3.3 Settlement and the material culture of daily life

Within these larger landscapes what has traditionally been the backbone of all archaeological enquiry must also be reviewed, namely the material repertoire of life in the form of household objects and the artefactual record in general. The pursuit of a cultural biography of objects, now an entirely commonplace approach in archaeology by contrast to its somewhat radical perception in the late 1980s, seems a productive road to travel here. As with the above considerations in charting the material worlds of early Scotland, regionality and ethnic variation must be prominent factors, and it must be seen that the full range of theoretical tools are employed to create multi-facetted interpretations that reflect this most diverse of regions. Published research on medieval standing buildings in Scotland has largely been limited to castles and ecclesiastical buildings due to the accidents of survival. Opportunities for recording and research should be taken where urban buildings may be hidden behind later facades and rural structures incorporated into later country houses or remain concealed in deserted or degraded farmsteads.

Aside from the recognition of regional variation in the post-medieval settlement pattern (Fenton and Walker 1981, Bruntskill 1981, Naismith 1989, Atkinson 1995) little is known about medieval rural settlement in Scotland. This has led to such statements as ‘archaeology of the medieval farming communities is one of the greatest mysteries of our past’ (Yeoman 1995, 108). One reason for this is that while numerous prehistoric sites are known and many have been excavated, very few sites of the medieval period relating to the rural society have been found and excavated, especially in the Highlands.

The term ‘Medieval or Later Rural Settlement’ (MoLRS) was coined in the 1990s to cover pre-Improvement /pre-Clearance settlement in Scotland and also the medieval and post-Clearance settlements (Hingley 1993). An initial Historic Scotland sponsored seminar included useful ‘where we are now’ type reviews (especially Corser 1993, Dixon 1993 discussed below). There was some discussion of the ‘way forward’ which consisted of the creation of an advisory group, the main priorities of which were the preservation and management of the known resource, which is essentially post-medieval in date. Research into the unknown was considered secondary. A review of progress in MoLRS research ’10 Years On’, based on a joint conference sponsored by Historic Scotland and the Medieval Settlement Research Group (Govan 2003) highlighted the issues affecting the location of medieval settlement in both highland and lowland areas (Lelong 2003 and Dixon 2003).  The term ‘MoLRS’ has now being replaced with ‘Historic Rural Settlement’ (National Trust Research Seminar held in March 2007) which reflects an attempt to widen the scope of the studies, and a research framework was published by the Historic Rural Settlement Group in 2008 (Dalglish and Dixon 2008). In addition, the RCAHMS project  (sponsored by Historic Scotland) to record all deserted rural settlements shown as such on the 1st edition of OS 1:10650 map produced a corpus of 28000 sites across Scotland and may represent a valuable research resource in this regard.

One of the main problems bedevilling medieval settlement is the ephemeral nature of the evidence. The majority of structures occupied by the Scottish population in the medieval and into the post-medieval period may have been predominantly of organic materials such as timber, turf, soil, branches and peat (Walker et al. 1996).

A variety of approaches have been applied in recent years to the study of medieval/Post-medieval, pre-Improvement/post-Improvement settlement and their associated field systems. Work at Bragar and Gásig in Lewis utilised geophysical survey and phosphate analysis to tentatively identify potential hearths, walls, ditches and anthropogenic enhancement of soils usually as a result of manuring (Banks and Atkinson 2000). A semi-automated classification of field systems utilising computerised image recognition techniques, has been coupled with examination of soil signatures within identified functional areas, and followed by radiocarbon dating of features to identify and characterise relict field systems associated with settlement (Chrystall and McCullagh 2000). Morphological analyses of medieval and later field-systems recorded by RCAHMS field teams have been carried out and published, including the varieties of rig and assessments of their method of formation and period of origin.  (Dixon 1994; Halliday 2003). The approach of historians and historical geographers (Campbell 2000, 2002, 2004; Boardman 2006; Crawford 1983, Dodgson 1980, 1993, 2000; Whyte 1981, 1995, 2000) has highlighted the potential contribution of the historical records and historic maps to an understanding of the location and nature of rural settlement, the organisation of society, and the medieval economy. An approach which combines traditional archaeological survey with historical geography has also been developed most successfully in Scotland by the Royal Commission (e.g. RCAHMS 1997a, 2001 and 2007 and others) and by individual researchers (James 2009). Other studies have examined historical documents, place names and compared surviving remains of settlement and house types on Islay (Caldwell et al. 2000), and undertaken topographic survey and excavation of a site at Easter Raitts, Badenoch (Lelong and Wood 2000; Lelong forthcoming).

