MAPPING LORDLY CENTRES

A first priority of ‘Where Power Lies’ was to define the signatures of lordly centres and create a corpus to be integrated into a single geographic information system (GIS).Footnote ³² Given that aristocratic complexes can be found across England, the project incorporated evidence from as many national datasets as possible. This allowed initial exploration of national and regional distributions, and cautious consideration was given to their potential significance. It is important to note that the resultant corpus of lordly centres cannot be taken as an accurate barometer of past activity, given that patterns also reflect varying research histories. In order to circumvent this as much as possible, the project targeted two macro regions for which the evidence for lordly centres was interrogated more rigorously: one focusing on south-west England (Devon, Dorset, Gloucestershire, Hampshire, Somerset, West Sussex and Wiltshire), and another focusing on north-east England (Lincolnshire and Yorkshire). Although differing in numbers of constituent counties, these macro regions are roughly comparable in size, covering approximately 26,477km² and 22,716km² respectively. This choice of scale permitted the compilation of a dataset that was large enough to facilitate identification of meaningful regional trends, while allowing comprehensive examination of sites and landscapes during the project’s duration. Additional benefits of this approach were that one macro region lay within the Danelaw and one without, and that a substantial cross-section of Roberts and Wrathmell’s settlement typesFootnote ³³ was included.

The primary data used to develop the macro study regions were Historic Environment Record (HER) databases and Historic England’s research records and Listed Buildings register, alongside other digital resources.Footnote ³⁴ These datasets were systematically searched for places that have evidence for lordly activity before c 1200, and, while this process identified many places of interest, interpretation was often far from routine. A particular challenge in generating an accurate database using this method was the inconsistent approach to categorisation used by HERs and variations in the quantity and quality of evidence used to recognise sites. While the lack of meaningful archaeological investigation encourages reliance on textual sources to identify many sites, their inconsistent use and interpretation had the potential to heavily skew the dataset. The utilisation of Domesday Book exemplifies this problem, with some HERs extrapolating the existence of an early lordly centre (commonly assigned an arbitrary location at the centre point of a modern settlement) from reference to a manor in the survey alone, whereas other HERs are more cautious.

The result of such discrepancies means that much of the evidence initially acquired for the macro regions was, once more, a product of varying research methodologies rather than reflecting an accurate profile of early aristocratic centres. Nevertheless, following a detailed programme of cleansing, corroborating and standardising data, ‘Where Power Lies’ produced a corpus of places for which there is unequivocal evidence for a lordly centre, as well as sites that are suggestive rather than definitive; other locations were rejected outright as spurious. A total of 870 unequivocal elite residences and 3,528 churches with origins before 1200 were recognised within the two study regions (fig 1). Within this corpus, the project assigned various chronological quantifiers to the database (ie earliest suggested dates and whether sites are pre-Conquest and/or pre-1200). In some cases, where further detailed research is needed to characterise and date a site more closely, only a conjectural phasing has been offered.

Fig 1. Distribution of elite residences and church foundations (c 800–1200) in the two macro regions of ‘Where Power Lies’. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

Although vagaries in the corpus prohibit rigorous quantification of statistical trends, anecdotal observations are revealing and potentially significant. It has long been recognised, for instance, that the archaeological evidence for England’s early lordly centres largely lies underneath manor houses and castles, many of which were built in the century or so after the Norman Conquest.Footnote ³⁵ A large proportion of these residences have a close association with a broadly contemporaneous church or chapel; in some counties 50 per cent of eleventh- and twelfth-century castles lie in the immediate vicinity of a parish church, and a significant minority are incorporated within baileys.Footnote ³⁶ Such spatial associations need not indicate examples of Norman imposition of power over local communities, however. Scholarly consensus is growing that in many regions the influence of lordship was already intensifying throughout the tenth and eleventh centuries, and the incoming Norman elite instead provided a visual makeover and a more military aspect to seigneurial foundations that were already a common feature.Footnote ³⁷ Likewise, the proliferation of churches in the ‘Romanesque revolution’ from the mid-eleventh century was a part of a longer-term phenomenon of local church foundation beginning in the late Anglo-Saxon period that was given impetus by Norman lords.Footnote ³⁸

In the first attempt to quantify the relationship between lordly residences and churches, but cognisant of the inherent limitations of the database noted above, ‘Where Power Lies’ undertook a GIS-led spatial analysis showing that approximately one-third (34 per cent) of elite residences within the two macro regions lay within 250m of a church (fig 2). By way of comparison, analysis of the nationwide dataset of early (eleventh- and twelfth-century) castles shows that 26.5 per cent lie adjacent to a church, with 3.6 per cent of these sites embracing the ecclesiastical site within its defences. Within the overall figure for elite residences in the two areas, there are regional distinctions: the church–residence relationship is slightly more pronounced in parts of south-west England, with Dorset (40 per cent), Gloucestershire (42 per cent), Somerset (42 per cent) and West Sussex (48 per cent) having more examples of this relationship than average. In Hampshire and Wiltshire, however, there are slightly fewer examples than average, at 29 per cent and 30 per cent respectively, while only 20 per cent of Devon’s residences lay within 250m of a church. Within the north-eastern macro region there exists a clear contrast between Lincolnshire and Yorkshire, with 42 per cent of the former’s elite residences lying within 250m of a church compared to only 29 per cent in the latter. There are also variations within Yorkshire, with the historic North Riding (24 per cent) having fewer examples of this close co-location compared with the East Riding (29 per cent) and the West Riding (32 per cent). There are also regional trends within the various counties as examples of close residence–church relationships are absent from Hampshire’s New Forest region, and are noticeably rare east of the Cotswold escarpment in Gloucestershire, the heathlands of south-east Dorset, the Culm Measures of Devon, the Fens of Lincolnshire and Holderness in Yorkshire.

