In: J. Hines (ed.). The Anglo-Saxons from the Migration Period to the eighth century: an ethnographic perspective.

(Studies in Historical Archaeoethnology 2). Woodbridge: Boydell Press 1997. 125-170.

Early Anglo-Saxon social structure

Heinrich Hrke This paper is intended to provide an overview of the main features of social structure in England during the fifth to seventh/eighth centuries AD. The term 'social structure' is applied loosely here, referring to all aspects of social organisation and differentiation. The emphasis is squarely on archaeological data, and given the state of research and the nature of the available evidence, burials and cemeteries figure prominently in this survey. Historical evidence will be brought to bear on the various questions where possible, but neither special expertise nor exhaustive treatment is claimed here for the evaluation of written sources. It hardly needs emphasizing that the two main types of evidence relate to different, but overlapping phases: the burial evidence dates mainly to the fifth to seventh centuries whereas written sources do not start before the seventh century. Archaeological evidence, in particular cemetery evidence, has been used since the 19th century for social inferences. While most archaeologists accept the validity of such an approach, not everybody agrees. Wilson has clearly stated his view that Anglo-Saxon archaeology provides "practically no clues to political structure, to national boundaries, to marital practices or to the rights of the individual. Occasionally a very rich grave may give an idea of social structure and the wealth of a particular person, but such indications are rare" (Wilson 1976, 3; for a radically different view published in the same year, cf. Addyman 1976). Even if we accept the possibility of social interpretation of cemeteries, their analysis raises two basic questions: To what degree does archaeological evidence reflect social relations? And in particular: does funerary evidence reflect social relations as they were active in living society, or does it reflect them in a way which is specific to ritual and the realm of the dead? This is not the place to tackle these questions in detail (cf. Hrke 1994a; forthcoming), but it would be wrong to ignore them entirely because they pose fundamental problems for the social interpretation of AngloSaxon funerary evidence. Some of these problems should become apparent below. The structure of this paper is determined by archaeological

methodology (i.e. possible approaches and available evidence) as much as by general considerations. It begins with a look at age and sex, both of them biological categories which assume cultural significance in a social context. They are followed by a discussion of social groups and aggregates up to the level of local communities. Social classes and hierarchies, categories which have often been central to archaeologists' perceptions of social structure, are considered next, supplemented by a glance at the role of ethnic divisions in early Anglo-Saxon society. Age groups In social anthropology, age classes or sets have long been recognized as important subdivisions of tribal societies, the transition from one age group to the next often marked by rites de passage (van Gennep 1960). In complex socities, the importance of such age classes is often obscured by the variety of other factors, but age distinctions persist into our own times and our own society, together with their attendant rites of passage. It seems that early medieval societies have rarely been analysed from this perspective, and most studies have tended to concentrate on children (e.g. K. Arnold 1986; Herlihy 1995, 1) 215-243; Ottinger 1974; Schwab 1982) . Following Aris' suggestion of the 'discovery of childhood' after the middle ages (Aris 1962), there has been some debate on the medieval recognition of childhood (K. Arnold 1986, 57; Herlihy 1995, 215-219), and on the differentiation of children from adults in the Anglo-Saxon period (Keufler 1991). Herlihy has distinguished Classical, barbarian and medieval childhood, with the barbarian childhood of the Migration and post-Roman periods being characterized by "an atmosphere of affectionate neglect" (Herlihy 1995, 225). However this, essentially psychological, question is answered, it seems sufficiently clear that several age distinctions can be recognised in the earliest Anglo-Saxon sources. The extant laws suggest that children were comparatively well protected (Keufler 1991, 827). They also identify the age of 10 as an important legal threshold: it marks the age of inheritance (Hlothere and Eadric, 6; AD 673-685?) and criminal responsibility (Ine 7.2, AD 688-694; later raised to 12 years, cf. II Athelstan 1, c. AD 925-930). The biographical information in saints' lives and heroic poetry suggests several age thresholds which do not coincide with the one in the laws: 7-8 years, 14-15 years and mid-20s. The first threshold signalled the end of childhood and the beginning of light work or

education, for high-status children not infrequently in fosterage outside the family (Felix's Life of St. Guthlac, I; Bede, HE V.24; Beowulf, lines 2428-2430). This threshold appears to have been a widespread feature in Northwest European societies, from Ireland to Scandinavia. For boys of noble background it meant the beginning of weapon training and hunting as a preparation for the life of a warrior (Davidson 1989, 1112, 20-21). The second threshold at 14-15 years of age marks the boy's entry into adult life (Eddius Stephanus, Life of St. Wilfrid, II.7; Bede, HE V.19; Felix, Life of St. Guthlac, II, XVII-XVIII). In the case of noble warriors, this phase may last about 10 years after which the mature man retires to the monastery (ibid., XIX) or to his acquired or inherited estate. A comitatus would therefore have comprised warriors mostly between the ages of about 15 and 25, with various levels of experience (cf. Beowulf, lines 2626-2628; Finnesburh; Maldon, Zeilen 2) 152-154), but made up of what is essentially one age set . Further distinctions are blurred; the sources refer to old warriors in hall as well as to noble advisers and friends of the king (Bede, HE II.13; Beowulf, Zeilen 356-357), both presumably older than the ordinary comitatus warriors. Archaeological evidence can supplement some of this textual information, and redress the social and gender bias in favour of noble males shown in the written sources. Most of the following observations are based on inhumation burials which offer better preservation of grave-goods, and an easier and more reliable age determination from skeletal remains, than cremations. There are some continuous trends which run across all age groups: the frequency of burial in multiple graves decreases with age, and the average number of grave-goods increases steadily. More specific evidence of age sets and thresholds can be found in several aspects of the burial ritual, and in types and sizes of objects deposited in graves. A first observation which may provide an idea of the attitude to infants in early Anglo-Saxon society relates to the possible underrepresentation of the youngest age groups in pagan cemeteries. In a sample of 47 cemeteries with just under 3000 aged individuals, only some 2.3% fall into the neonate and young infant categories (up to about 18 months of age; Hrke 1992a, 184 table 28). A separate analysis of 12 cemeteries arrived at a compatible figure of 6% under the age of three (Crawford 1993, 84). This low proportion represents a problem because infant mortality should be considerably higher if data from pre-industrial, Third World communities are anything to go by. This problem is common to virtually all early medieval (and many prehistoric) cemeteries across Western Europe. It has been recognised by Continental scholars several decades ago, and has been discussed

at some length in the archaeological and anthropological literature (for a summary, cf. Steuer 1982, 513-515). The alternatives are that either the comparative figures from Third World contexts cannot be applied to pre-industrial Europe, or that young infants were disposed of in different ways. Actually, both may apply to some degree: the former is now advocated by a number of physical anthropologists (Herrmann et al. 1990); and there seems to be some evidence (albeit not much) to substantiate the latter. Also, it appears that most Late Roman and Late Saxon cemeteries have somewhat higher proportions of infants (cf. Watts 1989). The disposal outside formal burial grounds of individuals under the age of two to three years of age would imply a low esteem of this age group, and a low status in society. This is underlined by the fact that the majority of individuals of this age were buried in multiple graves, and that two thirds of them were buried without grave-goods. A first threshold seems to have been reach around two to three years of age. Burial in individual graves and the deposition of gravegoods are more common from this age, and a weapon (mostly a single spear, occasionally an arrow) may now be put into the grave, although this still remains a rarity (Hrke 1992b, 156 table 4). Girls reached another threshold in the age group 7 to 14 years: there is a steep increase in the number and variety of objects in their graves, resulting from the increasing frequency of burial in full dress kit (a pair of brooches, necklace, keys or chatelaine/girdle hanger). A closer look at the relevant cases suggests that this point was reached at an age of between 10 and 12 years. The age of 12 certainly marked a threshold for boys: shields and swords were deposited occasionally from the age of 12 onwards (ibid.; Dickinson and Hrke 1992, 68 table 19). From the age of 15, there is a further, slight increase in the number of objects deposited in female graves, suggesting an elaboration of the dress kit in the juvenile age group. In male burials, the deposition of the body on its side in a flexed position all but disappears after the age of 15. The attainment of 'adult' age (in our modern definition), at around 18 to 20 years, is connected with further changes in the burial ritual. The orientation of bodies becomes more standardised, with the head pointing roughly west; coffins and other grave structures become much more frequent (about double their frequency in sub-adult graves), and the rare chamber graves make their appearance in this age group. But it is the grave-goods that most clearly distinguish the adult group from children and juveniles: artefacts such as musical instruments, scales, gaming pieces, horse harness, axes and seaxes have been found only in adult burials (Hrke 1992a, 160 fig. 26). Differentiation within the group of adult males is difficult to identify on the basis of the archaeological and skeletal evidence, but horse gear, axes and seaxes appear to be deposited preferentially with older adults (aged 30 and

over). A similar analysis is still lacking for female adult burials

3)

Skeletal evidence can provide additional information. Mortality peaks provide indirect evidence of particular stress at certain ages, but interestingly enough, the picture that emerges from mortality figures is too unclear for interpretation because of marked local variations. The study of Harris lines is of greater weight because it provides direct indicators of periods of arrested growth through stress, malnutrition and illness, but not many cemetery populations have been analysed for this feature. At the Late Saxon site of North Elmham, girls experienced most of their growth-arresting episodes after the age of seven, while boys underwent more such episodes, and most of them after the age of 12 (Wells and Cayton 1980, 298). This suggests gender differences in upbringing and socialisation, and it may even mean that from the age of 12, "boys were exposed to something like the full rigours and physical demands of adulthood" (ibid.). Although this would tie in with threshold ages suggested by other types of evidence, it is difficult to extrapolate from this one sample, and to extend this conclusion back to the Early Saxon period without further supporting studies of this kind. Settlement evidence does not contribute anything to our understanding of age divisions and age-related status in Anglo-Saxon society, except perhaps in the absence of any subdivisions or partitions in buildings and settlements which could be related to specific age groups. All that this absence of special residential or activity areas of children, adults or old people suggests that all age groups lived and worked side by side in social groups which cross-cut age divisions. The above outline provides us with three series of age thresholds from various types of evidence (Table 1). The age classes suggested, in particular, by the biographical data are strikingly similar to those encountered in societies with gerontocratic structures. For example, the Lokop of North Kenia have a system in which the males proceed through three age classes: boys (7 - 15 years), warriors (15 - 30 years), 4) and elders (over 30 years) . Age structures in other post-Roman societies appear to have been similar to the Anglo-Saxon pattern, but not identical. In the Visigothic laws, the ages of 1, 10, 15, 20, and 45/50 (females/males) mark changes in wergild levels, with the decline in its value after the age of 45/50 reflecting the low value of ageing members to society (Herlihy 1995, 222-223). However, the divergences between Anglo-Saxon literary perception, legal codes, and ritual expression of age classes is conspicuous. The biographical data appear to suggest four relatively

