Mead-Halls of the Oiscingas: A New Kentish Perspective on the Anglo-Saxon Great Hall Complex Phenomenon

Abstract

THE GREAT HALL complex represents one of the most distinctive and evocative expressions of the Anglo-Saxon settlement record, and is widely cited as a metaphor for the emergence of kingship in early medieval England. Yet interpretation of these sites remains underdeveloped and heavily weighted towards the excavated findings from the well-known site of Yeavering in Northumberland. Inspired by the results of recent excavations at Lyminge, Kent, this paper undertakes a detailed comparative interrogation of three great hall complexes in Kent, and exploits this new regional perspective to advance our understanding of the agency and embodied meanings of these settlements as ‘theatres of power’. Explored through the thematic prisms of place, social memory and monumental hybridity, this examination leads to a new appreciation of the involvement of great hall sites in the genealogical strategies of 7th-century royal dynasties and a fresh perspective on how this remarkable, yet short-lived, monumental idiom was adapted to harness the symbolic capital of Romanitas.

INTRODUCTION

Defined by monumental timber halls arranged in highly formalised spatial configurations, the class of settlement known as the ‘great hall complex’ has occupied a prominent position in Anglo-Saxon studies since the iconic excavations at Yeavering in the 1950s.² In the intervening years a great deal of illustrative weight has been placed on the shoulders of this rare category of site, crystallising its status as a defining metaphor for the emergence of kingship and elite authority in Anglo-Saxon England. Regularly cited across a wide spectrum of historical, literary and archaeological studies, the phenomenon has proved to be remarkably malleable and adaptable, in the process accumulating a series of meanings and associations drawn from different disciplinary traditions. Through comparison with literary sources, pre-eminently Beowulf’s vivid descriptions of Heorot, it has become conceptually bound with the rituals and preoccupations of early medieval kings and their mobile warbands — hospitality, formalised gift-exchange, and ritualised feasting³ — and, at a deeper level, with the idea of the hall as a unifying ideological construct across early medieval Germanic societies.⁴ Meanwhile, correlation with historical labels pertaining to places of royal authority transmitted through Bede and other textual sources has encouraged the phenomenon to be closely identified with the royal vill/tun, and its wider place in the extractive mechanisms of ‘extensive lordship’, including royal iteration and the consumption of royal tribute (OE ‘feorm’).⁵

In light of the powerful influence that this category of site has, and continues to exercise over Anglo-Saxon scholarship, it is important to emphasise that there are significant weaknesses in our understanding of the great hall complex as an archaeological phenomenon. Sites in this category first received comparative attention in the 1980s and 1990s as part of successive attempts to characterise and periodise the architectural forms brought to light by a rapidly expanding corpus of excavated Anglo-Saxon rural settlements.⁶ These examinations demonstrated that sites like Yeavering, while exceptional in terms of the scale and formality of their architecture, formed part of a much more extensive Anglo-Saxon ‘building tradition’ which had seemingly been adapted to create a distinctive class of elite residence current during the later 6th and 7th centuries ad. While these studies provided an important foundation for further research, they projected an over-generalised image of these sites and offered little in the way of interpretation beyond brief comments on their social and ethnic significance, the latter being subsumed within a broader debate concerning the cultural origins of the Anglo-Saxon house.⁷

Subsequent studies have begun to encourage deeper conceptual engagement with these sites, not simply as the social apex of broader Anglo-Saxon building tradition, but as ‘theatres of power’ — highly manipulated settings which provided emergent elites with an extravagant new outlet, or stage, for enacting the rituals of rulership and political authority.⁸ Under the influence of theoretical and socially informed perspectives, emphasis has shifted towards understanding the social agency and lived experience of these sites and decoding the complex symbolic messages which they were designed to convey as settings where sacral and ideological claims to political authority were actively channelled and animated. This shift has brought previously neglected dimensions of the phenomenon to the forefront of the research agenda — prehistoric monument reuse, the ritualised configuration of space, special deposits, and the strategic exploitation of wider landscape settings⁹ — while encouraging comparisons with related expressions of early medieval hall-culture from other parts of the early medieval North Sea world, in particular a burgeoning of central-place complexes from Scandinavia.¹⁰

While these studies have placed the field on a more sophisticated footing, much of the discussion has been framed around the site of Yeavering, which has had a disproportionate influence on the way in which great hall complexes have been conceptualised. This is partly a reflection of the deeply provocative and arresting qualities of Yeavering itself, laid bare in stunning detail by Hope-Taylor’s excavations and subsequent monograph.¹¹ But it is also a product of the fact that the majority of the great hall complexes recognised today represent poorly understood cropmark sites susceptible, at most, to fairly superficial examinations of plan form.¹² This has ultimately deflected attention back on to Yeavering and, to a lesser extent, Cowdery’s Down (Hampshire) as ‘type-sites’, while at the same time entrenching the homogenised view of later 6th-/7th-century elite residences portrayed by early syntheses of Anglo-Saxon rural settlements.

The simplistic application of monolithic categories of the ‘great hall complex’ type has been recognised as one of the key obstacles standing in the way of more complex and socially nuanced understandings of the medieval past. In a critique that could be applied to any number of sub-fields of medieval archaeology, but which is particularly pertinent to the current context, Kate Giles has recently observed,

buildings must be understood as the result of intimate, local negotiations and interpreted in the context, not of global, national, cultural models or ‘types’, but rather the lived lives of men and women inhabiting particular buildings, at particular times [and places] in the past.¹³

Indeed, some of the most enriching and stimulating advances made within medieval archaeology in recent years have involved exploring the subtleties and idiosyncrasies existing at the margins of established archaeological categories, whether in relation to landscapes of power,¹⁴ monumental sculpture,¹⁵ mortuary traditions,¹⁶ or portable material culture.¹⁷ As these studies readily demonstrate, such treatment lends itself to rigorous and closely contextualised examinations of the archaeological record that offer, at once, sustained and meaningful engagement with theoretical approaches (eg embodiment, materiality, social memory, and biographical approaches) and a critical platform for reformulating entrenched ‘ways of seeing’ the medieval past.

The topic of Anglo-Saxon hall complexes is ripe for re-evaluation of this kind. Recent years have seen a significant growth in the corpus of sites and relevant datasets encapsulating systematic excavation of hall arrays on a scale comparable to Yeavering and Cowdery’s Down (Lyminge, Kent),¹⁸ targeted investigation aimed at improving understanding of cropmark sites (Sutton Courtenay, Oxfordshire),¹⁹ detailed survey and reconnaissance work offering a landscape-scale perspective on the evolution of documented royal centres (Rendlesham, Suffolk),²⁰ and the recognition of new sites through the re-evaluation of old and neglected datasets (Benson, Oxfordshire).²¹ The point has been reached where it is possible to delineate patterns of similarity and difference across multiple parameters of the great hall complex tradition — spatial and monumental properties, landscape settings, life histories etc — both as a prerequisite for attaining a more holistic and nuanced appreciation of the phenomenon itself, and as a spur for informing a wider set of intersecting debates concerning the changing material expression of kingship, power and social complexity in early medieval England.²²

This article develops this research agenda by opening up a new regional perspective on great hall complex phenomenon grounded in the rich context of Anglo-Saxon Kent. It has its origins in a decade-long programme of excavation and research directed by the author at Lyminge, where the remains of a 7th-century great hall complex were systematically excavated between 2012 and 2015. While ruminating on these results I became convinced that Anglo-Saxon discoveries made during the urban redevelopment of Dover in the 1970s offered crucial parallels and context for the Lyminge halls, in spite of being published, in the most definitive of terms, as the remains of an Anglo-Saxon monastery.²³ This realisation was soon followed by reports of the unearthing of yet another Kentish great hall site at Eynsford, in the west of the county. Within the space of a few years Kent — a region long maligned for its dearth of Anglo-Saxon settlement archaeology — had emerged with one of the richest excavated datasets pertaining to the theme of pre-Viking royal residence in the country. Brought together here for detailed comparative examination, this paper harnesses the collective interpretative potential offered by the three sites by combining fine-grained interrogation of salient archaeological and architectural details, with theoretical approaches to access deeper realms of significance and meaning. In pursuing this agenda, advantage is taken of the foundational importance of Kentish historical sources for the study of Anglo-Saxon kingship: sources that provide an unusually detailed, multi-faceted portrayal of the ascendency and hegemony of the native royal bloodline, the Oiscingas dynasty.²⁴

The paper begins by appraising the contextual and archaeological evidence for the three Kentish sites, commencing with the most fully understood — Lyminge — to establish the parameters for subsequent comparison. This prompts a systematic re-evaluation of the published site sequence for Dover, bringing to the fore a number of shared architectural tendencies with Lyminge, permitting its identity as a 7th-century great hall complex to be appreciated fully for the first time. The second part of the article draws out comparative strands from the analysis as a basis for building new and enriched interpretations oriented by and through the local Kentish context. Drawing upon recent archaeological approaches to social memory, it is argued that these settlements were heavily implicated in the genealogical strategies of royal dynasties and that this role is directly manifested in their temporal rhythms and life histories. Meaning is also sought in the distinctive repertoire of architectural features and tendencies shared by the Kentish halls themselves. The argument is advanced that these constructions deserve to be regarded as hybrid monuments which fused traditional elements of the Anglo-Saxon hall with classicising idioms drawn from the architectural programme of the Augustinian mission, most strikingly, sophisticated opus signinum flooring of a type more familiarly associated with pre-Viking churches. In presenting evidence that both churches and elite residences in Anglo-Saxon Kent drew upon the symbolic capital of Romanitas, this study challenges long-standing dichotomies in early medieval scholarship that have emphasised distinctions between secular halls and ecclesiastical buildings on the one hand, and between timber and masonry architectural traditions on the other.

A COMPARATIVE SURVEY OF GREAT HALL COMPLEXES IN KENT

lyminge

Background context

The site of the great hall complex was discovered during a major scheme of research excavation directed by the author under the auspices of the University of Reading, 2008–15. The initial target of this research was a pre-Viking monastic foundation previously investigated by the antiquarian-cleric, Canon Jenkins, who uncovered the remains of an Anglo-Saxon masonry church of ‘Kentish’ type within the limits of the churchyard.²⁵ While early results were very much consistent with this identification, subsequent discoveries demonstrated that archaeological strata relating to an antecedent Anglo-Saxon settlement of the 5th–7th centuries ad lay preserved beneath the core of the modern-day village. Thereafter, the emphasis of the research was redirected towards understanding the character of this precursor settlement and its relationship to the monastic focus. This aim was advanced in a 3-year campaign of excavation sited within a large recreational space within the centre of the village known as ‘Tayne Field’, which brought to light the previously unattested great hall complex as part of a rich, multi-period palimpsest of past occupation and activity (Fig 1).

fig 1 Location of excavations in relation to the modern-day topography of Lyminge. Illustration by Lyminge Archaeological Project. Base map © Crown Copyright/database right 2016. An Ordnance Survey/Edina Service.