Fig 2. Locations of lordly residential sites and contemporary churches. Examples that exhibit a close residence-church correlation usually represent a single lordly centre. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

One of the most surprising patterns observable among the data is that across the macro regions, lordly centres located in close proximity to a church are more prevalent in areas of historic low settlement density (ie zones characterised by dispersed settlements, especially hamlets and isolated farmsteads):Footnote ³⁹ 9 per cent are found in areas of extremely low settlement density, 17 per cent in areas of extremely low to very low density, 6 per cent in areas of very low density and 18 per cent in areas of very low to low density (fig 3).Footnote ⁴⁰ While the prevalence in areas of low settlement density can be partly explained by the relative dearth in coverage of other settlement character types across the two study regions, closer examination of the data corroborates the overall pattern; instances of close church–residence relationships are in fact disproportionately more common in areas of lower settlement densities compared with areas of higher densities (fig 4: top).

Fig 3. Lordly centres within the macro study regions overlaid on medieval settlement density zones, as classified by Roberts and Wrathmell (Reference Roberts and Wrathmell2000) and Lowerre et al (Reference Lowerre, Lyonds, Roberts and Wrathmell2015). Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

Fig 4. Distribution of lordly centres (top) and early castles (bottom) relative to medieval settlement densities. These correlations suggest that landscapes with a variety of settlement densities feature lordly centres, and they were not the preserve of ‘champion’ countryside. Image: authors.

This finding is supported by analysis of the nationwide dataset of early (eleventh- and twelfth-century) castles and their relationship with settlement density. Castles adjacent to a church are disproportionately more numerous in areas of extremely low to very low settlement density and slightly disproportionately more numerous in areas of very low to low, and high to very high settlement density (fig 4: bottom). These patterns highlight that the places where church and lordly residence were closely juxtaposed were at the very least embedded within a diversity of settlement landscapes and were not a peculiarity of nucleated villages in ‘champion’ countryside, as has sometimes been assumed.Footnote ⁴¹ Indeed, this outcome suggests that scholarship has over-amplified the prevalence and impact of lordly centres in areas of high settlement density, perhaps as components such as ostensibly ‘planned’ villages and fields conform to an idealised blueprint of what a heavily seigneurialised landscape should look like.Footnote ⁴² A note of caution is required, though, given that these observations are based on modelling of historic settlement densities from nineteenth-century mapsFootnote ⁴³ and that areas characterised as being low settlement density do not necessarily show a total absence of nucleated settlements. As such, it is feasible that a number of lordly centres found in such areas may in fact have been associated with isolated nucleated settlements. Unfortunately, it has not been possible to determine the frequency of this phenomenon at any scale as grades, dispersal and size of nucleated settlements in Roberts and Wrathmell’s GIS is structured around point rather than shape data. The correlation between lordly centres and settlement densities identified here must therefore be considered suggestive rather than definitive and, as useful as a truly dedicated investigation of the topic would be, it fell outside the scope of the project.

Another pattern identified in data from the macro regions is the close spatial correlation between lordly centres and watercourses, with approximately 37 per cent across the two study regions being located within 250m of a watercourse. In fact, a large number of lordly sites were found to include or be delineated by a river or stream, sometimes clearly through hydrological engineering or modification (fig 5). These watercourses would have been invaluable as sources of drinking water and irrigation, as well as potentially offering transport in some cases,Footnote ⁴⁴ but many also provided power for watermills (fig 6). That there is an explicit link between lordly centres and watermills is evidenced by the ‘suit of mill’ custom, whereby tenants were bound to grind grain at their lords’ mills, usually for a toll. The question as to when this custom was introduced to England is yet to be satisfactorily answered; the most recent analysis suggests that it was a post-Conquest development, imposed upon tenants as a consequence of the tenurial reorganisation of the country in the decades after the Conquest.Footnote ⁴⁵ Indeed, it may have been possible that one of the criteria for defining the boundaries of Norman manors was that they contained at least one mill or were adjacent to one or more manors containing a mill.Footnote ⁴⁶

Fig 5. Comparative plans of lordly centres integrating watercourses. At Broadclyst, Rand, Sidbury and Southrop, later mills are likely to perpetuate medieval precursors. Image: authors.

Fig 6. Distribution of pre-1200 lordly residences and churches/chapels in relation to the major watercourses of Wiltshire and Hampshire. The close correlation of lordly centres with rivers and streams suggests that water power was a fundamental consideration in siting. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

That watermills were core features of lordly centres in the years immediately after the Conquest is apparently confirmed by the more than 6,000 mills recorded in Domesday Book, a number that is almost certainly an underestimate.Footnote ⁴⁷ Interpreting this metric is problematic, though, as it is not clear whether all these records represent watermills. Admittedly, clear evidence for windmills in England only occurs from the late twelfth century, but three late tenth-century references to mill oxen in Huntingdonshire, and a possible animal-powered mill at Cheddar’s royal palace, demonstrate the diversity of options available for operating mills in the Late Saxon period.Footnote ⁴⁸ In addition, it is not always obvious whether Domesday Book’s mills actually refer to individual buildings, shared millstones or rights to the number of days that an individual was permitted to use a mill.Footnote ⁴⁹ In fact, Domesday’s frequent recording of multiple mills within a single manor raises the question of just how such places would have genuinely benefited from supporting more than a single establishment.Footnote ⁵⁰

