Identifying outdoor assembly sites in early medieval England

Abstract Venues of outdoor assembly are an important type of archaeological site. Using the example of early medieval (Anglo-Saxon; 5th–11th centuries a.d.) meeting places in England we describe a new multidisciplinary method for identifying and characterizing such sites. This method employs place name studies, field survey, and phenomenological approaches such as viewshed, sound-mark, and landscape character recording. While each site may comprise a unique combination of landscape features, it is argued that by applying criteria of accessibility, distinctiveness, functionality, and location, important patterns in the characteristics of outdoor assembly places emerge. Our observations relating to Anglo-Saxon meeting places have relevance to other ephemeral sites. Archaeological fieldwork can benefit greatly by a rigorous application of evidence from place name studies and folklore/oral history to the question of outdoor assembly sites. Also, phenomenological approaches are important in assessing the choice of assembly places by past peoples.


Introduction
Temporary, popular gatherings in outdoor settings are common in societies of the past and present. Fairs, political rallies, festivals, sporting events, camps, theater, and battles are frequent events, but most leave few physical traces for archaeologists to recover. In some cases, outdoor events have taken on such importance that the sites where they took place are now invested with special significance and the need for heritage protection. The battlefields of Waterloo or Culloden (United Kingdom), sites of mass protest such as Greenham Common Women's Peace Camp (United Kingdom) or the Gdeim Izik protest camp (Western Sahara), and even the venues of music festivals such as Glastonbury (United Kingdom) or Woodstock (United States) possess great cultural and historical import, and have been duly commemorated and memorialized (Fiorato 2007;Schofield 2000: 144-148). In other cases, the site of an event has been lost or is only vaguely recorded, or several alternative locations exist in popular memory.
In many instances-whether the site of an outdoor event is memorialized or lost-archaeological examination of what survives has never taken place. Myriad difficulties exist in finding and defining the material remains associated with an outdoor event.
Identifying the venue may rely on folk memories or fragmentary descriptions, and less frequently on deduction based on archaeological field survey. While outdoor assembly encompasses a range of activities, its essential characteristic-the temporary gathering of people-may leave only ephemeral traces in the archaeological record. Often these are sites where recoverable materials, if they exist, are to be found in topsoil deposits rather than in stratified, buried contexts. Even where subsurface deposits exist, they are typically low-density horizons, and it is usually difficult to demonstrate stratigraphic associations between widely spaced features.
It follows that field techniques for analyzing horizontal relationships such as artifact spreads, landscape associations, or the topographic context of places may be the only ways of assessing these sites. Yet the intangible nature of these sitestransitory and rooted in communal experiencemay complicate their assessment, and it might be doubted that such sites can even be systematically analyzed and compared. Here we set out a new approach for the study of such sites by examining early medieval meeting places in England (5th-11th centuries A.D.). We argue, firstly, that a systematic use of place name evidence and folklore, alongside more traditional historical and cartographical analyses, is required. Secondly, we propose that the physical characterization and recording of these locations must include, in addition to archaeology, a range of supplementary non-intrusive methods including detailed toponymic and phenomenological analyses. In the following sections, we outline the methods used in recording over 250 such sites and summarize some of the key common qualities of these places as defined by the themes of accessibility, distinctiveness, functionality, and location. We argue that this innovative multidisciplinary method has important implications for future work in Britain and across early medieval Europe and elsewhere at open-air assembly places of all kinds, including judicial sites, fairs and temporary markets, battlefields, and places of religious gatherings, often known heretofore only from written sources (e.g., Baker and Brookes 2013c;Semple and Sanmark 2013;Williams in press).

Open-air assembly sites in early medieval England
In early medieval northern Europe several forms of public assembly are known, including royal and regional courts that sometimes met outdoors (the Frankish mallus or English witan), and those of local administrative assemblies that were held in regular intervals in open-air places, and often referred to as ''things'' or ''moots'' in English historiography. The latter were a fundamental element of government and society in early medieval (Anglo-Saxon) England. The earliest legal code produced in England in the court of the Kentish King AEthelberht (ca. A.D. 600) outlines an elaborate series of payments through which kindred could be compensated by the initial wrongdoer (and presumably his kindred) for injuries sustained as a result of robberies, brawls, or fights. Many disputes were settled without the intervention of officials; the role of kings was simply to administer justice when necessary and to uphold and clarify points of custom (Hudson 1996: 24-40). Significantly, the first clause of AEthelberht's code states that ''breaching the peace of a meeting [is to be paid for] with a two-fold compensation,'' (Liebermann 1903: 3;Whitelock 1968: 357) emphasizing the centrality of public assembly in mediating in-group conflict. Implicit in this clause is a sense of spatially and temporally defined parameters within which a meeting could be formally conducted (since it would be impossible otherwise to say when a breach had occurred), and it seems likely that such limits were tied to specific venues.
Many of these assembly sites can be identified by combining written, archaeological, and toponymic evidence. A crucial source is Domesday Book, the great survey of holdings and liabilities over much of England and parts of Wales completed in 1086 (Williams and Martin 1992). The information recorded by the Domesday survey includes geographic data on the estates, manors, and villas, and the administrative territories (hundreds and wapentakes) to which they belonged (FIG. 1). Each hundred had a meeting place, and while their locations are not explicitly stated in the Domesday survey, it is clear that many of the hundreds were named for the places where they assembled, usually open-air sites.
Open-air assemblies were important in the functioning of early medieval societies. As inferred from the Laws of King Edgar (943-975) and other written sources (Lieberman 1903: 192-215;Whitelock 1968: 393-401), they could fulfill either judiciary or legislative functions, were venues of local, regional, and national decision-making, and on occasion might be quasi-democratic or autocratic in form. As significant elements in the administrative, legal, and military institutions of the Anglo-Saxon state, openair assemblies, and by extension their venues, were also important places in the organization of the landscape. Yet questions remain about how they were chosen, how they were used, by whom, and for what purposes.
The Landscapes of Governance project, funded by The Leverhulme Trust, has begun to address these questions through the compilation of a list of more than 800 places of documented public assembly in Anglo-Saxon England. The project combines deskbased document research with groundtruthing to explore the nature and development of legal and political frameworks in an early state. From 2009 to 2012 fieldwork was carried out at more than 250 open-air assembly sites. Here we use case studies from this fieldwork to illustrate methods and characteristics for identifying the locations of early medieval assembly places. We conclude with a discussion of the relevance of these case studies to other cultural settings.

