Archaeological data, subcultures and social dynamics.

Matthews, Keith J.

The archaeological record is dominated by the repeated object and the repeated event, so we search for patterns that explain the regular in general terms. But human societies are not like that; the mass is actually made up of individuals, and the engine of change more often at the margin than at the centre.

Subcultures are an important but little-recognized element in every society. As soon as human communities exceed the nuclear family group, tensions may arise between separate power-groups (Argyle 1992: 199); groups which are not dominant will form embryonic subcultures focussed around any feature which gives them self-identification. This is, for example, how the modern homosexual male defines himself within a gay subculture (Freer 1987: 57; Giddens 1993: 198): his sexuality is what enables him to identify with other members, and gives the subculture a common 'purpose' or historical trajectory. There is literary evidence for a distinct gay subculture in London as early as the 1720s (Weeks 1991: 202), which it should be possible to identify from its material remains.

In the definition proposed here, every member of a society belongs to a number of subcultures, each with its own 'centre of gravity' to which each member is drawn, more or less strongly. As all members belong to many different subcultures, their 'pull' on each individual is in more than one direction; with all members of society being pulled in a multiplicity of directions, their separate and chaotic trajectories together give societies as a whole a single trajectory (Carrithers 1992: 199). Although I have used the term 'subculture', I do not wish to give the impression that they are in any way inferior to or dominated by 'mainstream' culture. Indeed, 'mainstream' culture as such does not exist; it is a construct, the point of meeting of the agreed subcultures to which dominant elites belong. In the late 20th-century United Kingdom this is the culture of the white, English, (nominally) Church of England, upper-middle-class, property-owning, heterosexual, married male; each defining adjective by itself is the descriptor of a separate subculture which includes a greater diversity of individuals than those within the elite.

Historical trajectories in this model become quasi-Marxist struggles between opposing forces. Because of the self-identifying nature of social subcultures, no single mechanistic cause, such as class struggle, drives their dynamics (Shanks & Tilley 1987: 210); they have no evolutionary direction. There will be cohesive forces - basically a form of feedback - but no system can remain either stable or on a single trajectory for long periods. The rise and fall of complex societies, which caused the processualist school so much anguish (Tainter 1988: 42), can be explained purely in terms of internal dynamics. This model of subcultures gives an explanation of culture change which relies entirely on itself: change is the norm, and it is the apparently stable culture (such as Bronze Age Egypt) which requires explanation. Society does not depend solely on economic, subsistence, ideological or any other subsystems or metasystems and their interaction: it is not some great mechanistic entity but a polythetic, fluid mediation of the concatenation of individual human behaviours.

The post-processual interest in subcultures and minorities has been criticised for encouraging the proliferation of 'unconstrained multiple readings of the past' (Kohl 1993: 16). Classic sociological interpretation views subcultures as 'deviant' behaviour, particularly associated with working-class and predominantly male youth (Hebidge 1979: 90; Brake 1985: 11); however, this seriously undervalues the variety of expression found even within mainstream elite culture. In any but the simplest societies there are social groups which fall outside the mainstream; an interest in these groups will not attempt to rewrite archaeology and history from their viewpoint but will seek to integrate the variety of experience of those marginalized within society (which usually means non-elites) into an understanding of how societies function as entities and to celebrate the importance of individual experience in the past and the present. These views also ignore the creative role many subcultures have played within society. The Bronze Age of Wessex, for instance, cannot be understood purely in terms of the complex burials which characterize its elites (Morris 1988: 71). Some subcultures have long been objects of archaeological study without an appreciation of their status as subcultures: medieval monasticism and the Roman military are two examples from British archaeology. A recognition that groups other than economic strata do exist within society is fundamental to this approach.

Subcultures and material culture change

Milroy & Milroy (1992) have proposed a model for rapid and far-reaching linguistic change: neologisms are transmitted between peripheral members of otherwise strongly bonded (and even mutually hostile) social groups, and the use of these neologisms is seen as a further reinforcement of group identity. A material culture parallel can be drawn from the behaviour of young adult men in 1970s Britain and their use of ear-rings. At the start of the 1970s the wearing of ear-rings by men was an expression of sexual deviance; the growth of punk in 1976 brought body piercing - particularly of the ears and nose - into prominence with 'forbidden signifiers' like safety pins and paper-clips used as jewellery (Hebidge 1979: 115). Their use was intended to shock those in authority, especially parents, teachers and other adults, and this appears to have worked.

