Challenging Classifications? Interpreting a “Difficult” Enclosure at Inchnadamph in Northwestern Scotland

ABSTRACT

This paper explores some challenges of archaeological interpretation and classification through an enclosure at Inchnadamph in Sutherland, northwestern Scotland, a site that has proven difficult to interpret. Despite a small number of archaeological interventions, including topographic and geophysical survey and excavation, the site remains enigmatic. We discuss the different interpretations suggested for the site in turn, concluding that the enclosure does not fit readily into existing classification schemes. This raises issues surrounding the use of classification systems, prior assumptions, and the need for critical thinking in interpretation. These are universal issues, applicable beyond the Scottish example chosen, and this paper highlights concerns and difficulties encountered by all who deal with the classification of sites and monuments. Ultimately, this challenges some of our preconceptions and sheds light on the limits of our knowledge, as well as the limits of our classification systems.

KEYWORDS:

• Classification
• interpretation
• archaeological record
• enclosure
• prehistoric
• Medieval
• post-Medieval
• Scotland

Introduction

Classification systems are central to archaeological practice (e.g. Read 1974, 216; Adams 1988; Pollock and Bernbeck 2010), no more so than when dealing with the interpretation of archaeological sites. In formal settings such as within the databases of national and regional bodies, sites are formally indexed and ordered according to a controlled set of values (e.g. Council of Europe 2009; Vergain 2013; Forum on Information Standards in Heritage 2023), enabling users to access and retrieve information more easily. At local and regional levels, records also utilize broadly similar systems of classification. In less formal settings, all archaeologists classify to some extent (e.g. Adams 1988; Edis, Macleod, and Bewley 1989; Lyman 2021), and some even introduce new systems of classification when rejecting official, accepted, versions (e.g., Russell 2002). Indeed, classifications create the framework for archaeological discourse (Boozer 2014, 96); without typologies and systems of classifications, we would struggle to discuss and make sense of the varied archaeological record. Therefore, classification is an important part of how we interpret and understand the sites we study.

However, as anyone who has ever worked within systems of classification is aware, not all sites fit neatly within these systems (Pollock and Bernbeck 2010, 39). The partial nature of the archaeological record means that many sites are difficult to interpret (Baines and Brophy 2006a, 213). For some, there is simply not enough information to enable us to firmly interpret or classify them, while others do not appear to fit within our developed systems of classification. By classifying sites, we are also applying an interpretation, but the interpretative process itself is not objective and is influenced by our own experiences, preconceptions, and values (Hodder 1992, 214; 2004, 30–65; Baines and Brophy 2006a). Together, these factors influence and affect the conclusions we reach and the way in which the sites we study are understood. Accepting and understanding this is relevant, as these factors play a part in the way the archaeological record is structured and how sites are discussed at the site, regional, national, and international level. Ultimately, this affects the nature of the archaeological record and the way it is constructed.

Such issues of classification and the application of typologies and archaeological interpretation have long been of concern to archaeologists (e.g. Spalding 1953; Binford 1972; Adams and Adams 1991, 310–311; Boozer 2014; Sørensen 2015). But while detailed discussions and disagreements about particular interpretations or classifications are common, explicit discussion of some of the issues and assumptions behind the use and application of classifications and the difficulties encountered are undertaken less often (Sørensen 2015, 91). This paper aims to explore some of those challenges through a case study focused on a site at Inchnadamph (Canmore ID 4664) in Sutherland, northwestern Scotland.¹ This is an enclosure for which several possible interpretations have been offered. Despite a small number of archaeological investigations (Hodgson 2005; Cavers and Douglas 2013), it has proven difficult to interpret, and, even in the writing of this paper, the enigmatic nature and difficulty of interpretation of this site has been illustrated by the different, and often contradictory, interpretations suggested to us by those who have read this paper. Superficially, the site could be viewed as having an “easy” interpretation; however, when one looks at the different strands of evidence, it is not as clear cut. Quite simply, the enclosure at Inchnadamph is not an easy site to understand or interpret and therefore presents an excellent case study of the classificatory and interpretative challenges presented by an apparently “simple” archaeological site.

As an earthwork enclosure, Inchnadamph is also of a type common to many regions across the world, and so the arguments developed here are of relevance beyond Inchnadamph’s Scottish context. Many earthwork sites, particularly those that have had little archaeological intervention, suffer from being placed in the generic and non-specific classification of “enclosure” or of being squeezed into broad classificatory groups. These sites often have few distinguishing features to allow easy classification, while further investigation can at times provide frustratingly little additional information. However, these issues are not specific to earthwork enclosures alone, and the arguments presented here are equally applicable to many types and forms of site, both earthworks and those recorded as cropmarks.