Early rural settlement studies were concerned with the architecture of upstanding vernacular houses and the numerous stone-built deserted settlements seen within the Scottish landscape – although these structures are now acknowledged to date no earlier than the mid-18^th century (Fenton and Walker 1981, Crawford 1983, RCAHMS 1992, 32-36). Some early work on vernacular architecture took place, fortuitously, in Argyll (e.g. Gailey 1962a; 1962b), including work prompted by an interest in contemporary thatched domestic houses (Sinclair 1953). Sinclair identified three ‘types’ which he termed (due to their broad geographic spread): ‘Dalriadic’; ‘Skye’; and ‘Hebridean’. Sinclair’s work was useful in that it recorded structures which have since disappeared and highlighted the variety of structural forms, (which he then over-simplified into three ‘types’) and thus recognised that rural vernacular housing was not a static remnant of the past. The more recent work by Bruce Walker and others in the field of vernacular building studies (Walker and McGregor 1993; Walker and MacGregor 1996; Walker, MacGregor and Little 1996, Dixon 2002a) has highlighted the range of building techniques and materials used in the late 18^th and 19^th centuries. This provides a useful way of understanding the likely requirements of earlier vernacular architecture.

Horace Fairhurst’s work provided a specifically archaeological element to the study of rural settlement that was lacking in earlier work (Fairhurst 1960, 1963).  His work at Lix in Perthshire (1971), Rosal in Sutherland (1969a) and Loch Glashan in Argyll (1969b) have been particularly useful in providing morphological information and dating evidence for pre-Improvement settlement. At both Lix and Rosal the visible ruins could be dated only as far back as the late 18^th and 19^th centuries despite the documentary evidence for earlier occupation in 1559 in the case of Lix (Drummond Papers 1569) and in 1269 for Rosal (Registrum Episcopatus Moraviensis 1837).

Fairhurst’s work was ahead of its time in that he was interested in the archaeology of rural settlement. Fairhurst identified a wide variety of structures in the landscape, as Gailey had done, but was only able to excavate small trenches across a small proportion of them and his analysis of the pottery assemblage was unsophisticated. However, he did identify stone foundations, turf-walled structures and sub-floor depressions which could have been the remains of medieval timber framed cruck-built buildings. The 1970s and 1980s saw increasing numbers of archaeological field surveys being undertaken in Scotland. While this greatly increased the amount of data recorded and enabled suggestions to be made about chronology and process based on the morphological differences, without excavation still no real chronological framework could be established.

Fieldwork was also undertaken in the late 1970s in Caithness and Sutherland (Mercer 1980) where three types of pre-Clearance settlement were identified: Numerous shielings near streams were set in large mounds of debris, which suggested a long period of use; Individual farmsteads of the immediate pre-Clearance period were characterised by compartmentalised long houses, some with bow-shaped walls perhaps indicating a Norse influence. These structures were associated with other smaller rectangular houses; finally there were small rectilinear and sub-rectilinear houses concentrated in large numbers and associated with large enclosures. This type was, however, found in only one area of south Sutherland at Dalchork where it is known from documentary evidence that ‘cottar towns’ existed.

