Due to their monumental construction and positioning, as much as being the product of the lordly aspirations of the leading kindreds and the creation of localised lordships, secular seigniorial buildings are perhaps the best surviving and most obvious features of the medieval landscape in Argyll. They include re-occupied or continuously used prehistoric duns, hillforts and crannogs, alongside new forms of monument such as mottes, castles, towers and island sites, which are sometimes built on top of earlier sites.
There is also another grouping of sites which may date to this period, but may equally belong to another period altogether: ringworks. The most upstanding examples are in Islay: Dun Nosebridge (CANMORE ID 37721) and Dun Guaidhre (CANMORE ID 21853). They resemble, but are not identical to, some of the large bailey earthworks in Galloway, such as the Motte of Urr (CANMORE ID 64982), but they lack the motte. They have a passing resemblance to Scandinavian ringworks, Manx promontory forts or Norse longphorts in Ireland, but are multivallate, inland and generally sited on hilltops. Whilst their date remains unclear it is entirely possible that these reflect a local interpretation of earth and timber castles situated within the early establishment of the Kingdom of the Isles or the power of the MacSorleys. Earthworks at Dunollie (Alcock and Alcock 1987) and Dunstaffnage (CANMORE ID 23036) (Breen et al. 2010), both containing late thirteenth-century material but which may be earlier, could also sit within this grouping. An additional possible example may be the ringwork shown in aerial photographs at Kilchoman Bay, Islay, but generally attributed to a later mansion belonging to John of Islay.
A handful of sites that have been classified as mottes have been recorded in Argyll, with their distribution largely restricted to Cowal and the upper stretches of the Clyde, with one example in the Mull of Kintyre. The Cowal mottes tend to be located at the base of valleys, presumably close to the best arable land and to control routeways. If these fall within the classical motte-building tradition it could be argued that these are a product of an Anglo-Norman form or approach to lordship, similar to those adopted both by colonising communities in England and Ireland and by native kinship groups elsewhere in Scotland, as part of a pan-European package informing how land should be held and exploited and how lordship should be expressed. In Scotland mottes seem to have some correlation with knights-fees, but the relationship is not clear, nor does it appear to be absolute (eg see Tabraham 1984, 1988). Scottish examples tend to consist of simple towers on mounds and lack the large baileys or enclosed or defended infrastructure necessary to house large military retinues that are common in England and Ireland. Those on the Clyde could potentially be attributed to the early influence of direct Scottish central authority. Those in Cowal, however, could equally be understood as a product of local lords positively assimilating themselves into Scottish society. The local lords and motte builders, the Lamonts, came to be stewards to the Stewarts in the 1200s, also exemplifying this process in action. Indeed, this may also highlight that it may be unhelpful to view all motte-type structures as the product of a unified cultural model. In this period Gaelic lords in Ireland were building ‘raised raths’, which have been interpreted as a Gaelic reaction to and interpretation of European motte building, rather than an adoption of the Anglo-Norman model. The differences between Scottish mottes and ringworks and those in the Anglo-Norman world could reveal similar cultural and social divergences, and more work is required in order to understand and specifically culturally situate Scottish motte building.
Whilst it is tempting to see mottes as setting a precedent for the masonry castles to come, paralleling the wider European narrative, their construction appears to have been contemporaneous with the re-adoption of prehistoric sites, such as crannogs and duns, which arguably sit more comfortably within a Gaelic milieu (in the central Highlands there is evidence for a degree of continuity throughout the tenth to eleventh centuries, but in the west and north, especially where Norse settlement appears to have been most absolute, there is a clear phase of abandonment: Ireland also shows a re-intensification of settlement on crannogs and ringforts in the same period). This indicates a fundamental change in perceptions of lordship, hierarchy and heritage in Argyll society. The pattern of monument-type adoption reveals that this process was mediated, informed, interpreted and expressed at different times and in different places.
