surveys provide a context for other workers to interpret their own data, and make it more likely that finds from disparate excavations will be used actively in interpretation. Many catalogues languish as undigested appendices in unremarkable reports. Without synthetic works to assist in making sense of them, their value is limited to an archive one.
The supply of future specialists is a major problem. Material culture no longer figures strongly in undergraduate teaching and there are few MA courses in artefact studies. A lack of career opportunities ensures that few of those who do complete material culture centred PhDs consolidate their expertise into a professional specialisation. Yet many areas rely on a few specialists with no obvious successors. Without enhanced funding it is hard to suggest a way round this problem, although training courses to allow generalists to acquire the basics of a specialisation might be one way forward.
The topic of decorative metalwork exemplifies many of the above problems. This material ought to fuel a wide range of debates, on chronology, continental contacts, hoarding, status, technology, and the nature of 'Celtic' art. Yet there is still no catalogue of decorative metalwork covering the whole of Britain (although the Jacobstahl/Jope volume is long awaited), while the last major general study of insular art is over 40 years old (Fox 1958). Individual aspects have since received thorough treatment (e.g. Stead 1991a, 1995), but many new finds are often unknown beyond a small and specialist audience. Indeed, even major 'art' objects are often known more by repute than by detailed publication: studies of the Battersea shield (Stead 1985) and the Deskford carnyx (Hunter in preparation) show just how much new information can be gleaned from such supposedly well-known objects. The detailed reanalysis of the Llyn Cerrig Bach metalwork now underway, will similarly augment our knowledge of this major assemblage (e.g. Macdonald forthcoming)
Scientific analysis is still under-utilised. Its value is perhaps best
seen in the study of metalwork, where a wide range of established techniques
are available. This is one of the few specialist areas to see the development
of syntheses and an individual research agenda (e.g. Northover
1987; Dungworth 1996; Bayley
1998). The characterisation of pottery fabrics is the other main area
where scientific analysis is routine, and the results often highly informative
(e.g. Morris 1994). Similar insights could be expected
from many other material types, but often work has not progressed beyond
promising initial studies, with little attempt at routine application.
Crew 1991; Halkon 1997), while excavations on Kimmeridge shale production sites and salterns are good examples of the insights which industrial sites alone can produce (e.g. Sunter and Woodward 1987). They have a value far beyond the local context, touching on much broader questions of exchange and interaction. There is mounting evidence, for instance around the Humber basin (e.g. Foster 1996; Halkon 1997), to suggest that certain productive activities were preferentially located in marginal areas of landscape, whether for cultural or economic reasons (Haselgrove 1989; Sharples 1990).
Broad-scale patterns of artefact production have been considered by Morris (1996), although with an inevitable bias to the better-understood material from southern Britain. Yet few detailed regional studies of the organisation of production exist. With many material types, existing data are perfectly adequate for first-level interpretative models which could provide a framework for further study. Metalworking is a good example, as even older excavations typically retrieved at least non-ferrous metalworking evidence. A broad-brush approach can produce valuable models of the organisation of production, as Northover's (1984) work on contrasting cast versus sheet copper alloy production showed (but see Morris 1996, 54); or Ehrenreich's (1991, 1994) innovative work on the different levels of iron production.
Distribution studies merit more work, having faded from their processual heyday as practitioners became emburdened with post-modern angst over interpretation. This remains an area where artefact study and scientific analysis can play an important role in revealing patterns, whatever interpretation is then put on them. Studies of salt (Morris 1985), iron (Crew 1995), and glass (Henderson 1989a) highlight the potential. Petrological examination of all medium to large (defined regionally) pottery assemblages should be standard, while pilot studies on materials like jet (e.g. Hunter et al 1993; Allason-Jones and Jones 1994) show the possibilities of such work for other categories of artefact. Given the explosion of new metalwork finds, distributional analysis ought now to be able to provide some insights into the output of individual workshops or schools (e.g. Cunliffe 1996).Crewe 1991). Such practical experience should be more readily incorporated in archaeological knowledge, particularly by guiding the efforts of enthusiastic amateurs (cf. Bishop and Coulston 1993, 38-41, for Roman military equipment). In some fields, science can provide insights; organic residue analysis of ceramics, for instance, ought by now to be routine (D3.2.2 below). Recent work has also shown that detailed study of the form, size and use of pottery can help us understand the ways foods were stored, prepared and consumed in the Iron Age (e.g. Woodward 1997, Woodward and Blinkhorn 1997, Hill forthcoming, Hill and Braddock forthcoming).