A survey on the shores of Loch Tay, Perthshire, incorporated a study of historic maps including Pont, Roy’s Military map and an estate plan of 1769 which enabled recognition of four types of structures in this area (Morrison 1980): shielings, round or oval in shape, built of drystone or turf in the hills above the head dykes; low, hip-ended, drystone longhouses or byre-dwellings clustered together with accompanying outhouses, barns, kilns and smaller ‘cottars’ houses as depicted on a 1769 survey of the estate; clustered or isolated 19^th-century buildings as depicted on the 1^st edition O.S. map, but not depicted on the 1769 survey; and low-turf covered ‘rectangular and sub-rectangular structures and straggling field dykes’ not shown on either the 1769 plans or later maps, which could be pre-18^th-century settlements.

The RCAHMS have undertaken extensive archaeological surveys in Argyll between 1971 and 1992 but included only a small number of the more outstanding or exceptional remains of pre-Improvement settlement in these publications (RCAHMS 1971, 1975, 1980, 1982, 1984, 1988 and 1992). Such exceptional shieling sites included Talatoll (Kintyre) and Douglas Water (Loch Fyne). The site of Talatoll was unusually extensive in that there were at least 43 structures lying across open moorland (RCAHMS 1971, 200). These surveys have shown that shielings display variable construction – some are clearly little more than temporary huts, while others are much more substantial and are similar to the structures found in the townships. This may be evidence for more permanent occupation of the shieling grounds at some point in time. The illustration by Pennant of beehive and tepee shaped shielings on Jura in 1772 show how varied the construction of such structures could be (Pennant 1790).

Eldbotle, East Lothian. View of a medieval building under excavation at the deserted village of Eldbotle. This site has survived due to sand-blow covering the site. The superstructure of the buildings are typically of perishable materials that do not survive © Piers Dixon

In highland Perthshire, long turf and timber houses of the late first millennium have been identified by survey (RCAHMS 1990) and dated by excavation at Pitcarmick (Barrett and Downes 1994a; 1994b). Outside highland Perthshire, early medieval buildings are rarely found, with Viking settlements providing the best parallels for date in the Outer Isles. Lowland equivalents, however, have been difficult to locate or identify and there is little idea of either the buildings or the pattern of settlement at this time. The post-set buildings used as kiln barns, at the monastic site of Hoddom, Dumfriesshire, dating to the 7^th to 9^th centuries, are akin to middle Saxon houses in southern England although the excavators interpreted the post-holes as slots for crucks (Lowe 2006). This important site shows the potential, but more excavation of sites of this period is urgently needed. The early medieval phases found in one part of the site at Eldbotle suggests that known medieval village sites are a good starting point for locating settlements of this period in the south-east Lowlands, if not elsewhere.

The Ben Lawers Landscape History Project has been very productive in providing multi-disciplinary information on the rural settlement landscape of the central highlands. Although an excellent achievement in getting to grips with shielings and peripheral settlements on the outfield areas, less attention was devoted to the settlement foci and it still remains an unfulfilled aspiration to excavate a highland rural settlement focus in extenso. Recent survey work by RCAHMS as part of the Project has updated and refined this analysis, and shown that there is a degree of continuity in settlement from the mid-18^th century through the Improvement Period to the later-19^th century in the settlement foci (Boyle 2003, 2009).

Research into Medieval settlement in the Northern and Western Isles has increasingly focused on Scandinavian settlement. Pre-Norse settlement is characterised by cellular buildings in isolated farmsteads (Parker Pearson 2004; Dalglish 2011), for example, Buckquoy in Orkney (Ritchie 1977; Brundle et al. 2003). It is not a simple story of replacement by Norse settlement however as at many sites there is continuity of location and use, such as at Skaill, Shetland (Buteux 1997) and Buckquoy (Ritchie 1977; Brundle et al. 2003). Similar long-term patterns of settlement and use of resources – isolated farmsteads focused on mixed farming/fishing – also persisted over the Norse ‘transition’ at sites in South Uist, including Cille Pheadair (Brennand et al. 1997) and Bornais (Sharples 2005 and forthcoming). There are exceptions however, at sites such as Norwick, Shetland, where an early Viking settlement (possibly pre 950AD; Ballin Smith 2007) had no precursor. The nature of the settlement record pre- and post- Norse impacts is a current area of debate.