clear-cut age groups which are at odds with the five groups which can be inferred from burial ritual. The only coincidence between written sources and archaeological evidence is provided by the age of criminal responsibility and the onset of burial with full female dress or sword and/or shield, respectively, somewhere between the ages of 10 and 12. There are, at least, two possible explanations for these discrepancies. We may be looking at two different chronological horizons, and Christianity may have had an impact on the later horizon represented by the biographical data (cf. Crawford 1993; Herlihy 1995, 225-228; Watts 1989). Alternatively, we may have to consider the possibility of a systematic divergence between the cultural perception of biographical age, on the one hand, and the biological age inferred from skeletal evidence, on the other. Finally, it is interesting to note a possible change in the attitude towards children in the seventh/eighth centuries. Children were included in the (largely symbolic) weapon burial rite which ended around AD 700 (Hrke 1992b). But the deposition of weapons in subadult burials ceased earlier than in adult burials: weapons were phased out gradually over the seventh century, first from children's burials, then from juvenile burials, and finally from adult burials (Hrke 1992a, 184). Also, children are never found on their own, i.e. without an adult burial, under barrows of seventh/eighth century date. Shephard (1979, 67) has interpreted this as evidence that adult status was part of the 'superordinate' social status represented in the barrows. If this reflects a change in attitude, it may be connected with the introduction and spread of Christianity (cf. above), or with the emergence of early state structures which might also have led to a change in the status of children through the formalisation of inheritance rules and the decline in importance of kinship. Sex and gender 5) Interest in aspects of sex and gender in Anglo-Saxon society started well before the recent attention to gender and feminist perspectives. For a number of reasons, the study of gender has usually been equated with the study of women's place in society. For AngloSaxon England, the starting point of the debate was Stenton's pioneering study of place-names which suggested that "women were associated with men on terms of rough equality in the common life of the countryside" (Stenton 1943, 13). This was echoed by his wife, Doris Stenton's book in which she concluded that "women were then more nearly the equal companions of their husbands and brothers than at any other period before the modern age" (D. Stenton 1957, 348). Later

writers on this subject tended to concur (e.g. Dietrich 1980, 43-44; Fell 1984a), but there has also been some disagreement. Klinck, in particular, has argued on the basis of the laws that women's rights increased during Anglo-Saxon period, but even so, Anglo-Saxon women had never been the equals of men (Klinck 1982, 115, 118). There appears to be evidence to support both sides of the debate, although to varying degrees. One aspect that all commentators have agreed on, though, is the clear assumption in virtually all extant sources that marriage was monogamous. In Beowulf, wives and daughters of men of high rank appear mostly as peaceweavers and cupbearers, and other women do not appear at all (Fell 1984a, 67). The Anglo-Saxon Chronicle and Bede present a very similar picture of a male scheme of social and political power relations. Where females are mentioned in the laws, from Ethelbert's (between AD 597 and 616) to Alfred's (between AD 885 and 899), a woman's status is almost invariably determined by that of her father: marriage did not lead to a change of social class (Loyn 1974, 206). Apart from her class, the most important status distinction of a free woman is her marital status (unmarried, married, widow); in the case of men, this is less often mentioned and clearly of much less importance (Fell 1984a, 62). And the laws of Ine mention almost in passing, as a matter of course, that a woman must obey 'her lord' (Ine 57), i.e. her husband. But there are also indications of less uneven economic and power relations between the sexes. Some female rulers are referred to in the Anglo-Saxon Chronicle (e.g. Queen Seaxburh of Wessex ruling after Cenwealh death, ASC s.a. 672; and later thelfld, 'Lady of the Mercians', ruling AD 911-918 after her husband's death, ASC s.a. 910, 913, 916, 917, 918; cf. Wainwright 1959). thelburh, the wife of King Ine of Wessex, is even mentioned as having destroyed Taunton (s.a. 722) during the reign of her husband, suggesting the possibility of military leadership by a noblewoman. This would tie in with the brief hint in Beowulf (l. 2060) that Freawaru, the daughter of Hrothgar, would bring her own followers abroad to the wedding. The poem, unambiguously male-centred as it is, also refers to several queens (Wealtheow, Hygd and Freawaru; Beowulf ll. 612-661, 1926-1931, 2015-2069) who distribute gifts to warriors, giving them an important role in the male world of the hall, and possibly implying some control over property. Bede's Ecclesiastical History portrays several women from noble or royal families as instrumental during the Conversion and in leadership positions in the early Church, mainly as abbesses (e.g. HE IV.21). This would be surprising if they could not have held similar positions of authority in contemporary secular society.

Fell (1984a, 59-61; 1984b) has argued that a law of Ethelbert (73) may mean that a woman could be in control of keys to treasure stores, and drew a connection to the keys (chatelaines, girdle-hangers) found in the burials of many female adults and juveniles (from the age of about 12). Certainly, the much later laws of Cnut (76.1a; AD 10201023) stipulate that the wife must look after the keys to her storeroom, her chest and her coffer. Earlier laws also define, or imply, a degree of independence of the wife from her husband. In Kent, rules were drawn up about the division of the marriage goods in case the woman went away, with or without her children (Ethelbert, 79, 80). And in both, Kent and Wessex, a woman's criminal responsibility was judged independently of that of her husband: she was not to be punished if she was not aware of his crime (Wihtred, 12, AD 695; Ine, 7, 57). Charters and wills (predominantly of Middle and Late Saxon date) provide a picture of male economic dominance punctuatued by instances of female equality or independence. Some 93% of all charters were granted to men (Meyer 1980). Charters granted to women did not differ in form or content, but they were predominantly given to women of royal families (ibid., 59-60). And while many charters were granted to husband and wife together, giving the impression of economic partnership, the reason behind this may be that a woman was not guaranteed possession of her husband's land after his death unless this had been specified in the charter. However, of the 39 wills compiled by Whitelock (1939), ten are by female testators alone, and another four by wives together with their husbands (Lancaster 1958, 362). In other words: women are represented in more than a quarter of the extant wills (cf. Sheehan 1963, 70-71), which is more than one would expect if Anglo-Saxon women had a uniformly low economic status. A certain measure of economic control, albeit over one particular item, is suggested by the regulations concerning bride payment and morning gift (Ethelbert, 81). Fell (1984a, 56-57) has concluded from laws and later charters that women kept control of land received as morning gift. And if a woman died childless, the morning gift would go to her paternal kinsmen (Ethelbert, 81), not to her husband, implying a separate economic identity at least in this respect. Middleton has pointed out that societies with a bridewealth system (involving the payment of bride price and/or morning gift to the wife and her family, as against the dowry system) are usually characterised by high female productivity (Middleton 1995, 57-58). Therefore, the Anglo-Saxon marriage system may be an indicator of the economic significance of female agricultural or craft production (ibid., 58-59). It is also worth mentioning that Cynetryth, the wife of king Offa, had coins struck in her name (D. Stenton 1957, 2), but it is not clear if this can be taken to

indicate a measure of economic independence. It has often been suggested that Christianity brought about a change in women's status because ecclesiatical writings protraying women as inferior and impure are assumed to have affected social attitudes. However, Fell (1984a, 13-14) has emphasized that these writings largely remained theory, and that - at least initially - little changed at the practical level. One may add that, if there was any change in the perception and status of women after about AD 600, this may have had as much to do with Christian theories as with the emergence of statehood in England. One concomitant of the development of state structures is the decline in importance of kinship and family links, and this is likely to have affected gender relations as it probably affected the status and treatment of children (cf. above). Early Saxon burial ritual put a remarkably strong emphasis on gender display through artefacts which formed part of the dress or were deposited separately in inhumation graves. While there are gender-neutral burials without diagnostic objects or without any gravegoods at all (these are the majority of infants and young children, and a proportion of up to 30% of juveniles and adults), the majority of adult burials show a marked male/female dichotomy. This phenomenon is more pronounced in inhumation graves (Brush 1988) than in cremations (Richards 1981), but it is present in both. The male 'kit' consists of weapons and tools (excluding textile-working tools); the female 'kit' is made up of dress ornaments (two or three brooches, bead necklace, pins), keys and girdle hanger (chatelaine) suspended from the belt, and textile-working tools (spindle whorls, loom weights, weaving batten). The mutually exclusive gender attribution of these artefacts has been supported, in the overwhelming number of cases, by skeletal sexing. There is a small number of exceptions where archaeological and skeletal sexing are in contradiction. It is not impossible that some of them are genuine cases of cross-dressing, but the very rarity of this phenomenon suggests that it may be the product of the error span inherent in skeletal sexing. This suspicion is largely confirmed by a current re-study of the archaeology and biology of key cases from this 6) group . Also, there is not a single unambiguous case of mixing of the gender-specific artefact kits in undisturbed graves. The rare instances of brooches in graves with weapons involve single brooches, reminiscent of Roman and Celtic male dress, but not of Anglo-Saxon female dress; and the few cases of weapons in graves with female kit show features which are best explained as secondary re-use of parts of weapons as tools (Hrke 1992a, 179-182). The only exception may be

arrowheads found with female children, an observation matched elsewhere in post-Roman Europe (ibid., 181). Even supposedly gender-neutral artefacts may show some degree of gender differentiation. The humble iron knife, by far the most frequent object in burials of either sex and all age groups, is a good example (Hrke 1989b). Adult women were never buried with knives the blades of which were longer than 126 mm (5"); such large knives were only found with adult men. A parallel for this may be found among the Lapps (Saami) of northern Scandinavia where knives are part of the display dress, with men carrying the longer knives. Among Anglo-Saxon children, there is no gender-related difference in sizes of knives, but children with a male kit (i.e. weapons) are twice as likely to be accompanied by a knife than are children with a female kit (63 and 31%, respectively). Other aspects of the inhumation ritual confirm the emphasis on gender display. While the standard deposition of the body was extended on the back, with the head pointing somewhere around west, females were more often deposited with the head to the south than males were (under 10% of male adults); and they were more often in a flexed position on the side than males (only about 5% of male adults; Hrke 1992a, 151-155). Both traits link the mortuary treatment of female adults with that of children (c. 20% with head to south, and c. 20% flexed; cf. above). In her analysis of the Cambridgeshire cemetery of Holywell Row, Pader (1982, 130) has suggested that this could mean that women and children were assigned a similar status, different from (and by implication, inferior to) men. It should, however, be noted that the gender-specific differences in burial treatment vary from cemetery to cemetery. This need not weaken Pader's interpretation, but it does indicate local variations in gender-linked ritual symbolism. It is conceivable that this also implies local and regional variations in gender relations. Gender differences in burial ritual, and in particular the marked gender dichotomy in grave-goods, are likely to relate to Anglo-Saxon perceptions of gender roles and statuses in society. The frequency of weapons in male burials (just under 50% of male adults; Hrke 1989a), and the exclusive appearance of textile-working tools in female burials is echoed several centuries later in the Will of Alfred which refers to the 'spear side' and the 'spindle side' of his family (Whitelock 1979, 537 footnote 1). This material culture symbolism should not, however, tempt anybody to draw seemingly straightforward interpretations: all males buried with weapons had not been warriors (Hrke 1990); similarly, all females buried with tools had probably not been weavers or spinners.