As much as these caveats urge caution when considering the prevalence of watermills, detailed local study often reveals the prolific exploitation of waterpower at lordly centres. For example, investigations at the neighbouring townships of Wharram le Street and Wharram Percy (North Yorkshire) found physical evidence for the locations of five watermills along a 4km stretch of the Wharram stream.Footnote ⁵¹ Two of these mills, albeit belonging to different landowners, were situated only c 100m apart,Footnote ⁵² suggesting high numbers of mills supported by short stretches of watercourse, even though such proximity would have hindered operating capacity. Such examples are even more impressive given the generally poor archaeological visibility of watermills; not only is excavation of these features rare, mills and leats fluctuated over time as channels became silted. The challenges of recognising mills archaeologically are neatly exemplified at West Cotton (Northamptonshire), where a full sequence of houses and leats adjacent to the lordly centre was only identified by chance. Understanding was further hampered by the removal of timbers that comprised the majority of the buildings when the site was abandoned in the twelfth century.Footnote ⁵³ Even though the picture derived from excavation and topographical evidence is thus only partial, it is evident that watermills had an economic and symbolic centrality to lordly centres as ‘the most complex of the technological advances of the Middle Ages’.Footnote ⁵⁴ Installation of a watermill at a lordly centre would have constituted an ostentatious projection of innovation and wealth, just as residences and churches themselves were theatres of consumption and display.Footnote ⁵⁵ In the same way, sundials and scratch dials frequently found in the fabric of churches at seigneurial sites rarely seem to have served their primary function as timepieces.Footnote ⁵⁶ Instead, they were probably seen as an appropriate investment for the church by elites, and may even have been tied to specific benchmarks of status for leading families and their churches, such as the acquisition of burial rights.

The correlation between lordly centres and administrative units was a further spatial relationship subject to investigation, revealing that 10 per cent of the corpus are located within 250m of hundred or wapentake boundaries. Within this global figure, an average of 12 per cent of lordly centres in the south-west counties are situated close to a territorial division, compared with 8 per cent in the north-east macro region. Once more, small sample sizes encourages caution in reading too much into the data; for example, Gloucestershire’s lordly centres are the most consistently sited close to hundred boundaries at 14 per cent, but this is derived from a sample of only fifteen sites out of a county-wide total of 108. A further consideration is the common use of watercourses to delineate Late Saxon units, meaning that even sites located at interfaces could just as easily reflect the significant attractions of rivers and streams, explored above, than territorial considerations. In spite of these qualifications, the evidence from more concerted study of individual sites shows that political geography sometimes played a role in location. At Saintbury (Gloucestershire), the lordly centre was located within a landscape featuring a number of assembly places that was a well-established ecotone between different groups; this liminal zone was not only where three shires intersected, but also on the boundary of the Diocese of Worcester and represented the limits of the former kingdom of the Hwicce (fig 7).Footnote ⁵⁷ In this contested territorial setting, a lordly centre was positioned mid-way up the Cotswold scarp, close to a kink in Ryknild Street; this location offered good opportunity to survey movement along the Roman road as it slowed to navigate the dramatic change in topography, as well as uninhibited 180 degree views northward over the Vale of Evesham. Just as tellingly, the mill associated with this enclave was not integrated into the enclosure, or even in close proximity, but situated over 800m north-east at the bottom of the hillslope.

Fig 7. The administrative landscape of the lordly centre at Saintbury (Gloucestershire). Both Kiftsgate Court and Kiftsgate Stone were early medieval assembly places. During the early seventeenth century, Dover’s Hill was selected as the venue for the Cotswold Olimpick Games; the choice may demonstrate the continued recognition of this as a landscape of assembly into the post-medieval period. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

The integration of national and regional datasets allows, then, some broad trends in the siting and character of lordly centres to be tentatively identified, and explanation offered regarding their significance. As illustrated by the evidence at Saintbury, though, more meaningful understanding of sites is often only feasible through detailed study of specific centres and their landscapes. Building upon the methodologies used during a pilot project at Laughton en le Morthen (South Yorkshire),Footnote ⁵⁸ ‘Where Power Lies’ conducted concerted investigation of seven lordly centres: four in the south-west macro region and three in the north-east.Footnote ⁵⁹ Each of these was subject to a range of analyses including standing building assessment, topographic modelling and geophysical survey, all of which were supplemented by desk-based research including interrogation of data from the Portable Antiquities Scheme (PAS), an official partner of the project. A summary of results from two of these investigations is offered here, from Bosham (West Sussex) and Hornby (North Yorkshire). These sites, one from each macro region, are chosen for a number of reasons. In some ways they help problematise ‘lordly centre’ as an umbrella term, illustrating the diversity of agents behind their development, and the inconsistencies of existing archaeological signatures. The two locations also serve to demonstrate the value of applying multifaceted desk-based and fieldwork methodologies; historic building assessment is central to understanding both complexes, but at Bosham it is geophysics and interrogation of previous interventions that is revealing, whereas photogrammetric modelling locates an early medieval high-status enclosure at Hornby for the first time. Finally, as neither site was developed into a castle in the post-Conquest period, these case studies seek to move beyond the fetishisation of ‘defence’ and ‘fortification’ that has been so dominant in scholarship, highlighting instead the potential of studying the full range of material investment detectable at early lordly centres.

