6.4 Enclosed Places

This section includes features conventionally termed forts and enclosed settlements,  (excluding crannogs, which it can, however, be argued are enclosed by water). The term ‘enclosure‘ has been used multifariously as a classificatory term (by e.g. RCAHMS). This section has not included different scales of enclosure - sub-divisions within sites, or wider land / landscape enclosure (including linear banks, pit alignments, field systems) except where they can be specifically associated with forts and enclosed settlements. The social contexts of forts and enclosed settlements, hierarchical or otherwise, have been mentioned above.

There is no absolute distinction between ‘forts‘ and ‘enclosed settlements‘ or ‘enclosures‘. The range of sites present within these classificatory categories represents a continuum, from tiny enclosed promontories (e.g. Auchmithie, Ralston 1986) and settlement enclosures (e.g. Biel Water, Lelong and MacGregor 2008) through to large enclosed hilltops (e.g. Eildon Hill North, Roxburghs. Rideout et al. 1992) and massive enclosed promontories (e.g. the undated Mull of Galloway). Where there is evidence of habitation (be it permanent, seasonal or episodic), settlements ranging from individual households / farmsteads through to sizeable communities / villages are represented (Cunliffe‘s (2005) ‘defended homesteads‘). The construction and use of these enclosed places spans two millennia, from the late second or early first millennium BC through to the late first millennium AD (c. AD 800?). The dataset is highly diverse, although there are some regional differences in the character and range of enclosed places. With this diversity in scale, time and space in mind, the question of whether hillforts / enclosures form a sensible unit of analysis should be discussed.

Defining forts

Recent definitions of ‘fort‘ or ‘hillfort‘ (e.g. RCAHMS thesaurus, Halliday and Ralston 2010; RCAHMS 1997; Ralston 2006, 12) make reference to topographic advantage and enclosing works, but embody subjective considerations of defensive potential (from both the enclosing works and topography) and the extent of enclosed ground. Cunliffe (2006, 152) noted the common denominators of Wessex hillforts as enclosure, visibility and communal functions (the meanings of which may change over time). There is a bewildering array of schemes by which scholars at different times have attempted to sub-classify ‘forts‘ in different parts of Scotland, by attributes such as site morphology, topographic location, date, and perceived function (sometimes in combination).

Given the continuum of enclosure forms, there is a case for viewing the search for a definition of a ‘fort‘ as a ‘lost cause‘ (Halliday and Ralston 2010). However, there is evidently a considerable number of later prehistoric enclosed places that were not settlements (c.f. Hill 1996; ‘not farmsteads‘) in any conventional sense, although it does not follow that they were therefore solely, or even primarily, fortifications. Can enclosed places be classified into distinguishable ‘types‘ in time and/or space? If so how, and at what scale (local, regional, national)?

Explanatory frameworks for Iron Age forts and enclosures

In response, possibly implicitly, to the development of post-processual perspectives Ralston (1996, 145) called for a reassessment of the roles of enclosed sites within Scottish Iron Age communities, while Armit (1999, 72-3) identified a need for ‘grand narratives‘ for a Scottish Iron Age freed from its ‘unhealthy dependency‘ on models generated in southern England. This dependency is arguably a difficulty that has arisen through Scottish enclosed places being located on the northern periphery of a sub-continental scale phenomenon. To what extent either of these researchers‘ aims has been delivered, and whether interpretative models of the Scottish Iron Age are really free of southern influence (or indeed whether they should be), are outstanding questions.

'Invasionist' models are now very largely rejected (following Clark J.G.D. 1963 Antiquity). The primarily defensive / military explanations of hillforts and fortified settlements that underpinned such models have been widely criticised and reassessed. However, the defensive properties of enclosing works are still widely cited within the context of Scottish sites, although the degree of emphasis placed upon those properties varies widely between both scholars and archaeological context (Ralston 2006).