Apart from Whithorn, Scandinavian settlement appears to concentrate in the Northern and Western Isles (see Section 2.3 for further summary of Viking/Norse presence and influences). Later Norse influences resulted in long rectangular houses, sometimes with byres under the same roofs. Several sites on Shetland have been extensively excavated, including Jarlshof (Hamilton 1956; Dockrill, Bond and Batey 2004), the Udal on North Uist (Sharples 2005; forthcoming), The Biggings, (Crawford 1999), and Kebister (Owen and Lowe 1999). On Orkney, excavation has taken place on sites including the Brough of Deerness, (Morris with Emery 1986; Barrett and Slater 2009) and in West Mainland (Griffiths et al. 2007).

In the Western Isles, excavations at Cille Pheadair (Brennand et al. 1997) and Bornais (Sharples 2005 and forthcoming) have traced settlement evolution over the 11th to 14th centuries AD, illustrating the development of regional architectural traditions from the earlier longhouse form. Some settlements became more complex, with the addition of new buildings to sites such as Jarlshof (Ritchie 2003) possibly reflecting an extended family. The group of settlement mounds at Bornais, (Sharples 2005 and forthcoming) probably suggest a local power centre, as does the Brough of Birsay (Curle 1982; Hunter 1986; Morris 1996 and forthcoming) in Orkney. Later Norse sites demonstrate the concentration of political power, as at the Earldom sites in Orkney, including the Brough of Birsay (Hunter 1986; Morris 1996), the  Earl’s Bu, (Batey 1993), Westness, (Kaland 1993), and Tuquoy (Owen 1993 and forthcoming). The complexity of settlement hierarchies has also been explored at sites such as Sandwick (Bigelow 1985; 1987), Hamar and Underhoull (Bond et al. 2007, 2008), all in the Shetland Isles.

Extensive survey of the machair plain on the island of South Uist has been able to plot the distribution of Norse settlement mounds (Sharples and Parker Pearson 1997). The distribution of these settlements is comparable with the Iron Age settlement and indicates the relative continuity of settlement location on the island from the first millennium BC across the Viking period. However, it contrasts dramatically with the post medieval settlement pattern which appears to almost completely avoid the machair plain. There appears to be a major disruption in the settlement pattern in the 14^th to 15^th centuries which may be explained as either the result of climatic or political developments.

Some pre-Improvement settlements are found clustered around medieval hall-houses.  Such examples are Finlaggan, Islay, Ardtornish Castle (Morvern), Aros Castle (Mull) and Dun Ara Castle (Mull). The RCAHMS has assumed that the structures are contemporary with the hall-houses, implying a close relationship between the lord and at least some of the local population. However, without excavation it is impossible to say what the relationship is between the hall-house and the townships. It is probable that several of these structures post-date the medieval lordly occupation of the hall-house.

Springwood Park, Kelso, Scottish Borders. View of the row of 14th century buildings under excavation. This site has been preserved because the buildings were set into the slope and it has been covered by plough soil. The superstructure of the buildings was largely made of perishable materials © Piers Dixon