Finding and Characterizing Early Medieval Assembly Places
In many cases archaeology may refine our understanding of the location at which assemblies took place and furnish important evidence about the character of a site, but is often not the starting point for such research. Such gatherings appear regularly in written sources (e.g., legal and manorial documents, narrative texts, folklore) and are suggested by place names, enabling their identification, but the ephemeral nature of the human activities which took place means that there is little or no material evidence. Their study, therefore, requires a multidisciplinary approach. used to define parts of a landscape in a way that is topographically meaningful. Though the frame of reference for a given place name may vary, place names preserve a precise nugget of data about an aspect of that site at the time of naming-its physical appearance, location, function, etc. Research on the boundary clauses attached to Anglo-Saxon charters and on place names (Hooke 1981;Kitson 2008;Gelling and Cole 2000) emphasizes the precision with which early medieval people understood and described their surroundings. Therefore place names allow us to understand not just the physical characteristics of medieval settlements, but also the contemporary perceptions of the landscape and functional relationships within it.
Generally Domesday Book provides the earliest onomastic record for this category of place and is therefore important for identifying meeting places. Not all Domesday hundreds were necessarily named for their meeting places, however, and sometimes it is an alternative name for a hundred preserved in later medieval surveys that provides the crucial information. This is especially true when the name of the hundred contains specific locational information (a description of a lake, barrow, standing stone, ford, etc.). Such hundred names are common and the features they describe are sometimes identifiable. The barrow at which the freemen of Brictwoldesberg met is still visible at the place it has stood since the early medieval period or before, while the distinctive landscape form anticipated in the hundred name Holeford, ''hollow ford'' or ''ford in a hollow,'' has been convincingly associated with the hamlet of Ford (Gloucestershire) (Anderson 1939a: 18;Pantos 2002: 299-300). Nevertheless, close scrutiny must be paid to late medieval or modern records and antiquarian accounts. The stones that are believed to have marked the meeting places of Tibblestone (Gloucestershire)  2) and Hurstingstone (Huntingdonshire) hundreds have probably been relocated, slightly in the first case, substantially in the other (Pantos 2002: 310-311;Anderson 1934: 109;Meaney 1993: 80-81).
The correlation of medieval hundred names with modern place names and, eventually, with specific positions in the landscape demands a rigorous historical linguistic approach, which may then be built on by those from other fields. Anderson (1934Anderson ( , 1939aAnderson ( , 1939b set out the first, and until now, the only systematic national catalog and etymological discussion of hundred and wapentake names, setting the subject on a sound linguistic footing, and helping to identify the meeting places of many districts. Great strides have also been made by the English Place Name Society (EPNS), whose survey now partially or completely covers all but seven of the traditional shires of England. This vast archive of place name data and interpretations provides a foundation for the investigation of hundred names.
Beyond the major names of hundreds and wapentakes, minor place names, e.g., the names of farms or fields, can also be valuable (Cox 1971(Cox -1972Meaney 1993Meaney , 1997Pantos 2002;Renaud and Ridel 2000;Hobaek 2013). Modern EPNS county surveys provide detailed information on minor names and field names, and these, alongside local maps and boundary charters, can contribute to identifying hundred meeting places. In addition to surviving microtoponyms (i.e., field names) associated with a hundred's meeting place, there exists a wide range of toponyms that seem to identify the hundredal center or more general assembling of people. These might include modern minor names such as Hundred House, Shire Hill, or Court Oak (FIG. 3), or place names containing elements such as (ge)mōt, þing, spell, maeþel and so on-terms that refer to assembly, discussion, or speech-making (Smith 1956a(Smith : 110, 268-269, 1956b. These elements may lie behind modern names using Mot-, Mod-, or Mut-(e.g., Motborow, Modbury, both in Dorset and Mutlow from various counties), Thing-, Ding-, or Ting-(e.g., Thingoe in Suffolk, Dinghill in Leicestershire, and Tingrith in Bedfordshire), although the modern forms alone are not conclusive evidence.
Moreover, although sites whose names contain elements such as (ge)mōt, þing, spell, maeþel are likely to have been the foci of gatherings, it is not a logical step to assume that they were the sites of hundredal moots (Pantos 2002(Pantos : 176-461, 2004a. It is clear that the premodern landscape had a wide range of meeting places of different kinds, and belonging to different periods (e.g., Guthlaxton wapentake in Leicestershire) (Pantos 2002: 326-330).
Place name evidence also provides detailed and accurate descriptions of the early medieval landscape at the time of naming. They characterize assembly sites by their environment and physical appearance, and by the types of activities that took place in their vicinity. Toponymy can be an important guide to local infrastructure and communications, sometimes indicating the use of a particular route during the Anglo-Saxon period, and giving an impression of the strategic appreciation of the landscape (Baker and Brookes 2013a). They can also reflect the general layout and function of a site, hinting at a focus for religious, commercial, or leisure activities; they can contain more detailed information about the presence of benches or platforms, pits or mounds, or the types of leisure activities that took place (Pantos 2004b;Baker in press). They also provide a glimpse into early medieval spiritual associations and perceived links with mythical or historical figures (Brink 2004: 213-215;Williams 2006: 207;Baker 2014, in press).

Historical and folkloric traditions
For this toponymic approach to succeed, legal, tenurial, or fiscal texts are often required (in addition to Domesday Book). Yet, folklore and antiquarian surveys can also help to identify meeting places. Eighteenthcentury county surveys by Hutchins (1773Hutchins ( -1774, Hasted (1788Hasted ( -1799, and Nichols (1795and Nichols ( -1815 contain evidence that helps identify meeting places, often recording the location of courts, ancient trees, and other relevant details as they existed in the early modern period, and perhaps preserving older traditions. An example of the importance of folklore is provided by Combs Ditch hundred (Dorset) (FIG. 4), which takes its name from a large earthwork of the same name. A crier continued to summon the hundred court at Combs Ditch near Goschen and at the nearby Bloxworth crossroads as late as 1905, even though district meetings were by that time held at Anderson manor (Hutchins 1773-1774Guest 1851: 149;Dacombe 1935: 32). A footpath leading north from Anderson manor intersects with Combs Ditch close to the junction of the Goschen, Bloxworth, and Anderson parish boundaries, where the earthwork is at its highest elevation. This may have been the original site of the hundred moot, before it moved to a manor house.

