7. Multiple Liminalities in Early Anglo-Saxon England: Age, Gender and Religion

can be useful in identifying complex patterns in many areas of study. This paper examines two elderly burials from the Anglo-Saxon cemetery of Mill Hill, Deal, Kent in England. These two individuals display multiple liminalities, particularly with respect to their advanced age, gendered identities and in their potential religious beliefs. Their advanced age suggests they spent the last years of their lives in the knowledge that they were close to the ultimate transition: death. Moreover, being quite elderly has likely aﬀected their gendered identities, rendering them muted, altered and/or de-gendered. The combination of their elderly status, their physical position in the cemetery and the phasing of their graves suggests they were the last two individuals buried at Mill Hill, possibly some considerable time after it had gone out of general use. Their graves are dated to the Final Phase or Conversion Period of Early Anglo-Saxon England when, among other changes, Christianity begins to become the dominant religion; this and the presence of an unexcavated cemetery nearby, with goods pointing to a later date than this one, suggests the possibility that these two individuals were the last pagans in their community. These are all factors which point to the multiple liminalities of the two individuals.


Introduction
Liminality is existence on the threshold of another reality. It is an ambiguous space, where two or more worlds may merge, a place of multiple inclusion and/or exclusion, and where opposites may co-exist. Liminal spaces can be formed by nature (as in the sea shore, alternatively wet and dry), forcing those who exist there to deal with both realities. They can also be the result of human activity and thought; humans define their societies through their differing cultures, philosophies and truths, thus creating liminal spaces within and without their communities.
An exploration of liminality can therefore illuminate hitherto shadowy aspects of a society. This chapter aims to explore liminal spaces and transitions with respect to Anglo-Saxon funerary rituals. Specifically, we examine and explore the evidence for multiple liminalities among two burials from the Anglo-Saxon cemetery of Mill Hill, Deal, Kent. Before we can examine the burials, however, we need to explore the meanings and uses of liminality. Then, to contextualise these burials, we will examine the cemetery on Mill Hill, its meaning, structure and chronology as well as those interred within it. Following this, we identify two somewhat unique individuals who, due to their multiple liminalities -their placement in the cemetery, their extreme and thus liminal age, and their gendered identities -are illustrative of multiple aspects of their culture. We will then discuss the changing religious beliefs of the people of Mill Hill, and the multiple liminalities uncovered in this process.

Liminality: Life, Death and Other Passages
The concept of liminality was originally formulated by van Gennep ([1908] 1960) over a century ago when considering rites of passage. It would be another half Cave, C. and Oxenham, M. 2018. Across the Generations: The Old and the Young in Past Societies. [91][92][93][94][95][96][97][98][99][100][101][102][103][104] Liminalities and ambiguities can be useful in identifying complex patterns in many areas of study. This paper examines two elderly burials from the Anglo-Saxon cemetery of Mill Hill, Deal, Kent in England. These two individuals display multiple liminalities, particularly with respect to their advanced age, gendered identities and in their potential religious beliefs. Their advanced age suggests they spent the last years of their lives in the knowledge that they were close to the ultimate transition: death. Moreover, being quite elderly has likely affected their gendered identities, rendering them muted, altered and/or de-gendered. The combination of their elderly status, their physical position in the cemetery and the phasing of their graves suggests they were the last two individuals buried at Mill Hill, possibly some considerable time after it had gone out of general use. Their graves are dated to the Final Phase or Conversion Period of Early Anglo-Saxon England when, among other changes, Christianity begins to become the dominant religion; this and the presence of an unexcavated cemetery nearby, with goods pointing to a later date than this one, suggests the possibility that these two individuals were the last pagans in their community. These are all factors which point to the multiple liminalities of the two individuals.
Keywords: Mortuary archaeology; archaeology of religion; gender; old age; pagan; Christian Address: School of Archaeology and Anthropology, Banks Building, Daley Road, Acton ACT 2601, Australia. Emails: Christine.Cave@anu.edu.au; Marc.oxenham@anu.edu.au century before van Gennep was rediscovered by Turner (1967), paving the way for a vast and diverse liminality literature. Van Gennep (1960) suggested that humans create distinctions between groups of individuals in terms of class, status, ability, qualifications and so on. A single distinction creates two categories of being, say 'in' and 'out', which in turn creates a third, 'between' (Metcalf and Huntington 1991, 30); the creation of distinctions facilitates events or ceremonies which enable individuals to move from one state to another -these are the rites of passage (van Gennep 1960, 3). Such rites may include the passage from childhood to adulthood, from unmarried to married, from apprentice to master and so on, where the previous identity ritually 'dies', and a new self is 'born' (Metcalf and Huntington 1991, 30). Van Gennep (1960, 11) observed that these transitions were themselves subdivided into rites of separation, rites of transition and rites of incorporation. It is with rites of transition, or liminal rites, that an individual moves on from their previous status, but prior to incorporation into their new status they reside in a liminal state, neither one nor the other, 'betwixt and between' (Turner 1967, 93-111).
