The building of castles is a tradition that was imported into Scotland in the early 12th century. Their spread across the land is usually seen as a reflection of the gradual extension of the authority of the Scottish kings over Scotland. Scottish Kings extended their power both through securing the allegiance of local potentates or by installing their own dependents as local agents of royal authority. Through this process, the castle became a symbol of intruded power and functioned as a defensive structure for newcomers in potentially hostile territories. Castles were also an expression of lordship that was swiftly adopted by indigenous rulers, who already had a long tradition of stone built defensive and residential structures, expressions of their status extending from late prehistory and the early Middle Ages. Each castle has a story on why such as structure was chosen to show status by both local and incoming builders. Some Iron Age hillforts were reused as castle sites in this period, for example Craig Phadrig, Inverness (MHG3809), Dun Lagaidh, Wester Ross, and Dun Creich, Sutherland (MHG39685). The use of coastal promontories for late medieval castles is a particular feature of Caithness and may have earlier origins.
Much of the early work on castles has portrayed their development as a sequence of building traditions, starting from the earthen motte with timber structure on top, moving onto enclosure castles, and finally towards the end of the medieval period to tower houses. These are generally false categories and do not conform to linear developmental models. For example, towers are present in multiple forms from the later 12th to the late 17th century. Similarly, stone enclosures date back into prehistoric times, but in the medieval period they are constructed with the use of mortar and occassionally ashlar masonry.
The castles of Skye and Lochalsh have been published by Miket and Roberts (2007), with the evidence for the Inverness area summarised by Meldrum (1975). The overview of medieval architecture in the Highlands by Dunbar (1981) is still valuable in providing regional and national contexts for castles. Regional architectural guides provide summaries (Stell 1986; Gifford 1992; Beaton 1992; 1995; 1996, Miers 2008) though there are gaps in the Highland coverage. Castle-building by specific noble families, including the MacDougalls, MacDonalds, Comyns, Gordons, Mackenzies and Sinclairs, are explored in the contributions to Oram and Stell (2005). More recently, the studies in Martin (2017) offer a reassessment of castles in the maritime contexts of the West Highlands and Hebrides.
Thacker’s recent work on dating building phases through radiocarbon dating mortar from Lochindorb Castle, Strathspey (Thacker 2020) and Caisteal Camus, Skye (Thacker 2020b) shows potential for similar projects at other sites. Such work has the potential to fundamentally transform researchers’ understanding of castles in the Highlands, where dating has previously been based on architectural details and plan characteristics, and it would allow phasing to be reassessed. His work also links to the source of supplies of wood, for use not only as building material but also as fuel for the limekilns needed for making of mortar.
Mottes (hills, sometimes artificial) with timber structures on top, defensive ditches and occasionally outer bailies, are generally attributed to the expansion of the Scottish kingdom, but all need not be. Some were perhaps constructed by local lords emulating the encroaching symbol of power (RARFA section 9.3). Confirmed examples of mottes in the Highlands are few, with surveys by the RCAHMS suggesting possible contenders such as at Proncy (MHG11745) and Invershin (MHG12913) in Sutherland. There are also early ringworks, for example at Auldearn, Nairnshire (MHG7064), and some important platform motte sites, such as Ruthven, Badenoch (MHG60286) and Dunskeath, Easter Ross (MHG8444).
A number of well-preserved castles with stone enclosures survive in the Highlands. Urquhart Castle (MHG3265), Inverness-shire Ormond Castle, Avoch (MHG8226; Case Study Ormond Castle) and Castle Roy, Badenoch (MHG4571) are good examples in the east; Inverlochy Castle (MHG4182), Mingary Castle (MHG454) and Castle Tioram (MHG4028), Lochaber are good examples from the west. Recent rescue conservation at Castle Roy prompted archaeological intervention. This showed the castle mound to be glacial in origin and archaeologists found evidence that the enclosure featured architectural elaboration and internal stone and timber structures (Spall 2012; 2013). Likewise, recent excavation and documentary historical analysis at Eilean Donan Castle (MHG45372; Case Study Eilean Donan Castle) has revealed the complex building-history of this important site; this is especially true for its multi-towered 13th to 15th century phases. THe results of this recent work underscores Eilean Donan’s attribution to the Gaelic Earls of Ross rather than the Scottish crown, illustrating the adoption of this alien cultural tradition by an indigenous regional power (Clarke et al forthcoming). More survey and excavation at similar sites would no doubt enable our understanding to be refined.