Perhaps because of its ability to shock adults and express the youthful rebelliousness of the wearer, the fashion spread rapidly among young working-class men. It may have been given further impetus by the adoption by gay men of an earlobe code in the late 1970s in which a single ear-ring worn in the right earlobe came to signify that the wearer was homosexual: the opposite implied heterosexuality. In the 1980s male ear-rings became relatively commonplace, and not only among the young and working-class; as male ear-piercing became more general, so the safety-pins were replaced by more conventional ear-rings, both studs and sleepers. This change in the signifier occurred as the subcultural practice was transmitted from punk subculture towards a more mainstream expression. As this occurred so the power of ear-rings to indicate membership of the original subculture became dissipated. The alternative youth subculture has continued to adapt by adopting nose-piercing, eyebrow-piercing and lip-piercing as badges of membership; there are signs that nose-piercing is on the fringes of 'respectability', at least among women, but eyebrows and lips remain areas whose piercings retain the power to shock. Since the supermodel Cindy Crawford had her navel pierced in 1994, navel-piercing has enjoyed popularity, but it remains to be seen whether or not it becomes a subcultural badge. As these badges become accepted into mainstream taste via often short-lived fashions, new signifiers of subcultural identity need to be adopted.

This process can be seen clearly among the Lokop of Samburu, Kenya, where Roy Larick (1985: 218) has shown how 'subpopulations' use different styles of spear-head as membership emblems. New styles, introduced by the younger warriors of between about 13 and 18, are designed to reinforce group identity by contrasting their styles with those of older warriors, elders and boys. Indeed, the last two classes use out-dated and worn spears. Choice of spear form is dictated by the desire for innovation as a mark of 'modernity' and the appropriation of styles from neighbouring groups as proof of 'cosmopolitan' taste. What is of particular relevance to the present theme is that Larick (1991: 327) interprets the vigour of innovation as a direct consequence of the marginal social status of both the consumers (young warriors) and producers (blacksmiths) of the spears.

Celticity as subcultural badge

Knowledge of Celtic identity in the 1st millennium BC comes entirely from the writings of the literate cultures of Greece and Rome. They give a clear impression of a group of peoples who shared a common identity but no political unity, and perhaps not even a language (Powell 1958: 17); Caesar regarded the Rhine as forming their northern border and Herodotos (II.34) thought them the most westerly people of Europe after the Kunesioi. Archaeologists, long identifying them with the La Tene culture complex, have sought their origins in the earlier Hallstatt and Urnfield cultures.

In recent years the relationship between ethnicity and material culture has become an important question in archaeology; it is recognized that there is no simple equation between the two (Shennan 1989: 10). This is evident in attempts to define a Celtic culture complex: its geographical extent varies according to the criteria used, as does the date at which it is supposed to have emerged. A linguistic definition of Celticity also causes difficulties in pre-literate societies: north of the Rhine, rulers of tribes labelled Germanic by Classical ethnographers have names which are linguistically Celtic, but to what extent does this reflect the ethnic self-identification of the individuals, not to mention their polities? Such is the confusion over Celticity that it has even been possible for a Celticist to write that 'Celtic art . . . is anything but Celtic' (Green 1989: 6).

A large part of the problem is that a so-called Celtic identity still exists - in contemporary Wales, Scotland, Ireland, Brittany and Cornwall - but that identity is a fusion between a subculture officially suppressed for centuries and its Romantic re-creation two centuries ago. Celticity as presently conceived is invested with many emotional burdens: other-worldliness, oppression, exoticism, nationalism, periphery, lost grandeur, bellicosity and so on (Delaney 1986: 219). Study of the Celts has particularly been linked with Welsh and Irish nationalism and, more surprisingly perhaps, with French nationalism (Kristiansen 1990: 827).