Burial wealth, i.e. the number and quality of objects in a grave, can be understood and interpreted in either symbolic or economic terms. Christlein (1973) has seen the wealth of male burials as directly reflecting the economic power of individuals, and this is a widespread assumption underlying many quantitative analyses of post-Roman burials. Comparisons of male and female burial wealth in Anglo-Saxon cemeteries agree on one key point: female burials are 'richer' than male burials (Arnold 1980, 110 table 4.6; Pader 1982; Shephard 1979). They are furnished, on average, with a larger number of objects, a larger variety of artefact types, and a higher proportion of objects made of, or decorated with, precious metals or other rare materials (amber, rock crystal, etc.; Table 2). This wealth differential is absent among infants and young children, but starts abruptly among older children (aged 7-14 years), and continues through all age groups. There is broad agreement that this wealth differential is a consequence of the differences between male and female dress, but interpretations vary. Arnold (1980, 132) has suggested that the wealth of a woman reflects not just her own status, but that of her husband as well. Shephard (1979, 58) has claimed that there was the difference between the sexes was also a difference between social status and economic power: male grave-goods were determined social status, female grave-goods by wealth. Pader, however, has pointed out that costume styles are not a suitable starting point for wealth analysis because costume may be determined by many other factors (Pader 1982, 132). On the Continent where the gender differential is similar, Jrgensen has come up with a different explanation: families expressed their status through the burial of the first deceased of the heads of families, and because of the lower life expectancy of women, the result was a preponderance of rich female burials (Jrgensen 1987; 1990, 66 fig. 48, 88; 1991, 30-31). Skeletal evidence provides some information on gender differences in living conditions and economic status. In the Early Saxon period, women had a shorter life expectancy: 33.1 years as against 34.7 years for men (Brothwell 1972, 83 table 25, calculated on adults only). While life expectancy appears to have risen in the Middle and Late Saxon periods (e.g. to 35.8 and 38.2 years respectively at North Elmham; Wells and Cayton 1980, 252), the gender differential persisted. The lower female life expectancy is significant because it must have led (as it did in early and high medieval societies across Europe) to a higher proportion of male adults in the living population, a trend which was only reversed in the later Middle Ages (Herlihy 1995, 59-60). Physical anthropologists have interpreted this mortality differential as a consequence of biological rather than economic

conditions, resulting from childbearing, in general, and a high rate of death in childbirth, in particular. Differential access to food cannot be excluded as a contributory factor, though. At the Early Saxon cemetery of Worthy Park, Wells has observed that 25% of women had enamel hypoplasia, but only 10% of men, suggesting that girls were more likely to suffer from stress and malnutrition than boys (Hawkes and Wells 1983). However, this seems to be contradicted by the fact that at this particular site, the mean age at death of male adults (36.2 years) is lower than that of female adults (37.0 years; unpubl. report by C. Wells, pers. comm. S.C. Hawkes). Gender differences in health patterns may provide further insights on differential living conditions, but skeletal data have not yet been analysed systematically from this perspective. Theoretical and anthropological considerations suggest what to look for in future research. If Anglo-Saxon society was the product of large-scale migration from the Continent (a view no longer as widely accepted as it was a decade ago), colonial societies might provide some clues. Stoodley's survey of gender roles in modern frontier societies suggests that distinctions of gender roles become more blurred, and gender differentiation of work more flexible among settlers than they were in the societies of origin: women have to carry out heavy, 'male' work whereas men have to care for children and do more household chores (Stoodley 1993). This finding is strikingly reminiscent of Doris Stenton's pronouncement on the "rough and ready partnership" (D. Stenton 1957, 348) between the sexes in Anglo-Saxon England, and of Dietrich's suggestion of "role flexibility .. for AngloSaxon women" (Dietrich 1980, 44). It is surprising in another respect: the theoretical expectation of blurred gender roles contrasts with the marked gender dichotomy in burial ritual. Is the 'frontier society' model wrong for post-Roman England? Or was ritual a way of coping with the erosion of traditional gender roles in real life? The clearly symbolic function of Anglo-Saxon burial ritual would lend considerable support to the latter interpretation, as do anthropological observations on the role of ritual in reconciling ideal and reality (Leach 1964). Settlement sites may provide information on the gender aspect of social relations (cf. Hastorp 1991), but there is not much genderspecific evidence in either the buildings or the finds from Early AngloSaxon habitation sites. The mostly undifferentiated interiors of the typical timber-built dwellings (sometimes called 'halls') do not allow to identify separate activity areas. However, about a quarter of all timber 'halls' on a variety of fifth to seventh century sites, from ordinary to royal, have a partition at one end (Marshall and Marshall 1993, 399). This provides a screened-off area of about one fifth of the interior

(usually less than 2 x 5 metres), without a hearth of its own. The purpose of these compartments is unknown (Rahtz 1976, 88), but it is often presumed that they were "private quarters" (Addyman 1972, 304; Addyman and Leigh 1972, 23 note 5) for the higher-ranking members of the household. Given the existence of such partitions in large halls on high-status sites such as Cowdery's Down (Millett and James 1983, Fig. 31) and Yeavering (Hope-Taylor 1977, 127 fig. 60, 130 fig. 61) where written sources suggest that the highest-ranking individuals may have had their sleeping quarters in separate buildings (Beowulf l. 1236; Addyman 1972, 304), it is conceivable that this feature provided a bower for the women. Other functions such as storage, byre or stairwell cannot be excluded, though (Marshall and Marshall 1993, 399; Powlesland, this volume). Heroic poetry gives the impression of the rest of the hall being an almost exclusively male activity area, but this may well be a biased representation, and in any case, it relates to high-status halls. There has been a good deal of speculation and discussion concerning the function of the numerous sunken-floor buildings (SFBs) in early Anglo-Saxon settlements. Only in exceptional cases are they still seen as dug-out dwellings, e.g. for the first phase of Mucking (Dixon 1993, 129; but cf. Hamerow 1993, 86). On most sites, they seem to have been used as ancillary buildings. Occasionally this may have been in connection with metalworking (sixth/seventh century Mucking; Hamerow 1993, 17); much more often, though, they contain evidence of textile working (Rahtz 1976, 76, 77 fig. 2.12; Arnold 1988, 25 fig. 2.3). The concentration of loomweights and spindlewhorls in SFBs varies from site to site (Bell 1977, 226; Hamerow 1993, 15, 1719, 64; Losco-Bradley 1977, 362; West 1985, 138-139), but overall there can be little doubt about the link between many buildings of this type with textile production. Significantly, the two largest SFBs found on Early Saxon sites (at Upton and Charlton, with sizes around 9 x 5.5 metres) have produced loomweights the association of which with the respective buildings is clear because both had been destroyed by fire (Addyman 1972, 281; Champion 1977, 365). The exclusive association of textile tools with female burials is strongly suggestive of textile production having been (or been presented as being) in the hands of women. The even distribution of SFBs within settlement sites means that weaving sheds were not concentrated to create a female activity area for the entire settlement. Rather, each household or farmstead had its own female activity area defined by its own weaving shed (if it had one). Other gender-specific areas in settlements of this period have yet to be identified.

Social groups: family, household, community The raison d'tre of this seemingly diverse section is the nature of the archaeological evidence in which kinship units, economic units and settlement units cannot always be distinguished easily, neither in cemeteries nor in settlements, and therefore require discussion side by side. It stands to reason that in Anglo-Saxon society, these categories overlapped substantially, anyway. Kinship was essential to the working of early Anglo-Saxon society: it determined personal status, provided genealogical links, and gave access to land; equally importantly, the kindred guaranteed individual safety by paying part of the wergild to avert feuds, and by conducting feuds on behalf of its members (Loyn 1974). Further discussion of kinship structures can be kept brief here because a separate paper is devoted to it (Charles-Edwards, this volume). The Anglo-Saxon kinship system was most likely bilateral, with a weak patrilateral bias (including, probably, virilocal residence; Lancaster 1958). Such a system which characterized early medieval kinship before the twelfth century (Herlihy 1995, 143-145) mitigates against the creation of large descent groups as they are typical of unilateral kinship systems (Lancaster 1958, 232, 359; but cf. Charles-Edwards 1972). In keeping with this theoretical expectation, there is indeed an emphasis in the Anglo-Saxon kinship terminology on the nuclear family, and an absence of evidence for territorial clans in the charters (Loyn 1974, 198, 202). In spite of these limitations, it is conceivable that kinship was also the initial basis of larger political units, as in most pre-state societies. Scull (1993) has suggested that regional structures, and eventually kingdoms, were created by the rise to dominance of individual lineages. The rise of territorial lordship and Christian kingship from the seventh century, in turn, suppressed the further development of kin groups into strong, land-holding institutions (Loyn 1974, 209). Textual references to households and other residential units are not very frequent, and they provide only limited information on their composition, sizes and structures. The early laws imply the existence of dependants, or other persons of lower status, in the households of kings (Ine, 33) and freemen (Ethelbert, 16, 25; Ine, 22, 63). In Kent and Wessex, the master of the house was responsible for the acts of free and unfree persons of his household (Hlothere and Eadric, 1; Wihtred, 14; Ine, 50). Thus, the evidence of the laws suggests that individuals of different status lived under one roof, or in very close and constant proximity, and that such households were not just residential units, but also to a certain degree legal units.

Other types of sources are not very helpful in the elucidation of this aspect. Charters are generally later than the period under discussion, and their nature, and the kinds of details they provide, have not encouraged an analysis of the demographic and social structures of residential units. The gift of 87 hides at Selsey to St. Wilfrid included an unspecified number of free men as well as 250 slaves (John 1964, 26). On the conventional assumption that one hide should supply one man for military service, these figures give us an idea of the proportion of slaves to free households of a certain status on this particular estate. But the absence of further information, here and in other cases, makes it impossible to draw more precise inferences of the kind facilitated by Carolingian records on the Continent (Herlihy 1985, 62-77; Le Jan 1995). Cemeteries provide more direct evidence for sizes of local communities and their structures. At the upper end of the scale, some cremation cemeteries (such as Sancton I or Spong Hill) may originally have contained several thousand urns, which has led to the suggestion that these were the cemeteries of several settlements or local communities (Arnold 1981; Hills and Penn 1981, 22). Inhumation, and mixed, cemeteries appear to correlate much better with the sizes of excavated settlement sites (cf. below). Arnold (1984a, 125; id. 1988, 166) has used the evidence from 13 cemeteries to calculate the sizes of living communities; he arrived at figures between 15 and 36 individuals. While the cemetery chronologies used for these calculations may need some adjustments, the figures nevertheless give an useful idea of the size of a community which buried their dead together. Internal structuring of cemeteries suggests the existence of distinct sub-groups within local communities. This structuring does not take the form of tangible partitions of cemetery areas, but of burial clusters identified on the basis of similarities in grave-goods or other aspects of ritual treatment (cf. Pader 1982), horizontal stratigraphy, spatial relationships between graves of men, women and children, or occasionally skeletal evidence. Cemeteries may be classified by the type of internal structuring observed (arranged in the order of their frequencies; cf. Hrke 1992a, 169-172): (1) polycentric cemetery, with independent development of several contemporaneous sectors (plots) which coalesce over time (e.g. Alfriston, Andover, Holywell Row, Polhill, Sewerby, and Spong Hill inhumations); monocentric cemetery, with several contemporaneous sectors radiating out from a common focus (e.g. Bergh Apton, Berinsfield

(2)

and Collingbourne Ducis, possibly Petersfinger; Spong Hill cremations); (3) (4) cemetery with simple horizontal stratigraphy, expanding in one direction (e.g. Finglesham); irregular development (possibly Abingdon I).