Castle Sween (CANMORE ID 39028) is widely accepted, on the basis of architectural analogy, as being one of the oldest masonry castles in Scotland. Indeed, throughout the 1200s there appears to have been a spate of early masonry castle construction throughout the west coast, including Argyll. Many historians have seen these as being built by local lords under royal patronage as a bulwark against the wild west (eg Grant 1988). This perhaps fits well for a good number of enclosure castles, built by lords who had a substantial mainland base and were heavily ingratiating themselves with affairs of Scottish Kings and, at the same time extracting themselves from the influence of the Kings of Norway, Man and the Isles and/or the MacSorleys. Castle Sween seems to have been erected in conjunction with the MacSweens gaining the Scottish king’s confirmation of their estates. Likewise, the MacDougall castles at Dunstaffnage (CANMORE ID 23036) and Duart (CANMORE ID 22662) appear at the same time as they courted the Scottish king and attempted to extricate themselves from Norway.
Further west, hall houses, or chamber castles, seem to have been more common. These appear to have been built within a different cultural context, perhaps outwith or as a rejection of Scottish royal influence. In Argyll, the architectural embellishments seen at and size of Aros (CANMORE ID 22272), coupled with its size, is perhaps the most indicative reflection of the western lords’ independence and their equality of status with royal magnates further east. Skipness (CANMORE ID 39798) or Fincharn (CANMORE ID 22777), in mainland Argyll, however, demonstrate that they were not solely a western phenomenon. Other forms of the hall house can be found throughout Scotland, but again they tend to be restricted to the very high echelons of court society. They can also be found in Ireland, providing an alternative source of inspiration for Gaelic lords. The earlier examples are relatively simple in form, showing no evidence for vaulting and containing a first floor entrance. The hall house form of castle is certainly different from the enclosures mentioned above, but is this simply just a different type of building or does it reveal something different about the messages these lords were keen to convey? If we interpret these, as has been common, as fortified or monumentalised halls, rather than ‘castles’, then there would seem to be a keen emphasis on display and the bringing together of disparate communities for activities, such as feasting and judicial assembly. However, such activities are equally likely to have taken place in free-standing or lean-to buildings within the enclosure castles. The insertion of masonry hall ranges at a later stage, though, may suggest that there was perhaps some recognition that the timber buildings at enclosure castles did not provide a suitable venue for some lords and a later converging of the two traditions.
It remains unclear where accommodation was provided in many early castles, Irish sources would suggest that this was mainly located outside the main castle and that the main chambers were only utilised for display and defence (Loeber 2001). The number of outbuildings surrounding Hebridean castles would support a similar situation in Argyll (Raven 2005). Outwith Argyll, at Dundonald (CANMORE ID 41970), the addition of chambers was only an afterthought (Ewart and Pringle 2004), and the development of the towerhouse, with internal chambers above the hall, would also seem to verify that the provision of accommodation within the castle itself was a later requirement.
The understanding of castles in the Lordship of the Isles and Scotland’s Gaelic seaboard remains very much in the dark, and this confusion extends to their dating. There are a few glimpses from the documentary record, a handful of charters, some poems and other written histories, often written long after the event. Yet it is difficult to directly relate the documents with the castles themselves, either in form or to phasing in the remains of the masonry. Some have tried to identify characteristics in building styles with particular dates (for example see Caldwell and Ruckley 2005), others have tried to rely on comparisons between more ornate features, such as moulding. However, this is problematic at the best of time and interpretation becomes even more complicated when dealing with a culture keen to emphasise their earlier antecedents and cultural distinctiveness and therefore all too ready to adopt already old forms of architecture. Simplicity and antiquated window forms do not therefore necessarily indicate backwardness but can represent a deliberately chosen conservatism, reflecting agency and choice. Dating consequently represents a key issue for castle studies in Argyll and beyond. Building on the work in England and Ireland Mark Thacker has pioneered innovative techniques to analysis, phase and date mortars that he is applying to west coast castles and which has the potential to transform the ongoing debate.
One linking feature shared by most early enclosure castles and hall houses is their location, which tend to be coastal and focussed on good harbours – this almost certainly reveals the importance of fisheries and the contribution of the payment of portage and for permission by local and foreign fishermen to exploit a lord’s fisheries to the economies of the west coast lordships. Warmer temperatures in the twelfth to fourteenth centuries saw an explosion in fish numbers, much closer to shore than previously. This resulted in a significant expansion of Atlantic fisheries, English and European fleets followed, and, presumably, Hebridean fleets also joined them, trading their catch for export. Indeed, it is the vast, new, monetary income that these fisheries generated that fuelled the Gaelic renaissance in this period and Gaelic culture’s artistic and architectural legacy.