Ethnicity is today regarded as situational and meaningful only in terms of relations between self-identifying groups (Cohen 1978: 389); it must be questioned whether the peasant farmers of the 2nd century BC from the Danube to Ireland could conceive of themselves as part of a single people. If there were any community of interest, it was shared between those engaged in the exchange network of prestige goods, the consumers of La Tene metalwork. These material signifiers of identity were not available to the majority of the population: they circulated among a restricted, self-identifying 'subpopulation', in other words a subculture. This allows us to see pre-Roman Celticity as part of an elite tradition, perhaps part of a growing self-awareness as north European economies moved from a largely local to pan-European phenomenon during the 1st millennium Be. We should not deceive ourselves in using the word 'Celtic' as archaeologists that the term would have had any meaning whatsoever to the pre-literate peoples of temperate Western Europe (Collis 1994).

The survival of pre-Roman Celtic artistic styles through the Roman period illustrates the effect of removing elite patronage from a sub-cultural tradition. Whereas during the pre-Roman period, so-called Celtic art is closely identified with the elites for whom prestige goods were produced and among whom they were traded, following the Roman conquest, goods bearing this style of decoration all but disappear. The re-emergence of the style as a dominant (elite) form after the 400 years of Roman rule, even in contexts which are supposedly Germanic (Bruce-Mitford 1993: 45), has been taken to demonstrate the continuity of the tradition in non-elite contexts, presumably on objects which have not survived to become archaeological finds. The phenomenon is easily explained in terms of the elite transferring its allegiance from Celtic to Mediterranean artistic styles following the Roman conquest: in order to become part of the European elites, it was necessary to patronize those craftspeople familiar with Roman styles. Once that need ceased, it became possible once again to appropriate native artistic traditions to express group identity, and perhaps (in view of the following) even to reinforce it.

The Germanization of England as subcultural change

A subcultural explanation can be adduced for the phenomenon in England known as the adventus Saxonum (Higham 1994: 3). Recent treatments of the subject (e.g. Hodges 1989: 25; Higham 1992: 15) have played down the numbers of putative immigrants, seeing the considerable changes observed in material culture in 5th- and 6th-century lowland Britain as the result of an increasing Germanization of taste among the native population as much as the migration of large numbers of individuals across the North Sea. Critics have asked how small numbers of immigrants effected change on the scale needed for the almost complete language replacement, transformation of material culture and disruption of settlement patterns evident in the record.

A subcultural model resolves the problem. The observed changes would involve not simply the acquisition of Germanic tastes as well as limited migration, but also the use of Germanizing material culture as a means of establishing and reinforcing group identity at a time of political crisis and social fragmentation. Transmission of the new material culture would be through those on the social (but not necessarily geographical) peripheries: a large-scale movement of the disaffected, similar to the Gaulish bacaudae, could be one of the channels of communication. Another channel - not to be underestimated in its impact - would be the very small numbers of north German immigrants, not all of whom necessarily arrived in the 5th and 6th centuries.

The areas in which the earliest Germanizing material culture appears are mostly in eastern England: east Kent, Norfolk and Lindsey. These are areas with small numbers of towns or villa-estates, both of which may be taken as indicators of social and economic prosperity and, in the case of the latter at least, of identification with southern European elites. We may suggest that these areas, indeed peripheral to the social life of the late Roman Diocese, had easy contacts with north Germany and Scandinavia. The exception is the upper Thames valley, where there are large numbers of villas and small towns but an early group of German material culture remains. This group has long been identified as anomalous (e.g. Leeds 1913: 53), as the invasion/settlement hypotheses are clearly inadequate to explain so massive a penetration so deep into central Britain at this date; furthermore, it is not identifiable as the core of a later Anglo-Saxon kingdom, despite valiant attempts to link it with Wessex (e.g. Stenton 1971: 26). Here is perhaps the best evidence for the Germanic mercenaries mentioned by Gildas (Higham 1994: 104).

To look at the 'problem' from the other end, the material culture of the former Roman citizens who must, on this hypothesis, have continued to be the majority of the population, may provide valuable clues about social patterning. The British population remains largely invisible, although the recognition of sub-Roman pottery types has helped the identification of continuing occupation in places such as Baldock, Hertfordshire. One particularly interesting example of a sub-Roman pottery type was found during rescue excavation of a late to sub-Roman settlement at Pirton, Hertfordshire. Employing decorative elements found on late 6th-century Anglo-Saxon pottery, it was found in a purely sub-Roman context; moreover, the fabric was not a recognized Saxon type and the juxtaposition of decorative elements was unusual, if not unique (Susan Tyler pers. comm.). Had the maker of the vessel seen pottery of Saxon type and was consciously imitating its decoration without understanding the semiotic grammar? Here we are perhaps witnessing the beginning of a process of Germanization in an otherwise conservative community of Britons.