All these types appear to occur all over England throughout the Early Saxon Period. While types 3 and 4 do not have any unambiguous social implications, types 1 and 2 clearly do. Their contemporaneous sectors, and other conspicuous clusters of apparently related burials, have often been intuitively interpreted as 'family groups' (Adams 1982, 145-146; Aldsworth 1978, 162; Cook and Dacre 1985, 54; Evison 1987, 145; Philp 1973, 200-201; Welch 1980, 266). But at Sewerby, Hirst (1985, 102) observed that burials within each plot showed remarkable differences in status (as indicated by burial wealth), and concluded that these clusters probably represent households rather than families. A similar situation may be given at Broughton Lodge (Willoughby-onthe-Wolds) where poor burials, perhaps the "slave element in the population" (Kinsley 1993, 72), were found on the fringes of the main burial cluster, but the incomplete excavation makes it impossible to judge how many such clusters there were at this site. However, internal wealth differentials within distinct burial clusters need not invalidate the 'family' interpretation: Steuer (1982, 518-525) has suggested that Merovingian society was an 'offene Ranggesellschaft' (open ranked society) in which social differences within families were as important as status differences between families. In the future, the combination of biological information (skeletal data or DNA) and archaeological data may provide new insights. In the cemetery of Berinsfield (Oxfordshire), three sectors radiate out from a common focal area in the centre. For one of the three sectors, the skeletal evidence is detailed enough to suggest that this was the burial plot of a social group consisting of between 10 and 15 individuals at any one time, and comprising two or three distinct sub-groups whose members did not intermarry (Hrke 1992a, 205, 206 Abb. 41; id. 1995). This structure would imply that the plot represents a large household; and if this interpretation is extended to the other two plots (where there is not enough skeletal evidence), one may conclude that the cemetery was the burial ground of three households, perhaps farmsteads, representing the community of a hamlet or similar settlement. Biological data may also help us to decide whether monocentric and polycentric origins of cemeteries imply further differences in the

social development of the communities buried in type 1 and type 2 cemeteries, respectively. Is a type 2 cemetery the reflection of a primary settler group splitting up to form several households in subsequent generations, while type 1 reflects several separate households without initial ties (kinship or otherwise)? Settlement evidence can supplement some of the above inferences. The main types of buildings found in Early Saxon settlements are sunken-floor buildings (SFBs), predominantly used as ancillary buildings (cf. above), and above-ground, timber-built 'halls' which are the normal dwelling of this period. In the fifth to eighth centuries, the majority of these 'halls' are between 6 and 12 metres long, and between 3.5 and 7 metres wide (Addyman 1972; James et al. 1984; Marshall and Marshall 1993; Rahtz 1976; Zimmermann 1988). Taking mean length and width (about 10 x 5 metres; Marshall and Marshall 1993, 381 ill. 10), and assuming that there was no upper floor (firm evidence for which exists only from the Late Saxon period; Rahtz 1976, 90; but cf. Powlesland, this volume), we arrive at average floor areas around 50 sq.metres (cf. also Hamerow 1993, 10). Given that this space was contiguous, it could have accommodated up to about a dozen or so people, assuming that several of them would have been children. Buildings with larger floor areas seem to have been a feature of later, high-status sites (cf. below). The internal partition found in many timber buildings may have provided a separate area, perhaps sleeping area (cf. Rahtz 1976, 69 fig. 2.8), for one group within the household, defined by status, gender or age. Alternatively, such separate areas may, on some sites, have been provided by a smaller type of timber building with only one door which some authors have identified with the bur (bower) known from later sources (Addyman and Leigh 1972, 23). A complete farmstead consisted of a main building, the 'hall', and a few (one to four) smaller, ancillary buildings which rarely appear to have been dwellings (Addyman and Leigh 1973, 19; West 1985, 168). Thus, if there were two or more groups of different status living and working in any given household or farmstead, in most cases they must have lived under one roof. Rural settlements were usually made up of several farmsteads, arranged in a loose conglomeration without any regular lay-out. Sizes range from a single farmstead to about ten or more, but the incomplete nature of most settlement plans of the Early Saxon period makes this aspect difficult to judge. The East Anglian 'village' of West Stow comprised three, or possibly four, farmsteads at any one time between the early fifth and mid-seventh century (West 1985, 150 table 68, 168). Further south, the site of Mucking is estimated to have

consisted of a minimum of eight to ten contemporary units (Hamerow 1993, 90), and the plan of Chalton indicates a sixth/seventh century hamlet of similar size (Champion 1977, 364). Catholme in the Trent valley shows a continuous development of five, or possibly seven, 'holdings', clearly demarcated by ditches, from the late fifth to the tenth century (Losco-Bradley 1977, 359). The settlement evidence, then, suggests local community sizes of between a dozen and well over fifty people, but rarely touching one hundred. This ties in well with the cemetery evidence, although there are few inhumation cemeteries which could represent communities at the upper end of this scale (cf. above). In the future, the analysis of settlements with adjacent cemeteries should provide more concrete figures. At Mucking, Hamerow has calculated on the basis of the grave numbers a population of 94 ( 10%) for the eight to ten farmsteads in the settlement (Hamerow 1993, 90), suggesting a household size of nine to 12 persons each. Unfortunately, there are few such cases so far (for West Heslerton, cf. Powlesland, this volume), and all too often, half of their evidence is incomplete (West Stow cemetery: West 1985, 6469; Bishopstone settlement: Bell 1977). Overall, the evidence seems to point to small communities made up of a few farmsteads each. It is suggested here that households formed the basic residential and economic units in the fifth to midseventh centuries. They comprised individuals and groups of different status, most likely the family of the master of the household, and unfree or semi-free dependants. This suggestion, based primarily on the archaeological and skeletal evidence, agrees well with Lancaster's analysis of the Anglo-Saxon kinship system which led her to postulate that local groups are more likely to have consisted of househoulds with dependants, rather than extended kin groups (Lancaster 1958, 373374). Moreover, simple calculations show that the low life expectancy (cf. above) would result in two rather than three generations under one roof, and an average number of, perhaps, three to four surviving children per family. The latter figure ties in closely with statistics produced by analyses of Carolingian sources which suggest an average family size of 5.79 persons (Herlihy 1985, 69, 70 table 3.2), or 3.52 surviving children per family, respectively (Le Jan 1994, 287). These figures belong to the ninth century by which time household structures had become transformed by social and religious change (Herlihy 1985, 78). If we want to account for earlier social conditions, we may add the one or the other unmarried brother or sister and surviving parent, plus a few unfree individuals, and we begin to see for Early Saxon England quite a good correspondence between theoretical expectations, written sources, cemetery data and settlement evidence.

Social classes and legal status The key issue in the historical debate on social status distinctions in early Anglo-Saxon society has been the status of freemen, the nature of their freedom, and (intimately linked with this issue) the nature of land tenure and the extent of military service (for a recent summary, cf. Pelteret 1995, 4-24). The disagreements have not been resolved, but the debate appears to have resulted among historians in a shift of opinion away from the assumption of an essentially free peasant society (as advocated, e.g., by Stenton 1971). Among archaeologists, the intellectual shifts in the discipline (cf. below) have resulted in a split: many of the younger archaeologists who have studied Anglo-Saxon social structures have eschewed the use of written sources, while many of their older colleagues who are more comfortable with the use of historical concepts and sources appear to be more familiar with the opinion of Stenton and his contemporaries than with more recent historical opinion. Some of the most useful textual evidence for vertical social differentiation is contained in the wergild clauses of the early laws. From their provisions for the various levels of compensation to be paid for murder or manslaughter, the classic model of a three-tier social hierarchy has been inferred (cf. e.g. Stenton 1971, 277-318; Whitelock 1974, 83-114): - nobility (eorl in Kent, gesith in Wessex; thegn from 10th cent.) - freeman (ceorl) - slave (eow, esne). In addition, the laws of Ethelbert of Kent (26) mention the lt who have widely been interpreted as semi-free persons, perhaps manumitted slaves (Loyn 1962, 10; Whitelock 1979, 392 footnote 7; Pelteret 1995, 294-296), but neither is this certain, nor can it be assumed that such a distinction existed in all Anglo-Saxon kingdoms. A number of other points emerge from the laws, the most important ones being the internal differentiation of these classes, and some mobility within and between them. In Kent, the unfree and semifree classes were divided into three sub-classes each (Ethelbert, 11, 26), and noble widows into four (ibid., 75). In Wessex, free peasants of lower status are mentioned (gafolgelda and gebur; Ine, 6.3; cf. Whitelock 1979, 399 footnote 4), suggesting subdivisions within the class of freemen. But status could change: a slave could be given

freedom (Wihtred, 8); freemen could go into slavery if convicted of theft (ibid., 26; Ine, 7.1) or found to work on a Sunday (Ine, 3.2); and service for the king or in the royal household increased the wergild of unfree (Ethelbert, 7; Ine, 33) and free individuals (Ine, 19). This conforms to expectations of a certain amount of upward mobility which is typical of stratified societies (Lancaster 1958, 366-367), but also to theoretical expectations of increased social mobility during migration and new settlement. On the other hand, Wormald (1978, 35, 91-92) has pointed out that the terms gesithcund and earlcund demonstrate the existence of nobility by birth as early the seventh century. Also from the seventh century onwards, the status of freemen changed as a consequence of the introduction of land charters, shifting the meaning of 'freedom' from the right to be granted the protection of the law to the power to leave an estate (Pelteret 1995, 251). Apart from the problem that the surviving lawcodes relate to Kent and Wessex only, it should be borne in mind that the wergild classes are a differentiation in law, and there is no guarantee that legal status and social status always coincided (James 1989, 38). This distinction becomes even more critical when we begin to look for the reflection of such classes in the archaeological record, in particular burial evidence which is, after all, the result of ritual behaviour and may not faithfully reflect legal and social differentiations. In German archaeology, the evidence from post-Roman cemeteries has routinely been interpreted in social terms since the 1930s, using primarily male burials and their weaponry (for a survey, cf. Steuer 1982; for summaries in English, cf. James 1989; Samson 1987). The reason for the predominant use of men's graves for this kind of social analysis is the smaller range of objects in them compared with female burials, facilitating a subdivision by 'wealth' and resulting in a much clearer differentiation than can be achieved for female grave assemblages, both on the Continent and in England. In Anglo-Saxon archaeology, there was no extensive debate on the social interpretation of burials before the 1970s, but in what would now be termed the 'traditionalist' (or 'cultural-historical') school, the equation of quality of grave-goods with legally defined social classes has always been widespread. The often implicit premise of this approach has been spelt out by Hawkes (in Philp 1973, 186-187): "There was nothing haphazard about the ownership of weapons: it was a matter of legal right and obligation." Accordingly, she postulated the equation of certain weapon types in graves with social classes mentioned in the sources: sword freemen of higher status seax freemen of intermediate status spear ordinary freemen and some half-free