The archaeological signature of assembly
The transient nature of outdoor assemblies means that the places where they occurred rarely feature in archaeological literature. Nevertheless, the investigation of two other types of open-air gathering sitestemporary prehistoric camps and battlefields-has generated hypotheses and methods relevant for the study of early medieval assembly places.
Archaeological materials from Palaeolithic camps are fragmentary and spaced widely across a landscape, an aspect relevant to the study of open-air assembly. As a result, it is difficult to demonstrate the contemporaneity of activities; some features may result from repeated activities, others from an isolated event (Stern 1993: 215;McNabb 1998: 15-16). In order to address this problem, archaeologists Botloe's Piece. Lying directly between these is ''Hundred Field.'' B) The second element in Botloe seems to be OE hlā w ''mound,'' apparently in reference to a lost tumulus, or perhaps to the distinctive mound-like form of Hundred Field. The field is dome-shaped. First Edition OS map ß Crown Copyright and Landmark Information Group Limited 1873. All rights reserved. might employ a landscape (or ''off-site'') approach to contextualize the excavated deposits from various sites (Stern 1993: 219;Potts et al. 1999: 786;Pope et al. 2009: 261). Following this method, research at Olduvai Gorge (Tanzania) has determined that different types of activities took place across the basin, with frequent reuse of some locations (McNabb 1998). At the Olorgesailie Basin in Kenya, such an approach has helped to define persistent activity patterns associated with certain geographic settings (Potts et al. 1999). Further explanation of why these locations attracted persistent activity has focused on microtopography and vegetation through detailed environmental reconstruction (Kroll and Isaac 1984: 27-28;Pope et al. 2009: 261).
These Palaeolithic studies utilized a two-pronged approach-detailed analysis of artifact assemblages and topographic reconstruction-which has been also advocated by battlefield archaeologists. Following Scott and colleagues (1989;Fox and Scott 1991;Fox 1993) in their work since 1983 at the Little Bighorn battle site in Montana Territory, emphasis has been given to understanding the spatial clustering of features and materials across the landscape; as with the Palaeolithic examples, human agency is read from differences in the frequencies, variations, and proportions of artifacts within individual scatters (Fox and Scott 1991: 94). For example, a plot of cartridge cases can be used to trace the positions and movements of individual weapons across the field of battle. Foard (2009) analyzed the English War of the Roses battlefield of Bosworth using detailed environmental reconstruction to locate the site of the battle. Ground and aerial surveys as well as environmental sampling (soil mapping, and analysis of peat deposits) enabled the reconstruction of 14th-century land use patterns that are consistent with the archaeological evidence and contemporary written accounts of the battle. The analysis of projectiles and dress fittings recovered from the plowsoil demonstrates the value of unstratified surface materials (Foard 2009).
Surface materials can also be used to identify open-air assembly sites. The hundred court was a place for legal and administrative activities, but Anglo-Saxon law codes also emphasize its role in regulating trade, with large transactions expressly forbidden in one of AEthelstan's codes unless done ''in the witness of the reeves'' in a public meeting (II AEthelstan 12) (Liebermann 1903: 156;Whitelock 1968: 384). As a consequence, hundred meeting places frequently became sites of fairs and trading too. Archaeologically, the signature for these activities might be reflected in patterns of casual coin loss; there is potential overlap with a class of sites from A.D. 650-900 known as ''productive sites'' of concentrated coins and metal materials identified by metal detectorists (Pestell and Ulmschneider 2003;Arthur 2000: 427). The Portable Antiquities Scheme (PAS) (2014) and Early Medieval Corpus of Coin Finds (EMC) (2014) can be used to plot single coin finds, and in some cases concentrations of metal materials can be positively correlated with assembly sites. At Tan Hill in Wiltshire, for example, medieval fairs were held at least by 1499. Reynolds (Pollard and Reynolds 2002: 254;Chandler 1991: 98) argues that the alternative name, Charlborough (1499; OE ceorlabeorg ''hill or mound of the peasants''), implies an even earlier association with groups of peasants, and may mark it as the early meeting place for the Domesday hundred of Studfold (note also Swanborough hundred in Wiltshire, from OE swānabeorg ''hill or mound of the herdsmen/peasants'') (Gover et al. 1939: 320). The PAS records a number of finds near Tan Hill, including a fragment of a penny of Edward I or Edward II, two Jetton (coin tokens), and other items of metalwork. Other examples have recently been reported by Green (2012: 140-147) from Lissingleys in Lincolnshire and Ewelme in Oxfordshire (Mileson and Brookes 2014). In both cases extensive metal detecting of the sites has recovered early medieval metalwork, which combined with historical, toponymic, and cartographic analyses, supports the interpretation that they were once important administrative centers.
Certain excavated sites provide clues about places that may have functioned as similar temporary markets. Excavations at a site 2 km west of Eton Wick, at Dorney (Buckinghamshire), have revealed evidence for Middle Anglo-Saxon activity (Foreman et al. 2002). Across the three areas of excavation, 123 features dating to the 8th century were recorded, a large number of which were pits of varying sizes filled with feces, animal remains, plant remains, and small finds (including exotic imports), all providing evidence of craft production and agriculture. The formal arrangement of the features, the material culture, and the scarcity of occupational structures suggested to the excavators that this was the site of an open-air trading place operating for a brief period (Foreman et al. 2002). Although there is no hundredal meeting place associated with this site, the evidence hints at the kind of open-air gathering place that might occasionally have served as one.
The Dorney site can be compared with Iron Age ''specialized'' cooking pit fields in Norway (Gustafson et al. 2005;Skre 2007: 385-406) and grain-rich pits found in various Late Bronze Age and Iron Age (1000 B.C.-A.D.100) hillforts across Europe. Archaeobotanical analyses of the latter (Jones 1984;van der Veen and Jones 2006;Kreuz and Schäfer 2008;McClatchie 2009) suggest that the cereals in large amounts as well as the weed diversity represent the labors of different communities. These sites appear to have acted as centralized locations for harvested crops from different environments, some of which were consumed in feasting events when many communities were assembled, and others were perhaps stored for later use and redistribution.
Although archaeological evidence may exist, other lines of evidence are generally required to confirm identification of open-air gathering sites. One example is of the early Anglo-Saxon cemetery of Saltwood, near Folkestone in Kent (Booth et al. 2011). Excavation revealed 219 burials of late 5th-to 7th-century date buried across four plots, three of which surrounding a Bronze Age barrow on either side of an Iron Age trackway (FIG. 5). Four centuries later the site emerged as the meeting place of the local Domesday hundred, Heane (Heane Wood still stands less than 250 m to the southwest). A number of pits in the western and eastern ends of the excavated area and scattered surface materials attest to the sporadic, non-funerary activities at the site through the Middle and Late Anglo-Saxon periods after the cemeteries had ceased to be used. These findings likely track the transition of the site from an early community-used cemetery to a judicial site, a role it held until at least 1279.