Lyminge is located 6.75 km inland of the south coast of Kent at the head of the valley of the River Nailbourne (Little Stour), — one of the most important natural corridors through the North Downs in south-eastern Kent. As well as enjoying a commanding position within this major N/S communication artery, Lyminge sits at the hub of a complex arterial network of downland routeways, those to the west offering convenient access to the Roman road of Stone Street, which connected the shore fort of Portus Lemanis with the civitas capital of Canterbury (Fig 2). While somewhat overshadowed by adjacent valley ridges and promontories, the site of the great hall complex — directly overlooking the source of the River Nailbourne and surmounting a low projecting spur encircled by its headwater — offers commanding views of approaches through the Nailbourne valley, as well as direct sightlines to nearby early medieval cemeteries.²⁶

fig 2 Topographic settings of Lyminge, Dover and Eynsford. Illustration by Matthew Austin. Base map © Crown Copyright/database right 2016. An Ordnance Survey/Edina Service.

Although not without ambiguities, the early historical documentation for Lyminge is considerably stronger than that for the other pair of sites examined in this article.²⁷ It first enters the historical record in ad 689 in connection with a charter of King Oswine granting land formerly appurtenant to the ‘cors’ (royal vill) of Lyminge to St Augustine’s Abbey, Canterbury.²⁸ Explicit mention of a monastic presence at Lyminge appears a decade or so later when King Wihtred granted an estate to the church (basilica) of St Mary, identifiable as a member of a closely knit family of nunneries established by the Kentish royal dynasty in the second half of the 7th century.²⁹ The monastic endowment was enriched considerably over the course of the 8th century with a succession of detached estates located in Romney Marsh and the Weald, a period when Lyminge and its sister establishment of Minster-in-Thanet were, for a time, placed under the rule of a single abbess serving as a proxy of the Mercian royal family.³⁰ References become much sparser in the charter record thereafter, but enough is known to indicate that a monastic community endured at Lyminge into the middle of the 9th century, before its estates and relics were finally, in a fate befalling Kent’s formerly independent royal minsters, absorbed by Canterbury.³¹

Lyminge’s documented status as a royal vill in the 7th century clearly derived from its long-standing significance as a regional administrative centre which, on the evidence of the place name itself and the information provided by two pre-Christian cemeteries, extended at least as far back as the 6th century.³² During this period it is appropriate to view Lyminge as operating as the head settlement of an extensive territory, or regio, broadly commensurate with the lathe of lemenwara documented in later historical sources.³³

Occupation sequence

As the most comprehensive investigation of a pre-Viking royal centre in Kent, the recent excavations at Lyminge have generated a measure against which other local sites of similar status can be compared and evaluated. This capacity is underpinned by a chronologically refined occupation sequence (calibrated by an abundance of diagnostic artefact types and a suite of almost forty radiocarbon dates), allowing Lyminge’s evolution as an early medieval ‘central place’ to be charted with unusual precision over the longue durée. A distillation of this narrative now follows to inform subsequent discussion.³⁴

The earliest phase of Anglo-Saxon occupation identified at Lyminge, dating to the 5th–6th centuries ad, appears to have been confined to gently sloping terrain bisected by the source and headwater of the River Nailbourne (Fig 2). The discovery of contemporary habitation features in a number of separate locales within the village, endorses the view that this early settlement covered a sizeable area, although not necessarily as a continuous swathe. On the evidence of the comparative intensity of occupation and a number of distinctive characteristics, the low plateau subsequently occupied by the great hall complex seems to have constituted the ancestral focus of early Anglo-Saxon Lyminge.

This setting was distinguished by a striking juxtaposition of natural and prehistoric monumental features — a Bronze-Age barrow positioned beside a pre-existing natural solution hollow, or ‘doline’ — which very likely structured the development of the Anglo-Saxon settlement (Fig 3). The site of the former was appropriated by an E/W rectangular timber building of post-hole construction which displayed an unusually complex constructional history and multiple foundation deposits hinting at a potential cultic significance. The neighbouring solution hollow was exploited variously throughout the life of the early settlement, including as a source of clay, a setting for metalworking and related high-temperature crafts, and intermittently as a receptacle for the disposal of large quantities of domestic midden material. Completing the settlement focus were four sunken-featured buildings, widely disposed across the Tayne Field plateau, and a spatially defined cluster of large pits cut with a level of care and precision, suggesting that their original purpose was connected with the storage of perishable food renders.

fig 3 Plan of Tayne Field excavations showing features relating to the two main phases of Anglo-Saxon occupation. Illustration by Lyminge Archaeological Project.

The great hall complex was constructed over a portion of the ancestral focus just described, in the process truncating a number of the sunken-featured buildings (Fig 3). Radiocarbon and associative dating places the currency of the great hall complex firmly within the 7th century, and there is no indication of a hiatus between it and the earlier phase of occupation.³⁵ In its fully developed form, the complex comprised three substantial halls with single annexes configured around a central open ‘courtyard’ (Fig 4). The two halls forming the southern and western ranges of the courtyard (Halls A and C), both orientated on an E/W axis, were built on a larger scale than the hall perpendicularly aligned to the east (Hall B), the former exhibiting lengths in excess of 20 m and the latter a maximum length of around 15 m. There is good evidence that the complex was constructed to a precise and highly formalised spatial template. Firstly, it displays signs of the ‘ritual symmetry’ seen at other great hall complex sites,³⁶ here clearly manifested in the intersection of doorways through Halls B and C (Fig 4). Secondly, Lyminge has been shown to constitute a very early deployment of a standardised form of grid-planning which grew to prominence over the 8th–9th centuries in the spatial articulation of monasteries and rural settlements serving as monastic dependencies.³⁷ As we shall come to see, this detail endorses the view that royal accommodation in 7th-century Kent evolved in close dialogue with the architectural programme of the Augustinian mission.

fig 4 Plan of the 7th-century great hall complex; arrows mark the position of entrances. Illustration by Lyminge Archaeological Project.

Constructed in the ‘post-in-trench’ technique, the halls at Lyminge display many of the constructional traits associated with the great hall tradition: rectangular timber wall posts (baulks) set in different configurations, wattle-and-daub wall panelling; the use of exterior raking posts to counteract the thrust of the roof; and diametrically opposed entrances positioned centrally within the long, and in the case of Hall C, both long and short walls (Figs 4 and 5).³⁸ Another idiom shared by Lyminge is the tendency for halls to be successively rebuilt on the same or overlapping footprints, as particularly demonstrated by Halls B and C, which each passed through three iterations, in the latter case accompanied by a change of constructional technique (Figs 4 and 6). While much of this detail is familiar from previously excavated sites, the Lyminge halls also present evidence for more unusual and idiosyncratic features, most distinctive being the deployment of opus signinum flooring, detailed consideration of which is reserved for a comparative evaluation of parallel evidence from Dover.

fig 5 A view of the western doorway of Hall B (Phase 3) showing structural detail at the base of the massive retaining pit. The rectangular planks of the two door posts are clearly visible in the configuration of stone packing material. Vertical scale: 1 m. Photograph © Lyminge Archaeological Project.

fig 6 Vertical view of Hall B and its three constituent phases. Photograph by William Laing. © Lyminge Archaeological Project.

The long sequence of early medieval occupation which unfolded on the Tayne Field plateau came to an abrupt end when the hall complex was taken out of commission, an event which is unlikely to have occurred much later than ad 700 on the parameters of the available dating evidence. Without a thorough reappraisal of the remains brought to light by Canon Jenkins in the churchyard (located 180 m south-west), it is impossible to know whether there was a chronological overlap between the royal complex sited on Tayne Field and the Christian cult focus which first enters clear historical light as a royal nunnery at the turn of the 7th century.³⁹ What can be said with certainty, however, is that within a relatively short period of time the latter had superseded the former as the monumental core and gravitational focus of the settlement: the great hall complex thus sits at the cusp of a decisive transformation in Lyminge’s monumental trajectory as a theatre of royal power.⁴⁰

dover

Background context

The unearthing and excavation in the 1970s of a major sequence of post-Roman occupation within Dover has not received the attention that it deserves in Anglo-Saxon settlement studies. As we shall come to see, the reason for this neglect largely resides in deficiencies with the final publication — the third in a series of reports based on excavations by the Kent Archaeological Rescue Unit in and around the Classis Britannica fort of Dubris and its late Roman-period successor.⁴¹ This should not detract from the achievements of what was, by any estimation, a vitally important series of archaeological interventions. As the report’s author makes clear, the circumstances surrounding the excavations could not have been less propitious: as was characteristic of the pioneering stages of ‘rescue archaeology’, funding was very scarce and the work had to be completed in the face of highly unsympathetic contractors and civic authorities. That such important, yet comparatively ephemeral, remains were recognised and recorded for posterity in such unfavourable conditions is testament to the persistence, fortitude and, most fundamentally, skill, of Brian Philp and his excavation team.

As made explicit in the report’s introduction, the ensuing programme of post-excavation analysis was self-financed and completed in-house by Kent Archaeological Rescue Unit volunteers. Qualitatively, The Discovery and Excavation of Anglo-Saxon Dover thus falls a long way short of most modern excavation monographs, most conspicuously in its lack of consideration of environmental remains, including animal bones, and the absence of a synthetic appraisal of the portable material culture. While some of these weaknesses can be excused as a product of various constraints both during and after excavation, it is more difficult to forgive the author’s uncritical approach to interpretation which relies heavily upon speculative inference to portray the site as a pre-Viking monastery. Given how much of this argument must be taken on trust it is impossible to recover the real meaning and significance of the underlying results without systematically dismantling the report’s interpretative edifice. Before doing so brief consideration needs to be given to the topographic setting and historical background of the site.

Dover is located 16 km east of Lyminge on the western edge of the now silted-up estuary of the River Dour where it cuts a steep-sided valley through chalk expanse of the North Downs. Offering the shortest sea crossing from England to the Continent, this natural harbour was of pre-eminent strategic importance (Fig 2). The late Roman-period shore fort which provided the setting for the Anglo-Saxon hall complex was itself constructed over an earlier sequence of Roman buildings located within what had originally been the civilian focus (vicus) of the Classis Britannica fort abandoned in about ad 208.⁴² Most notable of these was an opulent 3rd-century mansio known as the ‘Roman Painted House’, comprising a two-storey range situated within a more extensive complex of masonry buildings.⁴³

Several 5th-/7th-century sites are known from the town and its wider environs, according with the picture that the area of the lower reaches of the Dour valley ‘was quite extensively settled during the early Anglo-Saxon period’ (Fig 2).⁴⁴ Within the urban limits of Dover itself, the slopes of Priory Hill and Durham Hill provided prominent settings for two inhumation cemeteries. The former of these is likely to be twinned with a second occupational focus identified in the valley bottom on the west bank of the Dour, some 500 m north of the Roman shore fort.⁴⁵

While Dover is more poorly documented in pre-Conquest sources than Lyminge, available glimpses indicate that it shared kindred status as one of the early minsters of Kent, similarly founded under royal aegis, although somewhat later, by King Wihtred at the turn of the 7th century.⁴⁶ Strong royal connections are, of course, to be expected given Dover’s major strategic importance as a port, but this role only enters clear historical light in the 11th century when, as head of the confederation of Cinque Ports, its harbour was commandeered as a base for the royal fleet.⁴⁷ The establishment of a mint during the reign of King Æthelstan (ad 924–39) nevertheless indicates that Dover was serving as a centre of royal administration by at least the middle of the 10th century, complementing its strategic role as a late-Saxon burgh and bulwark against seaborne Viking raiding.⁴⁸

The published occupation sequence

Philp divides the Anglo-Saxon occupation sequence identified within the south-western sector of the late Roman-period shore fort into two chronologically consecutive phases which he terms ‘early Saxon’ and ‘late-Saxon’, accorded date ranges of 6th–8th centuries and 9th–11th centuries ad respectively (Fig 7).⁴⁹ Activity attributed to the ‘early Saxon’ phase was mainly concentrated in what the report refers to as the ‘North Area’, defined on the west by the perimeter wall of the late Roman-period shore fort and the underlying structures associated with the ‘Roman Painted House’, and to the south by the remains of a substantial 2nd-century bath-house attached to the Classis Britannica fort. The occupation in question comprised a tightly disposed cluster of four sunken-featured buildings (N1-4), together with a modest ground-level timber building of two structural phases (N6 and N7). These buildings were accompanied by an extensive midden deposit (N5) dumped into a declivity formed by the partly silted exterior ditch of the late Roman-period shore fort, against the wall of one of its projecting bastions.

fig 7 Plan of excavated Anglo-Saxon features within the late Roman-period fort at Dover. Redrawn from Philp 2003, fig 1, with additions by author.