Van Gennep saw universal similarities in ceremonies across cultures, especially with respect to liminality, and was able to relate these to the world around the ceremonial actors: 'the universe itself is governed by a periodicity which has repercussions on human life, with stages and transitions, movements forward, and periods of relative inactivity' (van Gennep cited by Thomassen 2009, 12). Turner (1987, 6-7) who, on recognising van Gennep's insight, liberated his rites of passage and introduced these ideas into archaeology noting that the individual in the liminal period is 'structurally, if not physically "invisible"', with a dual character, essentially neither one state nor the other, and yet of both states. This liminal condition is an ambiguity and a paradox, confusing all states; it is an essentially unstructured space, which is also de-structured and pre-structured, and one that is frequently regarded as limitless and unbounded (Turner 1987, 7-8). He also noted that often this in-between space has a negative tint -of death, decomposition or other pollutant (Turner 1987, 7-8).
The negative, polluting expression of liminality is expressed where objects and people are out of place or do not fit into society's constructed categories (Jackson 2005, 333). The idea of margins being dangerous and polluting is explored by Douglas (2003); liminal substances issue from the body -including saliva, blood, milk, faeces, urine, tears and sweat -and are often, but not always, regarded as polluting. Death is also a time not only of loss and separation, but also a time of fear: the corpse itself can engender fear and pollution, while decomposition is a liminal state (Turner 1987, 53;Metcalf and Huntington 1991, 23). Similarly, morality creates borders of behaviour from which errant behaviour, or moral pollution, can escape and which must be mitigated through ritual cleansing (Douglas 2003).
Van Gennep's work is particularly visible in anthropological writing, often used in exploring a range of topics, including sex and gender (e.g. Mageo 1996), migration (e.g. Wilson 2017) and eating disorders (e.g. Eli 2017). The state of liminality can be experienced by single individuals, groups of individuals, or whole societies over time periods ranging from moments to aeons, and confined to limited and specific places like doorways or lines, or may cover larger areas, or even countries or continents (Thomassen 2009, 16). These liminal expressions of people, time and space can function singly or in combination and are not necessarily related to a recognisable or identifiable rite of transition. There can also be differences in scale: the degree of liminality experienced is related to surrounding structures (Thomassen 2009, 17-18) Liminality has been co-opted by a diverse range of disciplines (Thomassen 2009, 18). For instance, psychologists and therapists recognise and use liminality in individuals, whose experiences may be felt and recognisable, or imperceptible and subliminal; some individuals may be diagnosed as 'borderline' due to the nature of their condition (Thomassen 2009, 18). In literature, post-modern and/or post-colonial writers create characters from the interstices, between common identities, expressing cultural hybridity (Thomassen 2009, 18). Other fields using liminality include, but are not confined to, business consultancy (e.g. Czarniawska and Mazza 2003;Thomassen 2009), large scale societies and modernity (Szakolczai 2003;, and organisational settings (Beech 2011). Beech (2011, 287) utilised the concept of liminality for use in organisational settings, where identity is being constructed and reconstructed. Liminality is utilised to classify people, occupations, hierarchies, organisations, events and even spaces. Self-identity is an internal self-view, constituted through social structures and discourses, but liminality is created where structure and agency intersect (Beech 2011, 286). Identity reconstruction can occur throughout life and work: for example, a student may change from an undergraduate to a postgraduate, but in that new position the reconstruction may be incomplete as the student suffers imposter syndrome thus making their self-identity liminal (Beech 2011, 286-7). Changes in identity also imply changes in meaning, both within the individual and from society; whether internalised or external, these dialogues also result in change, which again, creates liminal spaces (Beech 2011, 288). The changeable nature of identities, the multiple meanings they can present and any negative consequences can be emphasised and illustrated by the concept of liminality (Beech 2011, 287). Change can disrupt an individual's sense of self and their place within society (Noble and Walker 1997, 31); thus, any reconstruction of identity may result in liminality (Beech 2011, 287).
The idea of liminality is not restricted to human actors, although it is humans who categorise in this way. Particular animals can be perceived as liminalthe beaver moves between water and land; the bat is not a bird, yet flies; the bear lives on the earth and sleeps in the earth (Emerson 2003, 76). Figurines can be interpreted as liminal objects; in ritual, transition can be anchored into a compelling reality when expressed in (tangible) images (Haaland and Haaland 1996, 297). Humans also see the landscape, whether natural or created, as having liminal spaces, like the sea shore, or a demilitarised zone between warring parties (see, for example, papers in Andrews and Roberts 2012).
Although liminality can identify and embrace all these concepts, it is its relevance to death and mortuary practices that is explored in this chapter. Rituals, including ritual cleansing and rites of passage are enactments of social relations and their purpose is to provide a visible, tangible performance for the cohesion of society. Rites of passage involve a structural 'death' and a 'rebirth' of an individual, creating and recreating the community; shamans who manipulate the dead through death and rebirth for example, in the contexts of long barrows, are able to move and operate in the liminal space between the worlds of the dead and the living (van Gennep 1960;Emerson 2003). Thomas (2000, 662) sees British Neolithic long barrows and megalithic tombs as places of transition and liminality; these monuments were places where personhood was dissolved and recreated through manipulation of the skeletal remains; these practices not only memorialised the ancestors but also established their rebirth and presence in the surrounding landscape.
Manipulations like these are about rebuilding the social, cultural and spiritual fabric of a community that has been damaged by the death of a community member (Metcalf and Huntington 1991, 33-7;Emerson 2003, 74). In these communities, group identity revolved around common social practices and the surrounding landscape as the group moved seasonally between places in regular, predictable cycles (Thomas 2000, 665). The situation changed after c. 3000 BC, as identity became more elaborate and more asymmetrical, and the monuments were then used to exclude some and include others (Thomas 2000, 665). The dead were moved from the liminal space between life and death, and were relocated spatially into the past, by the closing of long barrows and the move to single graves (Thomas 2000, 664-5).