The value of local ethnography is also rarely appreciated (cf. Clark 1951). In much of Britain, there are good records of pre-industrial societies: this applies as much to areas like northern England (e.g. Hartley and Ingilby 1990, 1997) as to more traditional ethnographic haunts such as the Northern Isles (Fenton 1978). For instance, the manufacture of post-medieval handmade pottery in Lewis (Cheape 1988) provides instructive parallels for Iron Age pottery production in the area. While ethnographic analogy is an indication of possibilities rather than certainties, local parallels in similar circumstances may have far more validity than exotic ones. This is an area calling for interdisciplinary work. We also need to consider the social and symbolic use of artefacts. It is now commonplace that artefacts are best seen as active objects, their meaning varying in different cultural contexts: the increasingly complex interpretations of Roman artefacts in non-Roman contexts highlight this (e.g. Willis 1994, Hill 1997). Only by tackling artefact use with an expectation of complexity and subtlety will better-rounded interpretations emerge.
It is convenient to differentiate on-site and off-site deposition. Apart from opening the eyes of archaeologists to the problem of Iron Age deposition, Hill's (1995a) work on Wessex stands out as an attempt to explain as well as to identify what went on, however tricky this may be. Although similar phenomena have been noted in many parts of Britain (e.g. Hingley 1992; Willis 1999), there are few equally detailed studies (e.g. Gwilt 1997). Yet this is essential, to avoid the uncritical application of a good idea. What happens in areas where pits are less common and preservation less good than Wessex? Regional variation ought to be expected in such rituals. In the Western Isles, for instance, there are recurring deposits of animal bone in houses, but in contrast to Wessex, these often include wild animals (Armit 1996, 155-7).
Much artefact deposition may be structured by guiding cosmological principles, for instance within houses (e.g. Fitzpatrick et al 1995, 85-9). This shows the potential of artefacts in tackling complex problems, but equally raises taphonomic issues. The quantities of artefacts involved are often small, and the sites badly plough-damaged. How do such patterns stand up outside Wessex, or on sites with surviving floors, like brochs and wheelhouses (e.g. Parker-Pearson and Sharples 1999, 16-21)? While the concept of structured deposition has raised important new issues to be faced in excavations, we need more sites with preserved floor, yard or midden deposits to provide models of 'normal' disposal with which to 'test' the hypothesis (e.g. Needham and Spense 1996, Hodder and Evans forthcoming).
Deposition of artefacts - often valuable metalwork - in the landscape
has been studied for some time, with attempts to develop explanatory frameworks
(C2.2 above). Despite this interest, few findspots
have so far been excavated, although the value of such work is clearly
shown by those that have (e.g. Lynn 1977; Stead
1991b, 1998; Fitts et al 1999). Equally, the widespread
use of wet places and other significant natural locations evidently conceals
significant regional variations (e.g. Hunter 1997).
Votive deposits represent one of the key interfaces between humans and
gods, and are crucial to understanding belief systems; if the mechanics
of ritual are to be better understood and we are to move beyond pan-Celticism
in studies of Iron Age religion (cf. Fitzpatrick
1991), investigating regional and chronological variations in the nature
and setting of such deposits is essential.
A key point is the sheer quantity of bone. It was one of the most abundant resources in prehistory, and on sites where preservation is favourable is the most abundant find after pottery. This should make us reconsider the apparent poverty of much of the northern Iron Age. A society based largely on organic objects could be highly complex but leave little or no artefactual signal, as a rapid comparison of assemblages from the Western Isles (Hallén 1994) or crannogs (Munro 1882) with those from 'typical' northern sites (e.g. Jobey 1974; Watkins 1980) will indicate.
D3.1.2 Textiles and skins
D3.2 InorganicInorganic material survives better, and is correspondingly better studied. Areas requiring more work are set out in Table 2 and discussed selectively below.
Function is another key topic, especially for the ubiquitous 'coarse stone tool'. As with bone, these are commonly found but rarely understood. There is need both for regional syntheses to identify recurring types, and for experimental work to clarify function. A related issue is the nature and extent of flint working in the first millennium BC (Ford et al 1984). It needs to be resolved whether late lithic industries similar to those recognised on southern Iron Age sites (e.g. Gardiner 1993) existed in the rest of Britain - as seems likely - and how long these continued. As well as providing a new perspective on later prehistoric technology, defining these industries might, in certain circumstances, provide a means of detecting ploughed-out first millennium BC sites .
Although the study of moulds and crucibles is well-established (e.g. Foster 1980; Bayley 1992), detailed technological examination of objects is less well served, but is crucial to understanding artefacts and their cultural context, for instance the effort required in manufacture, leading on to questions of craft specialisation and patronage. Metalworking is often treated as an intrinsically high-status activity, but in truth we understand little of its detailed meaning or operation, for instance in the role of travelling smiths or the magical aspects of metal (Budd and Taylor 1995; Hingley 1997). As already noted, comprehensive modern databases of major and minor copper alloy objects are urgently needed (D1.2 above).