A number of castles throughout the Highlands were built on islands, either reusing Iron Age crannog sites or taking advantage of natural islands for example Lochindorb (MHG6778), Eilean Donan (MHG45372; Case Study Eilean Donan Castle). Sites of high-status medieval occupation need not be castles either: Loch Kinellan in Easter Ross, a probable Iron Age crannog, was re-occupied in the 14th and 15th centuries as shown by medieval pottery, and again in the post-medieval period, possibly as a hunting lodge (MHG6285, Fraser 1917). At Loch Vaa, Badenoch and Strathspey (MHG4691), the crannog was built in the Medieval period, with a modest building, clearly showing new construction in the medieval period and the enduring appeal of this site type (Stratigos and Noble 2019; forthcoming). Elsewhere documentary evidence shows that some crannogs were occupied by ecclesiastical as well as secular owners, with some crannogs suggesting seasonal occupation (Stratigos and Noble forthcoming). More detailed investigation throughout the Highlands as has been undertaken in northeast Scotland (Stratigos and Noble 2017) and around Loch Tay in the Living on the Water project, would allow comparative data for Highland crannogs.
In addition, some smaller stone-built structures, generally termed hall houses – itself another invented and contested category – are also known from the Highlands. Most are generally dated to this period based on architectural features. They were not designed for serious defence, but probably would have had a ditched enclosure with earthen banks topped by wooden palisades. The following sites have been proposed as hall-houses in Highlands
- Inverness-shire: Castle of Petty (MHG4362), Isle of Moy (MHG2869)
- Nairn: Rait Castle (MHG7030)
- Wester Ross: Strome Castle (MHG7648)
- Skye: Camus Castle/Knock Castle (MHG45401)
- Lochaber: Ardtornish Castle (MHG479)
Ardtornish, Morvern in Lochaber was an important centre for the Lordship of the Isles into the third quarter of the 15th century (Caldwell 2014). A survey by RCAHMS (1980) revealed what appears to be an extensive complex of buildings clustered around the dominant hall-house building. This layout can be compared with the excavated site of Finlaggan on Islay, the chief centre of the MacDonald lords of the Isles (Caldwell forthcoming), where physical defensiveness in structure was likewise largely absent for the later 14th and 15th century manifestations of the castle. Although Thacker has undertaken investigative work at Ardtornish, only for Camus Castle has detailed analytical work combining documentary research and environmental sampling been published,(Thacker 2020).
In the Norse area, there are few castles from the early part of the medieval period. Scrabster Castle, the bishop of Caithness’s castle which is set on a promontory, has produced pottery dating to the 12th to 13th centuries (MHG1434; Talbot 1973). Ring of Castlehill near Lythe (MHG1894; Talbot 1977) and Old Wick (MHG22344), Caithness have also been suggested as castle sites but dating evidence is needed to confirm these suggestions. Dendrochronological results at Old Wick are expected in later 2021 (Will Wyeth pers comm).
The penetration of Scottish kings into eastern Sutherland in the early 13th century saw settlement by members of the Moray-based de Moravia family, whose 12th century caput at Duffus seems to have influenced their principal Sutherland centre at Skelbo (RCAHMS 1911; Yeoman 1988). The important multi-period site of Skelbo, now in precarious condition, would reward survey, excavation, and a study of documentary analysis.
In Ross, the Earls of Ross were apparently early adopters of castles. They took on the castle as a part of their wider embracing of continental expressions of lordship, including the status and equipment of knights, the use of parchment charters and seals, and the promotion of mainland European monastic orders. In Easter Ross, their principal centre at Earl’s Allan near Fearn (Oram et al 2009, 25; Case Study Earl’s Allan) has not been subject to archaeological investigation, but in the west their powerbase in Kintail at Eilean Donan has been the subject of an extended excavation campaign and historical analysis (MHG45372; Clark et al forthcoming; Case Study Eilean Donan Castle). The form of Eilean Donan perhaps offers an indication of the possible form of the early castle at Earl’s Allan. A site deserving more attention is Dingwall Castle,. It started as a royal powerbase in the 13th century but was in the hands of the Earls of Ross from the early 1300s to 1475 and it remained contested between MhicDomhnuill pretenders and royal agents into the mid-1500s. It is a site where lowland Scottish, northern Highland and Hebridean traditions ought to meet.