Bosham, West Sussex

The representation of Earl Harold’s residence at Bosham, West Sussex, on the Bayeux Tapestry lends the exceptional significance as the only visual representation of a lordly centre from pre-Conquest England.Footnote ⁶⁰ One of only four places in England to be named on the Tapestry, Bosham was one of Harold’s chief seats and the site of his harbour, private fleet and probably a deer park.Footnote ⁶¹ Depicting events of c 1064, the Tapestry shows Harold riding to Bosham, attending church and feasting in a hall, before descending a staircase to the water to embark on his ill-fated journey to Normandy (fig 8). On his return, it is probably Bosham again represented by two towers either side of Harold mounted on a horse; the left of the two buildings is four stages high, with people looking out to sea for the earl’s arrival, indicating its use as a gatehouse-cum-watchtower.Footnote ⁶² Apart from Westminster and Hastings, Bosham is the only location on the Tapestry to be shown twice, underscoring its centrality in the narrative. Given this clear importance, it is surprising that little previous effort had been made to understand Bosham’s lordly complex, and even its location has not been established for certain.

Fig 8. Harold Godwinson at Bosham, as depicted on the Bayeux Tapestry. Harold is shown attending church and feasting in the upper floor of a hall, before departing by ship to Normandy. Later in the Tapestry, it is probably Bosham depicted again when Harold is shown returning from his journey. Engraved by James Basire (1769–1822), after Stothard; hand-painted by Charles Alfred Stothard (1786–1821). Image: © courtesy of The Society of Antiquaries of London.

Seeking to rectify this, a programme of fieldwork was initiated by ‘Where Power Lies’, consisting of an analytical record of the standing structures at Bosham Manor House, alongside ground penetrating radar (GPR) survey of the gardens, as well as other accessible locations in the village (fig 9). Our work at the manor house was not the first archaeological assessment, but follows a highly significant scheme of evaluation undertaken by West Sussex Archaeology in 2006, during which seventeen trenches both inside the house and in the gardens were excavated. Although the majority of the garden was devoid of archaeology, excavations did locate a rubbish pit and two enclosure ditches dated to the Anglo-Norman period and, crucially, a Late Saxon latrine apparently associated with an earth-fast timber building.Footnote ⁶³ The importance of this latter feature could not be fully recognised at the time, as it is only subsequent publications, first by Gabor Thomas in 2010Footnote ⁶⁴ and then John Blair in 2015,Footnote ⁶⁵ that convincingly illustrate how latrines came to be linked to the chamber-ends of tenth-century and later high-status houses. In this light, it is overwhelmingly likely that the Late Saxon latrine underlying the timber building at Bosham was appended to a long range in the same way, confirming the manor house as the site of a pre-Conquest elite residence. Although only elements of the western side of the enclosure defining this domestic space were located in excavation, it is reasonable to assume that the area was bound to the west by the stream perpetuated into the present day as a canalised mill leat. This rectilinear compound measures roughly 50m east–west, and 70m north–south, but how far the complex extended north, and the overall capacity of the enclosed interior, is less clear. The residence was clearly connected with Holy Trinity Church to the south, the earliest parts of which date to the eleventh century, as evidenced by a large posthole in the enclosure circuit interpreted by the excavator as a support for a bridge or causeway.Footnote ⁶⁶

Fig 9. Modern topography of Bosham, showing key locations mentioned in the text and the results of the GPR survey, and (inset) Bosham’s location in southern Britain. Image: authors using imagery © 2024 Landsat/Copernicus, Maxar Technologies.

The next feature in the chronological development of the site is a structure known as the ‘garden ruin’, situated 35m west of the manor house. Most previous observers have assumed this to be the remains of a post-medieval building;Footnote ⁶⁷ it is depicted as a substantial L-shaped structure on mapping from 1784 and 1839, but had been truncated to a rectangle by the mid-nineteenth century and then largely demolished according to the Ordnance Survey map of 1912. Little of the building survives above ground, and it is heavily overgrown, meaning that the assessment outlined here is provisional only. The structure consists of part of the western and northern walls of a suspected rectangular building approximately 9m east–west by 19m north–south. It stands to a height of up to 2.4m above present ground level on the eastern side and c 1.2m to the west, due to the rise in ground level up to the mill leat. The majority of the structure is of coursed flint rubble, with limited patching in stone rubble and brick. Importantly, GPR survey located a buried linear feature in the area of the garden ruin, projecting broadly north–south and running perpendicular to the extant west wall of the garden, abutting the extant northern wall. This seems likely to be the buried eastern wall, confirming the rectilinear arrangement of the structure before addition of an eastward projection to form the L-shaped building.

The earliest visible phase of the ruin is the footings of the west wall, which are cautiously assumed as extending its full 19m length. These footings appear to be wholly stone with no brick, and are probably medieval in date. The earliest upstanding part of the structure is the northern part of this same west wall, which incorporates a relatively coherent window of twelfth- or thirteenth-century date (fig 10). The key relationship here is with the north wall, which appears to be coherent with the west wall in this area, although with flint rubble it can be very hard to be sure. Significantly, the north wall also seems to abut the early stone footings described above. If both of these aspects are accurate, then the stone footings of the western wall are earlier than the north and west walls, which are themselves contemporary. Both have eighteenth-century brick quoins, but there is no certainty these are primary, meaning that we may be dealing with a twelfth- or thirteenth-century building erected on earlier footings. If so, this may represent the base of either a so-called first-floor hall or a chamber block, a type of building that is characterised by narrow lancet windows to the cellar and larger windows for the hall above.Footnote ⁶⁸ Given its location close to the leat, another possibility is that this building functioned as a mill for at least part of its history.