no weapons majority of half-free, and all slaves. She claimed that these equations should apply not just to Kent, but to all of England. Alcock (1981) has suggested a more systematic scheme which was similar to Hawkes' in that he tried to identify the various quality levels of grave-goods with equivalent social classes known from written sources. He concentrated on male burials, but also provided a tentative scheme for females. Below the royal level, indicated by graves with helmets and mail coats, he distinguished three social grades (Table 3). A closely similar scheme has more recently been drawn up by Evison (1987, 146-150) for the social interpretation of the cemetery of Dover, but she added a fourth level of 'findless' graves identified as slaves, thus elevating the Dover equivalents of Alcock's gamma grade to "poor relations in the free class" (ibid., 149). One of the implications is that no slaves were buried in the Dover cemetery, and that these must be buried elsewhere (ibid., 150). The main drift of Alcock's argument, though, related to regional differences. He observed a marked difference between southern England where the alpha:beta ratio of 1:3.6 suggests a broad stratum of free Saxon peasants, and Bernicia where the much rarer alpha burials suggest a small Anglian elite over a large, indigenous population, continuing the Celtic social system of the sub-Roman period (Alcock 1981, 177-179). This observation is important because it is an indicator of profound differences in social structure between the various polities of Anglo-Saxon England (but cf. Cramp 1983). However, Alcock's approach assumes a direct correspondence between ritual symbols (in this case, grave-goods) and social classes, with both categories treated as broadly uniform across England and over the Early Saxon period. Even if later sources appear to suggest a correlation between weapon types and social status, this related to the status of the living; and Brooks (1978, 83) has emphasized that, whatever the texts might suggest, in reality the differences between weapons of the nobility and of freemen were probably blurred. By the time Alcock put forward his views on Anglo-Saxon social structure, quantitative approaches to the social analysis of burials had come into fashion in Anglo-Saxon archaeology, inspired by the neoevolutionist 'New Archaeology' (processual archaeology). The basic processualist premise has been that of a more or less direct correlation between social complexity, and the complexity of the funerary ritual of any given society. In practice, virtually the only archaeological indicator used to infer complexity has been burial wealth, measured by a quantitative analysis of the grave-goods. The observed levels of wealth have been interpreted as social strata, but no attempt has been made

to compare them with the social classes in the written sources: the proponents of the processualist approach have tended to distrust the historical sources, stressing instead the information potential inherent in the archaeological evidence. Quantitative social analysis became a regular feature of most Anglo-Saxon cemetery publications of the 1980s, but there are only two studies which analysed large samples and attempted to draw inferences on Anglo-Saxon social structure overall. Shephard (1979) analysed some 400 burials in order to elucidate the social background of the barrow burials of the late sixth to eighth centuries. As a result of statistical analyses, he identified five levels of wealth (Shephard 1979, 62-63; Table 4). He observed that the proportion of category A burials is less than 10% in flat inhumation cemeteries, but just under 30% of barrow burials, and concluded that individuals in barrows belonged to a 'superordinate' class emerging from the late sixth century onwards. In spite of the fact that children have never been found under tumuli on their own, without adults, he suggested that the high rank displayed in barrows was ascribed, in contrast to the earlier phase when high social status could be achieved (Shephard 1979, 70). Arnold (1980) analysed a sample of 27 cemeteries, of variable data quality, in order to identify the social factors in the development of the southern Anglo-Saxon kingdoms. He found an average of six levels of wealth across his study area, but noted that their number varied from four to eight in individual cemeteries (ibid., 131). At the regional level, he found a difference in burial wealth between coastal and inland regions, which he attributed to differential trade opportunities and access to raw materials and imported goods: "The complexity of social structure, population densities, and levels of armament decrease with distance from commercial centres" on the coast (ibid., 138; Table 5). Hodges (1989) has also applied the economic argument to the regional differences in social structures, but with a different conclusion. He pointed out that the early Kentish law codes appear to imply a much narrower gap between nobleman and freeman than in Wessex, and this may be explained with a more plentiful access to imported goods in Kent (Hodges 1989, 39; for distribution maps of imports in Early Saxon contexts, cf. Huggett 1988). Such quantitative analyses of have since been criticized by postprocessualists for their functionalist and materialist outlook. But while post-processualist studies (e.g. Pader 1982; Richards 1987) have convincingly demonstrated the pervading symbolism in the Early Saxon burial ritual, they have shied away from attempting to present overall conclusions on Anglo-Saxon social structure. In that sense, Arnold's and Shephard's work still provides a useful starting point for

any discussion of vertical social differentiation from a purely archaeological point of view. One possible way forward is a multidimensional approach which combines the quantitative analysis of large samples with the consideration of symbolism in the various aspects of burial ritual. The analysis of male burials from 47 cemeteries of the fifth to seventh/eighth centuries, using archaeological and skeletal evidence (Hrke 1989a; 1990; id. 1992a; id. 1992b), has demonstrated that weapon symbolism, burial treatment and living conditions do not coincide neatly. This underlines that the simple equation of burial wealth levels with social classes (let alone legal classes) may be misleading. Perhaps more importantly, the results also suggest that weapon symbolism changed over time, from a predominantly ethnic to a predominantly social meaning. This makes it necessary to distinguish clearly between early (fifth/sixth century) and late (seventh century) graves. Within the group of males buried with weapons, a ranking by frequency of weapon types and associated burial wealth reveals a top group characterised by the presence of sword, axe or seax (singleedged battle knife), followed at some distance by a group identified by shield and/or spear (Table 6). In spite of all the complications of the evidence, and many local and regional variations, it is sufficiently clear that the male members of the social elite are to be found among the men buried with sword, axe and/or seax: they were accompanied by the highest average number of grave-goods; they had the highest proportion of drinking vessels (most likely symbolising hospitality and the feast); their graves showed above-average labour investment in grave construction; their group includes an unusually high proportion of men with strong physique (suggesting regular exercise, perhaps weapon training); and their stature remained above average throughout the Early Saxon Period (unlike that of men with other weapons; cf. below, next section). The changes over time are intriguing: while the proportion of male elite burials remained static throughout the Early Saxon period, the proportion of intermediate burials dropped dramatically in the seventh century, and the proportion of graves without weapons rose accordingly (Table 6). Religious and ritual change may well have played a role in this change, but given other archaeological indications of social change in the seventh century, such as the growing wealth differential between elite burials and others (Arnold 1982), social change is equally likely to have been a factor. This change appears to have involved a decline in status of many males of intermediate social standing, swelling the ranks of the lowest social orders. The historian Brooks has argued that in Anglo-Saxon society, as in other Germanic

societies, the possession and bearing of arms was a symbol of legal freedom (Brooks 1978, 83). If this argument has any relevance at all for the Anglo-Saxon weapon burial rite (which is by no means clear), we may be looking at a decline in the seventh century in numbers of ordinary freemen, with many of them descending to the status of semifree or unfree. The available settlement evidence provides some further clues as to the nature of vertical stratification in post-Roman England, and confirms the impression of social change in the seventh century. The mostly undifferentiated fifth/sixth century hamlets, with their haphazard lay-out seem to be made up of groups of equal social standing. Where there are fences or ditches within these hamlets, as at Chalton, Catholme and West Stow, they separate farming units of comparable size (Champion 1977, 367; Losco-Bradley 1977, 359; West 1985, 54, fig. 302); fenced enclosures on high-status sites are a different matter (cf. below). Accordingly, the timber-built 'halls' at Chalton have been seen as the "normal dwelling of the freeman, holder of one hide" (Addyman and Leigh 1972, 24; id. 1973, 19), and the entire settlement as a "village of ceorls" (id. 1972, 24). Such interpretations have gained widespread, explict or tacit acceptance, but they highlight the contrast to contemporary cemeteries which display distinctions of wealth and status. Attempts to identify farms of slightly higher status, on the basis of location within the settlement (West 1985, 169) or number of ancillary buildings (Losco-Bradley 1977, 363-364), have not produced very convincing results. Occasionally, small finds such as an enamelled mount from a hanging bowl and imported Frankish pottery at Chalton (Addyman and Leigh 1973, 19-20; Champion 1977, 369) or an iron tripod lamp at West Stow (West 1985, 169), provide glimpses of possible differentials in material belongings and access to goods. But they are not enough to dispel the growing impression that the social differentiation seen in cemeteries may reflect status differences within rather than between families and/or households. This picture changed from the late sixth century onwards, but most markedly during the seventh century. Parallel to signs of increasing social stratification in the burial evidence, we see in this phase the emergence of a settlement hierarchy and the appearance of larger buildings on high-status sites. The various elements of the settlement hierarchy develop from the early seventh century. By the end of the seventh century, at least three levels can be distinguished (Arnold 1984, 278-279): rural settlements concerned with food production, trading sites engaged in craft production and commerce (Scull, this volume), and royal residences and centres of

administration. Also from the seventh century onwards, ecclesiastical sites comprising churches, monasteries and episcopal palaces were forming their own settlement hierarchy. The recent analysis of Anglo-Saxon buildings has shown that from the end of the sixth century, some structures with a floor area above 80 sq.metres appear, sufficiently larger than average 'halls' (around 50 sq. meteres; cf. above) to warrant an interpretation in terms of elite dwellings (Marshall and Marshall 1993, 398). Annexes attached to timber buildings made their appearances at the same time (ibid.), hinting at a connection with elite accommodation or the functions of elite farms. The rise of an Anglo-Saxon elite is exemplified by two excavated high-status sites. The social status of Cowdery's Down (Hampshire) could only be inferred from the sizes of buildings on the site, from their superior building techniques, and the investment of timber and labour into their construction (Millett and James 1983, 247). Over its occupation period in the sixth and seventh centuries, the settlement grew from three to six, and finally to ten, major buildings, with the size of the largest residential building in each phase increasing from 108 to 194 sq.metres (ibid.). Two fenced enclosures attached to, and accessed through, buildings suggest the existence here of several social units. The size of the largest timber building (C12) at Cowdery's Down has led to its interpretation as the seasonal residence of a peripatetic elite (ibid., 249), or even the palace of a tribal leader (Marshall and Marshall 1993, 400). At Yeavering (Northumberland), a seventh-century royal estate of the kings of Bernicia has been identified from place-name evidence and the testimony of Bede (HE II.14; Hope-Taylor 1977). The 'township' was made up in each phase of a major timber-built hall, a few smaller buildings including a temple (later church), a timber grandstand, and a large enclosure next to the residential complex (Hope-Taylor 1977, figs. 75-78). The successive timber halls (A2 and A4; ibid., 127 fig. 60, 130 fig. 61) of the early and mid-seventh century are the largest domestic structures uncovered so far on any Early Saxon site. With their internal length of c. 25 m and their floor areas of 260 and 290 sq.metres, respectively, each could have accomodated the 60 men who defended the Finnesburh hall (Finnesburh Fragment), or the 30 warriors killed by Grendel in King Hrothgar's hall (Beowulf, l. 123). Fenced enclosures were attached to the main hall in each phase, and like the partitions inside the halls (two in A2, one in A4), they hint at social and functional differentiations among the residents. The grandstand capacities of approx. 150 (later 320) people indicate the size of the group or groups assembled here, probably for council meetings, highlighting the