In Ireland, the powerful Catholic Church regulated who could, and could not, be buried in consecrated ground; those denied this rite were unbaptised infants, suicide victims, the mentally disabled, shipwrecked individuals, criminals and strangers to an area (Murphy 2011, 409). Cillíní were burial grounds established for unbaptised and stillborn children as well as other categories of outsiders (Murphy 2011, 409). Research on cillíní has related these burial grounds to the liminal phase of the rites of passage, considering the ambiguity of the dead child, with the exclusionary nature of these burial grounds reflecting the spiritual liminality of a dead infant denied heaven (Finlay 2000, 408). While noting that those buried in cillíní have been rendered marginal by the teachings of the Church, Murphy (2011) asserts that this fails to take into account the attitudes of those who bury these children, especially with regard to the emotions involved. She argues that rather than marginal, unrecognisable places, parents have selected unforgettable places to bury these children, places that would remain in their thoughts, despite the liminal status prescribed by the Church (Murphy 2011, 417). The burial grounds have been located with pre-existing monuments and abandoned ecclesiastical sites in contrast to infanticide dumps, such as privies and dung heaps (Murphy 2011, 417). The choice of 'marginal' bogs or woodlands would ensure that agricultural activities would not disturb the remains (Murphy 2011, 418). Thus, it can be seen that although rendered 'liminal' and out of place by the hegemonic Church, individuals took care to bury their loved children in prominent places, thereby suggesting that the burials were ambiguous in that the children were both loved and excluded, barred and buried. Although liminality was prescribed by the Church, it was resisted by the parents of these dead children.
In early Anglo-Saxon England, while outcasts are generally included in community cemeteries, analysis by Reynolds (2009, 231) suggests that such individuals -those buried prone, decapitated, stoned or amputated -are found at the edges of community cemeteries or family clusters. Those not on the edge when the cemetery went out of use may have been on the border when buried (Reynolds 2009, 231). Later, as the Anglo-Saxon kingdoms gradually emerged, isolated deviant burials and judicial execution cemeteries are found at the limits of administrated territories, major boundaries and crossroads. Although not all 'deviant' burials are found on borders or at crossroads, burials in these places are deliberately liminal. While early burials on the edges of communal cemeteries may be commonly understood examples of 'community punishment', later burials were demonstrations of the elite's ability to control and regulate their populations, and displays of centralised power (Reynolds 2009, 236-8). Positioning of 'deviant' burials on borders is suggestive that these individuals were outside the community of the living and of the dead; the liminality can perhaps be considered a description of their nature, as well as an exclusion from both the communities of the living and the dead.
These human-created liminalities can come in many forms. A community's religion defines its belief systems and provides a range of regulatory mechanisms; entities which tend to contain and corral, thus resulting in the creation of liminal spaces at the edges of these enclosures. The spaces may relate to visible physical aspects of a person like age, sex or ethnicity, aspects which are etched on the body and its associated material culture and obvious to the relevant beholder. The liminalities can also be culturally constructed -relating to behaviour, belief or political identity. An important instance of these structures is religious belief.
As humans, we are cultural animals, meaning our behaviour cannot always be explained directly through natural or evolutionary perspectives (Reicher 2004). Religious beliefs, in their various forms, are arguably the defining blueprint of any culture, regulating thought, behaviour and meaning. It is most likely to be a long-lived entity, surviving for generations, and in the case of 'world religions' like Christianity, Judaism, and Islam (Hefner 1993, 4), perhaps for millennia. An individual is born into a religion, lives that religion, passes it on to their descendants and dies in it. A societal religious change is a major undertaking and a rare occurrence and when such a change occurs it creates liminal spaces.
Liminal spaces also exist in the context of gendered communities. Gender is often seen as a binary entity, but with critical examination that duality disappears (Ghisleni et al. 2016). Gender liminality can be viewed not only through the prisms of homosexuality, hermaphroditism and gender dysphoria but also through the perceived sexless identities of particular occupations or age cohorts (Hird 2000;Gowland 2002, 31;Drescher 2010). Although the binary feminine and masculine are frequently seen as essential, and probably have been considered as such for millennia, this is not always so; some societies do not distinguish gender and the biological sex of their children until puberty (Godelier 2011), while others allow gender shifting unrelated to sex when the situation requires it (Ghisleni et al. 2016, 771). Furthermore, ambiguous gender identities can also exist through societal attitudes, recognised or unrecognised (Martin 2004). Whether deliberate or subliminal, both infancy and old age can also be viewed as de-gendered (Silver 2003).
Infants and the elderly feature in another liminality, which although related to religion and gender, is worthy of its own designation. The liminality of age relates to an individual's nearness to a state of non-existence -those newly arrived in the world, and those likely to leave it soon: such individuals are of this world now, but their proximity to the world beyond (and before) may engender fear, wonder or antipathy (Welinder 2001, 170). This other 'existence' may be the 'afterlife', or another state created by the culture or religion to which they belong, whose construction fulfils a need to understand the great questions of life, existence and death. Those close to this transition may be feared, revered, both or neither, for this very liminal proximity and any presumed traits they may hold due to this recent or forthcoming contact. This is the liminal space of age.