Surveys have also identified crannogs and fortified islands as potential types of medieval site. Underwater surveys of crannogs in Loch Tay (Dixon 1982; 1984), Loch Awe (Hardy, McArdle and Miles 1973; Dixon 1984; Morrison 1985; Holley 2000; Taylor 2003) and the Lake of Menteith (Henderson 1994) have generally found crannogs to date from the later prehistoric to the early-medieval periods. Dry land surveys have also shown that crannogs extend into south-west Scotland (Barber and Crone 1993) and the central Inner Hebrides (Holley 2000). Of the 23 radiocarbon dates from Scottish crannogs in south-west Scotland only one site (Lochrutton) produced two medieval dates of the 11^th to 13^th centuries (Crone 1993, 246) which suggests that in the south-west of Scotland re-use of crannogs in the medieval period was not a common phenomenon. In the west of Scotland, however, there is more evidence that crannogs were utilised for settlement well into the medieval period. At the Moss of Achnacree, Lorn, a crannog was excavated in the 19^th century and produced artefacts including two wooden double-sided combs, a wooden ladle, fragments of antler and skin shoe soles which were thought to be medieval in date. Sites classified as fortified islands may also have been originally built on crannogs as well as on suitably located rocky islands (RCAHMS 1975, 94-95). At several sites in Kintyre, Mull, Tiree, Coll, Loch Lomond and Mid-Argyll there is evidence for crannogs and islands being occupied by stone buildings (RCAHMS 1963; RCAHMS 1971; RCAHMS 1980). Several examples have been surveyed by the RCAHMS, as at Eilean na Circe, N Knapdale (RCAHMS 1992, 303), Eilean Tigh, S Knapdale (RCAHMS 1992, 303), Loch a’Bhàillidh, S Knapdale (RCAHMS 1992, 304) and Loch an Daimh, Craignish (RCAHMS 1992, 305). At Loch na Buaile, Tiree (RCAHMS 1980, 122) the island was occupied by an oval-shaped, turf-built rather than a stone built structure (which is similar to one found at MacEwan’s Castle, Cowal (see below) which is medieval in date). Some sites had outer revetments walls offering some degree of defence, which otherwise was provided by their island location. Documentary evidence suggests that these lightly fortified islands were often associated with clan chiefs and used as chiefly residences in the medieval period and persisted in use as refuges into the 16^th and 17^th centuries. A more recent survey by the RCAHMS found an example of a late-medieval farmstead occupying an ‘island’ at Eilean a’Bharain, Loch Tromlee (NMRS unpublished archive). This potentially important medieval site provides an example of a chiefly residence which may once have been more common throughout Argyll.

Plan of Finlaggan castle and chapel, Islay. The principal court of the Lordship of the Isles, Finlaggan was of particular importance during the 14th century. There is evidence that the island of Eilean Mor was artificially extended in the Medieval period, © HES

Even within the Lowlands, the remains of rural settlement are still relatively rare. One such site was excavated at Springwood Park, Kelso (Dixon 1998). This site was originally found due to scatters of medieval pottery within a ploughed field. It consisted of three periods, the first included building terraces, post-set structures and ditches, and the later phases included cruck-framed structures with clay walls, stone footings and cobbled areas. The settlement has been dated from the pottery and four coins to between the late 12^th and 14^th centuries. The walls of one of the later buildings was 1.2m wide, but only survived one course high. This site was located close to the royal burgh of Roxburgh, Roxburgh Castle and Kelso Abbey, which may explain why it was unusually rich in material culture (Martin and Oram 2007).

The study of rural settlement in the lowland shires of Scotland has been ill served by archaeology. The techniques available are different to those in upland and highland areas since upstanding remains are relatively rare. Furthermore, the best-preserved lowland village sites that have been excavated survive as a result of sand blow or soil drift. Few of these will ever be excavated, and it is necessary to use the techniques of field survey and landscape analysis aided by documentary sources as has been done by RCAHMS in Donside (RCAHMS 2007). In Donside it has been possible to identify a phase of 12^th-13^th century colonisation of villages (villa) by Flemish and Anglo-French incomers into an established Pictish settlement pattern of davochs, the basic territorial units of the northeast. A late- and post-medieval expansionary phase was also revealed by the splitting of many touns and the establishment of new farms in waste areas such as former hunting forests. No late medieval contraction was evident. The settlement pattern was also shown to be more complex than expected with a socially based hierarchy of estate centres, fermtouns, cottertouns and crofts. More regional settlement studies that use these techniques are clearly needed (RCAHMS 2007.