Methods: Identifying Assembly Sites
Recent investigations of battlefields demonstrate the value of a phenomenological approach (Tilley 1994) to the study of open-air assembly sites. J. and P. Carmans' (2001, 2005a, 2005b) work on the battlefields of Europe-which identifies shared characteristics such as ''boundedness'' in the landscape, locations on low or high ground, and intervisibility with settlementspermits a comparative analysis of the topography of battlefields and the extent to which it reflects changing methods of war. Full area archaeological field survey, remote sensing, paleoenvironmental reconstruction, and off-site sampling can be used to understand the environmental context, but these need to be combined with the experiential qualities of certain places. Considering how people in the past ''read'' and understood these places is essential for interpreting the material remains. For example, in their comparative analysis of ancient Greek battlefields, Carman and Carman (2005b) assert that ''the high visibility from urban centers suggests…that battle was also seen as a form of 'display' and that to be seen to fight was as important as the fighting itself' ' (2005b: 42).
When places can be linked to recorded events, scholars must consider the goals and choices that guided people to produce and participate in these events, and as such, they must also consider the cultural and historical structures that constrained past actions (Giddens 1986);Brumfiel 2000). This ''process of observation require[s]… time and a feeling for the place'' (Tilley 1994: 75). Glimpses of the former can occasionally be gained from written sources, or explored through the use of ethnographic analogy, but all interpretation relies on the concept of the ''fusion of horizons,'' a dialectic between past and present (Gadamer 1997: 302;Shanks and Tilley 1987: 103-115). Understanding the latter-the material world of social encounters-requires not only an appreciation of the manmade spaces of archaeological monuments, but also of the landscapes in which they are situated (Tilley 1994;Bradley 2000).
The practical experience of open-air assembly by its participants may be partially identified through our own encounters with a place. The Landscapes of Governance project team visited more than 250 possible sites of early medieval assembly and recorded the lines-of-sight with other monuments and features, viewsheds to and from meeting places, patterns of access and movement, the forms of architectural spaces, and acoustic conditions. Combined with traditional methods such as field survey and GIS analyses, these ob servations can be compared with other phenomenological studies of prehistoric monuments (e.g., Tilley 1996;Bradley 2002;Hamilton et al. 2006), early medieval burial sites (Williams 2006), and high medieval churches (Graves 2000) and aid in the identification of sites, their areal extents, and associated features.
Many of these observations are subjective, based on our experiences. It is not surprising, therefore, that previous attempts at uncovering past perceptions of open-air assembly venues have been criticized (Jones 1998;Fleming 1999). While acknowledging the problems, we take from these debates two important requirements: critical examination of ''historical contingency,'' (that the past is culturally and contextually specific) (Barrett and Ko 2009) and adherence to a rigorous and explicit methodology (Hamilton et al. 2006). We are fortunate in our study to be able to draw on a range of sources that allow us to reflect on medieval perceptions of the world (Altenberg 2003;Franklin 2006;Pluskowski 2006). Regarding the second requirement, our recording can be given a measure of objectivity by the framework of a pro forma (recording form), which provides an overarching structure to the assessment of a site while incorporating opportunities for comments on aspects that are unexpected or do not conform to established patterns.