The Cemetery -Mill Hill, Deal, Kent The Mill Hill cemetery was excavated between 1986 and 1989 by the Dover Archaeological Group under the direction of Keith Parfitt, ahead of development of the site for housing (Parfitt and Brugmann 1997). As suggested by the name, the cemetery sat on high ground commanding much surrounding countryside as well as views over the sea; the burials were arranged around a prehistoric ring ditch surrounding a barrow, since ploughed away, but apparently visible at the time. The monument sat in the midst of Iron Age burials, possible satellite barrows and Roman ditches, quarries and graves (Parfitt and Brugmann 1997). This position, imbued as it is with the past power of the landscape and its people, links the cemetery, and thus the community, not only to the surrounding countryside but also to the authority of those who went before them (Price 2010, xv).
The cemetery is dated, largely through brooches and dress accessories, to the period AD 500-590 (Parfitt and Brugmann 1997, 100). Where grave inclusions were undatable or absent, the excavators used grave positioning or placement to infer dates where appropriate. In addition, selected later phase burials were utilised by Hines and Bayliss (2013), including radiocarbon dating of those burials, in their search for a new chronological framework for the later part of the Early Anglo-Saxon period. Their chronological findings do not differ greatly from those of the original report. We note that Parfitt and Brugmann's (1997, 124) report states that age and sex determinations were not influenced by associated gendered grave goods, a practice that has been common in the past; however, the particular methods used for age and sex estimation were not reported.

Spatial Patterning
As noted earlier, the Anglo-Saxon cemetery focused around a Bronze Age ring ditch; graves were placed on, in and next to both the north-east and south-west sides of the ditch. The original report labelled these two groups Plot A and Plot B (Parfitt and Brugmann 1997, 13-17), but this simple division conceals finer grained spatial patterning evident in the cemetery. Consequently, we have created five discrete plots, labelled 1-5 ( Fig. 1) that can be identified by a combination of spatial, orientation and temporal (phase) factors.
• Plot 1 encompasses those graves on the south-west side of the barrow ditch that are orientated south-west to north-east, which differentiates them from the Plot 2 graves. Plot 1 graves derive from the earliest phases of the cemetery and they sit in three parallel rows orientated towards the ring ditch.
• Plot 2 covers the same geographical area, but the graves are oriented generally east/west. Many have disturbed earlier Plot 1 graves, suggesting a sufficient time gap between the burial phases to have obliterated memory of the earlier (Plot 1) graves.
• Plot 3 encompasses the graves on the north-east side of the ring ditch, sitting inside, on and outside the ditch. Almost all of these graves are placed in distinct rows that radiate towards the barrow centre without intersecting each other. Only Burial 10, which alone in this plot comes from the latest phase of the cemetery, does not point towards the centre. Burial 10 also cuts through two other burials (Burials 71 and 72), again suggesting the passage of a certain amount of time between the burials in this plot.
• Plot 4 is a group of six graves inside the ring ditch situated between the east and south-west of the centre. Four of these graves are orientated roughly north-west to south-east and two south-west to north-east and they are all dated to Phase III/IV (see discussion of chronology below). None of these individuals was aged over thirty-five years, although two remain unaged. • Plot 5 is a double row of graves along the eastern side of the barrow ditch, orientated largely south-west to north-east. The graves are well-spaced, suggesting they may have been covered with barrows, a thesis supported by the observation that eight of the eleven burials in this plot are masculine, weaponed, burials. Each of these eight individuals was buried with a spear, and three have the full weapon set of a spear, shield and sword; one has a seax with his spear, but no shield. These putative burial mounds appear to reference and reflect the earlier monuments on the site (Carver 2010, 11). There are three burials which do not fit this masculine pattern, however, all of which are located at the southern end of the plot. Burial 68 is closest to the ring ditch, a young material culturally feminine burial which could perhaps be included in Plot four, while Burials 94 and 95 are both older feminine burials.

Chronological Patterning
In addition to spatial information, the grey-scale coding in Figure 1 highlights the chronology and phase ordering of the burials within the cemetery. The phasing of the nine graves examined by Hines and Bayliss (2013) is also shown (stippled, striped and cross-hatched). As they were unable to reconcile gendered graves into one general phase, their study identified separate phasing for feminine and masculine graves. Their categories are labelled AS (for Anglo-Saxon), F (for female/feminine), or M (male/masculine), with phases ranging from B (earliest) to C (latest), including liminal categories like B-C which come between B and C (Hines and Bayliss 2013, 460-1). All the Mill Hill female/feminine graves they examined were in their AS-FB phase, which is the earliest of their phases, but the latest in the Mill Hill cemetery chronology. The Mill Hill male/masculine graves fell into three phases, AS-MB, AS-MB-C and AS-MC which, like the feminine phasing, were early phases in their chronology, but the latest phases in Mill Hill cemetery chronology. Although not all the latest phase graves are in Plot 5, all Plot 5 graves come from the latest phases of the cemetery. Final phase graves in Plots 2 and 3 intersected with, and disturbed, earlier graves thereby indicating that the earlier graves were no longer visible when these graves were dug. Using the dating of Parfitt and Brugmann (1997) (see Fig. 1), it can be seen that Plot 5 graves all fall within Kentish Phase III/IV or Kentish Phase IV. Although Burial 79 (Phase IV) is close to the middle of the plot, the rest of the Phase IV graves are at the southern end, with earlier graves to the north. The phasing of Hines and Bayliss (2013) shows the three male/masculine graves they examined in Plot 5 are ordered from earliest to latest from north to south. Consequently, it is likely that the southernmost graves in Plot 5, Burials 94 and 95, were the last to be dug; this is supported by the differences (see below) apparent between Burials 94 and 95 and the rest of the burials in this plot. As this is the latest phase in the cemetery, Burials 94 and 95 were possibly the last two individuals ever to be buried in the Mill Hill cemetery.