Further Anglo-Saxon buildings and occupational features were encountered in the shadow of the medieval church of St Martin-le-Grand – what the report defines as the ‘South Area’. The largest and most imposing were two substantial timber structures of post-in-trench construction, S14 and S10 (Fig 7). These were spaced some 30 m apart in a broadly perpendicular disposition, the former being partially overlain by the southern transept of St Martin-le-Grand, and the latter situated hard up against the internal rampart of the late Roman fort’s western perimeter wall. Of these two structures, the report devotes most attention to S14, a building with a highly complex constructional history, unearthed in a series of separate interventions over a 5-year period between 1974 and 9 (Fig 8). On the basis of its perceived longevity and spatial proximity to the overlying Norman edifice of St Martin-le-Grand, S14 is argued to represent the minster church and cult focus of the documented 7th-century monastic community of St Martin’s. Neighbouring buildings (including S10), as well as those in the North Area are duly attributed to the hypothesised monastic enceinte ‘for use by the canons for domestic or industrial purposes’.⁵⁰

fig 8 Plan of Dover building S14. Redrawn from Philp 2003, fig 34, with additions by author.

On Philp’s reading S14 originated as a diminutive single-celled structure measuring 5 m × 2.7 m with a putative altar base at the eastern end (Period 1).⁵¹ After modest enlarging and remodelling, this primary core was subsequently encased by a substantial post-in-trench building (Period 2), itself rebuilt on three subsequent occasions over the 8th to the 10th centuries (Periods 3–5). In lieu of direct dating evidence, Philp tethers each end of the structural sequence on to two notional historical hooks — the foundation of the Anglo-Saxon monastic community of St Martin’s, and the Norman construction date of St Martin-le-Grand — to conjure a continuous sequence of building activity spanning the 7th to the 11th centuries.⁵²

The other post-in-trench structure, S10, was the more fragmentary of the two buildings, with only portions of its southern and western walls exposed during excavation. If the author’s conjectured minimum width of 7.80 m is accepted, then this would have been a substantial building with a length in the region of 20 m.⁵³ Three structural phases for the building are identified, only the second of which presented clear evidence for wall foundations in the form of shallow U-shaped trenches measuring 350–500 mm in width. The fill of the wall trenches displayed a continuous internal gulley, which Philp suggests held a horizontal wall plate; on analogy with other great halls, however, this feature more likely represents a non-structural feature associated with the occupation or dismantling of the building.⁵⁴ This structure appears to have been built on the footprint of an earlier building represented by what the report describes as a ‘stone and mortar’ floor preserved in several discontinuous patches extending over an area of some 6 m × 5 m. The floor of this earlier building appears to have been partially reused by its successor (accompanied by a new compressed clay floor in the northern portion of the interior) with evidence for a subsequent repair constituting Period 3 of the hypothesised structural sequence.

Accompanying these focal buildings was a smaller and more lightly constructed clay-floored structure (S8), axially aligned with S14 and projecting just beyond its south-western corner. While later structural activity is present in this part of the site, this structure has strong claim to be contemporary with S14 on the grounds of its spatial alignment and affinities with ancillary structures recovered from Lyminge.⁵⁵

Dover re-evaluated: the architecture of the great halls

Since it is the lynchpin of the author’s interpretation, it is appropriate to commence the present reappraisal with the putative church, S14. Before examining the details of the argument, it is worth emphasising at the outset that this building is completely out of character with the masonry apsidal churches associated with early monastic centres in Anglo-Saxon Kent, and bears little association with possible timber expressions of the same ecclesiastical building tradition which have been tentatively identified at certain 7th-century royal residences.⁵⁶ One of the principal arguments proffered in support of S14 being a church is its supposed temporal endurance. Here the author takes his precedent from premier monastic sites such as St Augustine’s (Canterbury), Old Minster (Winchester) and Glastonbury, the monumental cores of which expanded incrementally over many centuries fuelled by the desire to enshrine saintly founders’ tombs and their translated remains.⁵⁷

While there is no doubt that S14 was successively rebuilt on the same footprint, this analogy fails to stand up to critical scrutiny. Firstly, the longevity of each of S14’s structural phases (estimated in round centuries spanning the 7th to the 11th centuries ad) is unrealistically generous. Informed assessments of building life, based on scientific dating and experimental archaeology, suggest that Anglo-Saxon earthfast structures are unlikely to have endured much more than 40 years without major structural repair and rebuilding.⁵⁸ Such estimates should also allow for the fact that the life of early medieval timber buildings was frequently cut short by conflagration (whether deliberate or accidental), and here it may be duly noted that S14 suffered precisely this fate in the third of its identified phases.⁵⁹ With these considerations in mind, an overall duration of 100–120 years for S14 and its various iterations — rather than the proposed 400 — seems inherently more likely. Accepting that its initial construction date lies somewhere in the first half of the 7th century, there is no reason to believe that S14 endured much beyond ad 700, and a terminus within one or two decades either side of this date is entirely in accord with informed expectations.

A second problematic aspect of the structural sequence is the theory that the primary internal cell with its hypothesised eastern altar base was encased, relic-like, within the walls of the much larger post-in-trench building in the process creating lofty flanking aisles. It is very difficult to see how this arrangement would have worked in practical terms, particularly in respect of the structural mechanics of supporting the roof, given the significant variance between the very lightly built walls of the internal cell (constructed of flimsy stakes set in a shallow gulley) and the, by comparison, substantially founded external walls represented by the Period 2 and later wall trenches.⁶⁰ It is admittedly difficult to appraise this aspect of the author’s interpretation given the extreme rarity of aisled buildings at this early period. Yet the instances that do exist, all from Yeavering (Halls A2 and A4), merely serve to cast further doubt on the interpretation by demonstrating that, in contradistinction to Dover, the internal roof supports of such buildings were both substantial and carefully positioned to articulate with other load-bearing components of the superstructure.⁶¹

Revealingly, no stratigraphic proof is presented in the report for the Period 1 cell being primary; indeed, a close inspection of the relevant plans and sections suggests that the relationship may have been very difficult (if not impossible) to discern stratigraphically during excavation.⁶² On the basis of the evidence presented, other scenarios are equally plausible. For example, the internal cell could have been a later insertion rather than a primary core, resulting (as appears to be shown on the relevant plan) in the truncation of the central portion of the opus signinum floor extending across the interior of the Period 3 hall. Alternatively, if the interior traces do indeed represent a contemporaneous element of a larger post-in-trench building, then the character of the evidence is more consistent with a free-standing screened antechamber of some kind, rather than a substantial, structurally enduring core around which the exterior walls were configured.

Serious doubt has been cast over the existence of a Period 1 cell and the idea that it was subsequently encased within the footprint of a much larger aisled building. When this and the other central pillar of the church thesis — the unfeasibly long structural history — are taken out of the equation, then it is possible to interpret S14 in a very different light. Writing before the advent of the recent Lyminge excavations, and relying more on instinct than a systematic appraisal of the evidence, Martin Welch recast the Dover site as a complex of grand halls ‘maintained by a royal official for periodic visits by the king’s itinerant court to his royal castellum in Dover’.⁶³ The idea that the Dover occupation indeed represents the nucleus of a 7th-century villa regius can now be systematically tested through comparison with Lyminge, set against the wider backdrop of excavated great hall complex sites in England.⁶⁴

As we have seen, Lyminge’s 7th-century occupation was focused on a planned arrangement of monumental timber halls of plank-in-trench construction, the two largest of which (A and C) compare very favourably in scale with S14 at Dover. Both sites display examples of halls that were sequentially rebuilt on the same, or overlapping, footprints. The constructional life histories of Dover S14 and Lyminge Hall C offer a particularly striking parallel. One of the peculiarities of these two halls is each phase of rebuilding was accompanied by an alteration in constructional technique: in the former case progressing from single rows of planks (Period 2) to double rows in a staggered disposition (Period 4), through to individual post pits (Period 5); and in the latter case from single planks (with exterior raking posts), through to double planks (set in parallel, not staggered, disposition), culminating in the same constructional shift from continuous foundation trenches to substantial posts set in individual pits.⁶⁵

Comparison between Dover S14 (Period 2) and the third iteration of Lyminge Hall B brings yet another shared constructional idiom into focus: the treatment of the doorposts forming the principal long-wall entrances. In the majority of Anglo-Saxon timber buildings of post-in-trench construction, the principal entrances are marked by an interruption in the wall foundations with the doorjambs being positioned at the terminals of the adjacent lengths of trench, sometimes within deeper pits. In this particular case, however, the doorposts were accommodated within separately defined pits of substantial elongated proportions, in the case of Dover slightly offset from the main wall alignment (Fig 5). While not necessarily unique to Anglo-Saxon Kent,⁶⁶ this distinctive and unusual treatment adds credence to the idea that the Lyminge and Dover halls drew upon the same, locally propagated reservoir of architectural knowledge. From shared idiosyncrasies in the construction of doorways, we finally arrive at the most arresting and distinctive constructional affinity of all: the use, in certain of the identified hall phases, of sophisticated opus signinum flooring. To substantiate this observation and appreciate its full significance, it is necessary to undertake a detailed comparison of the excavated evidence.

Halls of timber, halls of stone: the use of opus signinum flooring

Taken in combination, the evidence presented from the two sites for the employment of this style of flooring is complementary and mutually supporting. Dover is crucial because the flooring was preserved in situ as part of deep urban stratification; whereas at Lyminge, in spite of being recovered in ex situ contexts, there are no underlying Roman structural levels capable of being confused with Anglo-Saxon buildings. The latter point is more pertinent than may first appear, because Welch has previously argued that Philp’s identification of S14 as a church was partly based upon such a stratigraphic conflation.⁶⁷ This view can be, however, firmly rejected on two counts. Firstly, it is very difficult to reconcile with the technical competence of Brian Philp and his highly skilled excavation team who, by the point of the relevant interventions, had grown highly adept at reading the nuances of Dover’s complex urban stratigraphy.⁶⁸ Secondly, and more fundamentally, neither S10 and S14 (contra Welch) were cut directly through Roman buildings: a close inspection of the report reveals that the former was constructed over the tail of the rampart bank of the late Roman-period shore fort formed from a 1.5 m deep accumulation of soil dumped over the walled yard of the earlier Classis Britannica bath-house; similarly, the foundations of the latter were cut through Roman period ‘deposits’, but not in situ masonry buildings (Fig 7).