Turning to the study of the later use of hillforts, work in the Moray Firth (Craig Phadrig, Portknockie, Burghead) remains unpublished and it would be good to reinvigorate this series of projects emphasising the prehistoric context (as at Traprain Law) and the protohistoric settlement context (with the SERF project beginning to address this).  Whether protohistoric people are re-using a hillfort for defence or legitimation or are ‘making reference’ to earlier religious loyalities is perhaps more likely to emerge from studying the surrounding territory (see above) rather than the site itself. However, if a hillfort site is to be examined small-scale excavation must be superceded by a range of new approaches, including the use of survey techniques such as Lidar^{[see note 16]}   and laser scanning, TST^{[see note 17]}  survey, and area excavation.

Aerial photograph of Craig Phadrig, near Inverness. The hillfort was constructed during the Iron Age but was likely re-used, although the nature and dynamics of this are not fully understood, © RCAHMS.

The excavation of collapsed and compacted rubble deposits has made huge strides since Alcock’s pioneering work at South Cadbury (eg Cadbury Congresbury, and see Lelong 2007 for an example from Lothian).  But for the technique to succeed, a tower, spoil extraction systems, large on-site shelters and a reliable water supply to allow constant spraying are all desirable.  Therefore investigating a hillfort for its early historic role is not to be undertaken without a proper commitment of time and resources.

Few manorial estate centres of the 12^th to 14^th centuries have been examined and unless marked by moated enclosures or mottes are often difficult to locate. While the earthwork and stone castle sites that belonged to the aristocratic elite of the 12^th to 14^th centuries are more easily identified, those of the lesser knights and landowners have been poorly researched. There has been some consideration of the social aspects of castle use (Tabraham 1988; Rutherford 1998) and of their historical and social context (Dalglish 2002). The detailed work undertaken at Castle Tioram has been unprecedented, prompted by the owners desire to redevelop the site (Murray et al. unpub, Stell 2006). Stell’s work with respect to Tioram is a good example of analysis of the cultural significance of castles. However very little work has considered castles in Scotland from a cultural landscape perspective – Dunstaffnage being a notable, rare example (Breen and Forsythe 2008) as is the RCAHMS study of the landscape around Hermitage Castle (Dixon 1997, 352-4). Ireland provides something of a model approach here from which Scotland could learn (see for example McNeill 2003; Loeber 2001; Kenyon and O’Conor 2003, O’Conor and Murphy 2006; O’Conor and De Meulemeester 2007 and O’Conor 2008).

Dunstaffnage Castle, Argyll and Bute © RCAHMS

Historic buildings such as Dunstaffnage Castle are invaluable documents of the ages that produced them. Through careful analysis of the structural and archaeological evidence, combined with a study of relevant written records, it is possible to build up a detailed picture of the sequence of domestic requirements, social aspirations and architectural tastes of the generations of individuals for whom they were built.

[Note 16] Lidar and Laser Scanner survey –  Lidar survey is carried out from the air and provides an accurate form of topographical survey over a wide area that can penetrate beneath the tree canopy (Bewley et al.2005; Devereux et al. 2005; Challis et al.2008). Laser scanning is a form of ground survey that is a more accurate form of topographical survey, but is more practical for a small area. However, ground observation in conjunction with the above is always preferable in the interpretation of field remains even in woodland.

[Note 17] TST, GPS survey: Survey with TST controlled to GPS can be as accurate as GPS. Both can achieve centimetre accuracy and the generation of DTMs (Digital Terrain Models) is routinely possible. This makes deposit modelling, ie with augers, practical. The result would be to indicate where strata have survived, and thus where a dated sequence would be possible (it mostly is not).