This spatial and chronological patterning is also consistent with the age estimation data. Parfitt and Brugmann (1997, 108-9) used the high proportion -50%, of 'mature' individuals (although not defined by them, it appears to be 40+ years) -in Phase IV to conclude that the cemetery was abandoned in the late sixth century. To better explore older age and mortuary patterns, we re-estimated age-at-death in the Mill Hill sample using the method outlined in Cave and Oxenham (2016). Table 1 summarises the revised age-at-death profile where 12% of the sample has now been identified as being aged between fifty-five and 75+ years. Given the importance of Kentish Phase IV burials in our discussion, Table 2 provides age estimate revisions for each burial aged twenty-five years or more at death. It is also worth noting that the average age-at-death of all those buried in Kentish Phase IV is considerably higher than those from all other phases (Fig. 2). The cemetery abandonment hypothesis, noted above, is also supported by the existence of a second unexcavated burial ground situated within 500 m of this one; a brooch and a glass vessel found there have given that burial ground a probable later date than the current cemetery (Parfitt and Brugmann 1997, 109).  Cave and Oxenham (2016), 2 number (%). We consider it likely that a considerable gap in time existed between the earlier phase burials, Plots 1-4, and the feminine burials at the south of Plot 5. The later burials in Plots 1-4 have disturbed previous graves, whose location had apparently been lost to living memory. The putative Plot 5 barrows would have been easily visible and, perhaps knowing previous mistakes, the burying party took the 'safe' option of placing Burial 94 next in line after Burial 93; when the woman in Burial 95 died sometime later, she was placed in the next available space along that line of burials.

The Burials -94 and 95
The two oldest individuals, as determined in our re-aging, are Burials 94 and 95. As suggested by the chronology and spatial patterning of the cemetery, these also appear to be the very last individuals to be interred in the Mill Hill cemetery. Burial 94 was sexed as 'Female?' by Anderson and Andrews (1997, 194) and aged 40-50 years. Beavan and Mays (2010, 104) estimated the age-at-death of this individual as 50+ years and the sex as male. The skeletal material was in poor condition, only 20-40% complete (Beavan and Mays 2010, 104), and with no pelvic elements available, the sex determination was based on the presence of male cranial features (Simon Mays, pers. comm.).
Burial 94 was buried with feminine grave goods: one debased silver Kentish disc brooch, decorated with garnets, zig-zag niello, and Type 6.1-2 animal decoration with a punched rim, as well as fused fragments of keys, a knife, a necklace of fifty-four beads and various unidentified metal fragments (Fig. 3). The individual lay in an extended supine pose with the head facing the right and the left forearm lying across the body, wearing their jewellery and with the knife carried at the waist (Parfitt and Brugmann 1997, 154, fig. 73). The individual had suffered some ante-mortem tooth loss and an abscess, but exhibited neither dental enamel hypoplasia nor cribra orbitalia (Anderson and Andrews 1997, 233;Beavan and Mays 2010, 104). Following our re-estimation of age-at-death (Table 2), Burial 94 was determined to be the second oldest individual in the cemetery (65-74 years old). The grave, at a depth of 0.74 m, was relatively deep, and more than one standard deviation deeper than the average Mill Hill grave (0.53 m). Mattock marks were visible on the base of the grave cut. Finally, an examination of a photograph of the body in situ reveals further details from a field anthropology perspective (Duday 2006;Willis and Tayles 2009). Poor preservation meant that the labial joints could not be assessed, however, it was clear the left femur was laterally rotated somewhat out of the acetabulum, otherwise a linear alignment of the left side of the body was evident. It is possible that further lateral rotation of the femur was restricted by the grave cut or some other long since decomposed structure. Although not well preserved, it was apparent that considerable movement or disturbance has occurred in the thoracic region. The mandible had fallen open and the skull was crushed. From these observations, it is likely the body was interred in a void, perhaps a coffin (Anna Willis pers. comm.) Burial 95 was assessed as female by both Anderson and Andrews (1997, 155) and Beavan and Mays (2010, 104), with an age-at-death estimate of 45-55 years and 50+ years respectively. Her grave goods consisted of a small broken copper alloy quoit or annular brooch with worn punched decoration, a necklace of twenty-two beads, a key, a knife, an iron ring and ring fragment, as well as a fragment of the foot of a Roman dark blue glass vessel (Fig. 4). Like Burial 94, she lay in an extended supine position; both arms were by her side and, while she wore her necklace around her neck, all other goods, including the brooch, were found at her left hip, possibly encased within a since decomposed bag. She displayed evidence of vertebral osteophytosis as well as Schmorl's Nodes (Anderson and Andrews 1997, 227); Beavan and Mays (2010, 104) noted there was no evidence of cribra orbitalia.