If the association between opus signinum floors and timber buildings at Dover can indeed be trusted, how closely comparable is the evidence from Lyminge? To answer this question it is necessary to examine the character and constituents of the flooring material represented at the two sites in greater detail. The floor in Dover S10 is described as comprising a base of rolled flints 100–120 mm in size, ‘laid in a compact bed across the building covered in a coarse cream concrete mortar containing small pebbles, varying in depth from 1 to 5 cm’.⁶⁹ The floor in S14 (Period 3) was very similar in overall composition (a basal layer of flints capped by a compact white mortar with small pebbles) and thickness (120–130 mm), although it displayed the notable distinction of having a surface of opus signinum comprising ‘a thin skin of crushed tile and dust, barely 1 mm in thickness, that fused with the white mortar and provided a polished pink-red surface, perhaps in imitation of marble’.⁷⁰ While the flooring in these two buildings thus appears to have been finished somewhat differently, the technical and aesthetic background is clearly one and the same.

At Lyminge the relevant evidence was present in two related forms. The first was consolidated lumps of flooring recycled as packing material in the foundations (trenches and post pits) of stratigraphically superimposed halls, a treatment precisely paralleled at S14 Dover (Tab 1 and Supplementary Material).⁷¹ These pieces were confined to the structural foundations of Halls B and C, the majority deriving from the western doorpost of the former and the partition doorway of the latter (Fig 9).⁷² Sharing the same general distribution was evidence of a more indirect form: concentrations of flint gravel matching the aggregate used in the lime concrete constituent of the flooring.⁷³ Two possible derivations account for this material: either it represents the decomposed raw constituents of further broken up fragments or, alternatively, on the reasonable assumption that the flooring was prepared in close proximity to the halls themselves, a by-product of spillages associated with the initial production process.

fig 9 Distribution of opus signinum flooring from Lyminge. Fragments are denoted by crosses. Illustration by Lyminge Archaeological Project.

The five best-preserved fragments from Lyminge exhibit the same general composition as the Dover S14 floor, namely a basal layer of coarse flints over which has been poured a substantial layer of lime concrete finished with a surface skim (Tab 1). In four of the five cases, the latter was prepared from crushed Roman brick or tile, but the results of detailed analysis indicate that the treatment was not uniform (Supplementary Material). One of the pieces (SF 164) displays a fine surface skim above the crushed terracotta layer with a darker red colouration compositionally related to red ochre (Tab 1 and Supplementary Material for compositional analysis, Fig 10).⁷⁴ Yet further diversity is introduced into the equation in the form of a fifth fragment (SF 682), of more slender proportions bearing a limewashed surface (Fig 11).⁷⁵ It is difficult to draw firm conclusions given the fragmentary, decontextualised nature of the evidence, but these results raise the distinct possibility that the floors concerned were variegated both in colour and texture, perhaps to emphasise and delineate certain parts of the hall’s interior.

fig 10 Fragments of opus signinum flooring from Lyminge. Photographs by Sarah Lambert-Gates. © Lyminge Archaeological Project.

fig 11 Fragment of opus signinum flooring, or possible internal walling, from Lyminge. Photographs by Sarah Lambert-Gates. © Lyminge Archaeological Project.

In terms of extant archaeological survivals, the Kentish halls find their closest affinities in the pseudo opus signinum flooring characterising the earliest generation of churches and monastic buildings from Anglo-Saxon England. These are represented locally at St Augustine’s and Reculver,⁷⁶ and further afield at Wearmouth-Jarrow (Tyne and Wear) and Glastonbury (Somerset).⁷⁷ The dominant type of opus signinum used in Kent (as represented in both great hall and ecclesiastical contexts) conforms to Wearmouth/Jarrow Type 3, where the use of the terracotta constituent is confined to a surface skim rather than being, in the classic Roman manner, more widely distributed through the matrix of the underlying lime concrete.⁷⁸ Taken in conjunction, Dover and Lyminge provide compelling evidence that opus signinum floors formed part of the architectural vocabulary of some 7th-century Kentish great halls serving as royal accommodation. The significance of this observation will be explored in greater depth at a later juncture, but here it is worth briefly considering the implications of this archaeological discovery for Beowulf scholarship, for it does introduce a significant new context for interpreting the lexicography of the poem. The term in question, ‘fagne flor’, used in descriptive evocations of Heorot, has attracted considerable interest as a possible reference to a stone or tessellated floor but, as with many of the elliptical references within the poem, not without considerable uncertainty as to its authenticity.⁷⁹ While it is impossible to know the exact precedents that shaped the poem’s imagery, the Kentish evidence assembled here at the very least suggests that fagne flor could have been inspired by the realities of mead-hall culture (whether contemporary or remembered), rather than representing mere artistic invention.

Dover re-evaluated: the occupation sequence

Having reconsidered the architecture of Dover’s great halls and their interpretive implications, it is now possible to offer a new narrative for the site’s evolution informed by comparisons with the Lyminge sequence. The natural starting point for this is the ‘early Saxon’ activity encountered in the North Area of the site. As we have seen, Philp assigns this activity to his notional monastic phase, in spite of an obvious contradiction posed by the early (ie pre-Christian) dating of some of the sunken-featured buildings (eg N3). A more informed reading would see this 6th-/7th-century occupation as a portion of an early settlement nucleus offering strong analogies with the ancestral focus identified at Lyminge. The parallels extend to a very similar repertoire of settlement components (a sprinkling of sunken-featured buildings in combination with evidence for contemporary ground-level buildings (N6) and an extensive midden deposit (N5), the latter offering a good parallel for the treatment of the solution hollow at Lyminge) and shared cultural signatures reflecting similar levels of material prosperity and specialist activities, such as fine metalworking.⁸⁰

At some point during the 7th century the settlement was reconfigured around a newly constructed complex of great halls, with S10 and S14 (accompanied by its ancillary, S13) disposed in either a co-axial or centralised layout.⁸¹ On analogy with Lyminge, further halls might be expected to lie beyond the excavated area: a site to the north of St Martin-le-Grand being a distinct possibility. The layout of Dover lacks the orthogonal precision and symmetry of Lyminge and other great hall complexes, but here one must take into account the structuring influence of the Roman backdrop, not least the imposing defensive walls which may well have stood to full height during the Anglo-Saxon period.⁸²

Although not a primary concern of the current discussion, the Dover site clearly experienced a significant afterlife, represented by glimpses of late-Saxon occupation within the Roman defences.⁸³ It is very difficult to attribute meaning to such fragmentary remains, but on analogy with the Lyminge sequence, and given the propensity for pre-Viking religious communities to be established within former Roman forts, it is tempting to associate this occupation with the monasticisation of Dover as a formerly royal centre.⁸⁴

Eynsford

The site of Eynsford provides the third of our trio of Kentish great hall sites. Excavated by the Kent Archaeological Rescue Unit and Bromley and West Kent Archaeological Group in two campaigns — again under the direction of Philp — it is the most incompletely sampled of the three sites and for this reason can be considered fairly briefly.⁸⁵

Background context

The site of the great hall occupies a low terrace on the north side of the River Darent, some 200 m north-west of the parish church of St Martin, in an enclosed parcel of meadow within the built-up core of the modern-day settlement. This portion of the Darent valley (some 10 km from the River Thames at Dartford) features one of the densest concentrations of Romano-British settlement anywhere in Kent, including several opulent villa complexes, the most well-known being the site of Lullingstone, located some 1 km to the south-west of Eynsford (Fig 2). The same tract of landscape also holds an impressive cluster of early Anglo-Saxon settlements and cemeteries, a significant proportion of the former being sited on, or in close proximity to villa sites.⁸⁶ Eynsford exemplifies this continuity for the Anglo-Saxon occupation recently identified within the village lies a short distance away from a substantial Romano-British mortared stone building, very likely to represent part of a more extensive villa complex.⁸⁷

Eynsford is located beyond the western frontier of the historic heartland of the Anglo-Saxon kingdom of Kent, marked by the River Medway, and for this reason is poorly documented in comparison to Lyminge and Dover. Notwithstanding this lacunae in historical sources, Alan Everitt has advanced the plausible theory that Eynsford formed one of a series of early river estates strung along the axis of the Darent valley, which had coalesced into a royally administered core territory, or regio, before the 7th century.⁸⁸ Eynsford’s trajectory cannot be traced with the same level of confidence as Darenth, the eponymous caput of the Darent regio, which remained in the king’s hands until the 10th century, but Everitt argues that it may have originally held jurisdiction over the neighbouring parishes of Farmingham, Lullingstone and parts of Chelsfield (Fig 2).

Occupation sequence

The Anglo-Saxon occupation described in the Eynsford report was initially encountered during trial trenching in 1991, prompting the excavation of a 12 m by 6 m window in the southern sector of the site. This uncovered a single sunken-featured building, together with the partial ground plan of one, or possibly more, ground-level earthfast buildings represented by external alignments of post holes. These structural components were accompanied by an intact occupation horizon (L-4), rich in domestic midden material, sealed beneath a metre of topsoil and overburden. A 6th-/7th-century date can be ascribed to this occupational focus on the basis of the sizeable assemblage of cultural material recovered which included over 500 sherds of Anglo-Saxon pottery, annular loomweights and other weaving equipment.⁸⁹

Excavated between 2009 and 2012, the site of the great hall was identified some 50 m north-east of the occupation just described. Rather than being exposed in full, the structure was targeted by a series of small trenches on a piecemeal basis, meaning that certain parts of the published ground plan remain conjectural (Fig 12). Notwithstanding the highly fragmentary nature of the excavations, a reasonably coherent impression of the hall has emerged supported by, in places, very revealing structural detail. The building was laid out on the NE/SW axis and measured 20 m by 10 m with an internal partition set 3 m inside the north-eastern end wall. The external walls of the building were represented by continuous foundation trenches measuring some 500–600 mm wide and 600 mm deep, each containing the impressions of rectangular posts set in a single row towards the exterior edge of the trench, and flanked by alignments of regularly spaced post holes representing external raking timbers of sturdy proportions.

fig 12 Plan of great hall from Eynsford. Redrawn from Philp 2014, fig 5.

The building appears to have been provided with two opposing pairs of external doorways, those on the long axis of the hall being aligned with a further doorway pierced through the internal partition. In each case the doorjambs comprised adjoining planks arranged in an L-shaped configuration set in substantial pits packed with flint and gravel; the widths of the five doorways was relatively consistent ranging from 1.06 m to 1.15 m. Additional structural detail was provided by copious quantities of walling material recovered from the hall’s foundations, described in the report as a form of ‘plaster’ comprising ‘a weak white mortar, perhaps in at least two different fabrics, sometimes showing a thin external added layer and sometimes a painted white surface’.⁹⁰ The foundations of the hall produced a scarcity of datable artefacts, the only items of note being two fragments of vessel glass, one amber in colour and the other pale green with applied ribbing, which could as easily derive from the adjacent 6th-/7th-century occupation as from that of the hall itself.