The main exceptions to this coastal distribution are those castles on Loch Awe, which remained both a main thoroughfare and boundary between lordships for most of the middle ages. What these have in common with their coastal cousins, however, is that they are on nodal points, where Gael met incomer, lords met vassals, and that it was here that lords needed to impress their credentials and most required to exploit the architectural grammar that castles offered.
Cairnburgh Castle (CANMORE ID 21823), in the Treshnish Isles, Dun Chonnuill (CANMORE ID 22374), in the Garvellachs, and perhaps Claig Castle (CANMORE ID 38131), Jura, however, depart from this, being remote from any settlement or landing and in the case of the first two almost entirely out at sea: all can only be interpreted as being primarily concerned with monitoring sea routes. In the case of Cairnburgh it is perhaps telling that despite being perceived as being important enough to get mentioned in Hakon’s Saga and regularly mentioned in charters, neither it, nor Dun Chonnuill, really conforms to the castle model, mostly relying on cliffs and stretches of masonry covering the main approaches for defence, and both are often interpreted as being relatively early.
Across Western Argyll from the twelfth century onwards reoccupied prehistoric fortifications – brochs, duns and crannogs – became an increasing feature of medieval landscapes. The excavated examples reveal a range of forms of reoccupation, from the post-built houses that fill the entirety of walled interior of Dun Fhinn in Kintyre (Fairhurst 1939), to the scattered huts built throughout the enclosures of Kildonan (CANMORE ID 38756), Kintyre (Fairhurst 1956), and MacEwen’s Castle (CANMORE ID 39861), Cowal (Marshall 1983). To this we might add the more substantial buildings outside Tirefour broch in Lismore (Stoddart forthcoming). Further east it is possible that there was more continuous occupation of such monuments. In Perthshire and Galloway, for instance, there is evidence for their continued use throughout the tenth and eleventh centuries (see Raven 2005 for a fuller discussion).
The reuse of sites that would have been seen as old and predating the Norse incursions should be seen in the context of medieval kindreds who were increasingly emphasising their native and Gaelic origins through manipulation of their genealogical pedigrees. This was further emphasised by associating many of these sites with founding figure s in clan pedigrees. Dun Mhuirich (CANMORE ID 39122) in Knapdale, was named after one of the founding figures in the genealogies of the leading families in the area, including the MacSweens (see Sellar 1971 and Meek 1998). It is unclear whether ‘Muirich’ ever lived at the site or later associations with the dun as their ancestral seat were fictitious. Nevertheless, it was occupied in the later Middle Ages (Regan 2012, 2013) and the continued use of the placename, directly linking it to a key ancestral figure, is surely significant, as is the deliberate visual relationship between it and Castle Sween (CANMORE ID 39028), itself named after another eponymous Figure in the same genealogies. The association of both sites with the same stretch of seaway and with good harbours is also revealing. The MacSweens continued the practice of using pre-existing fortifications as they established themselves in Ireland, after their eviction from Scotland, prior to a later phase of castle building (Breen and Raven forthcoming). Even where the nomenclature is not so evident the ancestral connections of some clans with prehistoric fortifications lasted throughout the medieval period. The Campbells seem to have associated themselves, their core origins and heartland directly with a fortified island in Loch Avich, Caisteal na Nighinn Ruaidhe (CANMORE ID 23170). Likewise, the MacDonald and MacDougall occupation of the Dalriadic capitals of Dunaverty (CANMORE ID 38302) and Dunstaffnage (CANMORE ID 23036) were intended to demonstrate their inheritance of the early kingdoms with links to Ireland. Caldwell (pers. comm.) has tentatively suggested an alternative location for the ‘Isleborg’ mentioned in Hakon’s Saga could be an inland site in Tiree built upon what appears to have been a crannog. Those sites chosen for reoccupation predominantly reveal a concern with monitoring and seaborne traffic and controlling harbours. The number of these sites that then became the focus for later castles is also surely significant.
Lesser kindreds continued to occupy duns and crannogs throughout the later middle ages and into the seventeenth century. Many of these continued to be coastally situated but many are also located inland and show an association with local resources, arable, woodland, fishing, routeways, grazing and hunting forests. Their lack of fortification may also reveal a willingness to express connections with their tenants and clansmen, rather than the need to impress them with architecture. The longevity of occupation remains unclear, however, the picture may be skewed by the nature of the documentary and map based evidence we have to date their occupation, which tends to be much later in origin.