Fig 10. Garden ruin, Bosham Manor House, lancet window to northern half of west wall. External view looking east. Image: authors.

Of apparently slightly later date is the earliest phase of the manor house itself, a building that again has been the subject of only brief surveys, which have exclusively dated it to the post-medieval period.Footnote ⁶⁹ Our assessment confirms that the present house does indeed date principally to the seventeenth century and later, but at its core is the remnant of a stone wall belonging to an earlier structure, probably a dwelling (fig 11). One of the evaluation trenches in 2006 located stone footings to the north of the upstanding stone wall that may relate to the same building.Footnote ⁷⁰ On the basis of its stonework, the wall may date to the earlier part of the medieval period (c 1100–1300), although it appears to overlie two ditches, also excavated in 2006, that were seemingly infilled from the fourteenth century. While a date of c 1300 for construction of the wall just about satisfies both pieces of phasing evidence, only scientific dating will be able to clarify its origins more firmly. During the later medieval period, a new dwelling was constructed immediately to the south of the first stone building; all that survives of this are numerous re-used timbers throughout the house, including smoke-blackened rafters and curved windbraces, confirming that it took the form of a conventional open hall. The earlier stone building appears to have been retained upstanding, perhaps downgraded to a service function such as a kitchen. The manor house was subsequently modified a number of times from the early seventeenth century, and the current structure is two storeys in height, plus an attic, comprising an east–west aligned front range on the southern side, with a rear wing and stair appended to the north.

Fig 11. Phased reconstruction of Bosham Manor House. Image: authors.

Complementing the results of new and previous fieldwork, the PAS database was interrogated for evidence of medieval activity across the Bosham peninsula. No medieval finds are recorded from the site of Bosham’s lordly centre itself, and comparatively few artefacts come from the wider parish when compared to the substantial cluster recorded in neighbouring Appledram (fig 12). Nevertheless, Bosham parish has yielded several pre-1066 items, including an eighth-century silver sceat, an eighth- to tenth-century copper alloy mount, a silver brooch made from a Short Cross penny issued in Cnut’s reign (1016–35), a copper alloy stirrup terminal dated c 1030–60 and an eleventh-century stirrup mount. Four further finds of potentially pre-1200 date consist of a cast copper alloy balance arm of a medieval trebuchet coin dated to between 1150 and 1450, two silver pennies of Henry ii minted between 1158–61 and 1180–9 respectively, and a silver penny of Stephen (1135–54). The pre-Conquest assemblage from Bosham is notable, pointing towards an uptick in elite activity in the eleventh century that neatly coincides with the establishment of the Godwinssons’ power centre. In other areas examined by the project, this kind of distribution – often representing a halo around a historic settlement – has led to the identification of previously unknown aristocratic sites, demonstrating the immense value of the PAS as a landscape-scale prospecting tool.

Fig 12. Distribution of PAS medieval data in Bosham and the surrounding area, and (inset) a selection of the medieval finds. (a) eighth-century sceat; (b) silver brooch made from a Short Cross penny; (c) stirrup terminal dated c 1030–60; (d) eighth- to tenth-century mount; (e) eleventh-century stirrup strap mount. Note: all to scale, apart from (b) as none was provided. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

A combination of new evidence generated by ‘Where Power Lies’, alongside reassessment of the material deriving from earlier excavations, offers important new insight into Bosham’s medieval lordly power centre. Reconsideration of the 2006 evaluation trenching suggests, beyond reasonable doubt, that Bosham Manor is the site of the aristocratic complex made famous by the Bayeux Tapestry – a scenario that has often been assumed, but never demonstrated with physical evidence. The Godwinssons’ centre was sited at Bosham’s once-wealthy monastery that had existed by at least the 670s,Footnote ⁷¹ an institution that itself was probably built close to a Roman site located on this part of the peninsula.Footnote ⁷² During the eleventh century, the lay church and residence would most likely have been carved out from the monastic precinct in a manner similar to that at Deerhurst (Gloucestershire), where the religious community retained the northern half of the compound, while Earl Odda developed the southern half as his own residence.Footnote ⁷³ The division at Bosham was more plausibly east–west, with the stream dividing the two zones and providing the Godwinssons with access to residence, church and waterfront (fig 13). The harbour seems to have been particularly important to the aspiring family, as captured by two entries in the Anglo-Saxon Chronicle, first in 1049 and again in 1051.Footnote ⁷⁴ Later in the medieval period the lordly complex expanded eastward from its Anglo-Norman core, with the building in the manor house constructed across the enclosure ditches formerly demarcating the high-status residence (fig 14). Only through further work can this approximate phasing be refined, but the establishment of the Godwinssons’ private complex had a lasting legacy; manorial function at Bosham endured into the twentieth century through the annual hundred court, latterly held in the dining room of the manor house until it ceased to assemble from 1914.Footnote ⁷⁵

Fig 13. Conjectural reconstruction of Bosham’s chronological development incorporating information from Kenny (Reference Kenny2004), with modifications by the authors based on new evidence. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

Fig 14. Suggested phase plan of the development of Bosham’s lordly centre. Expansion seems to have occurred eastward from the earlier Anglo-Norman core. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