As a postscript it may be noted that, as far as can be gleaned from the results of these excavations and other strands of evidence, the settlement appears to have shifted to a new site following the abandonment of the Eynsford hall. As at Lyminge, the use of the hall complex presages a major episode of restructuring in the location and spatial configuration of the settlement.

DISCUSSION

The foregoing analysis has situated Lyminge, Dover and Eynsford firmly within the Anglo-Saxon great hall complex tradition and, in the process, identified a repertoire of family traits embodied by the three sites as a distinct regional grouping. The following interprets the meaning and significance of these patterns, both in the local context and for a wider understanding of the great hall complex phenomenon, through three complementary thematic prisms: topographic context; temporality and social memory; and monumental hybridity.

at the water’s edge: the topographic context of royal residence

One of the most obvious similarities displayed by the three sites is their proximity to rivers, the distances involved being as little as 50 m for Lyminge, 80 m for Eynsford and about 100 m for Dover (Fig 2). The intimacy of this association is very much underlined by the fact the sites concerned occupy locally prominent settings within valley bottoms — river terraces and minor spurs — rather than the much more visually dominant and defensible positions offered by neighbouring ridgetops and promontories. In each of the three cases the watercourse concerned would have been directly visible from its adjacent hall complex and doubtless exerted a powerful influence on how the site was perceived and experienced.⁹¹

It is worth pausing to reflect on the significance of this correspondence in light of the distinctive role of the sites concerned as places of royal residence. Part of the attraction of a riverine locale must have lain in a ready supply of fresh water — an essential requirement for the gatherings of humans and animals that characterised the life at royal vills as places of periodic assembly — and ease of transportation. Yet it would be wrong to explain the association in such narrowly deterministic terms. As prime arenas for promulgating royal ideology, places of pre-Viking royal residence were contrived to harness sources of sacral and supernatural authority embedded in particular landscape settings. Rivers held a special place in Anglo-Saxon communities’ perceptions of, and interactions with, the numinous world, particularly as a traditional setting for the ritual-laden sphere of public assembly.⁹² There is good reason to believe that this association was taken to theatrical extremes in the context of the earliest sites of royal residence, as illustrated by the 36-day mass baptism orchestrated by Bishop Paulinus in the River Glen at the villa regia of Yeavering. ⁹³ This tantalising reference, combined with the trajectory of royal vills such as Lyminge which were subsequently ‘monasticised’, serves as an important reminder that contemporary sites of kingly residence were deeply entangled in narratives of conversion and the contingent process of resacralising the landscape. While difficult to visualise directly through the archaeological record, the ritual affordances of a riverside setting can only have enhanced pre-Viking royal residences’ efficacy as a prime locus for negotiating the process of Christianisation.⁹⁴

the social memory of royal landscapes: hall building as a genealogical strategy

The antecedent life of the three great hall complex sites can also be shown to be crucial in structuring their river-valley locations. Indeed, the appropriation of pre-existing settlements and the ancestral authority which such appropriation conferred, arguably constitutes the overriding factor in determining their placement within the landscape. All three sites occur within river valleys which, as can be adduced from a convergence of archaeological and place-name evidence, constitute some of the earliest concentrations of recognisably ‘Anglo-Saxon’ occupation in Kent. These ‘valley estates’, to give them the name adopted by Everitt, feature an unusually high density of 5th-/7th-century sites, displaying a characteristic spatial disposition with settlements typically confined to the floor of the valley and overlooked by cemeteries occupying adjacent flanks at fairly regular intervals of 1–3 km.⁹⁵

Tom Williamson has argued that river territories of the kind exemplified by the Darent, Dour and Nailbourne in Kent were highly significant in structuring patterns of contact and connectivity between early medieval communities and, as such, can be regarded as essential building blocks in the formation of social identities in early Anglo-Saxon England.⁹⁶ This national perspective can be grounded in insights drawn from localised examinations of Anglo-Saxon mortuary landscapes in southern England, including in Kent itself. Research conducted at this detailed analytical scale has elucidated the important role played by cemeteries in the visual, mnemonic and symbolic articulation of these territorial ‘heartlands’, taking advantage of the fact that they were traversed by major routeways responsible for channelling movement through the landscape.⁹⁷

It is thus clear that the three Kentish great hall sites were established within what might be thought of as core territories or ‘heartlands’, representing core zones of early medieval habitation within their respective localities. This relationship takes on heightened significance when we examine the spatial and temporal trajectory of the sites concerned at a more fine-grained level. While Eynsford is more equivocal, the great hall complexes at Lyminge and Dover were both superimposed on pre-existing settlements that were long established by the turn of the 7th century. Moreover, both sites demonstrate clear evidence for a central-place role in the 6th century, adducible archaeologically by shared evidence for conspicuous consumption, social display, and skilled crafting.

The pattern of appropriation seen in the group of Kentish sites, no less apparent in the lengthy sequence of high-status activity recently brought to light at Rendlesham (Suffolk), has importance for a wider understanding of the mnemonic strategies focused on places of royal residence in conversion-period England.⁹⁸ Understandably, previous commentators have been attracted by the very explicit cases of prehistoric monument reuse presented by the corpus of English great hall sites taken to the theatrical extreme at Yeavering, where the axes of the main hall array were fixed in relation to a pair of Bronze-Age funerary monuments.⁹⁹ Yet while the manipulation of the distant past undoubtedly represents an important characteristic of these sites, however, it has arguably drawn attention away from temporal interplay between great hall complexes and their immediate antecedent phases. The monumental core of Yeavering evinces this interplay very clearly in sequences of hall construction initiated by, and focused on, lightly built structures dating back to the embryonic (6th-century) phase of the settlement, and the same tendency finds expression in the excavated sequence from Cowdery’s Down.¹⁰⁰ Granted, the Kentish evidence is more consistent with general locational rather than precise structural continuity, but the underlying motivation and rationale is very likely to have been the same: monumentalising places of dynastic importance as a strategy for legitimating and reifying royal authority.¹⁰¹

One has to turn to the climactic 7th-century phases of the Kentish sites to find more emphatic reflections on the theme of sequential rebuilding: in the case of Dover, S14 was rebuilt on at least three, and quite possibly more, occasions on the same site and S10 at least twice, whereas at Lyminge Halls B and C passed through three iterations on the same or on narrowly diverging footprints (Fig 4). In the case of the latter, there is clear evidence that these cycles of reconstruction involved structural as well as locational continuity, reflected in the fact that foundation trenches were sometimes reused from one phase to the next.¹⁰² Helena Hamerow has previously drawn attention to the significance of this general practice, remarking that:

it is thus at [these sites] that we first see an interest in extending the longevity of important buildings’, as means to ‘embody and evoke links with ancestors’, further observing that ‘it was entirely possible that there was a connection between this desire to create long-term relationships with a place and the growing importance of landholding and inheritance.¹⁰³

This perspective can be developed further by considering how the biographies of the halls themselves, embodied in cycles of construction, use and destruction, may have evoked notions of permanence, ancestry and dynastic memory. As is well known, Kent enjoys a richer survival of origin legends and royal genealogies than any other Anglo-Saxon kingdom, and for this reason it offers a very instructive context for situating great hall sites within a wider body of mnemonic traditions which rapidly proliferated in conversion-period England.¹⁰⁴ As with their peers in other regions of Anglo-Saxon England, the Kentish royal house displayed an intense interest in the art of dynastic manipulation. At an early stage in what was likely a drawn-out process of elaboration, ‘Oiscingas’ had emerged as the dynastic family name for rulers of Kent.¹⁰⁵ This name was derived from the eponymous figure, Oisc, one of a cast of mythic characters through which the kings of Kent traced ultimate descent and legitimacy from Woden — progenitor par excellence for Anglo-Saxon royal houses.¹⁰⁶ Moreover, genealogical fabrication of this type was not only confined to the male members of the Kentish royal line. As research on the body of related texts known as the ‘Mildreth Legend’ has shown, dynastic memory, and the authority this conferred over property and landed resources, was also transmitted through origin myths woven around female members of the Kentish royal house in their capacity as abbesses and proprietors of substantial monastic endowments.¹⁰⁷

Returning to the archaeological evidence itself, the idea that the cyclical rhythm of hall building witnessed at excavated sites of royal residence may have been synchronised with the ebb and flow of dynastic power is hardly novel: such a conception forms the crux of Brian Hope-Taylor’s (ultimately flawed) chronological interpretation of Yeavering as the villa regia for a documented succession of Bernician kings.¹⁰⁸ Beowulf of course provides a famous and highly suggestive literary precedent wherein the construction of Heorot serves to inaugurate Hrothgar’s ascendancy to the Danish throne as the supreme personification of his kingship, power and right to rule.¹⁰⁹ Similar conceptions, drawing upon a wider body of literary allusions supplied by Icelandic sagas, have also been applied to the interpretation of Scandinavian magnate residences of the migration and Viking periods, characterised by lengthy sequences of hall reconstruction in some cases spanning hundreds of years.¹¹⁰ In the case of this particular sphere, the foundation deposits commonly discovered on hall sites provide added witness to the significance of hall construction as an intensely political act immersed in the rituals of inauguration and sacral legitimation.¹¹¹

More than simply broadcasting new rulership, however, the sequential rebuilding characteristic of the Kentish and other English great hall sites suggests that there was a strong retrospective and commemorative element to hall building — a conscious attempt to proclaim genealogical links with the past through the manipulation of monumental space.¹¹² This behaviour bears comparison with Howard Williams reading of the elite funerary landscape of Sutton Hoo as a monumental genealogy where ‘a chain of monumental episodes’ was composed to ‘create social memories and relations between past and present’ in relation to origin myths and genealogies.¹¹³ The clear difference in the examined context is that genealogical memory was transmitted through cycles of (re-)construction in a single locale rather than through a series of spatially discrete monuments which referenced each other as enduring elements of the landscape. Here the destruction and/or disassembly of halls and its natural corollary — the reuse and re-incorporation of building materials from earlier phases into new phases of construction — can be conceived as an active part of the embodied and communally enacted practices through which dynastic memories were recalled and transmitted.¹¹⁴

A final dimension on the mnemonics of such sites is provided by ritual deposits. Although often treated as a self-contained category in examinations of the great hall complex phenomenon, this aspect deserves to be considered as an integral part of creating memory and genealogy, since they can often be shown to be biographically entwined with the construction, use and abandonment of halls.¹¹⁵ As Hamerow and more recently Clifford Sofield have shown, English great hall sites manifest such practices more subtly and rather differently than early medieval magnate residences in Scandinavia, which attracted a rich panoply of ritual deposits, especially, as we have seen previously, in connection with the inauguration of halls.¹¹⁶ As will be shown presently, Kent very much conforms to the English picture and the relevant signals need to be interpreted at this level of sensitivity.