In a recent survey of Romanization in Britain, Martin Millett (1990: 228) sees that 'Roman Britain disappeared piecemeal'. This is easier to explain in terms of a gradual erosion of Romanized legal and economic structures from beneath or within; the adoption of Germanizing material culture by the underclasses and other subcultures may mean that we have dated the disappearance of Romanization too soon, as early dates for Anglo-Saxon pottery have been taken to indicate early dates for political change. In the other model proposed here, the Romanized state(s) could survive for longer, its control increasingly eroded as more of its members adopted the iconography of the new material culture and the habits of an emergent elite, including its language.

An example: the Iron Age-Roman interface in central Hertfordshire

In Hertfordshire during the late pre-Roman Iron Age, thanks partly to the richness of the region's material culture, it is possible to recognize quite contrasting settlement patterns - despite the evident homogeneity of the material culture and political allegiance - which can be attributed to subcultural differences. The area under consideration, around Welwyn, in central Hertfordshire, is dominated by enclosed settlement sites of middling economic status, the only nucleated settlement being the enclosure at Wheathampstead. To the north and east, the settlement pattern is more varied, with predominantly open settlements clustering around the oppida at Baldock and Braughing, although a few enclosures are also known.

None of the enclosures in the Welwyn area has been fully excavated, but a number have been sampled, particularly as a result of the growth of Welwyn Garden City (Hughes 1939; Arnold 1954; Rook 1968a; Rook 1968b; Rook 1970a; Rook 1970b; Burleigh et al. 1990). At these sites, recorded under rescue conditions, the enclosures themselves are better understood than the areas enclosed. It is evident that they are domestic habitation sites: the pottery and animal bone assemblages are clearly of domestic origin, while the presence of daub and tile at Hollard's Farm (Burleigh et al. 1990: 24) suggests buildings of substantial construction, the tile demonstrating a willingness to use Romanizing material culture. Traces of buildings were found at Welwyn Garden City Grammar School (Arnold 1954: 128), hearths at Crookhams (Rook 1968a: 55), hearths and occupation debris at Grub's Barn (Rook 1970b: 31). The sites are concentrated on the plateaux overlooking the Mimram and Lea rivers (Arnold 1954: 128; Rook 1968a: 51).

The ditches of these enclosures are generally substantial; it must be assumed that banks commensurate with them once existed, now disappeared. At Hollard's Farm the ditches appear to have been deliberately backfilled, some of the material perhaps deriving from the banks (Burleigh et al. 1990: 19). At both Hollard's Farm and Brickwall Hill, pottery vessels in the ditches, complete but crushed, suggest rapid abandonment (Burleigh et al. 1990: 24; Rook 1970a: 25).

This class of site is thought to be predominantly middle-status, on the grounds that the construction of substantial banks and ditches would have required considerable expenditure. Imported pottery in the form of amphorae and butt beakers were found at Hollard's Farm, and the evident association of the Grub's Barn enclosure with the Welwyn Garden City 'chieftain's burial' (Rook 1970b: 31) also evokes the nature of the occupation.

The dating of the pottery from all these sites falls, without exception, into the first three-quarters of the 1st century AD. It may therefore be suggested that the abandonments of all the enclosures were roughly contemporary, within about 10 years of AD 65. Tony Rook (1968a: 51) has suggested that Crookhams was abandoned following the rebellion of Boudica in AD 61; this attractive hypothesis cannot be considered proven, and the pottery from Hollard's Farm indicates a date of up to 15 years later, after AD 70 at the earliest (Burleigh et al. 1990: 19). That all the enclosed sites in the Welwyn area were abandoned at roughly the same time does suggest that some external decision was made to discontinue their occupation, whatever the wishes of the inhabitants.