Hornby, Richmondshire, North Yorkshire

The small village of Hornby was also the target of concerted study by ‘Where Power Lies’; investigation here consisted of a standing building assessment of St Mary’s Church and a topographic survey via an unmanned aerial vehicle (UAV) of earthworks in a field to its west. Previous, mostly cursory study of the church had identified its mid-eleventh-century tower as its earliest phase, and it was argued that this had originally been attached to an aisle-less nave.Footnote ⁷⁶ Concerted re-examination of the fabric, however, disputes this reconstruction, with the tower’s prominent clasping buttresses key to unpicking the development of the building (fig 15). The lower halves of the buttresses are of identical rubble construction to the main body of the tower, albeit with some larger, better-dressed stones to serve as quoins. The upper halves are different, comprising coursed and dressed ashlar, the top course of which is chamfered, so as to articulate with the tower’s string course. The Taylors suggested that the buttresses were later additions, yet they are clearly bonded into the main body of the tower, and the chamfered plinth is entirely coherent across the structure.Footnote ⁷⁷ It is not certain why the upper halves of the buttresses are constructed of finer work than the lower, but there is no reason to believe they are of different dates, and the explanation that the base of the church was built up against another structure is unconvincing.

Fig 15. Oblique photograph of St Mary’s Church, Hornby, showing the eleventh-century tower-nave, now integrated as the western tower and (inset) Hornby’s location in central Britain. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

A more likely scenario is that the eastern buttresses did indeed clasp the north-eastern and south-eastern corners of the tower, in identical fashion to their western counterparts – a hypothesis supported by the apparent similarity in the form and construction of all four features. Additional backing for this interpretation is the presence of an internal crack running vertically down the plaster of the interior of St Mary’s, which presently obscures the eastern face of the tower. This crack aligns with the location of the eastern corners of the tower, and it is also stepped-in above the first stage, exactly where the second stage is stepped back from the base of the structure. Together this evidence demonstrates that the tower is structurally independent from the rest of the building, and implies that, when Hornby was first built, it was a turriform or tower-nave church. Recent research into tower-naves, whereby the tower comprised the main body of the church, has shown that they were not uncommon in the landscape of eleventh-century England and were typically associated with lordly residences, where they are argued to have had a dual function as buildings of elite worship and symbols of secular authority.Footnote ⁷⁸

That Hornby’s tower-nave was originally located within an aristocratic compound is demonstrated by identification of a substantial enclosure, the western earthwork of which was mapped as part of the topographic survey via UAV. This survey collected images across 2ha of a field historically known as ‘Parsons Ground’, from which digital surface models were produced to gain a detailed picture of the topography. A series of earthworks of archaeological origin were mapped through this method, the most prominent of which is a large C-shaped enclosure, encompassing several other earthworks within its interior (fig 16). The enclosure is formed by an upstanding earthwork, up to 5m wide, along its northern and western edges and as a cropmark, up to 10m wide, along its south-easterly projection. On the western edge of the enclosure there is a clear gap in the earthwork, aligned approximately with the church, that may represent a former entranceway. A shallow ditched feature, previously identified as a trackway by Historic England’s Aerial Investigation and Mapping (formerly the National Mapping Programme), follows the exterior of the enclosure on all sides and accentuates its profile.Footnote ⁷⁹ Within the enclosure itself, three rectilinear features formed by a combination of upstanding earthworks and negative features are probably building platforms. The easternmost of these, though fragmentary, measures c 19m × 11m, the northernmost is c 30m × 9.5m and the westernmost c 15m × 23m. On the exterior of the enclosure, further features may represent a hollow way projecting further westward.

Fig 16. Greyscale digital terrain model of Parsons Ground. The bank of the enclosure, which includes a break in the western side, is located just left of centre. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

Significantly, the enclosure encompassing the platforms is perpetuated in property boundaries further east, forming a circuit that incorporates St Mary’s and the core of Hornby village (fig 17). The curving profile of this oval enclosure is preserved by the line of the northern edge of St Mary’s churchyard, and on its eastern side by a north–south aligned drain located c 60m east of the chancel. It is clear, therefore, that Parsons Ground preserves only the western element of a much more extensive enclosure, reconstruction of which suggests a footprint measuring c 200m east–west by 150m north–south at its greatest extents. Identification of this putative circuit is supported by the ongoing excavations of The Architectural and Archaeological Society of Durham and Northumberland that have recovered early medieval material from a northern trench (fig 17: Trench 4) – an area suggested as lying within the enclosure – but not from interventions further south (fig 17: Trenches 2 and 3).Footnote ⁸⁰

Fig 17. Interpretive plan of the main features of the lordly complex at Hornby, and locations of some of the trenches excavated by The Architectural and Archaeological Society of Durham and Northumberland. Image: authors © Crown Copyright and Database Rights. Ordnance Survey (Digimap Licence).