At Lyminge, three items of metalwork classifiable as intentional closure deposits — an annular brooch, dress-pin and an elaborate horse-harness fitting — were recovered from halls spanning both the antecedent and climatic phases of the complex.¹¹⁷ This complements a rather more spectacular case of ritual closure in the form of a plough coulter concealed in the abandoned shell of a 7th-century sunken-featured building associated with a peripheral portion of the settlement.¹¹⁸ Dover presents further compelling evidence for such behaviour, most notably a 7th-century glass bell beaker recovered from the Period 2 foundations of hall S14.¹¹⁹ Quite remarkably given its fragility, this vessel was recovered substantially intact as if it had been recovered from a grave, which speaks eloquently of the care and reverence invested in its concealment.

A useful way of pulling the various strands of the foregoing discussion together is by thinking of great hall complexes as nexuses where genealogical memory was manufactured through a convergence of mutually reinforcing practices. Seen in this light, these sites can be argued to embody an interface between the internal theatre of the mead-hall on the one hand — stage-sets where the highly stratified origin myths portrayed in royal genealogies were literally acted out as choreographed showpieces¹²⁰ — and the monumental affordances of the structures themselves which provided a powerful vehicle for evoking and manipulating genealogical memory.

hybridity and the genesis of a monumental tradition

This concluding thematic excursus begins with the premise that the great hall complex phenomenon needs to be understood and interpreted in the context of its own time (ie the 7th century), and accordingly that previous attempts to understand its significance in terms of the fusion of ethnic traits are flawed and limiting.¹²¹ As previous studies have shown, there were fundamental differences between the age of the great hall complex and the preceding two centuries in relation to the dynamics of the ethnogenesis process, with the 7th–8th centuries recognised as a period of accelerating acculturation associated with new, highly politicised, expressions of ethnic affiliation.¹²²

Attention instead needs to be shifted towards understanding the phenomenon as the product of particular, historically contingent circumstances which shaped how elite groups in 7th-century England chose to articulate and represent their power. Progress towards this aim can be made by engaging with recent studies which have explored the distinctive material strategies employed in the assertion of 7th-century kingship and elite authority. A unifying theme brought to attention by this work is the highly assimilative character of these strategies: the inventive fusion of traditional practices with external influences to generate new ways of articulating power that had meaning and resonance in local contexts — a process which was accelerated and given intensified expression by the conversion of the Anglo-Saxon aristocracy.¹²³ It is very much in this vein that royal hall culture in Kent needs to be explained and interpreted.

The study speaks directly to this theme by divulging how great hall architecture in Kent followed a distinct, classicising path under the radiating influence of Canterbury as the epicentre of the Augustinian mission. In the case of Lyminge this influence can be detected in the site’s precise grid-planning, a Romanising idiom first deployed in England in relation to the early churches at St Augustine’s, Canterbury.¹²⁴ However, it is in the distinctive vocabulary of architectural flourishes displayed by the trio of Kentish sites that this theme finds unified expression. In their different ways the great halls of Kent embody a reverence for Romanitas as a key plank in the legitimating ideology of the newly converted Kentish court: Eynsford with its white plaster cladding, and Dover and Lyminge with their elaborate opus signinum floors technologically identical to the flooring used in near contemporary ecclesiastical buildings. Of course, what makes these latter constructions particularly arresting as embodiments of the assimilative character of 7th-century hall culture, is that that they cut directly across the timber/stone dichotomy lying at the heart of Anglo-Saxon architectural studies.¹²⁵

This is not the first time that a dialogue between timber great hall complexes and early masonry churches has been observed. The similarities between Yeavering and ecclesiastical buildings in Northumbria have attracted repeated attention over the years, especially with regards to dimensions and axial plan-forms, although opinion is divided on the extent to which the former may have served as direct models for the latter.¹²⁶ While an interplay of sorts is unquestionably embodied in the monumental landscape of Northumbria, what we see here is more akin to translation or transposition from one context to another, a process eloquently expressed in Eric Fernie’s assessment of Escomb church (Co Durham) as a ‘bridge between two very different traditions, using the layout of one, the Germanic, and the masonry techniques of the other, the Roman’.¹²⁷

By contrast, the great halls of east Kent featuring opus signinum floors can be regarded as genuine architectural hybridisations which in their novelty and ingenuity are every bit as redolent of the ‘powerfully creative interaction of Mediterranean and indigenous traditions’ as are the manuscript art and stone sculpture of conversion-period England.¹²⁸ This insight can be developed further by thinking through the likely mechanisms that lay behind the appearance of these hybrid constructions. A natural question which emerges in this context is the extent to which the floors in question were directly inspired by Roman buildings given that Kent enjoyed a particularly strong Imperial inheritance.¹²⁹ While there is no reason to believe that these relic monuments exercised less of an influence on the cultural practices and imagination of early medieval communities of Kent than those in other parts of Anglo-Saxon England, indeed, quite the contrary,¹³⁰ it is important to remember that the construction of such a floor would have depended upon the mastery of a spectrum of techniques executed in a precise sequence of stages constituting a complex chaîne opératoire.¹³¹ The evidence presented by these halls fits much better in the context of a contemporary revival of antique traditions achieved through and in relation to the imported technologies of the Roman Church than local experiments in aping the Roman past. An acceptance of this argument leads to an arresting implication: that the Frankish and Italian stone masons enlisted to build the earliest churches in Kent were also deployed in the aggrandisement of contemporary royal accommodation.¹³²

CONCLUSION

By harnessing the collective interpretive potential offered by the great hall complexes of Lyminge, Dover and Eynsford, this article has opened a new archaeological perspective on royal hegemony in 7th-century Kent, and in the process enriched its status as a fundamental context for the study of kingship in early medieval England. The insights generated by this comparative study demonstrate that great hall complexes played a key role in the assertion of royal authority during the period of independent Kentish supremacy, and that the monumental idiom was adapted and given fresh expression under the influence of the Romanising programme of the Augustinian mission.

Integration of the rich material perspectives offered by the trio of Kentish sites with contextual sources and theoretical approaches has revealed new insights into how great hall complexes operated as arenas for asserting power, lineage, and authority in an age of rapidly accelerating political centralisation. Their efficacy as ‘theatres of power’ was grounded in the strategic exploitation of meaningful and highly resonant settings that held deep ancestral and dynastic significance, while at the same time fulfilling the practical and ritual requirements of royal assembly. The distinctive temporal qualities and rhythms of these sites, combining relatively transient and episodic monumental life histories on the one hand with longer-term and more persistent ‘central-place’ trajectories on the other, have been argued to offer a crucial perspective on pre-Viking royal landscapes as reservoirs of social memory. As has been argued in relation to contemporary elite mortuary landscapes, these places were deeply implicated in the genealogical practices of 7th-century royal dynasties, offering multiple, mutually re-inforcing outlets for channelling dynastic memory. These were mediated through the interior theatre of the mead-hall as a prime locus for the creation and manipulation of origin myths and through sustained, cyclical programmes of monumental investment on an unprecedentedly lavish scale.

This article began by critiquing the employment of the label ‘great hall complex’ as a homogenising category. Evaluation of the Kentish evidence against the wider backdrop of cognate settlements is an important step towards replacing this monolithic category with a more complex and contextually nuanced portrayal. While displaying some internal variations and divergences, the great halls of Kent share a distinctive repertoire of classicising idioms, the evaluation of which has shed light on the mechanisms behind the local genesis of this monumental idiom. This tendency is most marked in eastern Kent, the historic heartland of the kingdom, where the monumentalisation of royal vills was taken to new heights in the construction of elaborate timber feasting halls featuring opus signinum flooring of a type reproduced by contemporary churches, and likely executed by the hands of the same foreign ateliers. These hybrid constructions stand as eloquent witness to the creative ingenuity of elite monumental strategies in conversion-period England and also make the wider point (all too often lost in archaeological debates) that the boundary between secular and religious places and spaces during this period was fluid and permeable.¹³³

Table 1 Fragments of opus signinum/concrete flooring from Lyminge. See Supplementary Material for the results of detailed compositional analysis.

Context	Dimensions	Morphological description	Surface treatment
LYM 14; SF 164	Th 120 mm; W 130 mm; D 105 mm	The fragment preserves the original thickness of the floor and displays vertical exterior edge with a projecting lip at the base where the floor has been pressed against the flat surface of a vertical timber.	The surface skim is comprised of two discrete layers. The lower is orange-red and consists of tiny chips of brick/tile set in a sparse, effervescent lime-mortar. The upper layer is dark purplish red and composed of partly crushed and powdered (?) red ochre/burnt iron ore set in a whitish, effervescent lime mortar.
LYM 13; SF 773	Th 110 mm; W 105 mm; D 82 mm.	A fragment of concrete adhering to a large basal flint derived from the bedding layer of the floor.	A thick (2–3 mm), orange-red, uneven coating of coarsely crushed brick/tile within a matrix of dust.
LYM 13; SF 389	Th 75 mm; W 155 mm; D 85 mm.	Fragment comprising a layer of concrete the underside of which bears ghost indentations from a basal layer of flints.	A thick (2–3 mm), orange-red, uneven coating of coarsely crushed brick/tile.
LYM 13; SF 540	Th 60 mm; W 115 mm; D 75 mm.	An irregular fragment comprising the mortar constituent with its surface skim.	A thick (2–3 mm), uneven coating of coarsely çrushed, orange-red brick/tile.
LYM 13, SF 682	Th 25 mm; 135 mm; D 120 mm	This fragment is substantially thinner than those above, but the regularity of the lower surface argues against the hypothesis that this has sheared away from a thicker fragment. This must imply a different, less robust, style of stone flooring or alternatively part of a wall feature such as a niche. While of the same general composition as the above, the matrix of the lime concrete in this case has small chips of Roman brick/tile running through it.	A thin (c 1 mm) layer of lime-mortar parts of which are worn smooth.

Résumé

Les « Mead-Halls » des Oiscingas: nouvelle perspective venant du Kent sur le phénomène du complexe du Grand Hall anglo-saxon par Gabor Thomas

Le complexe du Great Hall, l’une des expressions les plus distinctives et évocatrices des traces d’établissements anglo-saxons, sert souvent de métaphore pour l’émergence de la royauté dans l’Angleterre du début du Moyen-Âge. Or, l’interprétation de ces sites n’a, d’une part, pas été suffisamment développée et est, d’autre part, fortement influencée par les fouilles du site bien connu de Yeavering, dans le Northumberland. Inspiré par les résultats de fouilles récentes entreprises à Lyminge, ce papier s’interroge en procédant à la comparaison détaillée de trois complexes de Great Hall dans le Kent ; cette nouvelle perspective régionale permet de faire avancer notre connaissance de l’agence et de la signification inhérente de ces établissements en tant que « théâtres du pouvoir ». Exploré à travers les prismes thématiques du lieu, de la mémoire sociale et de l’hybridité monumentale, cet examen conduit à une nouvelle appréciation du rôle des sites de Grand Hall dans les stratégies généalogiques des dynasties royales au 7^e siècle ; il apporte une nouvelle perspective sur la façon dont ce langage monumental remarquable, mais de courte durée, a été adapté pour tirer parti de la capitale symbolique de la « romanité ».