At this date, such a decision could have been taken at provincial level by the Roman governor, or - perhaps more likely - the council of the Civitas Catuvellounorum at Verulamium. There is no sign of a nucleated settlement at Welwyn before this period, and the abandonment of the enclosures may be part of a move from a dispersed settlement pattern towards a more nucleated pattern. This contrasts strongly with the rest of Hertfordshire, where the beginnings of nucleation at Baldock, Braughing and Verulamium can be placed in the middle of the 1st century BC (Stead & Rigby 1986: 84), if not earlier, and where open settlements occur alongside enclosures.

As a class of site, these enclosed farmsteads give every appearance of having been intended for defensive purposes. However, what we know of the political situation in the region at the time of their construction suggests that conditions were relatively peaceful. Perhaps the best comparisons for this class of site are the medieval moated farms which occur in large numbers throughout lowland England, mostly constructed about AD 1300, often in areas of new woodland clearance and agricultural colonization (Platt 1978: 111). Some, at least, appear to have been constructed as status symbols by newly-prosperous peasant farmers anxious to copy the fortified manorial sites of their social superiors. They were not primarily a response to political instability, but an expression of social mobility and change.

The situation of social change with a rising elite can be paralleled in southeastern Britain at the end of the Iron Age. From the 1st century BC, particularly after the Roman annexation of Gaul in the middle of the century, contacts with the Mediterranean world proliferated (Cunliffe 1988: 152). Coinage began to be used for transactions with the advent of potin coins, and there are signs of an increasingly standardized method of pottery production. The number of occupation sites increases dramatically, and agricultural expansion must have gone hand-in-hand with this. In addition, the first true urban sites, such as Baldock and Braughing, are found in Britain.

Although these changes were once attributed to 'Belgic' invasions, it is now clear that the economic changes were not sudden, but took place over a century or more. The increasing prosperity at all levels of society seems to have been exploited by a rising class of aristocrats below tribal chieftain level, whose elaborate burials at Welwyn and Baldock (Smith 1912; Stead & Rigby 1986: 51) attest to their wealth.

We can view the settlement patterns in Hertfordshire as part of this general social change: society at a level beyond the individual polity was moving along a particular historical trajectory towards open, nucleated settlements seen clearly in most of the Hertfordshire region, with its diversity of settlement types. At Welwyn, the historical trajectory followed by a sub-set of society took a different path, deviating in creating a tradition of enclosed farmsteads. In the terms of this paper's approach, it was subcultural. We need not explain these sites in purely socio-economic terms: as well as prosperity and competition, they were a statement about belonging to a social group. They were as much an emotionally-defensive response to massive social change as an expression of that change: they reflect against the direction of change within society as a whole. Here a subculture is actively involved in a process of change, subverting it and even perhaps attempting to reverse it, while continuing to share in the material culture available to all.

The Welwyn enclosures are not a local, geographically-limited type: these enclosures exist in all parts of the region, a common feature of Catuvellaunian settlement patterns. But here alone they are the only settlement type of this particular social group. In my explanation, the elite culture accepted by the land-owning classes in the Welwyn area had at its heart the subculture which was identified by enclosed habitations. It may be that this elite consisted of a small number of competing families or that the display element of competition between families was particularly fierce; whatever the reason, the elite bucked the trend of Catuvellaunian society and ignored the shift toward open settlement types. Its use of Romanizing material culture in the form of tiles and pottery demonstrates that it was not reactionary or conservative: this was a vibrant and dynamic subculture whose outward show of solidarity was suppressed, perhaps for political reasons, about AD 70.

An archaeology of social divisions

My roots are in the dirt of the archaeological site, the best foundation for a truly archaeological theory. Rather than borrowing theory from some other discipline and tailoring it to suit their own needs (Yoffee & Sherratt (1993b: 3) describe the technique as 'mining and bridging'), archaeologists should seek to base the theory underpinning their discipline on the special qualities of the data they acquire and interpret. One special quality of archaeological data is the patterning of the broad picture and the deviation from it of the individual context.

This is not a new approach, but it is unfamiliar to a generation which has been accustomed to systems theory, environmental determinism, mathematical modelling, structuralist theory, hermeneutics and now the linguistic theory of post-processual archaeology. The approach to archaeological interpretation presented here is based on a return to the primary data of our discipline and on a fuller reconstruction of past behaviour that does not ignore the deviational. The data do not speak for themselves, and we should not deny that our recording and reading of the data are anything but historically constituted. However, I do not set out with a paradigm from outside the discipline and select those data with the best fit to illustrate the theory.