'Precious' metals are better served: here gold and gold alloys are the main concern, silver being rare to non-existent in most of Britain until a late period in the south. The full publication of Snettisham will be a major advance, while debate over the date of ribbon torcs indicates there is life beyond East Anglia (Eogan 1983). Warner's (1993) work on gold analyses highlights a topic which merits further study, while the analysis of coin composition is already an area of some success (Cowell 1992; Northover 1992; Hobbs 1996). In general with coins, while practitioners may debate points of detail, there are few obvious lacunae; what we need to do now is think, not classify (e.g. Creighton forthcoming).
Lead is mentioned here as an example of the need to question absence as much as accept presence. In Scotland at least, its presence seems to be primarily an indicator of Roman contact (Hunter 1998b), as Mackie (1982, 71) previously suggested. The same applies to apparently simple items like the iron nail: the main metallic small find on Roman sites, it is all but absent on Iron Age settlements, which in itself speaks of a major cultural difference.
Iron products are in one sense well-studied: there are good site reports and established scientific methodologies, while Scott's (1991) research on a particular region or Fell's (1997, 1998) studies of specific artefact types point to ways to build on the framework we have. Work, however, has concentrated on a few parts of Britain, with others barely touched (e.g. Hutcheson 1997). As with other forms of Iron Age material culture, we should expect regional variation and treat it as a crucial area for study. The reasons for the initial adoption of iron; its roles and abundance; and its great increase towards the end of the Iron Age remain key questions, while the organisation of production also merits further study (F1-F2 below). Crew's (1991) work on the practical side is invaluable, but needs support by equivalent work elsewhere, and by more attention to the by-products of ironworking (e.g. McDonnell 1987). Even the basic smelting/smithing division is poorly understood. Hingley's (1997) foray into the potential symbolism of iron artefacts emphasises the variety of roles which they can play, while Ehrenreich (1994) has linked production to heterarchical forms of social organisation. The divide between practical and theoretical sides of archaeology is well-illustrated here: scientists rarely consider the social use of iron, while many archaeologists do not understand the practical constraints on metalworking.
Although regional differences have long been recognised, for example between the four provinces of the Scottish Iron Age (Piggott 1966), they are usually swamped by the notion that a single, uniform Iron Age existed, or explained mainly in terms of geographical and environmental factors. The belief that there was a single, coherent British Iron Age, with a distinct set of common features (specific settlement types, objects, ritual practices and forms of social organisation) has been extremely strong over the last eighty years or more (cf. Cunliffe 1991, 7-18). At the start of Iron Age studies, this approach was understandable given the need to extend typologies and interpretations from a small number of type-sites, usually in Wessex, to cover the whole of the country.
More recently, as more data have accumulated and underlying assumptions have been re-examined, the shortcomings of this approach have been exposed. In particular, we can now see that traditional models of the Iron Age are based very largely on features typical of limited areas of southern Britain (notably Wessex), which in turn have been used as a base line against which other parts of the island have been either explicitly or implicitly compared, often in a negative manner (cf. Bevan 1999). The result of forcing the peculiarities of particular regions to fit a standard model and ignoring other differences, or even whole regions, which do not fit, has been to over-privilege Wessex and south-east England both in research, and in general accounts of the period. This focus also tends to reinforce the view that those parts of Britain closest to Europe were more highly developed than the 'peripheral' areas to the west and the north.
This view is increasingly difficult to sustain in the light of recent research across Britain, indicating consistent variations in settlement types, or mortuary and ritual practices, as well as different styles and uses of material culture. Some of the variations probably are a factor of climate and geography, but the term 'regionality' specifically describes characteristics that have to be explained in other ways. Differences between regions, such as those between East Yorkshire and its neighbours, or between north-east and Atlantic Scotland, point clearly to the construction and maintenance of distinct regional identities, often at varying scales. They also hint at different lifeways and social institutions in adjacent areas. As yet, these patterns are better known in some parts of Britain than others; characterising them properly is one of the key priorities facing Iron Age archaeology in the next decade.
The existence of marked regional differences in the way Iron Age societies were organised makes it harder to generalise from one part of Britain to another. Equally, the supposed centrality of events or processes in any one region can be questioned. All regions are important to reconstructing the complex social mosaic of the period, and need to be understood in their own right. There are, however, certain dangers in this new emphasis. Too much stress on regions could increase structural trends in British archaeology to focus narrowly on particular areas, leading to parochialism. For some phenomena, meaningful patterns only emerge through comparison with other regions, including those across the sea. Too narrow a focus could overlook supra-regional features of the British Iron Age that are manifested locally in different ways. These include living in round houses and selective adoption of particular aspects of the La Tène cultural 'package' by individual groups. Equally, there are underlying economic and social trends, such as later Iron Age settlement expansion (F2.2 below), which can be recognised across many different regions
There are considerable problems in tracing the regional mosaic of Iron
Age Britain. The lack of a uniform level of knowledge throughout the island
is the most important, a product both of academic interests over the last
century, and more recently the uneven distribution of rescue excavations.