This paper will explore some of the challenges of archaeological interpretation and classification through the enclosure at Inchnadamph. It will outline the results of the archaeological interventions undertaken at this site to date and the different interpretations that have been suggested for the site in turn, examining each in the light of the available evidence. We will demonstrate that, in each case, some of the reasoning for rejecting or accepting an interpretation requires reconsideration. Ultimately, this raises wider issues surrounding the utility of classification systems, prior assumptions, and the uncritical repetition of assertions. All are relevant to the way in which we understand and structure the archaeological record and how we deal with “difficult” sites encountered across the world. Although the arguments developed here are likely to have heritage management and conservation implications, this is beyond the scope of this paper. Instead, we will focus on the enclosure at Inchnadamph and some of the implications for classification systems more generally.

Site History and Location

The site at Inchnadamph (Canmore n.d) lies 29 km north-northeast of Ullapool in northwestern Scotland, adjacent to the church at Inchnadamph, on low-lying pasture at the southern end of Loch Assynt (Figure 1). It was first recorded in the Royal Commission on the Ancient and Historic Monuments of Scotland (RCAHMS) Inventory of Sutherland (1911, 6) and classified simply as an enclosure. RCAHMS described the site as an oval enclosure surrounded by a bank some 2–3 ft (0.6–0.9 m) high and 25 ft (7.6 m) broad at the base, with an internal diameter of 134 × 86 ft (41 × 26 m) and two entrances on the northern and eastern sides. The Inventory entry also records that it had a reputation as the place where the Chiefs of Macleods slept in times of danger, though the poor defensibility of the location was noted. The site was surveyed by the Ordnance Survey (OS) in 1962 and visited and described again in 1974. The OS description from 1974 records three breaks in the bank and states that the “NE [break in the bank] has been subject to considerable mutilation” and that it was the “wettest place of the whole work.” A causeway across the ditch was described on the eastern side and interpreted as the probable original entrance, while a third break in the bank on the southeast was noted to be of recent origin.

Figure 1. Location map of Inchnadamph enclosure, showing its position in northwestern Scotland and detail of its situation (contains OS data © Crown copyright and database rights 2022).

While the OS in 1974 favored a Medieval date, the description states that the ditch was unlikely to be strong enough to warrant an interpretation as a homestead moat and acknowledged that a henge interpretation could not be discounted. By 1980, that position had changed, and a further OS description firmly interpreted the site as a homestead moat, based on plan and appearance, discounting an interpretation as a henge because of its irregular shape. This became the dominant interpretation, and the site is listed as a Medieval moated site in McNeill and MacQueen’s (1996, 431) Atlas of Scottish History to 1707.

Hodgson (2005) and Cavers and Douglas (2013) describe the enclosure as a flattened oval in plan measuring around 52 m across × 36 m internally, although they give slightly differing dimensions. Both sources agree that it is defined by a ditch with a low external bank. The ditch averages about 5 m across, while the bank stands to a maximum height of 0.4 m and is at most 7 m across (Cavers and Douglas 2013, 6).

At the date of a geophysical survey of the site in 2005, a period of prolonged wet weather meant that part of the site was flooded, and the western bank was inaccessible (Hodgson 2005, 15). Periodic inundation such as this seems likely and undoubtedly will have resulted in some erosion of the surviving features. Agriculture may also have played a part in denuding the surviving features, as evidence of past cultivation is visible on aerial photographs (Airbus, Bluesky International Ltd, through Pan Government Agreement, APGB-RGB-250mm-NC2421-1, 23 April 2020) to the south and east of the enclosure and is depicted on Home’s 18th century a.d. estate map of Assynt (1774).

The authors visited the site in September 2022, and we are in broad agreement with the previous descriptions (Figure 2). The enclosure is sub-oval with an interior which is raised slightly above the height of the surrounding ground. The northwestern section of the bank is more substantial than the rest and is notably straighter, which gives the appearance that it has been reworked.

Figure 2. Rectified aerial photograph and plan of the enclosure. The plan is drawn from the rectified aerial photograph (photograph DP388270 © Historic Environment Scotland).

It is in an area of landscape defined as Rocky Hills and Moorland (Scottish Natural Heritage 2019). Some of the key characteristics of this landscape type are a rough landcover with an abundance of scattered rocks, boulders, and rock outcrops, with numerous prehistoric and historic environment features, with concentrations around the straths (valleys) and coasts. The landscape is described as being currently largely uninhabited, with a feeling of containment and seclusion, increased by small knolls, dips, and narrow valleys, although it abuts more settled coasts and loch shores.

