Evidence for assessing the idea of settlement hierarchies was considered inadequate twenty years ago (Hingley 1992, 34), and arguably remains so:
“The absence of well-defined chronological spans for the various types and scales of settlements still bedevils attempts to establish definitively whether there was a clear settlement hierarchy at any time during the pre-Roman Iron Age; social, economic, regional and chronological variations remain difficult to disentangle and the ambiguities of the data continue to foster a wide range of interpretations” (Armit & Ralston 2003, 182).
“The potential exists, therefore, to allow settlements in the Tees-Forth region to be ranked, but without a programme of large-scale excavation and a firm chronological framework it is unlikely to be realised” (Cunliffe 2005, 318).
“Attempts at defining regional zones in terms of settlement patterns must be regarded as tentative at best, since most regions show considerable diversity of forms” (Harding 2009, 246).
The term ‘hierarchies’ carries an inherent assumption of a rigidly stratified society, although currentmodels of Iron Age society are not necessarily consistent with this (theme 7.4); the term ‘patterns’ is more neutral. Any understanding of settlement hierarchy must depend upon a programme of investigation to create a paradigm for the sequential development of settlement type mainly through the refining, cumulatively, of site chronologies. Only in this way will it be possible to suggest which components of the prehistoric landscape refer to each other and which do not. Such sequences will, of necessity, be localised as the range of settlement forms is more varied in some areas than others. Most regions demonstrate a diversity of settlement form that can be recognised even if it cannot be satisfactorily explained. A more uniform settlement pattern within a region or locality at a given time has been argued to occur in some places (Armit 2002 on Barra and North Uist) and this is interpreted as evidence of an egalitarian society of independent landowning families, or at least one where social distinction was not manifested in architecture and all potential classes lived in the same structures, from elites to tenant farmers. Others believe that such local societies must have had a ‘big chief’ even if it cannot be detected in the settlement record (e.g. MacKie 2000, 105). A key methodological problem here is the possibility of low-status buildings without earthfast foundations, which would be hard to detect.
Of immediate relevance to assessing social differentiation and cohesion among settlements has been the presence or absence of enclosure. Social variation has often been seen as implicit in the labour obligations embedded in the construction and maintenance of settlement and hillfort enclosures (Hingley 1992, 32). More complex enclosure and multivallation has often been interpreted as a reflection of higher social status, although there has been much debate surrounding this issue (e.g. Banks 2000 on Woodend, although based substantially upon negative artefactual evidence). More recently, alternative models of competitive architectural expression and the demonstration and/or mutual expression of co-operation though mutual dependency have been advanced (Frodsham et al. 2007 on the Cheviots). Enclosed settlements in these models may be the product of less stratified societies than previously envisaged, where hillforts were seen as elite centres.
This debate about the social significance of open versus enclosed settlement is rather simplistic. There is a growing recognition that settlement layouts sometimes evolved in a complex fashion, incorporating both enclosed and unenclosed layouts – influential excavated examples include Broxmouth (Hill 1982a), Braidwood (Gannon 1999, on the basis of re-survey), Braehead (Ellis 2007), and observations accrued during the Traprain Law Environs Project (Haselgrove 2009).
There is no need to assume that enclosure was inherently superior or more desirable to open settlement or vice versa (Harding 2004, 66, 290 contra Hingley 1992, 33 who assumed open settlements formed the lowest level of society in East Lothian). The Roman coin hoards from Birnie, Moray, associated with what might otherwise have been considered a typical unenclosed timber roundhouse settlement is a case in point (Hunter 2002). This emphasises once more the need to investigate settlement patterns on a reqional basis.
The ‘clusters of communities’ model introduced by Hill (2006) allows for less hierarchical societies than had been previously generally considered. The model has been introduced in order to elucidate the East Lothian record by Lelong (2008) and Haselgrove (2009), but to what extent can archaeological correlates actually be identified? Lelong’s (2008) claim for communities a few kilometres across seems plausible but is nevertheless an assertion. The Tweed Valley does exhibit apparently discrete clusters of homesteads (Wise 2000), but as yet they lack excavation and any dating and thus demonstration of contemporaneity. It might be easier to identify clusters of communities in zones of preservation with established geographic boundaries (e.g. islands and isolated glens).
Biases of style and scale in regional research fundamentally affect the extent to which changing settlement patterns in different parts of Scotland can be modelled – i.e. whether areas have ‘research frameworks’, are ‘unsorted’ or are ‘black holes’ (Haselgrove et al. 2001, 23). An example of a well-researched area is the Western Isles, which has a developed and tolerably well understood settlement sequence in many respects for the first millennia BC and AD (Gilmour 2000; Armit 2003; Henderson 2007a), and is an area with a strong established ‘research framework’. Nevertheless there are still fundamental questions relating to prehistoric social structure in this area, such as:
- The context and chronology of the adoption of Atlantic roundhouse architecture (Gilmour 2000, 2002; Henderson 2007a; MacKie 2008, 2010).