The character, size and plan of the Hornby enclosure corresponds to those identified at other lordly centres by ‘Where Power Lies’ (fig 18) and, together with the tower-nave, it is most plausible that it originated as part of an aristocratic complex of the late eleventh century. The site seems to have been arranged with a residential focus to the west and the church, perhaps situated within its own compound or another subdivision, further east. Both church and enclosure could well have been commissioned by Gospatric, whose father Arnketil was either killed or forced into exile following his involvement in the northern rebellion of 1069.Footnote ⁸¹ In Domesday Book, Gospatric is recorded as a subtenant of Count Alan (the Red) of Brittany, in contrast to his father, who had apparently held Hornby outright before the Conquest. Gospatric’s continued status as a landholder, albeit now as a subtenant, is part of a process seen across the Honour of Richmond as well as other compact lordships in Yorkshire, whereby pre-Conquest lords retained landed interests but were demoted down the tenurial hierarchy. These instances, reflected in the high proportion of pre-Conquest lords still holding land by 1086, are probably the result of a compromise made by William in his attempt to rapidly gain control of potentially volatile and problematic regions.Footnote ⁸² It has long been recognised in these areas that establishment of subtenants went hand in hand with castle construction,Footnote ⁸³ and it is likely that the lordly centre at Hornby was initiated with similar motivations in mind; Gospatric would have been keen to assert his lordly status at a time when his familial possessions had become unstable as a result of his father’s insurrection. By establishing ‘concrete symbols of lordship … in a tenuously controlled landscape’, as Aleks McClain terms it,Footnote ⁸⁴ Gospatric and others like him were hoping to make statements of stability and credibility at a time of acute uncertainty for their lineages. Hornby’s lordly centre, unlike that at Bosham, does not seem to have been especially long lasting; instead, the site seems to have been deserted in favour of the nearby castle at some point in the fourteenth century.

Fig 18. Comparative plan of lordly centres identified by ‘Where Power Lies’. Note the consistent size of the sites, which all enclose an area of approximately one hectare. Image: authors.

DISCUSSION: RECONCEPTUALISING THE ARCHAEOLOGY OF LORDLY CENTRES

Considering expressions of wealth in Late Saxon England, historians Robin Fleming and Ann Williams, among others, have stressed the diversity of social standing across a group that law codes recognise only as thegns.Footnote ⁸⁵ Such insights make apparent that legal standing provides us with only the most primitive of benchmarks to understand what was a varied social spectrum within which individuals had different motivations, family histories and resources, meaning divergent priorities concerning material expression. Archaeologists should therefore not be lulled into creating an artificially uniform ‘elite’, no more than they must recognise the potential for those lower down the social strata to invest, sometimes heavily, in the material world.Footnote ⁸⁶ This is neatly exemplified by the somewhat contradictory picture painted by the two case studies offered here. Bosham’s ownership by the Godwinssons ranks it as a product of the highest echelons of the aristocracy, and indeed it was in part a royal holding from Harold’s reign, whereas Hornby was probably commissioned by an aspiring local family seeking to consolidate their fragile status. Yet, it is Hornby that retains the more impressive material signature, boasting both a tower-nave church and part of its original enclosure upstanding as an earthwork. While the later use of the sites has obviously resulted in differing preservation, it is nevertheless clear that the archaeological signatures of lordly centres cannot be considered a reliable barometer for the wealth or status of the agents behind their development. On the contrary, the material record may sometimes exhibit something of an inverse relationship, whereby parvenus invested at levels reflective more of their aspiration than their actual status, as opposed to well-established earls and leading magnates who may not have felt the same consistent pressure to articulate their authority via grand material statements. A further factor in play is that lower-ranking sites such as Hornby would have seen a more permanent seigneurial presence than centres in the hands of wealthier but more itinerant individuals, who maintained a portfolio of sites.

A further disparity between Bosham and Hornby is their relationship to antecedent landscapes; there is nothing at Hornby to suggest occupation before the eleventh century, but at Bosham the appropriation of the minster by a lay lord was part of the increasingly common phenomenon of secularisation in Late Saxon England. This process saw the assets of once-powerful churches run down, and estates formerly under autonomous control taken over by an increasingly avaricious aristocracy.Footnote ⁸⁷ The archaeological evidence for what appears to have been a widespread occurrence is surprisingly slim, but to Deerhurst and Bosham can be added Little Ouseburn (North Yorkshire), a site also investigated in detail by ‘Where Power Lies’.Footnote ⁸⁸ Here, a small minster community seems to have been established in a bend of the Ouse Gill Beck in the eighth or ninth century, but in the Late Saxon period it was transformed by development of a lordly centre. In a strikingly similar arrangement to Hornby, the seigneurial complex at Little Ouseburn was delimited by an elliptical enclosure, with high-status domestic buildings apparently lying in its western half and a tower-have located to the east. A further annexation of a probable minster was explored at Broadclyst, Devon, where only the most fragmentary evidence for a pre-Conquest enclave was found at a site given a thorough overhaul by construction of a new church and manor house c 1300.Footnote ⁸⁹ The archaeology of secularisation is clearly an area of significant potential for future research, although minsters represent only one example of a vast array of pre-existing sites and landscapes that were appropriated by aspirant Late Saxon lords. While the full connotations of such reuse cannot be explored in detail here, these strategies appear deeply-rooted in the worldview of an emerging aristocratic class who looked to past landscapes in particular as a means of bolstering nascent power.

In spite of the clear divergences detected in the project’s case studies, the presence of consistent elements implies that many early medieval lordly centres had essentially the same basic spatial grammar. Fundamental to all, of course, were halls for the use of lords and their families. The size of the compounds around these residences is remarkably uniform, measuring 1–2ha, sufficient to accommodate a hall and other domestic buildings surrounded by one or more yards. Excavation shows an increasing, if not commonplace, tendency to arrange buildings around a courtyard from the mid-tenth century, with the location of the hall evolving over time.Footnote ⁹⁰ These elite foci were either fully enclosed, as at sites such as Goltho, or when set within a larger enclosure, distinguished from other areas with a combination of banks and ditches, fences or hedges, as at Bosham and probably Hornby. In some instances circuits were built around a far greater area of up to 8ha, as at Weaverthorpe (East Yorkshire) and Almondsbury (Gloucestershire). The size of these compounds, sometimes 200m plus in diameter, surely indicates enclosure of non-elite occupation. While chronologies are often unclear or imprecise, these larger compounds should probably be considered distinctive settlement types in their own right, rather than being seen as a byproduct of a lordly enclave.