Academic research has concentrated on a few 'key' regions like Atlantic
Scotland, East Yorkshire and Wessex, all of which share highly visible
surviving Iron Age monuments. In these regions, both the benefits and the
dangers of letting research build on previous work are clear. Outside these
regions, rescue archaeology has partly filled the gap, but this too has
been concentrated in certain areas, particularly southern and eastern England.
The lack of regional syntheses exacerbate the results of this uneven coverage
(C2.1 above). More have been produced in recent years,
especially in East Anglia (e.g. Davies 1996, Bryant
1997, Sealey 1996), but collation and presentation of the basic data is
still a key priority for future work.
Halkon and Millett 1999) and the Mersey basin (Neville 1999), both of which would have figured in this category only a decade ago. The evidence of pollen analysis is also valuable in allowing the extent of human impact on the environment in the first millennium BC to be compared with regions where the settlement patterns are known in reasonable detail.
Taylor 1993). Despite the marked break in depositional practice, subsistence, settlement and pottery all show considerable continuities across the transition in most areas (e.g. Barrett 1980; Barrett et al 1991) - although in Wales, pottery did apparently cease to be used altogether for a while in the earlier 1st millennium BC - so that a longer perspective, extending back into the second millennium BC, is essential to understanding both the adoption of iron and its various consequences
It should be stressed how little we know about social and economic developments in the first four or five centuries of the Iron Age in most of Britain. A great deal of our evidence relates to the later part of the period, from c. 300 BC onwards. The major excavation projects of the last 25 years have generally concerned later sites: even Danebury has relatively little evidence before the fifth century BC. This imbalance is itself a pointer to the differing character of earlier and later Iron Age settlements, the former often open and rather amorphous, the latter defined by banks and ditches, or multiple enclosures. The presence of enclosure is not however the only major difference. Other evidence - from the use of the landscape, to early Iron Age settlement distributions and pottery - hints at forms of social and territorial organisation different from the later period.
Whereas late Bronze Age, and to a lesser extent earlier Iron Age, metalwork
is well documented, our knowledge of settlement and pottery sequences outside
Wessex and the Thames Valley is relatively poor. In many regions, only
a handful of settlements dating to this period have so far been identified
or excavated (Haselgrove 1999b), although in
certain others, the picture has been significantly modified as a result
of a spate of new discoveries, as for instance, in the Severn estuary,
(e.g. Whittle 1989; Nayling and
Caseldine 1997). Thus it is hard to assess the character of much of
Britain in the earlier first millennium BC, except through the nature and
types of bronze and iron metalwork. In this context, any newly discovered
late Bronze Age or earlier Iron Age site is potentially of greater importance
than one of the later Iron Age.
When and why the transition to iron use occurred in different parts of Britain are key questions for future research. As yet, we have very little idea of the mechanisms of this process (D3.3.3 above). Iron was used to make some objects before the end of the Bronze Age and gradually becomes more common in the archaeological record between the eighth and third centuries BC (e.g. Salter and Ehrenreich 1984). There is however little close dating for early iron objects or working, while the recent redating of the end of the Ewart Park phase to c. 800 BC (Needham et al. 1998) merely emphasises the gap in our knowledge of how the industry developed. If the cessation of bronze hoarding and the adoption of iron for utilitarian objects are indeed related - as seems likely - this implies that iron was already common by the eighth century BC, but if so, the Llyn Fawr hoard (Savory 1976) and a limited number of iron socketed axes (Manning and Saunders 1972) are still virtually the only tangible evidence for the new technology. The context of early iron production and its relationship to bronze working is also poorly understood. To make progress in understanding the effects and time scale of this transition, one strategy would be to radiocarbon date metalworking residues - slags, moulds, crucibles and hearths - from potentially early contexts. As noted above (D3.2.1), the role of flint technology at this period is another topic worthy of investigation (Ford et al 1994).Champion 1999). These forms of ritual continued throughout the Iron Age.
The period up to 300 BC is characterised by an increasing variety of settlement forms across Britain. There is also increased evidence for the organisation and exploitation of the agricultural landscape, including linear boundaries, field systems, pit alignments and isolated wells and pits (e.g. Taylor 1997; Hill 1999). These features point to a dynamic pattern of agricultural intensification, with a number of innovations. Production of salt starts in the late Bronze Age and intensifies through the period (Bradley 1978), while there appears to be an increasing reliance on cereals in many areas, with spelt replacing emmer over much of eastern and southern Britain (Jones 1981; Van der Veen 1992). The exact nature of regional farming systems is often poorly known, but complex patterns of local interdependence and transhumance are to be expected.