The enclosure is adjacent to the church at Inchnadamph, on low-lying pasture at the southern end of Loch Assynt (Figure 3). It is located on a slightly higher and therefore drier location. The pasture is bordered to the west by the shores of Loch Assynt and to the south by the River Traligill. The surrounding field is undulating, with higher areas to the north and east, and has notable stands of rushes (Juncus effusus). These are particularly prevalent outside the northwestern edge of the enclosure but entirely absent on the eastern side. This type of rush often indicates poor drainage and shows where the ground is particularly wet and marshy (Balingall 2014). This is significant, as Hodgson (2005, 16) suggested that the breaks in the bank may have been related to “the water management system that supplied the moat.” However, it was noticeable that the internal ditch was predominantly rush free, with only a small patch at the northern gap and in places along the western side. This indicates that the ditch is drier than other surrounding areas and does not often contain standing water, although periodic inundation is likely.

Figure 3. Photograph of the enclosure showing landscape location, looking south.

The area surrounding the enclosure tends to be wet and subject to periodic flooding due to the proximity to the water courses. This is attested to both through folklore and recorded fact. Local legends explain the loch's water rise above normal levels as the tears of Eimhir (the weeping mermaid of Loch Assynt) mourning her life lost on the land (Johncock 2023). In 1902, it was noted that water levels rose some 16 in (41 cm) over the course of three days of heavy rain (Murray and Pullar 1910, 149), and it was also observed that the highest drift-mark was 4½ ft (1.37 m) above its normal recorded levels. It is unknown if the water level of Loch Assynt has been lower or higher in the past, and further work would be required to determine if water levels have remained the same throughout all periods in the past. Nevertheless, whether or not the water level has been different in the past, the presence of paleochannels identified on both aerial photographs and geophysics testifies to the fact that the river has shifted and that the enclosure is located on the floodplain of the river. This must have been a deliberate selection, as constructing the enclosure only a short distance to the north would have raised it above the level of any flooding from either the loch or the river. Indeed, the present Inchnadamph church is positioned in such a location.

The low-lying location of the enclosure means it is effectively hidden by the surrounding topography and is only visible from any distance on an approach from the loch or from the heights of Creag Sròn Chrùbaidh to the east. As such, the enclosure is tucked away, particularly when approached from the north and south, suggesting that it was never intended to have a noticeable presence in the landscape.

Archaeological Interventions

A small number of archaeological interventions have taken place at Inchnadamph. Resistivity, gradiometry, and topographic surveys were carried out in 2005 by Highland Council Archaeological Unit and a team of volunteers (Hodgson 2005) as part of activities running during Highland Archaeology Fortnight. The main aims were to define any internal structures or features associated with the enclosure and also to contribute to ongoing research into the wider historical and archaeological past of the local area. The survey was successful in characterizing the morphology and elements of the nature of the enclosure, as well as identifying possible features both within and outside the enclosure. However, flooding around the western bank prevented survey of the whole of the site.

Small scale excavation and topographic survey by AOC Archaeology Group took place in 2013 (Cavers and Douglas 2013) as part of the Assynt Fire and Water project, a community project run by local historical and archaeological group Historic Assynt. The survey and excavation aimed to better characterize, date, and interpret the site. The stated aims also included resolving a potential link to early Medieval cross fragments found within the nearby churchyard and identifying any evidence that might elucidate the origin or cultural identity of the occupants of the site. A single slot trench was excavated over the bank and ditch on the eastern side of the enclosure.

Both the geophysical survey and the excavation reports discuss possible interpretations for the enclosure. Hodgson (2005) discusses two possible interpretations, that of a moated site and a henge, while Cavers and Douglas (2013, 7) suggest four—a prehistoric ritual site such as a henge, an Iron Age fortification, a Medieval moated site, and an early ecclesiastical settlement. Both Hodgson (2005, 7) and Cavers and Douglas (2013, 7) make clear that a henge interpretation has been rejected by some scholars because of the “irregular” form of the enclosure. Cavers and Douglas discount interpretation as a later prehistoric defensive enclosure because of the configuration of the ditch within the bank, which is noted to be “highly unusual” (Cavers 2012, 3; Cavers and Douglas 2013, 7). Both reports discuss possible Medieval origins for the site, with Cavers and Douglas introducing the suggestion of interpretation as a possible early ecclesiastical settlement.