The Landscapes of Governance pro forma
As stated above, the study of early medieval assembly sites demands an interdisciplinary approach; therefore, it is necessary to establish a framework within which different disciplines can operate and that harnesses their various outputs. The Landscapes of Governance approach emphasizes the strengths of each discipline and seeks to use the findings in a complementary fashion. Here, for example, place names provide one of the best methods for identifying the locations of meetings, but archaeological fieldwork is more likely to link activities with specific points in the landscape. Place names provide a picture of a site's functions, while archaeology defines the physical expression of thosec functions. Data from disparate sources are recorded in the field by means of a pro forma and stored in a database that facilitates advanced analysis.
The Landscapes of Governance pro forma is on four sheets of paper (Landscapes of Governance 2010). The first sheet allows us to compile archaeological, toponymic, and other data, as well as background information on the locations of outdoor assemblies. This information permits source criticism of written and place name evidence on a site-by-site basis. The second sheet lists field-based observations. The scale of the landscape and the presence of topographic features visible from a locale are recorded by drawing ''circular views,'' a method adapted from the Tavoliere-Garagano Prehistory Project (Hamilton et al. 2006) involving the circular depiction of the skyline and horizons to mark key topographic features at specific compass orientations. The drawings are supplemented by a 360 degree composite image of the landscape from photos taken from the site on which key features are later annotated. Given that an underlying function of public assembly sites is communication, we also record some of the audio qualities of sites, for example the distances at which the sound of a bell, individual words, or whole sentences can be distinguished by observers spaced around a central location (FIG. 6). The site's physical attributes are also listed: the proximity to route-ways (routes) and water sources, the presence/absence of distinguishing topographic features and ancient monuments, and the types of resources in a site catchment. Some observations are encoded on a third sheet, which consists of multiple-choice descriptive terms that provide a structured visual and experiential assessment of the site today. Modified from the Landscape Character Assessment forms used by the Countryside Agency (2002), these terms describe the character of the landscape from topography to texture. The final, fourth sheet illustrates a circular depiction of the skyline and horizons to mark key topographic features at specific compass orientations in earlier sections. The pro forma is designed to be as nonproscriptive as possible; the third sheet introduces a uniformity of description that permits comparative analysis and the fourth sheet ensures that observers are not limited in their perceptual evaluations.
The data collected as a result of desk-based research and field observations are entered into a relational database (Microsoft Access, and archived in both proprietary mdb as well as csv formats), allowing comparative analysis of sites in GIS (ArcGIS and GRASS), and statistical environments (R). By incorporating additional geographic datasets from historical, archaeological, and cartographic sources (e.g., geology and soils), the English Heritage National Mapping Project, and LiDAR data (FIG. 7), users can explore, query, and analyze these data and, most importantly, investigate the relationships between them. Parametric and nonparametric statistical analyses of these relationships has illuminated the patterning of sites relative to ancient route-ways (Brookes 2007b), viewsheds and lines-of-sight (Brookes 2012), administrative boundaries, and other monuments (Brookes in press). These complementary methods-toponymic and textual source criticism, phenomenological recording, and GIS analysis-help us to identify patterns in the data that can be used to evaluate assembly sites in terms of location, function, and form.

Characterizing Early Medieval Assembly Places
By using the pro forma to combine observations from site visits with archaeological, toponymic, geographical and folkloric data, analysis of the corpus of openair assembly places has produced some generalizable qualities. The results of the approach outlined above, are summarized under the following headings below: accessibility, distinctiveness, functionality/practicality, and territorial centrality/liminality.

Accessibility
Perhaps the most important feature of an assembly site is its accessibility. In most cases, the function of such sites demanded proximity to major roads and local paths between central places and smaller estates. Thus, it is no surprise that access to the main routes of communication appears to have been one of the principal criteria that determined the locations of meeting places. In some cases, the central position of a meeting place within the surrounding infrastructure can be demonstrated. The meeting place of the hundred of Swanborough in Wiltshire-thought to be the Swanabeorh of a charter of A.D. 987 (Sawyer 1968: no. 865)-is a low earthwork close to the junction of several routes mentioned in a number of 10th-century documents (Semple and Langlands 2001: 240-241). Similarly, the stone marking the meeting place of Kinwardstone hundred in eastern Wiltshire stood at the crossroads of the major northsouth and east-west route-ways (Langlands 2013). Therefore, identifying the relationships of assembly sites with the communication routes is perhaps the best framework for categorizing them.
One category consists of hundred meeting places located at significant points on major routes. Significant points include the intersection of two or more routes or a road and a stream (fords routinely feature in the names of hundreds and their probable meeting places), or a marked change in the direction or incline of a path. The suggested meeting place of Northstow hundred (Cambridgeshire) is at a local apex on the Roman Road from Cambridge to Godmanchester (Route 24) (Margary 1973), while Normancross hundred (Huntingdonshire) seems to have had a meeting place in the vicinity of a major bend of Ermine Street, at the top of a relatively steep ascent (Margary 1973: 205-206).
A second category of assembly sites includes those that overlook or dominate routes. At Pathlow in Warwickshire, the (now lost) mound overlooked an important early route between Henley-in-Arden and Stratford-upon-Avon (both Warwickshire). The probable meeting place of Fernecumbe hundred is located on higher ground approximately 500 m to the south of the Roman road from Alcester to Stratfordupon-Avon (Pantos 2002: 445).  (Lennon 2012). One of these is named in a 9th century charter as ''Cuheard's path'' (Sawyer 1968: cat no. 756), and in 18th century documents as Hare Path (Herepath) (probably from OE here-paeð ''army path'') (Lennon 2012: 109-110). The junction of this track and a Roman road, also visible in the LiDAR image, is marked by an ancient oak, perhaps the location of the Kinweardstone which gives the hundred its name. Image courtesy of the Forestry Commission (FC), based on FC and Unit for Landscape Modelling data. A third category is formed by upland sites such as Wittantree (OE ''tree of the wise men, councillors'') in Gloucestershire, which was possibly the meeting place of Bisley hundred. Such sites were not inaccessible, but located a distance away from the principal lines of long distance communication. Wittantree lies 300 m north of the Calfway, part of the ''great road'' of the 13th century, and 600 m northeast of a medieval holloway (a way, path, or road through a cutting) between Painwick and Cirencester (Pantos 2002: 280). The upland sites were probably located in communal grazing lands that bordered on or were accessible from the main routes, perhaps by minor but long-established tracks for people and herds.
For some assembly sites inaccessibility might have been a criterion, providing a level of secrecy for sacred or ceremonial activities, or emphasizing the importance of the site by imposing an awkward and time-consuming approach on those seeking access to it. One class of such meeting places that illustrates concepts of spatial ''depth'' (Hillier and Hanson 1984) are ''hanging promontory'' sites such as Moot Hill Piece (Dorset) and Botloe (Gloucestershire) (Baker and Brookes 2013c). Access to these meeting places becomes increasingly reduced (FIG. 8). They are domed hillocks up to 100 m in diameter located on spurs of land protruding below crests of higher ground. The high ground is often the location of a junction of several parish boundaries, and the site is always marked by at least one well-worn holloway descending downslope beside the promontory. The domed platform is accessed by ascending the holloway and then negotiating the spur linking the hillside and the promontory. Hanging promontory sites resemble Scottish assembly sites, such as Law Ting Holm, the Tingwall lawthing in Shetland, where the assembly took place on a small piece of land linked by a narrow isthmus to the lake shoreline, and accessible only to the most important assembly members (Coolen and Mehler 2011: 9-11).