- The social and chronological relationships between complex Atlantic roundhouses and wheelhouses (Armit 2003, 135; 2006; Gilmour 2000; Henderson 2000, 121; Harding 2004, 261-2 and 2009, 287; MacKie 2007, 2010), and other contemporary settlement forms (Gilmour 2002);
- The dating and role of promontory forts (Burgess 1999; Henderson 2007a).
In East Lothian a nuanced research framework is under construction based upon a range of recent programmes of investigation (Lelong 2008; Cowley 2009; Haselgrove 2009). Nevertheless Haselgrove has been very cautious in his approach to interpreting any diversity of the scale and design of settlement as an indicator of social structure and settlement hierarchy (and considerably more cautious than Lelong).
Macinnes (1982) outlined broad changes in the character of the settlement record in eastern Scotland as evidenced from the largely untapped cropmark record at that time. This, she saw, could reflect contrasting types of social organisation, land tenure, or political centralisation. The complexities of the topic have been explored, inter alia, by Davies (2007) and Cowley (2009), and remain a key area for research; extensive areas of the cropmark record are essentially unsorted (e.g. the Moray Firth; though see Jones et al. 1993).
In many areas (Haselgrove et al. 2001, 23) there is evidence of a diversity of prehistoric building and settlement form, where plausible hypotheses for the local or regional organisation of settlement and society can be proposed but where basic issues of chronology and function need to be resolved before any satisfactory model of settlement development can be accepted, (e.g. Isle of Skye, MacSween 1985; Sutherland and Caithness, Cowley 1999; Angus, Dunwell and Ralston 2008). In Argyll Harding (1997, 2004) and Armit (2009) have moved towards a partial construction of a settlement sequence, while in South-west Scotland (Cowley 2000, Henderson 2007a, 165-6; Cavers 2008) and Strathclyde (e.g. Alexander 2000) the links between the diverse components of the Iron Age settlement record are still unclear. Recent intensive work in East Lothian has provided one model of a means to establish a regional sequence. Other approaches take more of a keyhole approach, such as Cook’s work on Strathdon hillforts (2010), focussed on enclosure sequences, or Martin Wildgoose’s extensive sampling of hut circles in southern Skye, which targets central hearths to extract dating evidence (Wildgoose & Birch, pers comm.). Such approaches do not provide a rounded picture, nor will they recognise complexity in a site, which would cause misleading results. However, they can provide a reasonably rapid and cost-effective first-stage framework for subsequent testing. This would help to frame debate, and encourage others to tackle and challenge the model. In areas of predominantly cropmark archaeology, where the sites are being abraded year on year, such programmes may offer the only hope of extracting some information on overall settlement sequences at a broad-brush level. It is, however, much more difficult to characterise an amorphous (and probably long-lived) open settlement in this way than the specific moments of enclosure construction. The keyhole approach is not ideal, but would provide a means of obtaining basic sequences.
The enquiry must move to a situation where regions can be compared on a more equal footing. Some key ‘black holes’ sit between other better understood areas and would seem, therefore, to be immediate targets for research (e.g. Fife, between the Lothians and Angus; the western seaboard between Galloway and Argyll (perhaps, arguably, Cape Wrath!); the central and western Highlands. Sampling and dating large numbers of sites provides a valid first step in characterising sequences.
Was there a lower status, peasant, slave or landless element to societies that leaves little recognisable archaeological trace? Procedures for the detection of such an invisible component would have a profound impact on how the demography as well as the spatial organisation of prehistoric societies are modelled (e.g. Armit 2002, 2003; Gilmour 2002 on the Western Isles).
Open settlements appear under-represented in many parts of lowland Scotland except in pockets where cropmark production is good (e.g. parts of Angus) – this is often perceived as a visibility bias (e.g. East Lothian, Cowley 2009), but requires verification, possibly by an alternative method of remote sensing.
There is no overall clear picture regarding the role of ‘hillforts’, whether as tribal capitals, (seasonal) meeting places, elite residences, or other functions and it is likely that, anyway, their role varied across time and space (Armit 1997a, 50). There is no proven reason to see them as apex of a social triangle (Harding 2004, 290). Some have houses (Eildon Hill, Rideout et al. 1992; Traprain Law, Jobey 1976; Hownam, Piggott 1948) but again there is little obvious distinction between house sizes within hillforts or in comparison with other types of settlement (Harding 2009, 268-9). Promontory forts are also poorly understood.
This question impacts directly on social models for the Iron Age, a key research topic which settlement patterns inform (see theme 7.4).
Upland settlement, the dating and character of ‘hut circles‘, and the relationship to lowland settlement (e.g. in Sutherland, Cowley 1999; in North-east Scotland, Dunwell and Ralston 2008, RCAHMS 2007) may be the result of research biases fuelled by the focus on lowland cropmark excavations. How can this be redressed?
The relationships between timber and stone crannogs and island duns (Harding 2000), and their relationship to land-based settlement remain important research topics. An array of reasons for building on the water has been postulated, ranging from defence to maximising agricultural land onshore; but each site requires analysis within its own context.