Finlaggan (see Case Study 12: Finlaggan) provides an interesting glass through which to view the broad spectrum and development of west coast and Argyll fortifications. The first phase of medieval use appears to have been the construction of a masonry castle built over the remains of a prehistoric dun on a small island at the end of a larger island. The castle appears to be very early and resembles Rushen in Man more than its Scottish neighbours and has been tentatively identified with the Crovan dynasty. At a later date the castle seems to have become redundant, and perhaps as a deliberate attempt to obliterate and supersede any Crovan connection, was then demolished and replaced with a much larger complex of buildings spread across both islands (see Caldwell 1993, 1998, Caldwell and Ewart 1993). This complex included dwelling houses and service buildings but also a large, well built masonry hall on the larger island. The smaller island was retained for assembly or lordly councils, housed in a series of rather diminutive huts. Whilst no longer a castle it housed all the same functions, but with all the units usually arranged within one castle structure, instead distributed in separate disarticulated buildings. Unlike most castles of the time it is not sited on the coast but inland, on the main glen dividing Islay, but surrounded by wide pastures and hunting grounds. This may have allowed the inhabitants to display their cattle wealth but, also, the degree of separation from the sea would have encouraged a sense of transition, for those whose journeys had begun with a boat journey outwith Islay. The change in symbology embedded in the transition from a masonry castle to a less enclosed island complex may reveal much about the changing nature of MacSorley and later MacDonald lordship, from one more feudally orientated to one more embracing a Gaelic world view.
This arrangement (inland, with a number of related islands, and associated with some form assembly or judiciary function) is one reflected throughout Argyll and the Isles. For instance, see the relationship between the castle on Eilean Tighe Bhain (CANMORE ID 23471) and the nearby complex of structures on Eilean a Bharain (CANMORE ID 185988), on Loch Tromlee.
The changes at Finlaggan (CANMORE ID 37708) were taking place at the same time as many Argyll lords were beginning to adopt a new architectural form: the tower. Towerhouses became increasingly common throughout Scotland and Ireland from the late fourteenth century onwards, initially as the main residences of the main lords, then, from the sixteenth century, more commonly at the centre of sub and lesser lordships. They were often inserted into and onto earlier structures, such as Dunstaffnage (CANMORE ID 23036), Castle Sween (CANMORE ID 39028) and Duart (CANMORE ID 22662), the result of extensive remodelling, such as at Innis Chonnel (CANMORE ID 23162), but many appear to have been built on virgin sites, such as at Carrick (CANMORE ID 40804) and Moy (CANMORE ID 22392). They comprise a varied collection, widely differing in size, complexity and the nature of internal features. The origins of the towerhouse have been the subject of much debate. In Argyll, some were clearly inspired by Irish traditions, if not built by Irish masons. As the MacDougalls re-established themselves they built their new castle at Dunollie using wicker centring (a method used to support a vault from underneath during construction and achieved elsewhere largely through using wider planks). Wicker centring is relatively common in Irish towerhouses and increasingly taken to be indicative of a date in the first half of the 1400s (Rory Sherlock pers. comm.). Carrick Castle also displays features, such as the gable or angled open headed windows (the most distinctive feature being a window head formed from two flattish stones leaning together), which would perhaps sit more comfortably within an Irish setting than a Scottish one There is a small corpus of other Scottish examples), although it also takes a lead from Gothic traditions. However, the influences for most fourteenth and fifteenth-century towerhouses are not so clear cut.