Distinctiveness
Another approach to conceptualizing assembly sites is by the natural or monumental dominance of their settings. Early medieval meeting places were often distant from the main areas of settlement, but located at recognizable points in the landscape. Place names emphasize their connection with distinctive topography, trees, or other vegetation, or in other cases with manmade monuments like mounds or crosses. Some of these features may have had a functional utility for the proceedings (see below),  but in most cases they appear to be signposts for specific locations in the landscape.
Meeting places of some hundreds were in upland locations with commanding views over surrounding districts. The possible meeting places of Bisley and Langtree hundreds in Gloucestershire (Pantos 2002: 279-281, 302-303) and Street in Kent (Anderson 1939b: 137) display this characteristic. A few other sites possess such dominating views that this must have been a consideration in the choice of location. Spelsbury, next to Kiftsgate Court is an impressive example of a hanging promontory and affords commanding views over large tracts of the Gloucestershire landscape. Mutlow, in Fulbourn (Cambridgeshire), is also in a prominent position. These may have been hundred meeting places (Meaney 1993;Pantos 2002: 287-288, 315), but closer scrutiny suggests that they were alternatively (or simultaneously) meeting places for wider districts, their administrative importance perhaps reinforced by their impressive locations.
In some instances, the location of a hundred moot was not directly on an impressive landscape feature, but in close proximity to one which provided a marker for people traveling to the site and a backdrop to the meetings. For example, the meeting place of Heane hundred in Kent is close to the foot of a distinctive natural conical eminence known as Summerhouse Hill (FIG. 9A). Picked Hill beside the meeting place of Swanborough hundred is similar in profile (FIG. 9B). In both cases these dominating landmarks are associated with a large number of ancient routes, datable at least to the Iron Age, which link chalk uplands with areas of more productive soils. The distinctive combinations of route-ways, meeting places, and dominating landmarks identify these as liminal places that lie between areas of contrasting economic activity (agricultural and pastoral), and the communities that engaged in these activities. For the seasonal flow of pastoral transhumance, these places were fixed points of transition from one landscape-defined by economy, settlement, and culture-to another.
Natural monumentality may be mirrored or enhanced by the creation or reuse of imposing manmade markers. Hundredal gatherings seem to have taken place in Iron Age hill-forts at Badbury Rings and Eggardon in Dorset, while artificial mounds were focal features at Brightwells Barrow (Gloucestershire) and Pimperne Longbarrow (Dorset). Less imposing today are the remains of stones, crosses, and wooden posts used as markers for meeting places, but they may represent only small pieces of what were once more impressive constructions. Fragments of carved stone may indicate an early medieval high cross, market cross, or boundary stone. For example, a cross-arm fragment dating to the 9th or 10th century, recovered during the demolition of an old residence known as Styles' House, close to the crossing of the River Piddle in Puddletown Dorset, may be associated with the meeting place of Pydelan mentioned in a 10th-century source (Sawyer 1968: no. 830). Many charter boundaries and hundred names reference crosses, stones, and similar features and fieldwork can occasionally identify these as markers around which open-air gatherings occured. Folkloric memory might also be reflected in (and perhaps created by) the naming patterns of hundreds, including direct or indirect references to supernatural beings and deities (e.g., unor in Thunderlow [Essex], Wōden in Wenslow [Bedfordshire]). Heroic or mythical individuals were often commemorated in the names of meeting places. The Cwichelm of Scutchamer Knob (earlier Cwicelmeshlaewe A.D. 990-992) may be a reference to the early West Saxon king of the same name (Gelling 1973(Gelling -1976Williams 2006: 207; Baker in press), a significant person on the local psyche.

Functionality/practicality
The practical function of assembly sites, or the degree to which they were ''user-friendly,'' is a third analytical framework. As arenas of discussion and decision-making, assembly sites required a range of natural or artificial zones and structures in which separate groups could confer, or from which important announcements could be made. Moreover, assembly sites needed large catering capacities. An 11th-century account of the proceedings of a shire court assembly held in 1075 or 1076 on Penenden Heath near Maidstone reveals how long such gatherings could last-in that case, several days (Douglas and Greenaway 1953: 481-483). Given that all freemen were expected to attend hundredal moots and that some required an entourage of support staff, there were high demands for housing, feeding, and watering such large numbers of people and animals.

TOPOGRAPHY
Only a dozen or so English assembly sites have been investigated through detailed archaeological survey and excavation, but these have demonstrated the importance of mounds of prehistoric and medieval date. Excavations in 1977-1978 at a site now behind the public library in Milton Keynes city center have provided good evidence for a ''moot mound'' (Adkins and Petchey 1984: 243-251). In Domesday Book the area of Milton Keynes belonged to the Buckinghamshire hundred of Secklow-the meeting place of which was known to 18th-century antiquaries as the tumulus of Selly Hill. Excavation revealed a flattened mound of around 25 m in diameter, encircled by a ditch about 1 m across. The mound probably once stood at least 2 m high, but there was no evidence that it was ever used to mark a grave. The Secklow evidence suggests that some meeting mounds were artificially created, perhaps in the 10th or 11th centuries, when West Saxon kings probably implemented a range of administrative reforms and the creation of such a platform perhaps facilitated the delivery of pronouncements. Practical concerns are detectable in the topography of assembly sites. The many upland meeting places were probably situated in expanses of open pasture that could accommodate large crowds of people. Woodland locations might be included here too; hundred names with lēah, denn, or grāf, for instance, offer clues. Meetings that took place within zones of pastoral activity may have been preferred in order to avoid damaging crops. The locations of some meeting places at the gates of major settlements or at ecclesiastical and other high-status compounds (e.g., Westgate, Canterbury, Kent) may have been symbolic, but may also have involved a practical decision to keep potentially unruly crowds outside (Baker and Brookes 2013d).
Defining a space as being suitable for large gatherings is not simply a matter of assessing its size. For a site where speech-making is anticipated, acoustics are a factor, and it is likely that natural amphitheaters were selected for this reason. Aristotle claims that ''an urban space of assembly should be only as large as a shouting human voice can make itself heard in'' (Sennett 1990: 135). Natural bowlshaped arenas, or wider areas in which shallow depressions provide smaller, sheltered venues for closed discussion (Pantos 2004b: 161) may have appealed to those seeking to create assembly sites. Hutchins (1773Hutchins ( -1774Anderson 1939a: 111-112) reports that the meetings of Uggescombe hundred (Dorset) took place at some pits. At the probable site of these gatherings, a welldefined depression, perhaps the remains of an early quarry, is still identifiable in the southwestern corner of Benecke Wood. Similarly, until the 18th century, meetings of the Bingham hundred (Nottinghamshire) took place in a bowl-shaped depression beside the Foss Way known as Moothouse Pit (Anderson 1934: 42;Pantos 2004b: 161, 163).
A water supply would also have been important, especially if traveling delegates required animals for their own transport and to carry paraphernalia. The many references in hundred names to features associated with water-fords, bridges, water-meadows-support this interpretation. Five hundred names end in OE mere (''lake, pool''), and 21 refer to running water (OE burna, brōc, ēa, and welle, which all mean ''stream''). If considerable numbers of animals were present (for transport or to be slaughtered for feasting), stock enclosures would have been necessary; a remarkable number of hundred names refer to such features (Baker 2014).