Geophysics

Both resistivity and gradiometry were carried out at Inchnadamph (Hodgson 2005). These surveys confirmed the presence of the bank and ditch and identified the central area of the enclosure as a general area of high resistance (Figure 4). The encircling bank was visible as an anomaly of high resistance, through which three breaks, on the north, east, and southeast, were identified. The ditch was recorded as a corresponding area of low resistance, with apparent breaks in two places. One of the breaks in both bank and ditch corresponds with the gap in the bank on the east, and the surveyor suggested this could represent a causeway and therefore potentially an original entrance (Hodgson 2005, 16). The response of the ditch on the south was higher than elsewhere along its circuit, and the surveyor suggested this may be because the ditch is shallower at this point or the bank constructed of more resistant material and, once eroded into the ditch, producing a correspondingly higher resistance response.

Figure 4. Resistance results collected in 2005 (from Hodgson [2005]; copyright The Highland Council).

The break in the bank on the north matches the gap visible on the ground, as does the break on the southeast. Examination of the published resistivity results shows an area of high resistance immediately east of the southeastern break. The character of the response is very similar to that of the enclosure bank. Given the similarity of response and position of this anomaly, it may represent bank material that has been moved, probably from the area of the break in the bank. If this is the case, this suggests that the gap on the southeast is not an original feature and was created later, as suggested by the OS in 1974.

Within the general area of high resistance in the interior of the enclosure, the surveyor picked out a broadly rectangular area and interpreted it as the possible compacted floor surface of a dwelling. Gradiometry identified possible curvilinear features of enhanced magnetism within the enclosure, though the results were unclear.

Outside the enclosure, a linear high resistance anomaly was considered to represent a possible path between the enclosure and adjacent churchyard, though it was acknowledged it might be natural in origin. Finally, a series of low resistance anomalies were interpreted as paleochannels of the River Traligill, some of which are also visible on aerial photographs of the site (Hodgson 2005, 8). Some of these anomalies were less obvious, and it was suggested they might be feeder channels for a moat. A modeled pseudosection, extending from the center of the enclosure and across the bank and ditch, identified two wide channels underlying the enclosure, which were interpreted as further paleochannels (Hodgson 2005, 18–19). Based upon the interpretation of the geophysical survey results, Hodgson concludes that the enclosure “represents a rare survival” of a moated site (Hodgson 2005, 25).

Excavation

Small scale excavation, along with a topographic survey, was undertaken in 2013 (Cavers and Douglas 2013). A single slot trench was excavated over the bank and ditch on the eastern side of the enclosure. The excavation demonstrated the ditch was wide, U-shaped in form, and more steeply profiled on the outer face. The ditch repeatedly flooded during the excavation, and the excavator felt that, based on his experience, the layers excavated may have been secondary to the use of the enclosure (G. Cavers, personal communication 2022, 2023). A small bank or wall of rounded boulders was encountered on the inner edge of the ditch, possibly representing upcast from the digging of the ditch. It is unclear if this represents a continuous wall around the whole of the interior.

The excavations produced a limited range of artifacts and environmental finds, including metalworking, pottery, and iron artifacts. Perhaps the most notable piece of environmental evidence was the presence of the spider beetle, Tipnus unicolour. This is found either in oak woods or birds’ nests or within long-lived, often high-status, buildings. No other evidence of insects associated with trees was found, which would appear to suggest the presence of occupation within the enclosure.

Among the finds were a fragment of a disc-shaped quern, a folding iron knife, and a fragment of burnt clay which may have come from a hearth or furnace. The pottery is of uncertain date but most likely related to craggan or crogan ware. This is handmade, domestic ware found across the Hebrides and Atlantic Scotland. It was in use from the Iron Age through to the early 20th century a.d. and is therefore a difficult dating tool. Interestingly, the excavations also recovered a large volume (16.9 kg) of vitrified waste material relating to ironworking. This material appears to have been deliberately re-used as wall core material. The robust character of bloomery slags often results in their re-use as building materials either as building blocks for walls, rubble fills, or wall core material, as well as road metaling (Bayley, Dunworth, and Paynter 2001). Importantly, the excavation showed that the bank deposits from which many of these finds derived were very mixed and denuded at this point. Consequently, the security of the finds within the context is open to question.