There were considerable differences within Britain. The best known regions, Wessex and the Thames Valley, may have only superficial similarities to other parts of the island (Hill 1995c). Large roundhouses are a feature of several areas at the time of the Bronze Age to Iron Age transition (e.g. Hingley 1995), although the apparent presence of long houses on some early first millennium BC sites should not be overlooked. Hillforts and other large enclosures were first constructed at this period in several areas of southern, northern and western Britain, although the precise timing varies (Cunliffe 1991; Hill 1995c; Musson 1991; Needham and Ambers 1994). Other regions lack hillforts altogether. In parts of eastern England, smaller heavily enclosed ringworks were occupied (Champion 1994). Midden sites such as East Chisenbury, Potterne and Runnymede in central-southern England should also be mentioned (McOmish 1996). Although not necessarily identical types of site, large, thick accumulations of cultural material dating to the early first millennium BC may be commoner than presumed.
It is important to recognise that other changes were taking place in
this long period apart than those involving metalwork and hillfort construction.
The social implications of 'the end of the Bronze Age' have only been discussed
in fairly general terms, but even so have received decidedly more attention
than social changes in the earlier Iron Age. In part this is due to the
lack of detailed chronologies to illuminate the degree of change and continuity
in different regions. At present, late Bronze and early Iron Age pottery
chronologies are only well defined in Wessex and the Thames Valley; elsewhere
in southern England, broad changes in pottery styles are apparent, but
more local refined sequences are needed. Hampering this is the lack of
associations with dateable metalwork, and the problems of the calibration
curve between 800 and 400 cal BC. Critical use of radiocarbon and other
scientific dating methods, based on whatever materials are most appropriate,
is essential for sites both with and without pottery (B2
Traditionally the period has been divided into the middle and late Iron Age, the latter being distinguished by new forms of material culture such as coins and wheel made pottery, as well as new settlement types and ritual practices (cf. Cunliffe 1991; 1995). Many of these changes are, however, confined to certain parts of southern and eastern England. In other areas, such as East Sussex, Norfolk and the Fenlands, the 'middle Iron Age' - as defined by its pottery - continues up to the Roman conquest and beyond (Haselgrove 1987, 58-63; Hill 1999). Moreover, many 'late Iron Age' developments can now be seen either to have started earlier, or to be rooted in developments prior to 100 BC (Haselgrove 1989). They cannot be understood purely in terms of external causes in the first century BC. It is thus increasingly hard to sustain the traditional separation between a 'middle' and 'late' Iron Age even in southern Britain, let alone the island as a whole (Hill 1995c). Rather, the traits that typify parts of southern and eastern England in the first centuries BC and AD are another example of the regionality which typifies Iron Age Britain. In their turn, these features can be seen as part of a broader pattern of change which began c.400-200 BC and intensified in many regions towards the end of the millennium.B1 above). Neither the causes of this change in ceramic styles, nor the nature of broader social changes between c.500 to 300 BC have so far been investigated in any detail. In all areas, there is need for more radiocarbon dating, as well as additional associations between pottery and other dateable objects (B2.1-2.5 above).
One of the most striking features of the later Iron Age is the sheer abundance of evidence of different kinds, whether numbers of settlements or the quantities of material culture found on them, as well as in burials, on religious sites, and in a variety of 'off-site' locations. Although this material provides plentiful data for research, only rarely are the reasons behind this abundance questioned (Hill 1997). Equally, while regional variations in material culture and social practice are particularly visible in the period after 300 BC, we need to establish how much this is due to the greater quantity and variety of evidence, how much to a genuine intensification of regional differences at this time. By the first century BC, however, the distinctions in material culture, ritual, and settlement types across Britain do seem sufficiently clear cut to imply real differences in social organisation between regions.Haselgrove 1984; 1989; Hill 1995c). The closing centuries of the first millennium BC saw settlement expansion into previously sparsely settled areas, and the infilling of others, so that by the first century AD, large parts of the lowland landscape were virtually 'full' of settlement (Hill 1999). It seems likely that prior to this period, many areas were relatively sparsely occupied and exploited; one reason being that the fertile but heavy soils which characterise many of the relevant areas were previously probably relatively difficult to cultivate. The use of iron-tipped plough shares and cereal crops suited to heavier soils and the introduction of the rotary quern, together with the climatic improvements after 400 BC, undoubtedly assisted the expansion (Haselgrove 1984, 17-19), although as some writers have noted (e.g. Hill 1995b; Tipping 1997), the impetus to agricultural expansion may have come from changes in social organisation.