As the other fundamental component of lordly centres, churches were clearly of central importance to leading families, and in some parts of England they furnished nascent churchyards with funerary monuments. Private churches were rare before the 940s, but foundations grew rapidly thereafter as elites sought to demonstrate not only their piety, but also their rank and status through religious investment.Footnote ⁹¹ ‘Where Power Lies’ has amplified the observation, made by scholars previously, that churches associated with lordly centres characteristically occupy peripheral positions, either on the very edge of the compound or sometimes across the precinct boundary itself. ^{Footnote 92} Such positioning, it has been argued, indicates use of the church by the wider community, facilitating at least periodic admission for the purposes of worship and perhaps functions such as burial. In this case the ‘seigneurial zone’ was not hermetically sealed, but instead operated as a living and working part of, and interface with, the surrounding rural community.Footnote ⁹³ Yet, little explanation has been offered as to how wider communities may have utilised proprietary churches such as tower-naves, a building class consistently recognised as unsuitable for public worship almost since they were first identified in the late nineteenth century.Footnote ⁹⁴ One possibility is that towers were used by priests to orate to congregations gathered outside; the tower-nave at Earls Barton (Northamptonshire), which was part of a lordly complex, was used for just such a purpose until the 1950s (fig 19).

Fig 19. Louis Ewart, Vicar of Earls Barton, delivering a sermon to a crowd assembled in the churchyard of All Saints, 1954. The tower-nave may have been used by the wider community in such a way from its construction in the mid-eleventh century. Photograph: Earls Barton Museum of Village Life.

The example of lordly tower-naves cautions against interpreting liminal locations as a de facto indicator of communal use, given that on occasions they doubled up as gatehouses and arguably were first and foremost deliberately visible symbols of power and wealth.Footnote ⁹⁵ Indeed, the sense that lordly display may have influenced siting on the edge of compounds is furthered by evidence from places such as Barton-upon-Humber (Lincolnshire) and Chithurst (West Sussex), where churches were constructed upon prominent artificial earthen platforms, apparently to emphasise their topographic settings.Footnote ⁹⁶ From the late eleventh century, churches in these sorts of positions began to be incorporated into castles, typically surrounded by a bailey.Footnote ⁹⁷ Understanding location and use of churches at lordly centres is therefore more complex than it first appears, and siting buildings at the interface of private and public spheres is not an exclusive indicator of use by a wide population. Nevertheless, on balance it seems likely that even the most private churches still served at least some pastoral functions and that, somewhat paradoxically, the needs of emergent parishes was met by buildings whose genesis partly lay in personal or familial self-aggrandisement.

To the essential components of residence and church can be added watermills, often on the perimeters of complexes and serving as a means of social and economic power. Churches associated with residences were also furnished with one or more sundials, the minimal usability of which suggests that they were primarily commissioned as statements of lordly power. While features such as watermills and sundials are notoriously hard to date, from the wider landscape the pattern of casual losses of diagnostically high-status goods allows for slightly better chronological precision. For instance, within the PAS data a step-change is observable in the twelfth century, during which a greater number and diversity of prestige aristocratic objects such as seals and horse harness pendants begin to appear. These changes, observable in metal-detected assemblages of contemporary date across western Europe, have been explained by Robert Webley as being the product of ‘wider processes of lordly self-definition and social stratification’.Footnote ⁹⁸ Equestrian equipment often represents the majority of these assemblages and, while horse ownership was not always restricted to lordly classes across the period of study, decoration of animals with ornate fittings was evidently an elite practice alone. The regular dearth of metal-detected finds from lordly sites and their immediate surroundings is a reflection not only of the endurance of centres themselves that often continued to be occupied, but also their resilience as historic settlement foci. Instead, halos of artefacts are often found slightly further afield, that not only provide a window on evolving elite consumption but can be vital in locating previously unrecognised sites. Indeed, the potential of metal-detected finds to act as a rough proxy for lordly centres has been recognised by Gabor Thomas, who demonstrated the correlation between Late Saxon metalwork assemblages and the spring-line settlements of south-central Sussex.Footnote ⁹⁹ Thomas has also interrogated a number of hoards found in association with lordly centres, such as Bishopstone (East Sussex) and Poppleton (North Yorkshire), considering the possible symbolism behind their composition and deposition.Footnote ¹⁰⁰

Although watermills, sundials and prestige portable objects are not found as consistently as evidence for churches or residences, their regular association with lordly centres makes them important additional emblems of seigneurial activity. Their identification also serves to emphasise how a shift away from the traditional focus of defensibility facilitates a better understanding of the materiality of lordly centres in general. Instead, by considering these sites in the round, ‘Where Power Lies’ has begun the process of considering other, often overlooked, aspects of lordly self-promotion. These ‘technologies of power’ are typically poorly understood in contrast to the churches, castles and manor houses into which they were integrated; medieval sundials, for instance, lack even a rudimentary typology and are in desperate need of further research. Such efforts will integrate lordly centres more fully into the mainstream of academic dialogue and help to underscore the multifaceted and evolving role that these ‘theatres of consumption’ had in shaping medieval aristocratic life.