At only 0.29 m deep, this burial comprised one of the shallower graves in the cemetery, and was more than one standard deviation below average. Observations derived from an in situ photograph of the burial provide further information. As with Burial 94, the labile joints were not well represented except for the right hand, which had disarticulated, and the right foot. There was a linear alignment of the lower limbs, especially on the left, but also on the right to a degree, and the right foot was plantar-flexed in an unnatural position against either a long since decomposed wrapping, other structure/object or the edge of the grave cut. The pelvis was articulated, but some slight disarticulation was visible at the sacroiliac joint. Curvature of the vertebral column was evident and the ribs had expanded laterally, suggesting they were not constricted. Something had disturbed the left forearm enough to turn the distal fragment of the radius around (assuming this did not occur during excavation). No constriction was visible at the level of the shoulders, but evidence of curvature was present that indicates the bottom of the grave was not flat. The right arm appeared to be higher, and the skull was also propped up higher than the rest of the body, both supporting the interpretation of a curved grave base. The general movement of the smaller labile elements, although not well preserved, could indicate the individual was interred in a void, perhaps a coffin (Anna Willis pers. comm.). Although there was little movement of the larger elements, this is largely a product of the container or walls of the grave.
The only cranial material from Burial 95 available for the authors to examine was the mandible, and this showed extensive ante-mortem tooth loss. Noting the findings of Mays (1998, 62), that individuals older that 50-60 years have generally lost at least half their teeth, Burial 95 was placed at the furthest end of the seriation. Burial 95 was the southernmost burial in Plot 5 and it appears likely she was buried after Burial 94, and probably after the cemetery had largely gone out of general use. Burial 95 was most likely of the same, or a similar, cohort to Burial 94 but she was buried later, while an age re-estimation (Table 2) indicates that Burial 95 was slightly older than Burial 94 at 75+ years old.
These two graves display multiple liminalities: they are liminal in time, as they may have been the last buried in the cemetery; they are liminal in space, as they were granted slots on the edge of the cemetery. Additionally, they display liminality in their great age, their gendering and in their probable religious and/or spiritual beliefs.

Liminality and Age
Burials 94 and 95 lived the final years of their lives at an age where they, as the longest living members of their community, were closest to the ultimate transition: death. Here we examine how their advanced age may have affected the manner of their burial as well as perhaps the final years of their lives. They were both buried with arguably standard Anglo-Saxon feminine grave goods. At Mill Hill, females and/or feminine gendered individuals were interred with up to six brooches, but by the time this pair was buried, feminine burial fashion had become less ostentatious with only a single brooch worn at the neck being usual (Parfitt and Brugmann 1997, 108, 116). These two elderly individuals were probably born during the first phase of the establishment of the cemetery in the early decades of the sixth century. We cannot say whether they began their lives in the community in which they died, however, or moved to it as children or upon marriage. Notwithstanding, they likely grew to adulthood at a time when Kentish women, including women they must have known and associated with, wore many brooches. It is likely that they too, at one time in their lives, had more brooches than those that accompanied them to their graves; begging the question, what happened to those brooches? Martin (2012) suggests that brooches were the inalienable possessions of their owners and could not be disposed of except through the burial display, but it seems that by the time this pair died, fashion or other societal changes may have altered such funerary prescriptions. Elderly people were most likely prepared for their graves by those younger than them, perhaps daughters or granddaughters. As such, did these family members make sure their relatives were buried in the appropriate fashion, even if they did not dress that way in life? Did the burial party take any excess brooches for their own use, discard or gift them? Or had these elderly individuals themselves disposed of the objects prior to death?
The brooch interred with Burial 95 was broken and unable to be worn, likely carried in a bag at her hip. It was a relatively plain brooch, with punched decoration and no precious materials or inlay, in a cemetery with many highly decorated brooches. Only four individuals from the earlier phases of the cemetery (Kentish Phases II-III) were buried with only one brooch: two under twenty years old (one female, one of indeterminate sex), one aged 35-44 years (female) and one of uncertain age and sex. Considering the age-at-death of these single-brooched individuals, and the period in which they were buried, they all (including Burials 94 and 95, but possibly excluding the unaged Burial 63) were probably contemporaneous. The grave goods of Burial 95 do not compare particularly favourably with other individuals with brooches of debased silver and/or garnets, apart from Burial 63 whose undecorated copper alloy annular brooch was offset by a fine decorated copper alloy buckle and belt plate (Parfitt and Brugmann 1997). Was Burial 95 a relatively poor member of the community and/or of a lesser status? Were any usable or higher quality brooches shared among her descendants, with only this one unwanted or deemed appropriate as a burial good?
Burial 94's brooch was of relatively higher quality to that of Burial 95: like many Mill Hill brooches, it was crafted of debased silver, extensively inlaid with garnets and other materials, with a punched decoration. Was it also unfashionable or the least desirable of the original suite of brooches of Burial 94? Conversely, was it the richest and most beautiful, and considered the most appropriate for the grave or the afterlife? Minimally, this individual appears to have been wealthier than Burial 95, and perhaps of higher status.
Although we can provide no definitive answers to questions about the lives and status of these two individuals, age liminality is probably an important factor. Although these burials demonstrate marginality, in their brooches, age and burial plot location, they also demonstrate their ultimate inclusion, with both individuals being buried normatively.
