Figure 29: Medieval shielings at Ben Gunnary, Barra (Branigan and Foster 2002, 113)
1 Environmental and Research Background | 2 Post-medieval Buildings | 3 Earlier Vernacular Buildings | 4 Conclusions and Discussion
In section 2, the principal characteristics of the post-medieval vernacular in the Western Isles were highlighted, along with information regarding the sourcing and manufacture of materials and the use of the buildings. It is suggested that this knowledge could provide a possible model for earlier periods that are commonly studied in isolation. This section explores the nature of the evidence for vernacular architecture in the Western Isles in the medieval, Norse and Iron Age periods, and provides comparison, where appropriate, with the post-medieval period.
The Western Isles were ceded to the Scottish Crown by Norway in 1266, and by the mid-14th century they were under the control of John of Islay, Lord of the Isles, and, beneath him, regions were dominated by local chiefs and landowners. In Barra, Clan MacNeil could muster enough labour and wealth to erect two large fortifications, Kiessimul Castle in Castlebay, a typical highland 'castle of enceinte', and Dùn Mhic Leòid, a three-storey tower in Loch Tangusdale, 2.4km away. The dating of these buildings, to the 14th or 15th century, has caused some debate (MacNeil 1964; Dunbar 1978; Gifford 1992), but they perhaps represent the first phase of clearly non-vernacular architecture in the Outer Hebrides. Further examples include Breachacha Castle, Coll, and Borve Castle, Benbecula, as well as the re-occupation of Iron Age sites such as Dùn an Sticer and Dùn Ban in North Uist (Armit 1996, 218). These high-status buildings shun the use of turf and drystone and are instead built with lime-mortared coursed masonry and possibly slate roofing, some of which was imported (Turner and Dunbar 1970, 159; Dunbar 1978). Their angular design can be contrasted with the rounded, aerodynamic vernacular and it is possible that they incorporated imported structural timber, no evidence for which survives, and were perhaps built by mainland masons.
In contrast, the houses of MacNeil's subjects and their peers have proved hard to locate, with excavated evidence in the Western Isles generally limited to shieling huts rather than winter settlements. The only exception to this is the work at the Udal, North Uist (Selkirk 1996), which is yet to be published in full. Survey by Sheffield University (Branigan and Foster 1995 204; 2000) has tended to attribute a medieval date to buildings that have 'blackhouse' characteristics but which are small, reaching only 7–8m in length, with walls 1m thick. This simplistic analysis, supported by Armit (1994), needs to be tested by excavation, although it is to some extent borne out by known examples from St Kilda, where early irregular rectilinear buildings probably date from the later medieval period.
Two small buildings which form part of a long sequence of shelters at Ben Gunnary, Barra, were stratigraphically dated to the medieval period. Both were oval huts, with a stone-faced earth/turf wall foundation and a probable superstructure of turf (Branigan and Foster 2002, 112), exactly similar to post-medieval shielings discussed in section 2. Another small sub-rectangular structure at Eilean Olabhat, North Uist, was associated with pottery of possibly 13–15th century date (Campbell 1994). This building was separated into two parts, and although of small proportions (4.5 by 2.6m internally), contained a small hearth (Armit and Dunwell 1993). Further, the drystone construction and thick walls of the building are reminiscent of later buildings, although the diminutive size suggests a use as a temporary shieling.
Figure 29: Medieval shielings at Ben Gunnary, Barra (Branigan and Foster
2002, 113)
Further examples of shielings dated to the medieval period by stratigraphy, pottery, and condition, were found throughout Barra and the Southern Isles, suggesting that the culture of transhumance was as strong in the medieval period as later, and, further, that the buildings themselves were of similar character, being small, oval or sub-rectangular in plan, and constructed of a mixture of drystone and turf. In an important contribution to the analysis of changing settlement patterns in the Highlands, Dodghson (1993, 423) suggested that many medieval buildings in the area may have been built entirely of turf, and may have been regularly taken down and used as manure on the fields. Dodghson follows similar reasoning to that of Fenton and Walker (1981) who, as mentioned in section 2, argued that a high proportion of buildings up until the early 19th century may have been built of turf. If true, this would strongly affect the recognition of medieval settlement in the Hebridean landscape – one that is particularly difficult to survey (Branigan and Foster 1995).
Nearby to Olabhat, at Druim nan Dearcag, excavations at a site discovered by survey in 1986 (Armit 1990a; 1996, 210) revealed a rectilinear building measuring 6 by 4m externally. Its inner walls were lined with upright slabs, and the outer walls were of simple stone coursing with a turf superstructure. In the better preserved second phase, this building, dated to the 15th or 16th century, was divided into two and contained a hearth at one end (Armit 1996, 211). Comparative examples at Clibhe on the Bhaltos peninsula in Lewis were discovered during survey (Armit 1994, 86-87). These included a rectangular turf building (site 47) approximately 11m by 6.5m overall, and another group of turf buildings (site 49) with one structure 19m by 8m overall with an internal partition. Both these may have been winter longhouses of medieval date, perhaps with a separate byre and living area, though the building at Dearcag seems too small for permanent occupation.
Crucially, Armit's assertion (1996, 214), that the Hebridean vernacular tradition could not be traced through the medieval, occurred prior to the published survey of Barra and the Southern Isles (Branigan and Foster 1995; 2000). Contra Armit, I would suggest that these buildings form part of a vernacular building tradition: one defined by use of local materials such as turf and stone, a bipartite use of space, universal by the 18th and early 19th century, and incorporate design features such as the thick walls and rounded corners that are ubiquitous in later traditions. Their small size is also comparable to extant 19th-century buildings in St Kilda and smaller examples found in survey throughout the Hebrides. Indeed, it is precisely because the medieval vernacular is so similar to the post-medieval in both design and location that it is so difficult to discern. Given the apparently similar design and function of these buildings, it is possible to hypothesise that the roof structures and materials, and the traditions surrounding their procurement and construction, were also similar. From the limited evidence highlighted above, we can therefore tentatively suggest that the vernacular traditions of the medieval period were related and similar to those of the post-medieval in terms of design, construction and use, a hypothesis supported by the evidence for a longstanding tradition of transhumance and pastoralism. However, since the examples are few, particularly for the houses that formed the main winter settlements, I have included the longhouses at Clibhe in the analysis of post-medieval longhouses (see Appendices 1, 2 & 3).
© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions
| File last updated: Tues Feb 28 2006