CEMETERIES, SHRINES, AND TEMPLES
Apart from judicial and administrative activities, it is likely that these places also hosted symbolic and ritual assemblies. This is demonstrated by the coincidence of many documented meeting places with pagan burial sites of the 5th-8th centuries (e.g., Saltwood, discussed above) (Brookes in press). A similar continuity of symbolic functions might also explain the close association of some hundred meeting places with sites suggested on archaeological and toponymic grounds to be Roman temples or former pagan shrines. Place names containing the elements wēoh, wīg ''idol, shrine''-including Wye hundred, Kent-and hearg ''heathen temple'' may support this interpretation. The relationship with ancestral beings (supernatural or real) is also invoked by some hundred names and local traditions (e.g., Thunderlow and Wenslow, discussed above), while Easwrithe meeting place in Sussex may mean ''thicket of the gods'' (Anderson 1939b: 80).
These symbolic associations remain conjectural. Indeed, sites associated with local folk beliefs probably took on new associations over long periods of use. The reuse of ancient monuments as assembly places might be interpreted as a dialogue with the past, conferring legitimacy and authority to proceedings or oath-taking rituals. The erection of new monuments on ancient sacred sites, by contrast, might represent the reincorporation or realignment of the landscape within civil society. Certainly, regular meetings for judicial and administrative purposes could be associated with other types of communal activities such as marketing and sports. Fairs are documented at some early medieval meeting places, such as at Hinckford (Essex); the hundred name Gainfield (Berkshire) is probably from OE gamenafeld ''open land of games'' (Anderson 1939a: 211-212;Gelling 1973Gelling -1976. Military mustering, whether for campaigning or to review the troops under arms, also appears to have taken place at hundred and shire moots, and there is a correlation between places of assembly and recorded battlefields (Baker and Brookes 2013a).

Territorial centrality/liminality
Domesday Book tells us that much of England was already subdivided into administrative districts by 1086. These administrative territories can be reconstructed from evidence in the same source, supplemented with estate records and parish boundaries, to create a map of the administrative subdivisions of England at the time of William I (Thorn 1992). Locational analyses demonstrate that meeting places may be in central locations, and/or close to the boundary of a hundred, shire, parish, or estate. Gelling (1978) argues that a hundred meeting place was typically located in ''a sort of 'no-man's-land,' as far away as possible from the settlements of the community it served and on the boundary between two or more estates' ' (1978: 210), and indeed certain types of meeting places are correlated with district boundaries. For instance, in Guthlaxton wapentake (Leicestershire) both Spelthorn and a possible moot mound at Shackerstone are located on the boundaries of neighboring wapentakes. One reason may have been to ensure the neutrality of places whose core function was for mediation; these locations were perhaps considered common to all parties but particular to none (Pantos 2002: 129-134).
Considering this spatial pattern, it is worth noting that few hundred names refer to boundary locations (Baker and Brookes in press). This has two important implications. First, it probably indicates that the parish boundaries, which are often close to meeting places, are administrative features and postdate the establishment of the assembly sites themselves. Anglo-Saxon charter boundary clauses demonstrate that some fixed, linear estate boundaries existed by the 11th century, but much land must still have been divided in terms of the limits of exploitation rather than territory. Secondly, and perhaps more significantly, the locations themselves, though neutral, were not perceived as peripheral; they were places on the edge, but were nevertheless focal, and were not described as borderlands by those coining the names.
Different scales of liminal neutrality are evident in the data. The boundaries of kingdoms and shires were sometimes the locations of large political assemblies, such as military musters, major church councils, and meetings of royalty and nobles. The site of a meeting between King Cnut, King Edmund, and the witan in A.D. 1016 appears to have taken place on an island called Olanige in the middle of the River Severn on the boundary of Worcestershire and Gloucestershire, which was only accessible by fishing boat (Anglo-Saxon Chronicle 1016 manuscripts DEF; Harris 1992). Meeting places at the boundaries of less important units lay within the limits of hundreds in places that were deemed to be neutral. For example, the probable location of Wetherley hundred meeting place (Cambridgeshire) is close to the odd junction of Orwell, Little Eversden, Harlton, and Barrington parishes. The importance of communal land may be reflected in the post-Domesday hundred name of Manhood (Sussex), earlier Manwuða (1170), Mannewude (1230), ''common wood'' (OE (ge)maene and wudu) (Anderson 1939b: 74-75).
Retrogressive analyses (comparison of maps drawn up at different dates) of the form and size of administrative territories in conjunction with the study of meeting places themselves hint at additional patterns. The form and regularity of hundredal geography in Northamptonshire and Surrey, for example, suggests a deliberate policy in the layout of hundreds, perhaps during the 10th and 11th centuries. Huntingdonshire is subdivided into four equal parts with meeting places of similar types in each (FIG. 10). It is unclear whether this is the product of Mercian, Viking, or West Saxon authority, but it seems to reflect state-level administration (Baker and Brookes 2013b). In parts of England, on the other hand, the hundredal geography is related to older administrative units, which suggests a long evolution. In these areas the administrative organization of the 11th century may have been superimposed onto earlier groupings. Padel's (2010)  Open-air assemblies are one example of this development. Sites of judicial practices are another, for which archaeological evidence from early medieval England, including execution burials and gallows, has recently been collated (Reynolds 2009a). Reynolds (2009b) argues that the earliest manifestations of such practices are correlated with the boundaries of major political entities of the later 7th and early 8th centuries. Sites of capital punishment are closely associated with those of legal assembly; they are often separate from meeting places, but in highly visible locations within their viewsheds at the edges of hundred territories. Lawmaking, trials, ordeals, and executions were fundamental to the experiences of early medieval peoples.