Zusammenfassung

Methallen der Oiscingas: Neue kentische Perspektive auf das Phänomen angel sächsischer Großhallen-Komplexe von Gabor Thomas

Der Großhallen-Komplex ist eine der charakteristischsten und plastischsten Ausdrucksformen angelsächsischer Siedlungsgeschichte und wird weithin als Metapher für das Aufkommen des Königtums im frühmittelalterlichen England zitiert. Doch die Interpretation dieser Stätten ist noch nicht hinreichend entwickelt und legt großes Gewicht auf die Ausgrabungsfunde an der bekannten Fundstätte in Yeavering in Northumberland. Inspiriert von den Ergebnissen neuerer Ausgrabungen in Lyminge, widmet sich dieser Artikel einer detaillierten vergleichenden Untersuchung dreier Großhallen-Komplexe in Kent und nutzt diese neue regionale Perspektive, um unser Verständnis für die Funktion und die darin verkörperte Bedeutung dieser Siedlungen als „Theater der Macht“ weiterzuentwickeln. Diese Untersuchung betrachtet die Fundstätten durch die thematischen Prismen „Ort“, „gesellschaftliches Gedächtnis“ und „bauliche Hybridität“ und führt zu einer neuen Wertung der Stätten von Großhallen in den genealogischen Strategien der Königsdynastien des 7. Jahrhunderts und einer frischen Perspektive darauf, wie dieses bemerkenswerte und doch kurzlebig Bau-Idiom so angepasst wurde, dass es das symbolische Kapital der Romanitas nutzbar machte.

Riassunto

Le sale dell’idromele degli Oiscingas: dal Kent una nuova prospettiva sul fenomeno del great hall complex anglo sassone di Gabor Thomas

Il great hall complex rappresenta una delle espressioni più caratteristiche ed evocative nella documentazione dello stanziamento anglosassone e viene largamente citato come metafora della comparsa del potere sovrano nell’Inghilterra altomedievale. Tuttavia l’interpretazione di questi siti resta insufficientemente sviluppata ed è pesantemente influenzata dai ritrovamenti degli scavi del ben noto sito di Yeavering nel Northumberland. Ispirato dai risultati dei recenti scavi a Lyminge, questo saggio intraprende un’investigazione comparativa dettagliata di tre great hall complex nel Kent e sfrutta questa nuova prospettiva regionale per promuovere la nostra comprensione riguardo all’azione e ai significati intrinseci di questi stanziamenti quali “teatri del potere”. L’esplorazione, condotta attraverso i prismi tematici di località, memoria sociale e ibridità dei monumenti, porta a una nuova consapevolezza della partecipazione alle strategie genealogiche delle dinastie reali del VII secolo da parte dei siti che avevano una great hall e offre una nuova prospettiva sul modo in cui questo idioma monumentale, notevole ma di breve durata, sia stato adottato per imbrigliare la capitale simbolica della Romanitas.

Acknowledgments

The Tayne Field excavations could not have happened without the generous permission of Lyminge Parish Council who have been a continual source of support and encouragement throughout the decade-long existence of the Lyminge Project. Of the many individuals who assisted in the 2012–15 excavation campaign, I would particularly like to thank my co-director, Dr Alexandra Knox, and the following members of the supervisory team: Dr Matthew Austin, Rosie Cummings, Helen Harrington, Emily Harwood, Zoe Knapp, Andy Macintosh, Dr Simon Maslin, Roo Mitcheson and Keith Parfitt. I am deeply indebted to Professor John Allen and Sara Wilson for accepting my invitation to undertake the analysis of the Lyminge flooring material and for writing up the results so expeditiously. Thanks also to Brian Philp for kindly clarifying details relating to the archaeological discoveries made at Eynsford. Dr Matthew Austin, Sarah Lambert-Gates, and Dr Simon Maslin generously provided assistance with the production of illustrative material. I should also like to express my gratitude to Professor John Blair, Professor Roberta Gilchrist and Professor Helena Hamerow for providing very helpful critical feedback on previous versions of the text. This paper is published with financial assistance from the Arts and Humanities Research Council under the awards AH/J006068/1 and AH/N00218/1.

Notes

2 A number of labels have been applied to this class of settlement in the past, including ‘palace complex’, but this examination adopts the rather more neutral term ‘great hall complex’ as used in recent syntheses such as Hamerow 2012, 102–109. Some scholars (eg Sawyer 1983, 274; Welch 1993, 50) have expressed reticence in equating all such settlements with royal status. While we should indeed avoid making this connection automatically, the evidence assembled from Kent very much supports the view that the great hall complex was primarily exploited by, for, and under the direct aegis of, kings and their royal kindred.

3 Cramp 1957; 1993; Hills 1997; Bazlemans 1999; Webster 2002; Alcock 2003, 234–54; Pollington 2003; Rollason 2009.

4 Herschend 1998, 14–51.

5 Jones 1976; Sawyer 1983; Campbell 1986; Charles-Edwards 1989; Yorke 1990, 7–9, 162–7; Blair 2005, 275–9; Roach 2011.

6 James et al 1984; Marshall and Marshall 1991; 1993.

7 See Hamerow 2012, 18–22.

8 Rollason (2016) provides a penetrating investigation of this theme across a broad chronological and geographical canvas. The ideological and performative aspects of royal residence have been explored in a number of parallel early medieval contexts, including Carolingian Frankia (Samson 1994), early medieval Ireland (Gleeson 2012), northern Scotland (Gondek and Noble 2010) and Scandinavia (eg Hedeager 2001). Within an English context, Matthew Johnson’s research on castles and other classes of later medieval seigneurial architecture (2002; 2007) offers the most theoretically developed treatment of elite settlements as theatrical stage-sets for enacting the rituals of power.

9 Reynolds 2003; Ware 2005; Frodsham 1999; Walker 2010; 2011; Hamerow 2012, 102–9; Semple 2013, 97, 207–12; Rollason 2016, 36–7.

10 Herschend 1998, 14–51; Hamerow 2010; Walker 2010.

11 Hope-Taylor 1977; Frodsham and O’Brien 2005.

12 The most recent survey (Blair 2018) places the current total at 18, only five of which have seen more than keyhole excavation.

13 Giles 2014, 15. See Beck (2014, 127–30) for a consideration of the homogenising influence that taxonomic approaches have had on the study of Scandinavia hall culture. For the wider historiographical context see Gilchrist 2009 and McClain 2012.

14 Reynolds 2009.

15 Williams 2016.

16 Gilchrist 2008; Maldonado 2013.

17 Hall 2001; Thomas 2012.

18 Thomas 2013; 2017.

19 Hamerow et al 2007; Hamerow and Brennan 2015.

20 Scull et al 2016.

21 McBride 2016.

22 For wider discussion of this theme see essays in Dickinson and Griffiths 1999; Webster 2000; Maddicott 2003; Reynolds 2003; 2005; Carver 2011; Scull 2011; Loveluck 2013, 76–150; and Blair 2018.

23 Philp 2003.

24 Brooks 1989; Yorke 1990, 25–44.

25 Kelly 2006; Thomas 2013.

26 For an extended examination of Lyminge’s landscape context, see Austin 2017.

27 Kelly 2006; Brooks and Kelly 2013, 28–35.

28 Kelly 1995, no 8, 33–6.

29 Brooks and Kelly 2013, no 5, 286–93; see also Brooks 1984, 175–90; Yorke 2017.

30 Brooks and Kelly 2013, 31–2

31 Ibid, 34–5.

32 Richardson 2005 vol 2, 48–9; Thomas 2013.

33 Everitt 1986, 77; Brooks 1989; Brookes 2011.

34 The following is based upon Thomas 2013; 2017. The underlying site archive can be interrogated online via the Lyminge Project website, <www.lymingearchaeology.org> [accessed 16/08/18]

35 One of the sunken featured buildings truncated by the great hall complex yielded artefacts diagnostic of the 7th century (ie amethyst beads) suggesting that the transition may have been very swift indeed.

36 Reynolds 2003; Hamerow 2012.

37 Blair 2014; 2018; Blair et al forthcoming, ch 5.

38 Hope-Taylor 1977; Millett and James 1983; James et al 1984.

39 The date of a foundation of a royal nunnery at Lyminge is traditionally ascribed to the year ad 633 under the auspices of the Kentish queen, Æthelburga, but scholarly consensus would now place this event in the final third of the 7th century (Kelly 2006; Yorke 2017).

40 Thomas 2017.

41 Philp 2003.

42 Philp 1981.

43 Philp 1989.

44 Parfitt and Anderson 2012, 4.

45 Ibid.

46 Brooks 1984, 183, 193; Brooks and Kelly 2013, no 8, 303–13. Much ink has been spilt over the origins of a monastic community at Dover provoked by the (spurious) claim that King Wihtred refounded a pre-existing religious community established during the reign of King Eadbald c ad 630, in the process causing the site to be moved from ‘the castle’ into the town (Rigold 1977, 73). Few, if any, historians would accept the authenticity of this tradition.

47 Tatton-Brown 1984.

48 Ibid; Baker and Brookes 2013, 353–6.

49 Philp 2003.

50 Ibid, 124.

51 The altar base is highly dubious. The report describes its principal identifier as ‘a mortar base, originally rectangular, constructed of small pebbles set in white mortar’ (Philp 2003, 59). Much mutilated by the later Norman foundations of St Martin-le-Grand, an eye of faith is required to conjure this small irregular patch into the rectangle described; moreover, it is uncannily similar in composition to the concrete floor assigned to the Period 4 building so could alternatively be a truncated remnant of the same surface.

52 Dominated by residual Roman material, there was very little in the way of direct dating evidence from the foundations of the hall. In light of this it is somewhat surprising that more was not made of the complete glass vessel recovered from the Period 2 hall which is diagnostically mid-/late 7th century and must have been fairly fresh when deposited in the ground (Philp 2003, 121, fig 64, no 187; Evison 2008, 13, fig 9).

53 Philp 2003, 47.

54 Horizontal wall plates are alien to the constructional repertoire of great hall complexes. Structure C12 at Cowdery’s Down also had a gulley that ran along the inner edge of the foundation trench, but this was inside the main alignment of vertical wall timbers demonstrating that it was non-structural (Millett and James 1983, 217, fig 57).

55 Lyminge provides evidence for smaller, more lightly built, ancillary structures set on axial alignment with and almost adjoining the major halls, although here they shared post-in-trench construction (Thomas 2017, 109, fig 10).

56 Shapland (2015a, 497–500, 511–12) provides a general survey of putative churches on sites of great hall complexes.

57 Biddle 1986; Gittos 2013, 90–4; Shapland 2015b, 100–5.

58 Loveluck 2007, 50; Darrah 2007, 54–5; Hamerow 2012, 34–5. The longevity of the distinctive plank-walled construction represented by S14 and its analogues could have been somewhat shorter than estimates based on more traditional wall-posts formed from oak roundwood.

59 Philp 2003, 127.

60 The sequence of construction embodied by the consecutive phases of building D2 at Yeavering does, however, demonstrate that halls were sometimes purposefully constructed to encase earlier buildings (Hope-Taylor 1977, 97–8, 158, 277–8).