function of this estate as an administrative centre (Hope-Taylor 1977, 279). It is interesting to note that social distinctions apparent in the seventh-century settlement evidence are mostly between sites, rarely within sites. This mirrors the appearance, in the seventh century, of isolated barrow burials in which the elite buried their members, away from 'ordinary' community cemeteries. Enclosures on high-status settlements which suggest some kind of internal differentiation are so far ill understood in terms of their exact functions and implications. Defended sites are virtually non-existent in Early Saxon England, with the exception of the strange, empty enclosure at Yeavering which was located next to the royal complex, not around it. It seems that social differentiation, same as gender dichotomy, was expressed much more clearly in burial ritual than in domestic architecture or settlement layout. In this respect, Anglo-Saxon England differed from the Celtic areas of Britain where we find an earlier, and more pronounced, settlement hierarchy (Alcock 1987), but no social differentiation in the burial evidence. Ethnicity and social structure In many post-Roman societies, social structure and ethnic differentation were interrelated, usually as a result of immigrant groups achieving control over native populations. Nowhere is this interrelation clearer than in the early Anglo-Saxon laws where the Old English term wealh could mean either 'Welshman' or 'slave' (Whitelock 1979, 402 footnote 5; Pelteret 1995, 43). This illustrates at the same time the ambiguity and fluidity of the concept of ethnic identity in Migration Period Europe (cf. Pohl, this volume). The 'tribal' distinctions between Angles, Saxons, Jutes and others do not seem to have had any social significance, and while they provide interesting case studies in their own right, the discussion will be limited here to native Britons and their relationship with the immigrant Germani and their descendants. The narrative sources report the wholesale slaughter of the natives by the invaders, with the survivors going into "perpetual servitude" (Bede, HE I.15, following Gildas, 24.3-25.1). The best evidence for the status of Britons in an Anglo-Saxon kingdom is provided by the late seventh-century laws of king Ine of Wessex (extant as an appendix to Alfred's laws). They list six different wergild levels for the Welsh, four of them below the wergild level of a freeman (Ine, 23.3, 32); but the king's Welsh horseman had the wergild of a freeman (Ine, 33), and a Welshman with five hides had three times the wergild of a freeman, or half that of a nobleman (Ine, 24.2). One clause

implies that a Welsh slave could have free kindred (Ine, 74.1). But it appears that in some respects, all Welsh were considered to have a lower status than the English: if accused of cattle theft by an Englishman, the accused had to deny it with double the oath than if the accusation came from a Welshman (Ine, 46.1). Kentish laws do not mention the Welsh as such, but it has been suggested that the semifree lt in the laws of Ethelbert (26) may have been Britons (cf. above). The laws imply that, while some Britons could be wealthy freemen, the majority were of low legal status. On the other hand, not all slaves were of British extraction: Englishmen as well as Welshmen could become penally enslaved (Ine, 24, 54.2). At the same time, the laws also imply a close co-existence of both groups within the same kingdom and under the same jurisdiction. This co-existence is confirmed by Celtic river names and place names in England, the analysis of which has suggested that bilingual Britons had transmitted Celtic names to Germanic-speaking immigrants (Jackson 1956, 220246). However, the small number of Celtic loan-words in the English language indicates a low level of intermarriage between the two groups (Loyn 1962, 12-14), except possibly for Northumbria where personal names suggest a larger number of mixed marriages (Whitelock 1962, 18). Thus, the assimilation of Britons into an Englishspeaking language community must, at least initially, have been the result of cultural rather than biological processes. This large-scale assimilation and its one-sided liguistic outcome would have been facilitated by the low social and political status of the Britons and their language (Charles-Edwards 1995, 729-730). Among archaeologists, the recent re-assessment of the AngloSaxon 'invasion' (Arnold 1984a; Garwood 1989; Hodges 1989; Higham 1992) has led to renewed interest in the question of British survival which is crucial to the revisionist argument of minimal immigration. However, the search for natives in the archaeological record has not been very successful so far. The identification of native, post-Roman settlements has been rare, tentative, and limited to the northern and western fringes of the Anglo-Saxon settlement areas (Higham 1992, 106 fig. 4.11, 107). But even if the search for British farmsteads or villages were successful, their identification may not tell us much about their position within the social structures of post-Roman England. Any attempt to find in the Anglo-Saxon settlement sites the low-status Britons mentioned in the sources comes up against two problems: the ambiguity of internal subdivisions within houses and settlements, which may be economic, gender or status related (or a combination of these factors); and the absence of a distinctive, native material culture.

The latter problem also affects the interpretation of grave-goods in terms of ethnic affiliation, but other aspects of the burial ritual as well as skeletal evidence may provide further clues. The native subRoman burial rite is known from sites such as Cannington (Somerset; Rahtz 1977): inhumation without grave-goods, the body extended on its back, head approximately to the west. This is a useful contrast to Anglo-Saxon cremations, but it raises the question of how to distinguish Anglo-Saxon and native inhumations. The analysis of the archaeological and, where available, skeletal data of a sample of 1600 male and juvenile burials from Early Saxon cemeteries has suggested that in the fifth/sixth centuries, burial with weapons was used as an ethnic marker by the immigrants and their descendants (Hrke 1990; id. 1992a). The main arguments leading to this conclusion are (Hrke 1992a, 195-200): (1) the origin of the weapon burial rite on the Continent, and the absence of weapon deposition in graves in Roman Britain and Celtic post-Roman Britain and Ireland; the stature differential of 1" to 2" (2 to 5 cm) between men with and without weapons (compared separately for each cemetery).

(2)

While the latter may, in principle, be a purely social phenomenon created by preferential diet and living conditions of the elite, this is unlikely to be the case here. For a start, the two groups do not show any differential at all in indicators of stress and malnourishment. Secondly, the proportion of men buried with weapons (Table 6) does not encourage an interpretation in terms of an 'elite'. Thirdly, the stature differential disappears in the 7th century, but social differences do not - on the contrary (cf. above). Finally, the differential in question is virtually the same as that between British Iron Age and RomanoBritish male stature, on the one hand, and post-Roman male stature in Northern Germany and England, on the other hand. The key implication of the argument is that in the fifth/sixth centuries, up to half of the men (i.e. those without weapons) buried in 'Anglo-Saxon' inhumation cemeteries may have been native Britons. The weapon symbolism suggests that the Germanic population was expressing its dominance in ritual; and its higher average number of grave-goods suggests that it had more disposable wealth. It is as yet impossible to say if the proportion of natives to immigrants was similar among the females. In her thesis, Brush (1993) was unable to unambiguously identify British females on the basis of dress items. It may be that a combination of archaeological and skeletal data, as used

for males, will be more successful, but the wider spectrum of female grave-goods and burial wealth may make it difficult. In addition to the considerable proportion of natives in AngloSaxon cemeteries, further Britons must have lived in their own enclaves in southern and eastern England. Evidence for them is provided by river, hill and place-names (Jackson 1956, 234-241); by the Upper Thames cemetery of Queenford Farm (post-Roman, but without Anglo-Saxon artefacts; Chambers 1987); by an East Anglian cluster of enamelled dress ornaments (Scull 1985); by Anglo-Saxon pottery scatters on Romano-British settlement sites on Salisbury Plain (pers.comm. M.G. Fulford and R. Entwistle); and indirectly by a general drop in male stature in the seventh/eighth centuries which was most pronounced in Wessex. The latter is best explained as the consequence of the acculturation and assimilation of British populations which had previously been invisible in the archaeological record. Two other developments suggest that the increasing assimilation of the native population and their integration into Anglo-Saxon society during the seventh/eighth centuries may have been connected with changes in the social structure. During that period, burial with weapons became less frequent, was more and more concentrated in rich burials, and was no longer used as an ethnic marker (Hrke 1992b). The other development was the location of barrows and other, contemporaneous burial sites next to, in, or around, prehistoric barrows or Roman sites. A case in point is the mid-seventh century barrow on Lowbury Hill (Oxfordshire; Hrke 1994b). It is located next to a Roman temple site, and only a few miles from Wallingford the name of which indicates the existence here of a British enclave. The rich male burial in the barrow contained two objects (a hanging bowl and a unique spearhead) with enamel decoration which is widely accepted as a Celtic marker. The seventh/eighth century evidence suggests that the social elite transormed the weapon burial rite into a mere status symbol, and attempted to establish a link to the indigenous, pre-Saxon past, conceivably to legitimate their rule over an increasingly mixed population. There can be little doubt that ethnic factors played an important role in Early Anglo-Saxon social structure. Archaeological, historical and lingustice evidence are in agreement on this point. Parallels for this overlap of social and ethnic differentiations may be found in tribal and early state societies around the world where such situations have often been the consequence of immigration and conquest. Conclusions

Even though this paper draws on a variety of data, it turned out to be easier to construct static descriptions of certain aspects of society in certain phases, than to infer something about the actual workings of society. Also, the various types of evidence surveyed above give uneven and differential weight to the key factors of Anglo-Saxon social organisation before the ninth century. The paramount importance of kinship is quite clear in the written sources: it provided the fabric of society, and determined to a large degree an Anglo-Saxon's status and, literally, his or her value. By contrast, the archaeological evidence, by its very nature, puts the emphasis on the material expression of status, relegating kinship to being one of the possible inferences. Interestingly, though, archaeology highlights a number of aspects which are less prominent in the sources: household, gender and age. In fact, only a sophisticated analysis can achieve the distinction between household and family in the archaeological record, but the units we perceive in the organisation of cemeteries are better explained in terms of residential units with internal social differentation (i.e. households) than kinship units (i.e. families). The marked display of gender and age in Early Saxon burial ritual poses problems. Not only are these factors less pronounced in the historical sources and virtually invisible in the settlement evidence: the emphasis on gender is in stark contrast to theoretical expectations concerning blurred gender roles in settler societies; and the evidence for age groups is inconsistent, with the written sources stressing different age thresholds from those apparent in burial ritual. But both, written sources (in particular the laws) and the inferences possible from cemetery evidence agree on the key role played by ethnic affiliation in the Early Saxon period. Britons seem to have made up a large proportion of the population buried in 'AngloSaxon' cemeteries, and most of them appear to have been of unfree, or otherwise low, status. Differences between the kingdoms in the proportions of Britons are likely to have played a role in creating regional differences in social structure. Increasing intermarriage with Britons in Anglo-Saxon settlements, and increasing assimilation of Britons in enclaves may have been an important part of social change in England. The shift of emphasis in burial ritual from ethnic affiliation to social differentiation in the seventh century appears to mark a crucial transition: from ethnically divided conquest societies to the formation of early states.

Acknowledgements I am grateful for the incentive the CIROSS invitation gave me to clarify in my own mind a number of issues relating to early Anglo-Saxon society, and to bring into better focus those questions for which clarification could not be achieved. I benefitted immensely from the seminar discussions at San Marino even though this paper may not show it: points discussed or corrected during the meeting were not altered for this version as it would have rendered the publication of the discussion pointless. Helena Hamerow (Durham) kindly read and commented on the draft paper.

Footnotes 1 Anglo-Saxon children were the object of a recent doctoral thesis (S. Crawford, Age differentiation and related social status: a study of Early Anglo-Saxon childhood. Unpubl. D.Phil. thesis, Oxford 1991). Regrettably, permission was not given to quote from the thesis in this paper. Is it coincidence that in our days the age set 15 to 25 is the most prominent group in traffic accident and crime statistics? For a recent confirmation of this well-established and recognised trend, cf. the US murder statistics for 1994 (Lacayo 1996, 34). A detailed analysis of Anglo-Saxon female inhumations has been carried out by N. Stoodley (Reading) in his doctoral thesis; it was completed in April 1997. Unpubl. paper by R. Larick, held at the Theoretical Archaeology Group (TAG) Conference, Lampeter 1990. In contrast to a current theoretical trend which views both, sex and gender as culturally determined (so in several papers given at the Theoretical Archaeology Group conference, Reading, December 1995), 'sex' is taken here to be a biological category, 'gender' a cultural category. This distinction makes sense when dealing with skeletal as well as archaeological evidence: the latter would normally reflect gender as expressed in material culture and ritual, whereas skeletal evidence gives information on the biological sex of individuals (in terms of a probability statement rather than in absolute terms). It also follows that written sources would reflect either sex or gender, depending on the type of source and the context of the reference. This re-study deals with the 'cross-gender' cases from the cemeteries of Dover (Kent; Evison 1987) and Norton (Cleveland; Sherlock and Welch 1992), and is being carried out by N. Stoodley (archaeology; Reading), T. Molleson (physical anthropology; London) and J. Bailey (DNA; Oxford). The results to date (early 1997) suggest that in some cases, the original skeletal sexing was faulty; and where it was not, it is mostly contradicted by the DNA analysis (pers.comm. J. Bailey and N. Stoodley). As a result, there are no certain cases of females with weapons, and only a tiny number of probable males with jewellery from Anglo-Saxon cemeteries.