Conclusions
The study of outdoor assemblies is of major importance to understanding past societies. The common occurrence of significant events at places of open-air assembly is not incidental. Sociologists, political philosophers, historians, and anthropologists have emphasized the importance of public gatherings in shaping civil society and political order. Outdoor gatherings for political purposes are archetypical of the ''public realm'' and broader notions of political discourse (Habermas 1989). Public assemblies are free, open, and accessible to all members of society (at least in principle) and provide places to foster notions of civility and cultural community. For Arendt, the formal development of the public realm and government emerged from the gathering together of people. ''The political realm,'' she suggests, ''rises directly out of acting together'' (Arendt 1958: 198-199). The corollary of this development is the notion of public space. The Greek polis, the medieval city, and urban planning during the Enlightenment encompassed places of human action (Sennett 1990: 135). Public space enabled participatory democracy and in the story of western political development, the creation and use of these places was of pivotal importance (Neal 2010: 4-10).
Open-air assembly is just one manifestation of the public realm. To Thomas Jefferson (1955[1785: 26-72), breathing freely was a metaphor for public freedom and this link between outdoor public space and an open society is a recurrent theme of sociopolitical theories since the Enlightenment (Sennett 1992). Ethnographic and historical sources are replete with examples of outdoor gatherings at the heart of the political community. Many tribal societies had elaborate dispute settlement systems (van der Dennen 1998;Fukuyama 2011: 255), sometimes with formalized places of assembly where feuds could be settled and the business of the tribe enacted, such as the kgolta of the Tswana (Schapera 1994: 80-83) or the bora grounds that ''anchored'' ceremonies of Australian Aborigine Dreamtime (Mulvaney 1970: 211-215;Flood 1983: 274). Studies of contemporary public spaces have emphasized how these sites also become facilitators of civil order as locales of power and resistance, theater and performance (Orum and Neal 2010;Low 2000). Thus, ceremonial gatherings of the Australian Aborigines enable participants to interact with Dreamtime through dance, music, and costume, while open spaces such as the National Mall in Washington, D.C. or Trafalgar Square in London have become sites of political resistance and activism.
The occasional and temporary nature of an outdoor assembly in Anglo-Saxon England makes its material signature difficult to identify. We have sought to demonstrate that its impression on the landscape, on the contrary, is deep and long-lived and can be detected using a multidisciplinary approach. The locations of meeting places and of the types of communal activities they hosted can survive in place names and in historical and folkloric traditions for centuries. The signature of these sites is not only physical, but also verbal too. Communal memory, preserved as place names and folklore in the oral and written landscape can be used alongside topographic and archaeological observations to identify such places. Each discipline reveals information about an element of the historical landscape and can reinforce interpretations while maintaining important independence.
Assembly sites were selected by people in the past with specific functions in mind, and within established historical and ideological parameters. This may have produced uniformity in the criteria by which a locality was judged to be appropriate for assemblies. Practical and ceremonial needs-including visibility and communality, ease of access and identification, acoustic and topographical compatibility-may all have contributed to the choices of locations and created a distinct typology of assembly sites. Some of the characteristics of meeting places, for example their accessibility or recognizability, had little to do with the public meetings, but may still have given their names to administrative territories. In other cases, different activities associated with a meeting may have taken place in areas around a named locale.
Our method for studying relatively intangible aspects of the human past, such as political, spiritual, and sociocultural processes, is not just through historical records, which are sometimes nonexistent, but also through archaeology and local linguistic and folkloric traditions. This approach can also contribute to the study of pilgrimage, adding to our understanding of sites of veneration as well as in the wider ritual and logistical landscapes in which these acts took place (Coleman and Elsner 1994: 77-78;Silverman 1994: 13;Stopford 1994: 59-61, 63-68;Webb 2000Webb : 215-232, 2002Petersen et al. 2012: 213). Pilgrimage is likely associated with the kinds of folkloric and toponymic commemorations that we highlight here (Hammond and Bobo 1994: 19;Webb 2000Webb : 215, 2002. Our method has a wide range of applications within Anglo-Saxon studies, and more broadly in the study of poorly documented, proto-or prehistoric societies. We focus on Anglo-Saxon and Anglo-Scandinavian examples, but even in England some place names contain pre-English references to assembly locations. Liss in Hampshire, for example, which is a pre-English place name containing British *lisso-''main place in a district, a court'' (Smith 1956b: 25;Coates 1989: 109) may indicate that traces of early administrative organization can survive significant political and cultural changes. Folk memory, place names, adaptations to the landscape, and material culture are elements of any society at any given period. The multidisciplinary method set out here is transferable across cultures, time, and fields of study.

Acknowledgments
We are grateful to Andrew Reynolds, Barbara Yorke, and Jayne Carroll for reading and commenting on drafts of this manuscript. In addition, we would like to express our thanks to all those who have provided feedback on various aspects of the project at conferences and other public events. The Landscapes of Governance project website is: http://www.ucl. ac.uk/archaeology/research/projects/assembly. John Baker (Ph.D. 2001, University of Birmingham) is Senior Research Fellow at the Institute for Name-Studies, University of Nottingham. His research interests include landscape history, and in particular the role of place names in multidisciplinary approaches.
Stuart Brookes (Ph.D. 2003, University College London) is Honorary Senior Lecturer at the UCL Institute of Archaeology and Research Associate in the Oxford University Faculty of History. His research interests include comparative landscape studies and state formation.