In the Highland areas influenced by the Lordships of the Isles, castle building tends to start in the 13th century where it was arguably more to show status than to be defensive (Miket and Roberts 1990; Addyman and Oram 2012, 9; Martin 2017). The RARFA noted that castles in the Lordship of the Isles controlled areas still remain obscure (RARFA section 9.3). That observation can be extended to those parts of the western Highland mainland and adjacent islands of Skye and Raasay where the MacDonald Lords of the Isles and their kinsmen exercised lordship down to the end of the 15th century. This is exemplified by the castles at Strome (MHG7638), Caisteal Maol (MHG5416), Dun Scaich (MHG5171), Duntulm (MHG5215) and Brochel (MHG5707 ; Miket and Roberts 1990). Apart from some trial excavation at Strome (Cullen and Driscoll 1994) none of these sites have seen detailed modern survey or excavation.
In the southern Highlands, the earliest castles date to the 12th century, as identified in documentary sources. Most castles tend to be coastal, reflecting the likely concentrations of the economic resources upon which lordship was sustained in all periods and, perhaps importantly, this distribution seems to continue patterns discernible from the later prehistoric and early historic eras. There are examples of castles on interior routeways, notably on the Great Glen and the central Highland passes from Strathspey into the Tay basin, including Loch an Eilean (MHG4417), Ruthven in Badenoch (MHG60286) and the island site of ‘Laggankenze’ or King Fergus’s Island/Eilean an Righ in Loch Laggan (MHG2583; Maxwell 1951).
Very few unaltered castles survive in the Highlands, and some are known only from documentary records. This is testimony not simply to the turbulent times but also to later salvage and recycling of expensive worked stone and the fashions for the replacement of older castles with modern mansions during the era of ‘improvement’ in the 18th and 19th centuries. Dating is difficult in many cases, especially given multiple episodes of rebuilding, and it is often difficult to determine what the dates are for sites based on (Addyman and Oram 2012, 8–9). Advances in radiocarbon dating of building mortar offers a route to closer dating in many such cases. A number of castles in the Highlands have been excavated to varying levels of investigation, including Eilean Donan Castle (MHG45372, Case Study Eilean Donan Castle), Castle Sinclair Girnigoe (MHG417; Case Study Castle Sinclair Girnigoe) and Inverlochy Castle (MHG4182). At Urquhart Castle opportunities to investigate areas outwith the walls were presented when new carparks were constructed, adding to information from previous excavations (MHG3265). Some castles like Castle Tioram (MHG4028; Stell 2014), Mingary Castle (MHG454; Addyman and Oram 2012) and Fairburn Tower (MHG7784) where work is currently underway, have had detailed architectural investigations as well.
Setting aside the problems presented by the label ‘tower house’ (Oram forthcoming c; forthcoming d), this expression of castle-building has a widespread distribution throughout the Highlands. Remains often stand in modern isolation, for example at Caisteal Bharraich/Castle Varrich (MHG12350; Cheape 1993), or incorporated into later buildings, for example Tulloch (MHG8897). This might have distorted past perceptions of the status of these structures and their social significance, as well as their form, function and role as elements as larger complexes within the landscape. The extensive areas of attendant buildings and enclosures that adjoin Ardvreck Castle (MHG12103; Noble 1969), which survive as grass-covered stone-and-turf footings and foundations, for example, point to the potential scale and possible function of these complexes and the central role of the tower within them.
Stylistic similarities in plan or use of particular architectural details might reflect networks of family connections or artistic patronage (Oram and Stell 2005). However, little analysis of this has been undertaken within the Highlands to compare, for example, with Maxwell-Irving’s compendious recording of towers in the Borders region of Scotland (Maxwell-Irving 2000; 2014). Restoration of late medieval towers as modern homes, eg Muckrach near Carrbridge (MHG4638) or Ballone near Portmahomack MHG8493), saw close examination of the standing fabric carried out but this was not accompanied by excavation of the adjoining outer courts and closes.
The castles and tower houses of the Highlands retain enormous, and largely untapped, potential for enlightening the medieval and post-medieval transition.
Homestead moats are certainly not common in the Highlands. They are difficult to identify with certainty, and none have been dated. The best preserved is David’s Fort in Easter Ross (MHG8986), where the overgrown location may hold evidence of further water management. It would certainly deserve further investigation. Another possible candidate is at Inchnadamph in northwest Sutherland (MHG12107). The site has been surveyed by Historic Assynt but needs good dating evidence. A third possibility, at Achnasoul Wood, Easter Ross, is a mound enclosed by a moat, with footings of two buildings, surveyed by NoSAS (MHG29192; Marshall 2005). The possibility of waterlogged deposits with good preservation from any of these monuments provides potential for the dating and evidence of daily life.