61 Ibid, 125–34.

62 Philp 2003, figs 34 and 35B, 61–2, 66.

63 Welch 2007, 203.

64 While a non-ecclesiastical interpretation is preferred for S14 and its allied occupation, the Roman shore fort still remains the most convincing location for the documented monastic community of St Martin’s: this is a classic setting for a pre-Viking monastery and its former presence is further suggested by the existence of a ‘church-group’ in this part of the town (Blair 2005; 191–204).

65 Thomas 2017, 108.

66 A similar approach to door construction is featured by a putative timber church (Building 7098) at Brandon, Suffolk (Tester et al 2014, 48, 51, fig 4.14).

67 Welch 2007, 202–3.

68 Keith Parfitt pers comm.

69 Philp 2003, 46.

70 Ibid, 65, 127.

71 Ibid, 127.

72 A natural question which emerges from the relatively restricted volume and distribution of the material, is whether the halls were only partially floored in this way. While this is certainly a possibility, the evidence from Dover S14 indicates that much, if not all, of the interior of this comparable hall carried stone flooring. The fundamental point about Lyminge is that the survival of these fragments is entirely contingent upon the vagaries of site formation processes (ie the practice of using broken up flooring as packing material).

73 The flint aggregate is alien to the geology of Lyminge and its immediate environs. Compositional analysis of the flooring (see Supplementary Material) demonstrates that the aggregate was almost certainly sourced from a shingle beach on the southern coast of Kent.

74 It is possible that the other four fragments originally carried an upper surface of this kind to produce a polished/marbled effect as described in relation to the opus signinum floors at Dover and Reculver. The extant terracotta finish seems rather coarse and unworn to constitute a floor ‘surface’, so it may be that the original has eroded away in the depositional environment.

75 There is a possibility that this fragment is derived from an internal wall surface, perhaps the surrounds of a wall feature such as a niche, rather than a floor given its slender proportions and the fact that the limewashed surface would have been very easily worn underfoot (Roberta Gilchrist pers comm). Alternatively, if indeed derived from a floor, it may be associated with an area of the building with restricted or limited access.

76 St Mary’s Reculver, Period 1 nave and chancel and Period 2 porticus: Dowker’s description (1878, 259) reads, ‘The chancel, as far as the semi-circular apse, is paved with a remarkable concrete floor, consisting of a basis of boulders overlaid with mortar like that used in Roman work, but faced with red pounded tile, the surface of which appears to have been polished, the thickness exceeding six inches’; see also Peers 1928, 246–7; Taylor and Taylor 1965–78, ii, 508; Philp 2003, 129. For St Peter and Paul, Canterbury, narthex of the Period 1A church (7th century) see Saunders 1978, 44–52 and Blockley 2000, 45. For St Pancras, Canterbury, Period 2 nave constructed in the mid- to late 8th century see Hope 1902 and Blockley 2000, 71.

77 Wearmouth Building B; Jarrow Building A (Cramp 2005a, 98, 193–4; Cramp 2005b, 18–19) and Glastonbury (Gilchrist and Green 2015, 388, 418).

78 Cramp 2005b.

79 Cramp 1957, 76.

80 Among the artefacts recovered from Dover, an ornate gold and garnet ring stands out as a signifier of an elite presence in the later 6th century (111, fig 57, no 134); the discovery of clay moulds in the infill of one of the sunken-featured buildings (N3) (p 13) indicates metalworking activity. The rather more extensive metalworking evidence from Lyminge is discussed in Thomas 2017.

81 The contemporaneity of the halls deserves comment especially if, as argued here, S14 has a much more compressed chronology than proposed in the report. The report assigns S10 and its three constituent periods a broad 9th–11th century date largely on the evidence that its foundations cut an earlier sunken-featured building, S9. Contra Philp (2003, 48) there is no reason to assume a lengthy hiatus between the abandonment of S9 and the construction of S10; Lyminge provides a parallel scenario for an intercutting disposition between a great hall and a sunken-featured building, and here the evidence suggests that the transition was swift. The contemporaneity between S14 and S13 is difficult to establish in more certain terms than that they respect each other spatially.

82 Philp 2003, 3.

83 Relevant discoveries include, from the South Area, a sunken-featured, or cellared, building (S11) and a neighbouring metalled road (S12), and, from the North Area, a ground-level structure with a sequence of burnt floors (N8).

84 Rigold 1977; Blair 2005, 191–5.

85 Philp 2014. While only one great hall has been identified at Eynsford, there is a very high probability that the number would expand with further excavation.

86 Richardson 2005, vol 2, 32–3, 42–3; Tyler 1992; Simmonds et al 2011, 85–9, 197–200.

87 Philp 2003; Meates 1971.

88 Everitt 1986, 77–8.

89 The Eynsford report shares many of the limitations of its sister publication based on the Dover excavations, not least a very cursory finds section.

90 Philp 2003, 131. Highly rendered plastered finishes (as opposed to the more ubiquitous form of daub walling found widely on Anglo-Saxon settlements) has been inferred at other great hall sites, including Yeavering (Buildings A2 and A4, Hope-Taylor 1977, 53, 61) and Sutton Courtenay, Oxfordshire (Building 500, Hamerow and Brennan 2015, 338). To these examples can be added the impressively preserved portion of plastered panelling recovered in association with a 10th-century timber building from the site of Eynsham, Oxfordshire (Hamerow and Blair 2003). While strictly speaking of daub rather than plaster, the delicate painted fragments of wall finishing from the monastic site of Hartlepool, Co Durham (Daniels 2007, 130, 219–21), offer a good analogy for the wall treatment seen at Eynsford.

91 The overwhelming majority of great hall complexes replicate this topographic tendency. For a more detailed consideration of this aspect see Austin 2017; and Blair 2018, 104–8, 111–13.

92 Semple (2010) provides a good summary of the ritual connotations of rivers in Anglo-Saxon England as the loci for decision-making, boundary disputes and oath-taking; for related explorations see Everson and Stocker 2003 and Lund 2010.

93 Bede HE II, 14, Colgrave and Mynors 1969, 188–9. On the persistent association between places of royal assembly and open-air baptism in pre-Viking England see Morris 1991.

94 See Thomas 2013 and Thomas et al 2017 for discussion of Lyminge’s role in negotiating the process of Christianisation. The source of the River Nailbourne immediately suggests itself as an enduring pre-Christian cult focus given the long duration of occupation and activity brought to light by the Tayne Field excavations. Named ‘St Eadburgh’s Well’ in later medieval documents (Baldwin 2017, 206), the source may well have been resacralised in connection with the foundation of the adjacent minster (Morris 1989, 84–91; Blair 2005, 226–8, 462–3).

95 Everitt 1986, 75–6; Parfitt and Anderson 2012, 369; Tyler 1992; Simmonds et al 2011, 85–89, 197–20; Richardson 2005, 69–77, vol 1.

96 Williamson 2013, 56.

97 Brookes 2007, 149; Williams 2006, 185–7; Semple 2013, 22.

98 What follows is heavily indebted to the work of Howard Williams, in particular an unpublished paper presented at a workshop at Durham University in June 2016 under the auspices of an academic network on the theme of early medieval royal residence: <http://royalresidencenetwork.org/> [accessed 16/08/18].

99 Hope-Taylor 1977; Bradley 1987; Blair 2005, 54–7; Semple 2013, 207–12.

100 Hope-Taylor 1977, 157–61; Millett and James 1983, 205–12.

101 Smith 2003 provides a useful cross-cultural archaeological examination of the monumental and geopolitical strategies exploited in the reification of power and authority.

102 Thomas 2017, 108.

103 Hamerow 2010, 64; see also Semple 2013, 209 for similar reflections. While clearly an important aspect of this phenomenon, sequential rebuilding of this type appears selective in the sense that is not seen at all great hall complexes, or in relation to all the large halls represented at these sites. For example, none of the halls so far investigated at Sutton Courtenay (Oxfordshire) display such a tendency (Hamerow and Brennan 2015, 346) and Lyminge Hall A had a much less complex (and by implication shorter) life history than its counterparts. While difficult to explain, such temporal distinctions contribute to a growing awareness of localised and intra-site diversity within the great hall complex tradition.

104 Williams 2001; Fleming 2011, 90–102; Yorke 2008. As Williams has argued (2001), genealogical manipulation is a field of discourse where archaeological and historical perspectives can reflect upon and inform each other.

105 Brooks 1989.

106 Dumville 1977; John 1992.

107 Rollason 1982; Hollis 1998. See Hamerow (2016) for archaeological expressions of the same theme illustrated through the context of high-status female burials.

108 Hope-Taylor 1977; Scull 1991. Alcock 2003, 254 has also viewed hall building as a symbolic valorisation of royal authority following victories on the battlefield.

109 Herschend 1998, 36; Rollason 2016, 38.

110 Herschend 1998; Carstens 2015.

111 Carlie 2006; Fabech and Näsman 2013.

112 For Scandinavian perspectives on hall (re-)building as a form of ‘spatial remembrance’ see Stenholm 2006 and Eriksen 2016.

113 Williams 2006, 161; see also Williams 2001.

114 In this sense great hall complexes provide an instructive illustration of destruction/forgetting as an active part of the discourse of remembrance, see Connerton 1989, 72–89 and Williams 2001.

115 For wider discussion of the relationship between ritual actions and the biography of Anglo-Saxon buildings see Hamerow 2006; 2012, 127–42; Thomas et al 2016.

116 Hamerow 2006; Sofield 2015. See Walker (2010, 96) and Austin (2017, 11–13) for comparisons and contrasts between English great hall sites and parallel phenomena in Scandinavia.

117 Thomas 2017, 101; Thomas 2013, 127.

118 Thomas et al 2016.

119 Philp 2003, 72, 121.

120 Brooks 1989, 74. For complementary archaeological reflections on the mead-hall as the locus for genealogical performances see Price and Mortimer 2014.

121 Hope-Taylor 1977; James et al 1984. For a critique of this approach see Scull 1991.

122 Hines 1994; Scull 1999; Härke 2011.

123 Yorke 1999; Webster 2000; Williams 2001, 67; Carver 2001, 2–4. Blair (2013) has argued that British influence is likely to be one significant strand in the genesis of the great hall complex tradition and this might explain why the current distribution of sites appears to be on the periphery of the Anglian heartlands of eastern England facing Scandinavia and Frisia.

124 Blair et al forthcoming, ch 5.

125 For the dichotomy between stone and timber in Anglo-Saxon architecture see Fernie 1983, 9, 58–9, 174; Shapland 2013; 2015b, 94–9.

126 Fernie 1983; Gittos 2013, 73; Cramp 2017, 38–9.

127 Fernie 1983, 59.

128 Webster 2012, 97.

129 This argument has previously been advocated by Hope-Taylor in his interpretation of the Yeavering buildings (1977, 235–6) and in more general terms by James et al 1984, 206.

130 Bell 1998; Eaton 2000.

131 Dobres 2000.

132 Cambridge 1999.

133 For recent reflections on the problems of the secular/religious dichotomy in Anglo-Saxon settlement archaeology see Ulmschneider 2011, 161–163; Gittos 2011, 827–8.