Archaeology, reduced to its fundamentals, consists of the recognition of many different traces of past human activity. To be made meaningful, they need to be organized into patterns, analogous to the way in which individual human actions combine to form the patterns which define society (Giddens 1993: 31; Carrithers 1992: 35). Until recently, archaeology has used recursive patterns of material remains to define 'cultures' which were assumed to have been associated with particular societies and ethnic groups (Clarke 1978: 229). This emphasis on material remains - generally taken to mean artefacts - is an antiquarian mode of behaviour and increasingly inappropriate to archaeological theory. All the traces archaeologists detect in or above the ground - patterns composed of far more information than the artefactual element alone - are clues to past social behaviour.

The archaeological manifestation of social behaviour is constituted not in artefact design alone but in repeated patterns of use of both the artefact (as an extension of the human body and expression of the psyche) and the site (as the locus of social transactions). However, it is not enough simply to consider the 'context' of an artefact if that consists merely of setting that artefact against others which may be spatially or chronologically linked (contra Hodder 1991: 149); the evidence for use and discard of the object, the circumstances in which it came to be an archaeological object, are the most immediate guides to the social behaviour which underlay its production and use and to the individual behaviour which used and discarded it. This individual behaviour cannot be seen in isolation from society, though, nor does a single instance necessarily reflect the workings of all parts of society.

Archaeological data are, at best, incomplete. The identification of recursive patterns, necessarily the cornerstone of model-building in this situation, has served us well. This focussing on the readily recognized patterning at the broad scale has often led to the rejection of anomalous data, although anecdotal examples of extraordinary data (the tomb of Tut'ankhamun, for example) demonstrate its levelling effect. Sites are produced by human activities - especially repetitive activities and it is the physical traces of these, rather than the discarded elements of material culture, which give us the information by which we can attempt to reconstruct the dynamics of those societies (compare Asad 1986:12 for a parallel within ethnography). Strong patterns represent dominant (which usually means elite) subcultures: the weaker patterns and anomalies are not noise but the products of 'minority' subcultures. The challenge to method is to learn to recognize multiple patterns - not all of the same nature or strength - in which the multiplicities within a society declare themselves.

Such patterns do not correspond to traditional archaeological cultures: they show a far greater synchronic diversity than archaeologists have been willing to recognise. The reason for this is that individual societies are not homogeneous, but composed of interlocking sub-sets. In modern society we call these subsets 'minorities' and 'subcultures'; the term 'subpopulation' has also been used in archaeology (Johnson 1977: 484). Archaeologists have usually taken elite culture to be the dominant, the defining, the only culture - exactly as historians have done until recent years. Instead, we can see, from our archaeological remains, that elite culture remains an irrelevance much of the time to the majority of the population. By concentrating on the material culture, the artefacts, we have allowed the elite cultures - those whose remains include the largest numbers of the most complex artefacts used within a society - to define archaeological entities.

Conclusion

By my definition, a subculture is part of a human society with its own rituals and behaviour patterns which define it as a subset of that society (cf. Giddens 1993: 763). The subculture can be constituted by any self-identifying subset or, indeed, by oppression by dominant subsets; this may be based on class, elites being easily identified by material culture; it may be based on religion, as with early Christian communities in the Roman empire; it may be based on ethnic origin, whether real or perceived, as with the European Romanies of the 2nd millennium AD; it may be based on gender or sexuality, as with gay men in late 20th-century San Francisco; or it may be based on some less obvious communality of purpose, as with the New Age Travellers of 1990s Britain.

Each of these example groups has its own distinctive material culture; its artefacts, generally available to the whole of society, are chosen and combined in a way that produces a new and materially identifiable subculture. Some have long been characterized by their distinctive material cultures. Christian communities in Roman Britain are identified by the iconography employed on some of their prestige goods, such as silverware, and by buildings whose lay-out and relationships to other structures are distinct - martyria, baptisteries and so on. Christian burials may be distinguished by specific rites such as the gypsum burials of southwestern England (Woodward 1993: 228), although this example is by no means certain.