The influence of Cunliffe's ‘Danebury model‘ of hillforts, and before that Childe‘s ‘hill-top towns‘ and Feachem‘s ‘minor oppida‘, as elite-controlled central places within defined territories, forming the apex of economic, political and social hierarchies and with redistributive functions (e.g. Childe 1946; Feachem 1966), influenced the research of scholars working in Scotland at around the same time (e.g. Macinnes 1982, 1984a, 1984b). These processual approaches were set within the ‘Celtic‘ model of a hierarchical society (e.g. James 1993, 52-53). The size and scale of earthworks were discussed as socially controlled conventions (Cunliffe 2004, 50).

The origins, development and social context of enclosed places are now commonly explained with reference to a rather diffuse post-processual paradigm that has developed since the late 1980s, largely developed through reassessment of the Southern English evidence, and initially emerging from a rethinking of the ‘Danebury model‘ (reviewed recently by e.g. Armit 2007; Lock 2007 and Sharples 2010b from different standpoints). Post-processual perspectives incorporate several overlapping strands such as the rejection of the ‘generalizing‘ ‘Celtic Society‘ model (e.g. Hill 1996) and the ‘window on the past‘ approach, and the consideration of less hierarchical, more segmental later prehistoric societies with ‘clusters of communities‘ (Hill 2006). Others do not dispose of the ‘window on the Iron Age‘ approach, but regard it more as ‘of distorted glass‘ (Harding 2004, 294-7). Enclosing works have been interpreted as physical representations of metaphysical categorization (Lock 2007), with a wide potential range of practical and symbolic meanings (Collis 1996; Ralston 2006, 10-11), such as defining communities, as social defences, displaying status or isolation, and as expressions of power through the mobilisation of labour. Hillforts have been interpreted as ‘not farmsteads‘ (Hill 1996), as communal expressions or arenas with a potentially wide range of episodic / seasonal / semi-permanent uses (as suggested by the results of the Wessex Geophysical Project, Cunliffe 2006, 154) or as ‘extraordinary‘ places (Lock 2007). Cosmological and phenomenological approaches to understanding hillforts have also been attempted (e.g. Hamilton and Manley 2001), although not so far sustained for Scotland. This ‘pacification‘ of the Iron Age has been criticised by others (Sharples 1991; Armit 2007, James 2007), who variously consider warfare, other physical violence and related insecurity as endemic at this time.

Aspects of post-processual thinking can be seen influencing excavation site interpretations (e.g. Port Seton East and West, East Lothian, Haselgrove and McCullagh 2000; A1 excavations, Lelong and MacGregor 2008; Brown Caterthun, Angus Dunwell and Strachan 2007), and regionally based assessments of the social context of enclosed places and the structure of societies based on synthesised survey and new excavation data (e.g. East Lothian with ‘clusters of communities‘, Lelong 2008 and Haselgrove 2009; Anglo-Scottish Borders, Frodsham et al. 2007). They also appear in the interpretation of the potential functions of groups such as ‘oblong enclosures‘ and ‘promontory / headland‘ forts (Harding 2004).

Why enclosure?

The adoption of enclosure was a deliberate choice, and not all areas of Scotland enclosed places to the same degree. Armit and Ralston (2003, 193), for example, suggested that a perceived trend to enclosure could have been associated with factors such as an increasing emphasis on pastoral farming brought about by climatic deterioration, or a result of social change, but without offering a firm proposal. A lack of chronological control certainly hinders attempts to model contexts for the adoption of enclosure in different parts of Scotland, by not allowing identification of contemporary changes in patterns of settlement, society and land-use (following the approach adopted by Thomas (1997)). Can these issues be addressed in the Scottish Iron Age, and if so how? Or, as Lock (2007, 341) suggested for Wessex, does the ‘why hillforts‘ question lead to an inevitable dead end anyway?