This expansion into thinly-settled areas and the social processes underlying this phenomenon are increasingly emerging as one of the crucial features of the later Iron Age (Hill 1999). Frequently, the expansion process is linked with developing craft specialisation (Haselgrove 1989), for instance in the working of iron, pottery and glass , as well as with new kinds of settlement, which might in turn indicate new forms of social organisation. In some cases, the colonisation of new land was apparently accompanied by the laying out of extensive field systems as in East Anglia and the North Midlands (e.g. Williamson 1987; Chadwick 1999) - although better dating evidence is needed - while in others, settlement expansion may have promoted agricultural innovation. All these phenomena require further investigation, both to understand the mechanisms at work locally, and to assess to what extent similar processes were operating throughout different regions of both lowland and upland Britain.Haselgrove 1982; Cunliffe 1991). This simple model has, however, come under sustained critique over the last 15 years and developing more sophisticated alternatives is a priority for future synthesis and excavation. By the mid second century BC, southern Britain saw the re-emergence of gold, absent since the late Bronze Age. Torcs and coins were clearly among the media used to articulate social relations (Haselgrove 1987), and their appearance could have had a disruptive effect upon existing systems. Parts of southern Britain also show an increase in the use of horse trappings at about the same period. As yet the relationship between the various changes are poorly understood - work at sites like Danebury, Hengistbury, Maiden Castle and Westhampnett having raised as many questions as they answered.
Chronologies, artefact assemblages and sites are better known from the later first century BC onward. Rapid changes took place, with the foundation of oppida like Colchester, Silchester and St Albans, and intensified contacts with Gaul and the Mediterranean world. Many of the principal oppida were established in areas where earlier settlement was sparse (Hill 1995c; 1999), indicating the need to consider their emergence in relation to processes of settlement expansion in other parts of Britain. There were notable changes in the forms, imagery and distribution of coinage (Creighton forthcoming), as well as significant alterations in personal appearance, ways of eating, and in the nature of domestic architecture. These transformations are associated with the development of 'kingdoms' - large scale polities with clear signs of social hierarchy and elites - and are essentially confined to south-east England. Explaining them remains a key challenge. Here, the scale and nature of contacts with the Roman world, and with the imperial aristocracy in particular, is an area for considerable debate.
The next few years should see publication of several major excavations directly relevant to these issues (e.g. Silchester basilica, Hayling Island temple, Heybridge 'small town'). Nevertheless, major lacunae exist even for this relatively well known period. Oppida remain poorly understood and evidence for spheres such as agricultural and craft production, rural settlements, and territorial organisation is limited. Even in the heartlands of the two major kingdoms north and south of the Thames, the adoption of new practices, artefact types and social forms seems to be both selective and variable in date (Hill forthcoming), while in adjacent regions, the chronology of wheel-turned pottery and other late Iron Age forms is poorly known. Outside south-east England, the new burial traditions and ritual sites which emerge in this period have received little recent attention. In East Anglia and Lincolnshire, for example, the quantity of metalwork and coinage recovered from many probable settlement sites implies cult activity there (Willis 1997), but none of them have as yet been excavated on a large scale. There are also signs that many of the larger aggregations which typify various parts of eastern England are actually a product of fluid, shifting settlement patterns, similar to those found in other regions bordering the North Sea (Hill 1999), rather than representing stable nucleated villages. This would have important implications for the social organisation of the communities in question and requires further work.Foster 1989; Armit 1999). Here again, routine use of absolute dating is essential for progress. In some areas, enclosed sites - having multiplied during the final centuries BC - appear to have given way to predominantly open forms of settlement before the Roman occupation (e.g. Haselgrove 1984). As in parts of southern and eastern England, there are indications of greater emphasis on the individual at the expense of the community, such as larger numbers of personal ornaments or the use of formal burial. The conspicuous consumption of wealth through deposition of metalwork and other goods also increases in many areas at this time. Not all regions show such changes, however, a reminder of the difficulty of generalising across Britain at any stage of the Iron Age. The impact of the Roman occupation on Iron Age settlement forms and the often highly selective uptake of Romanised material culture both within and beyond the frontier remain key areas for further research.
The three major types of resource necessary for the study of the Iron
Age are, firstly, the archaeological record, as already known and as yet
to be revealed; second, the human resource of archaeologists with the necessary
skills; and third, the non-human resources of funding, facilities and equipment.
The previous sections have concentrated on the first of these elements
and have indicated various deficiencies in our existing knowledge. In assessing
how such an agenda can be put into practice, it is essential to take a
rather broader view of the state of Iron Age studies in Britain.