Other aspects of age and liminality are also apparent. The elderly is the cohort most likely to be buried with grooming items, especially tweezers (Cave and Oxenham forthcoming), but neither of these individuals took such items to their graves. While high status Anglo-Saxon male or masculine burials signal martial rank, female or feminine graves exhibit aspects of beauty (Cave and Oxenham 2017). Burial 94, although interred with a highly decorative brooch, was unlikely to have been beautiful: either because they were male (cranially sexed as male -see above), masculine in appearance or due to their advanced age (it is worth noting that females can develop male-like cranial features with increasing age (Walker 1995, 36)). Burial 95, on the other hand, retained her gracile cranial features until the grave. Yet neither individual was interred with an item used for creating or maintaining their looks. Although fewer brooches had become common, these two individuals were likely to have owned more than one brooch in their youth. Is it possibly the case that burial with a single brooch also occurred because elderly females had moved beyond conventional beauty into a liminal space reserved for those whose looks cannot be redeemed?
The very old in Anglo-Saxon England were also the cohort that was least likely to be awarded a nonnormative burial (Cave and Oxenham forthcoming). Although these graves each have features which differentiate them in small ways from most of their community, neither could be considered non-normative. In terms of grave construction, one is relatively deep, while the other is relatively shallow. Furthermore, field anthropological observations of the in situ photographs of the burials suggest they were both coffined. While we were unable to carry out field anthropological assessments of all burials, archaeological evidence for coffining has been noted for Burial 38 (female adult), Burial 70 (unsexed adult) and Burial 92 (elderly female of high status) (Parfitt and Brugmann 1997). Recent work at Ryburgh, Norfolk, suggests that coffining of Anglo-Saxon graves was not uncommon and that coffins could consist of hollowed out tree trunks (Hilts 2016), which is consistent with field anthropological observations of Burial 95 (possible curved base of grave/burial structure).
Turning to grave goods, these elderly individuals have fewer than usual brooches found in Mill Hill burials (up to six brooches, average of 2.8), but follow the norm; they also display beauty (or femininity at least) via brooches, but their advanced age (and in Burial 95 masculine features) suggests they lacked conventional beauty. A possible difference in status between the two individuals has been suggested; the older Burial 95 was interred in a shallow grave with a relatively plain brooch in a bag and wearing fewer beads than her neighbour, who had a fashionable Kentish disc brooch, highly decorated with precious substances, and was interred in a relatively deep grave with more goods, including keys and a shield on tongue brooch. Yet in death they are buried near each other, both similar and different, following a line of high status men. These aspects give the graves an ambiguous character: both fashionable and unfashionable, both beautiful and not beautiful, both kept and given away, both old and young, both more feminine and less feminine. These liminalities are likely related to their extended age spans, but can also speak to their gendered identities.

Liminality and Gender
Although Burial 95 was sexed (twice) as female, carried a brooch and wore a necklace, her gendering is muted. She had twenty-two beads in her necklace, while the average for the site was seventy-two, and the median, forty-seven. This small necklace was the only feminine item visible in the burial display. Her one brooch was broken, unable to be worn, and found in a group with other items; her only other gender signalling grave good, a probable key, was also hidden in the bag. In addition, the brooch was of a relatively unusual type, simply decorated, differentiating it from the showier Kentish and Continental types found at Mill Hill. Nonetheless, we should not ignore the foot of the Roman glass vessel -a translucent ring of dark blue glass of similar size and shape to the brooch. Was this item used to decorate the body in the same way as a brooch? Was it placed in the bag alongside the brooch because of this function? This woman was undoubtedly buried in a normative way, but the display was subdued, and most gendered items were hidden from view. Is this because her great age de-gendered her (Silver 2003)?
Burial 94 is a different case altogether. Although Anderson and Andrews (1997) skeletally sexed him/ her as possibly female, Beavan and Mays (2010) sexed him/her as male. Although the biological sex of this individual remains a puzzle (see above regarding age effects on cranial sex estimation), gender, in this instance, is well defined: the burial assemblage can be considered undoubtedly feminine, due to the presence of a showy brooch, fifty-four beads and fragments of possibly four keys.
Did this individual go through life as an unquestioned female, only becoming skeletally masculinised as they reached elderly status? Or were they an intersex or androgyne individual? While it is unclear how rare or common, for that matter, intersexed individuals were during this period, their existence in a site like this is possible. Sex estimation using cranial morphology is relatively accurate (Walrath et al. 2004;Walker 2008), but there are no standards for determining intersex or androgyne individuals and there are further complicating sex estimation factors related to Burial 94 being elderly. Alternatively, was this just an unusually skeletally robust woman? Or were they a male who was accepted as a feminine member of the community? The burial, including a deep grave and a silver brooch, attests to acceptance of feminine identity, but gender liminality remains.
These two relatively plain burials, one of muted femininity and one of potentially feminised masculinity, stand in contrast to the wealth of the male/masculine burials from this latest phase. Elderly men are more generously treated in death than elderly women (see Cave and Oxenham 2017) and this pair reflects this finding. Were their relatively modest graves representative of their liminalities? Or was their modesty a result of other factors? However, they do conform largely to the Anglo-Saxon and Mill Hill norms; their liminality sits inside, not outside, the border.