4 5

References Adams, B.D. 1982. The Anglo-Saxon cemetery at Wakerley, Northamptonshire. M.Phil. thesis, Birkbeck College, London. Addyman, P.V. 1972. The Anglo-Saxon house: a new review. Anglo-Saxon England 1. 273-307. Addyman, P.V. 1976. Archaeology and Anglo-Saxon society. In: G. de G. Sieveking, I.H. Longworth and K.E. Wilson (eds.). Problems in economic and social archaeology. London: Duckworth. 309-322. Addyman, P.V. and D. Leigh 1972. Anglo-Saxon houses at Chalton, Hampshire. Medieval Archaeology 16. 13-31. Addyman, P.V. and D. Leigh 1973. The Anglo-Saxon village at Chalton, Hampshire: second interim report. Medieval Archaeology 17. 1-25. Alcock, L. 1981. Quantity or quality: the Anglian graves of Bernicia. In: V.I. Evison. Angles, Saxons and Jutes: essays presented to J.N.L. Myres. Oxford. 168-183. (Reprinted in Alcock 1987, 255-266). Alcock, L. 1987. Economy, society and warfare among the Britons and Saxons. Cardiff: University of Wales Press. Aldsworth, F.R. 1978. The Droxford Anglo-Saxon cemetery, Soberton, Hampshire. Proc. Hants. Field Club Arch. Soc. 35. 93182. Aris, P. 1962. Centuries of childhood. A social history of family life: London. Jonathan Cape. Arnold, C.J. 1980. Wealth and social structure: a matter of life and death. In: P. Rahtz, T. Dickinson und L. Watts (eds.). AngloSaxon cemeteries 1979. The Fourth Anglo-Saxon Symposium at Oxford. Oxford. (Brit.Arch.Rep. 82). 81-142. Arnold, C.J. 1981. Early Anglo-Saxon pottery: production and distribution. In: H. Howard und E.L. Morris (eds.). Production and distribution: a ceramic viewpoint. Oxford. (Brit.Arch.Rep. S 120). 243-255.

Arnold, C.J. 1982. Stress as a stimulus to socio-economic change: England in the seventh century. In: C. Renfrew and S. Shennan (eds.). Ranking, resource and exchange. Cambridge University Press. (New Directions in Archaeology). 124-131. Arnold, C.J. 1984a. Roman Britain to Saxon England. London. (Croom Helm Studies in Archaeology). Arnold, C.J. 1984b. Social evolution in post-Roman western Europe. In: J. Bintliff (ed.). European social evolution. West Chiltington: Chanctonbury Press. 277-294. Arnold, C.J. 1988. An archaeology of the early Anglo-Saxon kingdoms. London und New York. Arnold, K. 1986. Die Einstellung zum Kind im Mittelalter. In: B. Herrmann (ed.). Mensch und Umwelt im Mittelalter. Darmstadt. 53-64. Bell, M. 1977. Excavations at Bishopstone. Lewes. (Sussex Archaeological Collections 115). Brooks, N.P. 1978. Arms, status and warfare in Late-Saxon England. In: D. Hill (ed.). Ethelred the Unready. Oxford. (Brit.Arch.Rep. 59). 81-103. Brothwell, D. 1972. Palaeodemography and earlier British populations. World Archaeology 4. 75-87. Brush, K.A. 1988. Gender and mortuary analysis in Pagan AngloSaxon archaeology. Archaeological Review from Cambridge 7:1. 76-89. Brush, K. 1993. Adorning the Dead. Unpubl. Ph.D. thesis, Cambridge University. Chambers, R.A. 1987. The Late- and Sub-Roman cemetery at Queenford Farm, Dorchester-on-Thames, Oxon. Oxoniensia 52. 3569. Champion, T. 1977. Chalton. Current Archaeology 59. 364-369. Charles-Edwards, T. 1972. Kinship, status and the origins of the hide. Past & Present 56. 3-33. Charles-Edwards, T. 1995. Language and society among the insular

Celts AD 400-1000. In: M.J. Green (ed.). The Celtic world. London and New York: Routledge. 703-736. Christlein, R. 1973. Besitzabstufungen zur Merowingerzeit im Spiegel reicher Grabfunde aus West- und Sddeutschland. Jahrbuch des Rmisch-Germanischen Zentralmuseums 20. 147180. Cook, A.M. und M.W. Dacre 1985. Excavations at Portway, Andover, 1974-5. Oxford. (Oxford University Committee for Archaeology, Monograph 4). Cramp, R. 1983. Anglo-Saxon settlement. In: J.C. Chapman and H.C. Mytum (ed.). Settlement in North Britain 1000 BC - AD 1000. Oxford. (B.A.R. 118). 263-297. Crawford, S. 1993. Children, death and the afterlife in Anglo-Saxon England. Anglo-Saxon Studies in Archaeology and History 6. 83-91. Dickinson, T.M. and Hrke, H. 1992. Early Anglo-Saxon shields. London: Society of Antiquaries. (Archaeologia 110). Dietrich, S.C. 1980. An introduction to women in Anglo-Saxon society. In: B. Kanner (ed.). The women of England from AngloSaxon times to the present. London: Mansell. 32-56. Dixon, P.H. 1993. The Anglo-Saxon settlement at Mucking: an interpretation. Anglo-Saxon Studies in Archaeology and History 6. 125-147. Evison, V.I. 1987. Dover: The Buckland Anglo-Saxon cemetery. London. (Historic Buildings and Monuments Commission for England, Archaeological Report No. 3). Fell, C. 1984a. Women in Anglo-Saxon England. London: British Museum. Fell, C. 1984b. A friwif locbore revisited. Anglo-Saxon England 13. 157-165. Garwood, P. 1989. Social transformation and relations of power in Britain in the late fourth to the sixth centuries A.D. Scottish Archaeological Review 6. 90-106. Hrke, H. 1989a. Early Saxon weapon burials: frequencies,

distributions and weapon combinations. In: S.C. Hawkes (ed.). Anglo-Saxon weapons and warfare. Oxford. (Oxford University Committee for Archaeology, Monograph 21). 49-61. Hrke, H. 1989b. Knives in Early Saxon burials: blade length and age at death. Medieval Archaeology 33. 144-148. Hrke, H. 1990. 'Warrior graves'? The background of the Anglo-Saxon weapon burial rite. Past & Present 126. 22-43. Hrke, H. 1992a. Angelschsische Waffengrber des 5. bis 7. Jahrhunderts Cologne and Bonn: Rheinland-Verlag and Habelt. (Zeitschrift fr Archologie des Mittelalters, Beiheft 6). Hrke, H. 1992b. Changing symbols in a changing society: the Anglo-Saxon weapon burial rite in the seventh century. In: M. Carver (ed.). The age of Sutton Hoo. Woodbridge: Boydell Press. 149-165. Hrke, H. 1994a. Data types in burial analysis. In: B. Stjernquist (ed.). Prehistoric graves as a source of information. Stockholm. (Kungl. Vitterhets och Antikvitets Akademiens Konferenser 29). 31-39. Hrke, H. 1994b. Lowbury Hill, Oxon: A context for the Saxon barrow. Archaeological Journal 151. 202-206. Hrke, H. 1995. Weapon burials and knives. In: A. Boyle, A. Dodd, D. Miles and A. Mudd. Two Oxfordshire Anglo-Saxon cemeteries: Berinsfield and Didcot. Oxford: Oxfordshire Archaeological Unit 1995. (Thames Valley Landscapes Monograph 8). 67-75. Hrke, H. forthcoming. The nature of burial data. In: K.H. Nielsen and K. Jensen (eds.). From burial to society. Aarhus Universitet. Hamerow, H. 1993. Excavations at Mucking. Vol. 2: The Anglo-Saxon settlement. London: English Heritage. (Archaeological Report 21). Hastorp, C. 1991. Gender, space and prehistory. In: J.M. Gero and M.W. Conkey (eds.). Engendering archaeology. Oxford: Blackwell. 132-159. Hawkes, S.C. and C. Wells 1983. The inhumed skeletal material from an early Anglo-Saxon cemetery in Worthy Park, Kingsworthy, Hampshire, South England. Palaeobios 1. 1-36.

Herlihy, D. 1985. Medieval households. Cambridge, Mass., and London: Harvard University Press. (Studies in Cultural History). Herlihy, D. 1995. Women, family and society in medieval Europe: historical essays, 1978-1991. Providence (USA) and Oxford: Berghahn. Herrmann, B., G. Grupe, S. Hummel, H. Piepenbrink and H. Schutkowksi 1990. Prhistorische Anthropologie: Leitfaden der Feld- und Labormethoden. Berlin. Higham, N. 1992. Rome, Britain and the Anglo-Saxons. London: Seaby. (The Archaeology of Change series). Hills, C. und K. Penn 1981. The Anglo-Saxon cemetery at Spong Hill, North Elmham. Part 2: Catalogue of cremations, Nos. 22, 41 and 1691-2285. Gressenhall. (East Anglian Archaeology 11). Hirst, S.M. 1985. An Anglo-Saxon inhumation cemetery at Sewerby, East Yorkshire. York. (York University Archaeological Publications 4). Hodges, R. 1989. The Anglo-Saxon achievement: Archaeology and the beginnings of English society. London: Duckworth. Hope-Taylor, B. 1977. Yeavering: an Anglo-British centre of early Northumbria. London. (Department of Environment Archaeological Reports 7). Huggett, J. 1988. Imported grave-goods and the early Anglo-Saxon economy. Medieval Archaeology 32. 63-96. Jackson, K.H. 1956. Language and history in early Britain. Repr. Edinburgh. (Edinburgh Univ. Publ., Language & Literature 4). James, E. 1989. Burial and status in the early medieval west. Transactions of the Royal Historical Society (5th series) 39. 23-40. James, S., A. Marshall and M. Millett 1984. An early medieval building tradition. Archaeological Journal 141. 182-215. Jrgensen, L. 1987. Family burial practices and inheritance systems: The development of an Iron Age society from 500 BC to AD 1000 on Bornholm, Denmark. Acta Archaeologica 58. 17-53.