It does not follow, however, from the age or the poor quality of much other data, that the questions posed in a newly reformulated agenda can only be answered by evidence newly extracted from the ground. There is still a wealth of information to be won from the material already out of the ground, and currently held in museums and elsewhere. There is no need to rehearse all the research questions raised in the previous sections, but if we look only at the physical material that exists, it is clear that significant aspects remain unexploited.
One very basic lack is our failure to develop a tradition of serious study of the historical formation and validity of the archaeological record (Quellenkritik), except possibly in the field of Iron Age numismatics. We have little idea of the circumstances in which much of the record was unearthed or recorded, and hence little idea of the confidence that can be placed on it. The lacunae in our knowledge of decorated metal artefacts exemplifies the situation for other artefact types (see D1.2 above). Similarly, although the limited evidence of production has been exploited, the few works devoted to metal composition or to the output of individual workshops or schools simply emphasise the potential of this line of approach (cf D2.1 above). Pottery has fared better, at least partly because of its abundance and continuing importance for regional chronologies. Even so, much remains unpublished, including some very important assemblages. In England alone, the recent later prehistoric pottery survey (Morris et al 1998) recorded 7138 recognisable pottery collections from a total of 6977 sites, of which only 2032 (28.5%) are actually published. It is evidently desirable that this survey is extended to Scotland and Wales at the earliest opportunity and that urgents steps are taken to reverse this imbalance.
Turning to settlement archaeology, the quality of many earlier excavations
means that there is only a limited amount to be learnt from the surviving
records without renewed fieldwork. In addition, a number of important sites
from more recent years remain unpublished. One collection of material which
is being transformed into a usable archive is the backlog of aerial photographs
now being processed in English Heritage’s National Mapping Programme. In
some parts of England a considerable part of this photographic evidence
is likely to belong to the later prehistoric or early Roman periods. Completion
of regional maps such as those undertaken for the Yorkshire Wolds (Stoertz
1997) will provide a powerful tool for investigating the Iron Age landscape.
The same will also hold true for similar parts of Wales and Scotland, and
a continuing commitment to their production should be encouraged.
The causes of this growth are complex, but we can identify two major factors which will particularly influence our knowledge of the Iron Age. The first is the fundamental change in the nature of professional archaeological practice since the introduction of PPG16 in England and its equivalents in Wales and Scotland. These reforms have produced not only an increase in the amount of funding devoted to archaeology, in this case by developers, but also a major change in the pattern of expenditure, with a decline in the number of major excavations and an increase in small-scale evaluations. The pattern is not strictly random, since it is dictated by pressures of development and by the known archaeological evidence, but the number of evaluations is now very large and their distribution is becoming sufficiently widespread for reliable conclusions to be drawn. This activity has produced, in effect, a sampling of the landscape on a scale that would have been have been previously unthinkable, and certainly far beyond the resources of any archaeological funding agency.
The challenge is now for us to find ways to use this vast body of data. So far, little use has yet been made of it, although recent work on Bronze Age field systems in eastern England shows what is possible: working mostly from evaluation reports, Yates (1999) was able to identify a wealth of new field systems in the Thames Valley. We need preliminary studies to determine the quality and compatibility of the records compiled by the various contractors in an area, as well as working to promote greater standardisation in recording techniques. The next stage would be to conduct surveys by region and by period to assess how these reports contribute to our understanding. Data from this type of intervention will almost certainly transform our perception of the distribution and nature of Iron Age settlement and landscape organisation (C2.1 above), about which there is still much we do not know.
The second main source of the information explosion is the activity of metal detectorists (D1.1 above). The saga of the Salisbury hoard (Stead 1998), although an extreme example, demonstrates both the importance of this activity as a source of new, and often striking, information about the Iron Age, and also the ease and speed with which information about a discovery can become distorted or lost. Nevertheless, its importance for the Iron Age is immense. The recent CBA survey (Dobinson and Denison 1995) took as one of its detailed case studies the quantitative and qualitative growth in our knowledge of East Anglian coinage, and demonstrated that it was due almost entirely to coins found by metal detectors. The impact on our understanding of other types of Iron Age metalwork could be just as great.
The first attempt to turn this wealth of data into archaeologically useful information has come with the trial introduction of a voluntary reporting scheme in England and Wales. The first annual report for England (DCMS 1998) demonstrates a considerable success in terms of the numbers of objects reported in the trial areas, and the breaking down of mistrust between archaeologists and detectorists, but it also shows the scale of the challenge to be met. The earlier estimate of 400,000 metal-detector finds per annum (Dobinson and Denison 1995) can now be seen as a gross under-estimate: a figure in excess of 1,000,000 may be a better approximation. It is a matter of some urgency to assess how best to deal with a flood of finds of this magnitude to ensure the optimal rate of reporting and recording, and to identify a lasting source of funding to enable the reporting scheme to be established on a permanent basis at national level.