New Age travellers are a very visible subculture in contemporary England (Lowe & Shaw 1993: x): in the popular press they have been demonized and are caricatured by reference to their hair in dreadlocks and wraps, dogs on strings and broken-down, garishly-decorated buses. Within their communities, however, there is a general rejection of many of the elements of material culture which do not fit in with their philosophy, such as environmentally-unfriendly plastics. Their material culture can be characterized as much by what the subculture does not have as by what it does.

An acceptance of the subcultural model of societies I have presented will allow archaeologists to explore the richness of the archaeological record. We can begin to look not just at broad patterns through time or space, but also at those sets of data which have always appeared anomalous. It gives us the opportunity to explore individuality, creativity and change. And above all, it gives us - I believe for the first time - an explanation of the chaos of our data.

Acknowledgements. I am grateful to Mark Turner, for the lengthy discussions which led to the initial idea for this paper, to Mike Morris, Anthony Sinclair, Bill Sillar and Christopher Chippindale for reading and making comments on earlier drafts and to Gil Burleigh for his help with information on the late Iron Age in Hertfordshire.

References

ARGYLE, M. 1992. The social psychology of everyday life. London: Routledge.

ARNOLD, B.J. 1954. A Belgic settlement at Welwyn Garden City, Transactions of the East Hertfordshire Archaeological Society (1952-4) 13: 128-35.

ASAD, T. 1986 The concept of cultural translation in British social anthropology, in Clifford & Marcus (1986): 141-64.

BINTLIFF, J. (ed.). 1988. Extracting meaning from the past. Oxford: Oxbow.

1991. The Annales school and archaeology. Leicester: Leicester University Press.

BRAKE, M. 1985. Comparative youth culture: the sociology of youth culture and youth subcultures in America, Britain and Canada. London: Routledge & Kegan Paul.

BRUCE-MITFORD, R. 1993. Late Celtic hanging-bowls in Lincolnshire and South Humberside, in Vince (1993): 4570.

BURLEIGH G.R., D.A. WENT & K.J. MATTHEWS. 1990. An archaeological evaluation at Hollard's Farm (Mimram Valley Golf Course), Codicote, Hertfordshire. Letchworth: North Herts. District Council, Field Archaeology Section. Evaluation report 5.

CARRITHERS, M. 1992. Why humans have cultures: explaining anthropology and social diversity. Oxford: Oxford University Press.

CLARKE, D.L. 1978. Analytical archaeology. (2nd edition revised by Bob Chapman.) London: Methuen.

CLIFFORD, J. & G.E. MARCUS (ed.). 1986. Writing culture: the poetics and politics of ethnography. Berkeley (CA): University of California Press.

COHEN, R. 1978 Ethnicity: problem and focus in anthropology, Annual Review of Anthropology 7:379-403.

COLLIS, J. 1994. Celtic fantasy, British Archaeological News (11): 5.

CUNLIFFE, B. 1988. Greeks, Romans and barbarians: spheres of interaction. London: Batsford.

DELANEY, F. 1986. The Celts. London: Hodder & Stoughton.

DUBERMAN, M.B., M. VICINUS & G. CHAUNCEY (ed.). 1991. Hidden from history: reclaiming the gay and lesbian past. Harmondsworth: Penguin.

FARWELL, D.E. & T.I. MOLLESON. 1993. Excavations at Poundbury 1966-80 2: The cemeteries. Dorchester: Dorset Natural History and Archaeological Society. Monograph 11.

FREER, P. 1987. AIDS and ..., in Hanscombe & Humphries (1987): 52-69.

GIDDENS. A. 1993. Sociology. (2nd (edition.) Cambridge: Polity Press.

GREEN, M. 1989. Symbol and image in Celtic religious art. London: Routledge.

HANSCOMBE, G.E. & M. HUMPHRIES (ed.). 1987. Heterosexuality. London: Gay Men's Press.

HEBIDGE, D. 1979. subculture: the meaning of style. London: Methuen.

HIGHAM, N.J. 1992. Rome, Britain and the Anglo-Saxons. London: Seaby.