Liminality and Religion
As noted before, Burials 94 and 95 are the two oldest individuals in this cemetery, were buried during the last phase of the cemetery, and their positioning suggests they may have been the last individuals buried here. These can all be considered liminal aspects of the burials, but together they point towards another, larger liminality shared by the two of them. The latest graves in this cemetery, including Burials 94 and 95, have been designated as 'Final Phase' or 'Conversion period' burials (Parfitt and Brugmann 1997;Hines and Bayliss 2013). This is the period where furnished burial begins to decline, when the cremation burial ritual ends, where regional, ethnic identities or fashions become standardised across Anglo-Saxon England and where Christianity begins to become the dominant religion (Welch 2011, 266-77). Kent is the closest landing point to the already Christian Continent and, at Mill Hill at least, this change appears to be occurring before the arrival of Augustine in AD 597 (Parfitt and Brugmann 1997, 109), an event often used to designate the beginning of Christian Anglo-Saxon England (e.g. Pluskowski 2011, 765).
Anglo-Saxon culture displays considerable regionality and temporal change occurs throughout Anglo-Saxon territory (Price 2010, xiv). Whether 'paganism' or even 'religion' are suitable terms that accurately describe the cosmological beliefs of all, or part, of the peoples covered by the Anglo-Saxon umbrella is difficult to say (Price 2010, xiv). Price (2010, xiv) notes that any examination of early English religion soon finds the need to examine 'almost every aspect of society and culture'. Definitions of 'paganism', the 'dead ghost behind Christianity' (Carver 2010, 3), largely revolve around its not being Christian but little else. 'Pagan' customs and beliefs persisted in England right through the first millennium, however, despite Christianity being well established, and despite efforts by the clergy to stamp them out (Price 2010, xi-xiv).
Conversion from one religious belief system to another on a population level is rarely straightforward but it may provide a reason for a new, possibly Christian, cemetery some 500 m from Mill Hill (Pluskowski 2011). Death is often a time of religious observance, whether followed stringently or ignored in life; it is a time of liminality, when a person moves from one state into a new one, where the living engage with mortality, the afterlife, the supernatural, as well as the transformation of the body and the soul of the deceased (Williams 2010, 68). As differing religions or cosmologies can see this change in different ways, and these may emphasise and re-emphasise the manner in which they see the world, the funeral ritual may be a demonstration of the religious beliefs of the burying party and probably of the deceased as well. Followers of a religion (Christianity) which at its heart rejects other religions, through its belief in a single god with access to eternal life in Heaven reserved for those who have been baptised and who have followed the rules set down by the Church, may wish to separate themselves from those who went before.
Archaeological evidence provides clues to the conversion and can demonstrate that, while Christian ideas were often welcomed alongside the pagan, Christian political dominance was resisted (Carver 2010, 15). The period under review here, the late fifth to early sixth centuries, are many years before the 'dark curtain of Christianity' closed around Europe, 'inhibiting original thought about the supernatural there for the next 1000 years' (Carver 2010, 16). The evidence provided by these two liminal burials suggests they were not beholden to the new faith (whether Christians were banished to a new cemetery, or pagans were banned from it is uncertain). This pair may have remained steadfast in their traditional pagan beliefs and their burial parties recognised and accepted this, despite the probability that they themselves had converted to Christianity. Religious conversion on a personal level involves a reorganisation of personal identity and, at the very least, an acceptance of a new belief system as well as the actions and controls that go with this new or reconceptualised identity (Hefner 1993, 17). Whether the new converts took up this new moral authority because of a rational understanding of the new doctrine, the vigilance of priests and their insistence on a new and sanctified dogma, or punishment of those who refused, we cannot perceive through the archaeology of a pagan cemetery. It does appear that a certain degree of acceptance was granted to those who stayed with their old faith, however, in that they were buried among their kind. It seems possible, or even likely, that those individuals in the community who had converted to Christianity were buried in the new cemetery and those who stayed with the religion they had known all their lives were buried in the old cemetery. It also makes sense that it was the older members of the community who refused to 'move with the times' and convert to Christianity and so continued to be buried in the old cemetery. These oldest individuals appear to have been the last of their kind.

Conclusions
This chapter has examined the burials of two individuals from the Early Anglo-Saxon cemetery on Mill Hill, Deal, Kent, in the context of substantive social and cosmological changes across space and time. These two individuals, the longest lived and probably the last to be buried in the cemetery, speak to multiple liminalities, that highlight not only aspects of their own lives and deaths, but also aspects of the community that buried their dead on the hill.
As the oldest in their community, these two spent many years visibly close to the ultimate transition. Their great age would have been etched on their bodies and perhaps their minds, and would have affected their relationships with other members of their community and, ultimately, those who buried them. Although they were buried on the edge of the cemetery, they were included among a high status burial group consisting mostly of weaponed males. Although Anglo-Saxon feminine burials usually celebrate beauty, or at least highlight it, these two were unlikely to be beautiful, and were buried without the tools used for maintaining and enhancing that beauty. Although they were feminine burials, with beads, brooches and other feminine items, their femininity was muted; this is especially the case with Burial 95 whose feminine identity was simple and almost austere, largely concealed in a bag; while the remains of Burial 94 suggest that this individual may not have been biologically female, despite the silvered and garneted jewellery. Finally, these two burials, possibly the last in this burial ground, appear to symbolise the transition from paganism to Christianity. This is supported by the presence of another (probably Christian) cemetery, within 500 m from this one; although it has not been excavated, grave goods found suggest a slightly later date than that of Mill Hill (Parfitt and Brugmann 1997). These two individuals, of a great age and of differing wealth and status, may have been the last pagans on Mill Hill.