Reporting and recording are only the first stage. It will be increasingly
difficult to win and to retain the co-operation of metal detectorists,
and to justify the allocation of public funds, if we as archaeologists
do not take steps to use the information being acquired in this way (D1.1
above). Like the results of field evaluations, metal-detector finds
provide a source of information of a quality and quantity that would have
been unthinkable until recently. With the completion of the first year
of the reporting scheme, it is an appropriate time to take stock of the
archaeological significance of the finds by trial evaluations of the information
acquired on a period and county basis, so as to gauge the scale and nature
of the appropriate archaeological response. This will require more people
equipped to deal with artefactual data.
These publication figures seem intuitively to be a fair reflection of
activity over the period. The overwhelming dominance of settlement studies
(category G) reflects the growth of developer-funded archaeology, where
the nature of evaluations encourages reporting of sites found. The high
numbers in categories H and K reflect recent academic interest in topics
such as structured deposition, hillforts and brochs. Category F (artefacts)
may seem large, but many of the items are reports on single objects rather
than larger-scale studies. The comparatively low numbers in category D,
especially regional surveys, reflects the comparative lack of attempts
to synthesise the rapidly growing mass of detailed evidence (C2.1
Universities play a critical role, and are central in training of future experts. At present, this sector includes approximately 15 lecturers who claim particular expertise in the Iron Age of Britain and western Europe (including four of the co-authors of this report). Staff have some available research time, as well as access to equipment and laboratories; and have student research time in the form of undergraduate and postgraduate dissertations. Against this, many of the relevant staff are individuals of some seniority, with many other calls of their time; the number of recently appointed staff is small and virtually none are primarily artefact experts. Additional staff with expertise in the period (including one of the co-authors) are based in the contracting arms of University departments, providing access to library and other resources, but the time available to them for personal research is strictly limited. The number of specialists in Universities is unlikely to increase significantly in the foreseeable future. Where Universities could have a major impact, however, is in encouraging more students to work on the period, as well as raising wider awareness of advances in research, for instance through organising conferences and introducing extra-mural and part-time courses.
A number of practitioners are based in museums, especially the national museums (including three co-authors), many of them with particular artefact expertise. Most of these individuals are not directly involved in student education and training, although this is encouraged in some cases. Their research time is subject to institutional constraints, which can be more severe in smaller museums and departments. The time available for research is nevertheless generally greater than for specialists working elsewhere in the public and private sector and and they also have a vital role in communicating the results of research to the public through new displays and exhibitions. Like their counterparts in universities, practitioners working in museums generally have some choice over where they conduct fieldwork. This freedom is likely to be increasingly relevant if the imbalances in coverage caused by regional variations in the volume of excavations conducted in advance of modern development are to be kept in check. It is also desirable that as far as possible national and university museums step in to fill the vacuum which the shift in undergraduate teaching away from material culture has created, both by encouraging students to work on their collections and by providing courses of their own. In general, links between universities and museums need to be strengthened to make the most effective use of the resources which do exist for research and training.
The remaining areas of the public and private sectors include some individuals who originally trained in later prehistory as well as many more who have built up considerable expertise in Iron Age sites and material culture, particular pottery. Their involvement in developing the research agenda would normally require funding. For the most part, they are not directly involved in training, and have limited time to pass on their skills to junior workers, or to publish more general research on the material (D1 above).
In sum, this brief analysis suggests:
Apart from developers, the main funding sources for research include
the three national archaeological agencies, as well as more limited funds
available from various national and local societies and private trusts.
Also available to staff of higher education institutions (HEI) are somewhat
larger grants from bodies such as the Arts and Humanities Research Board
(AHRB), the British Academy and NERC. The AHRB offers a potentially important
source for funding major archaeological projects and syntheses at an appropriate
scale, but given its broad remit, only a few grants are likely to be for
research on the Iron Age. As yet, no provision exists for collaborative
projects with non-HEIs, but this may come. The remit of NERC is limited
to science-based archaeology, but does allow for collaborative projects
between organisations, including Case Studentships; it also provides staff
in HEIs with free access to a range of relevant scientific facilities,
notably AMS dating. The British Academy disburses rather smaller grants
(currently up to £5K), but may well be the principal source of funds
for research excavations for the foreseeable future, as well as providing
funding for post-doctoral fellowships. A certain amount of research funding
is still available internally to archaeologists working in national museums
and the universities, for excavation and for other purposes, although in
universities such funds are increasingly restricted to pump-priming or
allocated as matching funding for external grants.