1994. The English conquest: Gildas and Britain in the fifth century. Manchester: Manchester University Press.

HODDER, I. 1991. Reading the past: current approaches to interpretation in archaeology. (2nd edition.) Cambridge: Cambridge University Press.

HODGES, R. 1989. The Anglo-Saxon achievement: archaeology and the beginnings of English society. London: Duckworth.

HUGHES, W.R. 1939. The Belgo-Roman occupation of the Welwyn area, Transactions the East Hertfordshire Archaeological Society (1938) 10: 141-9.

JOHNSON, G.A. 1977. Aspects of regional analysis in archaeology, Annual Review of Anthropology 6: 479-508.

KOHL, P.L. 1993. Limits to a post-processual archaeology (or, The dangers of a new scholasticism), in Yoffee & Sherratt (1993a): 13-19.

KRISTIANSEN, K. 1990. National archaeology in the age of European integration, Antiquity 64: 825-8.

LARICK, R. 1985. Spears, style, and time among Maa-speaking pastoralists, Journal of Anthropological Archaeology 4: 206-20.

1991. Warrior and blacksmiths: mediating ethnicity in East African spears, Journal of Anthropological Archaeology 10: 299-331.

LOWE, R. & W. SHAW. 1993. Travellers: voices of the New Age nomads. London: Fourth Estate.

MILLETT, M. 1990. The Romanization of Britain: an essay in archaeological interpretation. Cambridge: Cambridge University Press.

MILROY, L. & J. MILROY. 1992. Social network and social class: toward an integrated sociolinguistic model, Language in Society 21: 1-26.

MORRIS, M.N. 1988. Changing perceptions of the past: the Bronze Age - a case study, in Bintliff (1988): 69-85.

PLATT, C. 1978. Medieval England: a social history and archaeology from the Conquest to 1600 AD. London: Routledge & Kegan Paul.

POWELL, T.G.E. 1958. The Celts. London: Thames & Hudson.

ROOK, A.G. 1968a. Investigation of a Belgic occupation site at Crookhams, Welwyn Garden City, Hertfordshire Archaeology 1: 51-68.

1968b. A Belgic ditched enclosure at Nutfield, Welwyn Garden City, Hertfordshire Archaeology 1: 121-3.

1970a. A Belgic and Roman site at Brickwall Hill, Hertfordshire Archaeology 2: 23-30.

1970b. Investigation of a Belgic site at Grub's Barn, Welwyn Garden City, Hertfordshire Archaeology 2: 31-6.

SHANKS, M. & C. TILLEY. 1987. Social theory and archaeology. Cambridge: Polity Press.

SHENNAN, S.J. 1989. Introduction: archaeological approaches to cultural identity, in S.J. Sherman (ed.), Archaeological approaches to cultural identity: 1-32. London: Routledge. One World Archaeology 10.

SMITH, R.A. 1912. On late-Celtic antiquities discovered at Welwyn, Hertfordshire, Archaeologia (1911-12) 63: 1-30.

STEAD, I.M. & V. RIGBY. 1986. Baldock: the excavation of a Roman and pre-Roman settlement 1968-72. London: Society for the Promotion of Roman Studies. Britannia Monograph 7.

STENTON, F.M. 1971. Anglo-Saxon England. (3rd edition.) Oxford: Clarendon Press. The Oxford History of England 2.

TAINTER, J.A. 1988. The collapse of complex societies. Cambridge: Cambridge University Press.

VINCE, A. (ed.). 1993. Pre-Viking Lindsey. Lincoln: City of Lincoln Archaeological Unit. Lincoln Archaeological Studies 1.

WEEKS, J. 1991. Inverts, perverts, and Mary-Annes: male prostitution and the regulation of homosexuality in England in the nineteenth and early twentieth centuries, in Duberman et al. 1991: 195-211.

WOODWARD, A.D. 1993. Discussion, in Farwell & Molleson (1993): 215-39.

YOFFEE, N. & A. SHERRATT (ed.). 1993a. Archaeological theory: who sets the agenda? Cambridge: Cambridge University Press.

1993b. Introduction: the sources of archaeological theory, in Yoffee & Sherratt (1993a): 1-9.

KEITH J. MATTHEWS Chester Archaeology, Grosvenor Museum, 27 Grosvenor Street, Chester CH1 2DD, England.