The end of enclosure

There are various phases and places when enclosures were not used. Reasons proposed for the end of enclosure have included: the adoption of new forms of displaying status or ranking through e.g. personal ornaments and monumental houses; changes to the structure of society resulting in the reassertion of authority at a higher level; rejection of beliefs; and direct and indirect Roman influence. What factors lay behind the move away from enclosure in the first half of the first millennium AD, particularly as regards enclosed settlements, require more research.

The evidence base

Archaeologists are probably aware of most Iron Age ‘forts‘ surviving as earthworks in most parts of Scotland, but the gaps in the knowledge of smaller enclosures may be significant (e.g. Cowley 2000 on Galloway). Existing field survey plans of sites are of varying age and varying quality, accuracy and completeness. Resurvey of sites can lead to recognition of important details and nuances of character, complexity and sequence not previously appreciated (e.g. Gannon 1999 on Braidwood; RCAHMS 1997 and 2007). With a dearth of excavation in many parts of Scotland thorough and detailed survey evidence is crucial, although there is much in terms of complexity and dating that can be provided only through excavation (e.g. Brown Caterthun, Dunwell and Strachan 2007). There is a substantial cropmark record of forts and other enclosures, including in areas beyond the traditional excavational ‘honeypots‘. The synthesis and interpretation of this data is more advanced in some areas (e.g. East Lothian, Cowley 2009) than others. There has been very little geophysical investigation of forts and enclosed settlements, with a few notable exceptions (e.g. Traprain Law Environs Project, Newstead Research Project, Dent forthcoming). This is a cost-effective method for delivering important information on the character of sites (and particularly their internal arrangements), as the Wessex Hillforts Geophysical Project showed (Cunliffe 2006), albeit in different geological conditions.

For hillforts, Feachem (1966, 60) identified that much survey and little excavation had taken place, and this, broadly speaking, remains the case. Few sites have seen any excavation, and the excavation sample is biased towards south-eastern Scotland. The areas of the sites conventionally termed ‘hillfort‘ that has been examined by excavation is pitifully small, with exceptions such as Broxmouth (Hill 1982a), the sample proportion is probably substantially less than 5%. Most excavations have focused upon examining the enclosing works, with less emphasis on interiors and less still on exteriors (important because enclosing works need not delimit the extent of activity). Recent and ongoing research projects are undertaking limited sampling of numbers of ‘hillforts‘, usually within one localith, focused on enclosing work, in order to gather dating information (e.g. SERF, e.g. DES 2007, 155-6; 2008, 144; Hillforts of Strathdon, e.g Cook 2010; Rampart Scotland) - an unfashionable (when it is misconstrued) but useful and cost-effective approach (Haselgrove et al. 2001, 5). A considerable number of smaller enclosed sites have been more intensively examined, including settlements and ‘promontory forts‘. Some have been substantially or fully investigated in recent years as a result of development pressures (e.g. Port Seton, Haselgrove and McCullagh 2000; Braehead, Ellis 2007; Woodend, Banks 2000). An audit of all these categories of investigation, both published and as yet unpublished, may provide useful information on approaches to examination, focus of investigations, and discoveries made. Results of some important projects remain unpublished, such as the Newstead Research Project and work on the promontory forts of the Moray coast; publication of the critical Broxmouth sequence is expected soon.


  • Attempts to classify enclosed places have not been successful, but regionally-focussed studies seem to offer the best way forward rather than attempts at national-scale classifications.
  • The discipline would benefit from an overall review of the social context of enclosed places, based upon a detailed review of the evidence.
  • o    The lack of dating evidence for enclosed sites in many areas is a severe constraint in understanding them.
  • o    The lack of evidence for activities within such sites, due to limited work in enclosure interiors, is another severe constraint, as are the difficulties in connecting interior activity to enclosure sequences.
  • o    Geophysical survey offers a cost-effective approach to assessing the interiors of enclosures in favourable circumstances.
  •  Is the move away from enclosure in the early first millennium AD a general phenomenon, and what lay behind it?