Scottish towerhouses have not been subject to the same intensive study in terms of architectural and spatial analysis as their Irish siblings (perhaps best exemplified by Sherlock, O’Keeffe and others) and such a study is long overdue. Whether built anew or on sites with a pedigree, the predominating focus continued to be coastal and associated with harbours (and again this has become the focus of much Irish scholarship, see Sherlock, Kelleher, Naessens, Breen, Gardiner, McNeill amongst others). Whilst it is probably no coincidence that the commencement of this secondary phase of castle building coincided with increasing foreign exploitation of west coast fisheries, the wealth this exploitation produced helped contribute to the wider Gaelic renaissance (see Raven 2005). However, it continued into the fifteenth and sixteenth centuries, when the Little Ice Age resulted in a collapse in off shore fish numbers and their relocation to deeper waters. Pelagic fishery exploitation dramatically reduced, as did the income generated. However, leases and legal disputes show that lords, and notably the earls and bishops of Argyll, became suddenly very concerned with defining and maintaining control of river and estuary salmon fishing, often using fish traps in conjunction to boats. Trade in these fish species is likely to have made a contribution to lords’ coffers. Colin and Paula Martin have done much to identify the remains of fish traps and early harbours, and have noted their relationship with castles, but more work is required to date and understand their use and phasing (Martin and Martin, Martin 2008, Martin 2009). Richard Oram (pers. comm.) has noted that in Iceland the scale of industrial fish processing resulted in specialised sites and that it seems likely such sites existed in Argyll, centred on Loch Fyne and other nodal points where shelter and access coincide.
The tension between royal patronage and burgeoning independence also continued. For instance, the MacLeans established new castles as they emerged from the shadow of the Lordship of the Isles. Many of the older castles remained under direct royal ownership, albeit operated by a constable or captain with the position held by the head of the main local lordship, as at Dunstaffnage (CANMORE ID 23036), Dunoon (CANMORE ID 40729) and Tarbert (CANMORE ID 39316). Some of the largest and ostentatious towerhouses were clearly built to make a statement of newly established power: Carrick Castle was built on the site of an earlier hall but in its final form is proclaimed the Campbell’s arrival as a dominant power on the upper stretches of the Clyde, whilst Moy, seemingly built on a previously unoccupied site, announced the arrival of a newly empowered MacLean kin group.
Caldwell (pers. comm.) has recently suggested the early use of artillery fortifications throughout Argyll, notably by the MacDonald South at Dunyvaig (CANMORE ID 38002) and other sites in Islay and Colonsay as well as by the MacLeans at Breachacha (CANMORE ID21576). This reflecting an early adoption of artillery in the Irish Wars, and perhaps fuelled by guns recovered from the Spanish Armada.
Elsewhere, the later sixteenth century witnessed more gentile approaches to towerhouse building, more familiar with Lowland Scottish styles and, in some cases, perhaps demonstrating protestant identities. This is probably most demonstrable at the decorative Carnassarie Castle (CANMORE ID 22835), built by the new Protestant Bishop of the Isles, under Campbell patronage. Despite the renaissance mouldings, resembling politer houses in Stirlingshire, their place in the Gaelic world was emphasised through inscriptions and also through the siting of the towerhouse alongside an earlier dun. However, its domesticity may be overstated. It sat on the main pass into Mid-Argyll and was the focus of activity during reprisals against the Campbells throughout the 1600s. Likewise, Gylen Castle (CANMORE ID 22936), Kerrera, built by the heirs to the then protestant MacDougall lordship shows an understanding of the grammar of polite landscapes and manipulated approaches. From the sea the seaward face appears stern and unadorned, but once landed the visitor would climb up and perambulate around the hillside until the castle’s, previously obscured, gentile opposite face was revealed, exhibiting an oriel window and sculptures of ladies and gentlemen adorned in the most up to date fashions. These details are much more similar to the towerhouses of Ayrshire than its Argyllshire neighbours, and a study of the comparisons may bear further fruit.
The polite and gentile form of towerhouse became increasingly common from this period onwards, and many, such as Duntrune (CANMORE ID 39147) and Kilmartin (CANMORE ID 39515) would be not be incongruous anywhere else in Scotland (also see Dalglish 2005).
Argyll also possesses a number of lesser known smaller towers, which may be very simple diminutive towerhouses or represent something quite different. Some appear to be the centres of much smaller lordships, such as Island Muller (CANMORE ID 38787) or Dun Ara (CANMORE ID 22069) on Mull. Others, however, are less clear, such as Caol Chaorann (CANMORE ID 22795), Torran, built near a good landing on Loch Awe and just below a potentially reoccupied dun with a name linking it to a position in medieval lordship, Dun Toiseach (CANMORE ID 22771).
The individualism expressed across four centuries of castle building in Argyll may preclude against the usefulness of dividing them into broad categories – the differences between each example of hall house are almost as broad as between them and contemporary enclosure castles. This is a subject that would bear considerable further analysis.
Read the related case study Case Study 12: Finlaggan