Jrgensen, L. 1990. Bkkegrd and Glasergrd: Two cemeteries from the Late Iron Age on Bornholm. Copenhagen: Akademisk Forlag. (Arkologiske Studier 8). Jrgensen, L. 1991. Castel Trosino and Nocera Umbra: A chronological and social analysis of family burial practices in Lombard Italy (6th - 8th cent. AD). Acta Archaeologica 62. 1-58. John, E. 1964. Land tenure in early Enland: a discussion of some problems. Leicester University Press. (Studies in Early English History 1). Keufler, M.S. 1991. "A wryed existence": attitudes toward children in Anglo-Saxon England. Journal of Social History 24. 823834. Kinsley, A.G. 1993. Broughton Lodge. Excavations on the RomanoBritish settlement and Anglo-Saxon cemetery at Broughton Lodge, Willoughby-on-the-Wolds, Nottinghamshire, 1964-8. Nottingham: University of Nottingham. (Archaeological Monographs 4). Klinck, A.L. 1982. Anglo-Saxon women and the law. Journal of Medieval History 8. 107-121. Lacayo, R. 1996. Law and order. TIME 147 no. 5 (January 29, 1996). 30-35. Lancaster, L. 1958. Kinship in Anglo-Saxon society. British Journal of Sociology 9. 230-250, 359-377. Le Jan, R. 1995. Entre matres et dependants: reflexions sur la famille paysanne en Lotharingie, aux IXe et Xe sicles. In: Campagnes mdivales: l'homme et son espace. Etudes offertes Robert Fossier. Paris: Sorbonne. 277-296. Leach, E.R. 1964. Political systems of Highland Burma. Repr. London. (London School of Economics, Monograph 44). Losco-Bradley, S. 1977. Catholme. Current Archaeology 59. 358-364. Loyn, H.R. 1962. Anglo-Saxon England and the Norman Conquest. London. (Social and Economic History of England). Loyn, H.R. 1974. Kinship in Anglo-Saxon England. Anglo-Saxon England 3. 197-209.

Marshall, A. and G. Marshall 1993. Differentiation, change and continuity in Anglo-Saxon buildings. Archaeological Journal 150. 366-402. Meyer, M.A. 1980. Land charters and the legal position of Anglo-Saxon women. In: B. Kanner (ed.). The women of England from Anglo-Saxon times to the present. London: Mansell. 57-82. Middleton, P. 1995. Weaving Destiny: changes in the significance of textile production to the economy of Anglo-Saxon England and its impact on female labour and status. Unpubl. M.A. thesis, University of Reading. Millett, M. and S. James 1983. Excavations at Cowdery's Down, Basingstoke, Hampshire, 1978-81. Archaeological Journal 140. 151-279. Ottinger, I. 1974. Waffenbeigabe in Kindergrbern. In: G. Kossack and G. Ulbert (eds.). Studien zur vor- und frhgeschichtlichen Archologie (Werner-Festschrift). Vol. 2. Munich. (Mnchener Beitrge zur Vor- und Frhgeschichte, Ergnzungsband 1/II). 387410. Pader, E.J. 1982. Symbolism, social relations and the interpretation of mortuary remains. Oxford. (Brit.Arch.Rep. S 130). Pelteret, D. 1995. Slavery in early mediaeval England: from the reign of Alfred until the twelfth century. Woodbridge: Boydell. (Studies in Anglo-Saxon History 7). Philp, B. 1973. Excavations in West Kent 1960-1970. Dover. (Research Reports in the Kent Series 2). Rahtz, P. 1976. Buildings and rural settlement. In: D.M. Wilson (ed.). The archaeology of Anglo-Saxon England. Cambridge: Cambridge University Press. 49-98. Rahtz, P.A. 1977. Late Roman cemeteries and beyond. In: R. Reece (ed.). Burial in the Roman world. London. (Council for British Archaeology, Research Report 22). 53-64. Richards, J.D. 1987. The significance of form and decoration of Anglo-Saxon cremation urns. Oxford. (Brit.Arch.Rep. 166). Samson, R. 1987. Social structures from Reihengrber: mirror

or mirage? Scottish Archaeological Review 4. 116-126. Schwab, H. 1982. Bemerkenswert ausgestattete Kindergrber der Merowingerzeit. Archologisches Korrespondenzblatt 12. 251-262. Scull, C. 1985. Further evidence from East Anglia for enamelling on early Anglo-Saxon metalwork. Anglo-Saxon Studies in Archaeology and History 4. 117-124. Scull, C. 1993. Archaeology, early Anglo-Saxon society and the origins of Anglo-Saxon kingdoms. Anglo-Saxon Studies in Archaeology and History 6. 65-82. Sheehan, M. 1963. The will in medieval England. Toronto: Pontifical Institute of Mediaeval Studies. (Studies and Texts 6). Shephard, J. 1979. The social identity of the individual in isolated barrows and barrow cemeteries in Anglo-Saxon England. In: B.C. Burnham und J. Kingsbury (eds.). Space, hierarchy and society. Oxford. (Brit.Arch.Rep. S 59). 47-79. Sherlock, S.J. and M. Welch 1992. An Anglo-Saxon cemetery at Norton, Cleveland. London: Council for British Archaeology. (CBA Research Report 82). Stenton, D.M. 1957. The English woman in history. London and New York: Allen & Unwin and Macmillan. Stenton, F.M. 1943. The historical bearing of place-name studies: the place of women in Anglo-Saxon society. Transactions of the Royal Historical Society (4th series) 25. 1-13. Stenton, F.M. 1971. Anglo-Saxon England. 3rd ed. Oxford. (Oxford History of England 2). Steuer, H. 1982. Frhgeschichtliche Sozialstrukturen in Mitteleuropa. Gttingen. (Abhandlungen der Akademie der Wissenschaften in Gttingen, Phil.-Hist. Klasse, 3. Folge, 128). Stoodley, N. 1994. Gender relations within frontier societies. Unpubl. typescript, Reading University. van Gennep, A. 1960. The rites of passage. London: Routledge & Kegan Paul. Wainwright, F.T. 1959. thelfld, Lady of the Mercians. In:

P. Clemoes (ed.). The Anglo-Saxons: studies presented to Bruce Dickins. 53-70. Watts, D.J. 1989. Infant burials and Romano-British Christianity. Archaeological Journal 146. 372-383. Welch, M. 1980. The Saxon cemeteries of Sussex. In: P. Rahtz, T. Dickinson und L. Watts (eds.). Anglo-Saxon cemeteries 1979. The Fourth Anglo-Saxon Symposium at Oxford. Oxford. (Brit.Arch.Rep. 82). 255-283. Welch, M.G. 1983. Early Anglo-Saxon Sussex. Oxford. (Brit.Arch.Rep. 112). Wells, C. and H. Cayton 1980. The human bones. In: P. Wade-Martins. Excavations in North Elmham Park, 1967-1972. Gressenhall: Norfolk Archaeological Unit. (East Anglian Archaeology 9). 247-314. West, S. 1985. West Stow: The Anglo-Saxon village. 2 vols. Ipswich: Suffolk County Planning Department. (East Anglian Archaeology 24). Whitelock, D. 1930. Anglo-Saxon wills. Cambridge: Cambridge University Press. (Cambridge Studies in English Legal History). Whitelock, D. 1962. The beginnings of English society. Repr. Harmondsworth. (Pelican History of England 2). Whitelock, D. 1979. English historical documents. Vol. 1. Repr. London. Wormald, P. 1978. Bede, Beowulf and the conversion of the Anglo-Saxon aristocracy. In: R.T. Farrell (ed.). Bede and AngloSaxon England. Oxford. (Brit.Arch.Rep. 46). 32-95. Zimmermann, W.H. 1988. Regelhafte Innengliederung prhistorischer Langhuser in den Nordseeanrainerstaaten: Ein Zeugnis enger, langandauernder kultureller Kontakte. Germania 66. 465-488.

Primary Sources Athelstan, Laws of: F. Liebermann 1903. Die Gesetze der Angelsachsen. Halle. [Engl. translation in Whitelock 1979] Alfred, Laws of: (cf. Athelstan) Beda Venerabilis, Historia Ecclesiastica Gentis Anglorum (HE): B. Colgrave and R.A.B. Mynors 1969. Bede's Ecclesiastical History of the English People. Oxford. Beda Venerabilis, Life of St. Cuthbert: B. Colgrave (ed.) 1940. Two Lives of Saint Cuthbert. Cambridge. Beowulf: M. Swanton 1978. Beowulf. Manchester. (Manchester Medieval Classics Series). Cnut, Laws of: (cf. Athelstan) Eddius Stephanus, Life of St. Wilfred: B. Colgrave (ed.) 1927. The Life of Bishop Wilfred by Eddius Stephanus. Cambridge. Ethelbert, Laws of: (cf. Athelstan) Felix, Life of St. Guthlac: B. Colgrave (ed.) 1956. Felix's Life of St. Guthlac. Cambridge. Finnesburh Fragment: E.V.K. Dobbie (ed.) 1958. The Anglo-Saxon poetic records. Vol. VI: The Anglo-Saxon minor poems. New York. Gildas, De Excidio and Conquestu Britanniae: M. Winterbottom (ed.) 1978. Gildas: The Ruin of Britain and other sources. London and Chichester. (Arthurian Period Sources 7). Hlothere und Eadric, Laws of: (cf. Athelstan) Ine, Laws of: (cf. Athelstan) Maldon, Battle of: Dobbie 1958 (cf. Finnesburh) [Engl. translation in Whitelock 1979] Wihtred, Laws of: (cf. Athelstan)

Tables

biographical data

laws

burial evidence 2/3

7/8 10 (later 12) 14/15 18/20 mid-20s+ 30+ (males only) 12 (10/12 for females)

Table 1: Threshold ages

REGION cemetery KENT Lyminge II * Polhill UPPER THAMES Abingdon I Berinsfield Long Wittenham I WESSEX Petersfinger Pewsey (Blackpatch) Worthy Park * Winnall II *

average wealth scores males 26.8 22.5 16.4 30.4 29.1 32.7 34.8 39.7 8.6 females 25.3 33.6 69.8 54.1 41.3 81.8 65.3 43.8 25.9

average number of artefact types males 2.0 1.7 1.2 2.5 3.1 2.5 2.7 2.8 0.5 females 1.4 2.0 3.0 3.0 2.1 1.5 3.4 4.0 1.4

late cemetery (7th/8th cent.) Table 2: Male and female burial wealth (after Arnold 1980, table 4.6)

grade alpha

males sword

females gilt brooches gold/silver bracelets amber or crystal beads bronze bowl glass vessel weaving batten brooches necklace knife or buckle or 2-3 beads

social class thegn

beta gamma

spear knife

ceorl unfree

Table 3: Burial grades and social classes after Alcock

category A

males broad range of of rich finds, incl. ring sword, bronze bowl; later: seax, bronze bowl

females gilt brooches, gold braiding, crystal ball, bracteates, bronze bowl; later: union sets, cabochon garnets, biconical gold beads (poorer versions of A objects) ) ) ) (continuum; difficult to distinguish)

sword, shield, spear; later: seax + spear shield, spear spear knife or unfurnished

C D E

knife and/or buckle or unfurnished

Table 4: Wealth categories after Shephard

coastal regions wealth levels general wealth external trade weapon burial frequency up to 7 rich much high

inland regions up to 5 poor little low

Table 5: Regional differentiation of wealth and social structure after Arnold

Proportion of male adults in: weapons 5th/6th cent. sword axe (5th/6th cent.) seax (6th/7th cent.) shield spear no weapons 7th cent.

6%

6%

42%

17%

52%

77%

Table 6: Male grave equipment over time