Radiocarbon dates from waterlogged oak (Quercus sp) recovered from the ditch (cal a.d. 1456–1635 [95% probability; SUERC-45121]) and alder (Alnus) charcoal within the outer bank (cal a.d. 1473–1638 [95% probability; SUERC-45120]) range from the 15th–17th century a.d. An unpublished date from waterlogged Scots Pine (Pinus sylvestris) (cal a.d. 1476–1639 [95% probability; SUERC-46136]) recovered from the lower fill of the ditch and available since the publication of the report, also falls within this date range. A further date from a waterlogged sloe fruit stone (SUERC-45399) recovered from the same context broadly supports this dating, though there are problems with the calibration of this date. Together, the dates calibrate in the range a.d. 1456–1639 (Figure 5). The excavator felt, however, that the samples within the ditch may have derived from secondary deposits which may not relate to the primary use of the enclosure (G. Cavers, personal communication 2022, 2023). One was recovered from a silting deposit interpreted as representing a period of abandonment and the other two from the layer underlying this.

Figure 5. Calibrated radiocarbon dates from material recovered during the excavation of the enclosure in 2013 (calibrated using OxCal 4.4.4; Bronk Ramsey [2021]; r5Atmospheric data from Reimer and colleagues [2020]).

Based on the recovered dates and proximity to Inchnadamph church, the excavators conclude that an interpretation as a possible early post-Medieval high-status residence, closely associated with the church and possibly the Macleod family, is most likely (Cavers and Douglas 2013, 23), although they concede that “the work raises as many questions as it answers” (Cavers and Douglas 2013, 22).

What does this tell us about Inchnadamph?

Together, the geophysics and excavation have provided a good picture of the enclosure at Inchnadamph. The geophysics has added detail to some of the observations on the ground, indicating, for example, that the break in the bank on the southeast may be of more recent origin and not original. The excavation has characterized the enclosure further, identifying a possible internal wall or bank and providing dating material.

Some interpretations of the geophysics, however, including that of the floor of a rectangular building and feeder channels for a moat, would appear to be predicated on an assumption of a Medieval date. The feature interpreted as a possible floor certainly represents an area of more compact material, but this may not necessarily be the floor surface of a structure. Other possible interpretations not considered include the remains of a cobbled yard, compaction due to foot traffic, or even the remnants of a robbed cairn. Overall, though providing some more detail, the geophysical surveys do not provide any evidence to support one interpretation over the other.

Neither does the excavation satisfactorily answer the question of the date and function of the enclosure. Although the radiocarbon dates from the excavation range from the 15th–17th centuries a.d., they may not date the construction of the enclosure, coming from either what were felt to be secondary deposits or an insecure context. There is evidence that metalworking (iron smelting) was taking place at or near the site. The metal objects are primarily vitrified slag which was largely recovered from a built-up bank, suggesting that these bulk slags may have been deliberately re-used as wall core material. The material, however, is not closely datable, as bloomery technology has a long chronology of use, from later prehistoric times through to the post-medieval period. All that we can say with certainty is that the dates and material culture represent a phase of activity at the enclosure but not necessarily the date of its construction. This leaves the door open to other interpretations. It is common for henge monuments, for example, to be re-used in later periods (Younger 2015), and so dates from what may be secondary contexts should not lead to an earlier date being discounted.

So, having looked at the evidence from the archaeological investigations, how should we understand this site? The enclosure at Inchnadamph defies easy interpretation. Neither geophysics nor excavation have provided evidence that allows us to definitively interpret or date the site. Therefore, at present we have no evidence to draw upon that would place a clear date on the origin of this site nor have we much that would allow us to firmly rule out many of the varying interpretations that have been suggested. To better understand those interpretations and to consider those most likely, we will now examine each of the interpretations suggested for this site in turn, in chronological order.

Henge

Henge monuments are circular or subcircular enclosures of Late Neolithic or Early Bronze Age date, usually defined by a ditch and external bank and one or two entrances (Harding and Lee 1987; Harding 2003). The configuration of internal ditch and external bank at Inchnadamph lends itself to an interpretation as a possible henge. The bank and ditch of the enclosure, at around 5 m and 7 m wide, respectively, are also relatively broad in comparison to the area enclosed, a feature typical of henge monuments (Harding and Lee 1987; Younger 2015, 23). However, as noted above, interpretation as a henge has generally been discounted since the OS’s 1980 assertion (Canmore n.d) that its irregular shape excluded it from consideration as a henge monument. This contention has carried through all further discussions of the site, usually uncritically. Although the OS did not define what was meant by “irregular,” presumably this refers to the sub-oval, rather than circular, form of the enclosure. This argument is difficult to sustain, as henge monuments encompass a considerable variety of forms (Harding 2003; Younger 2016) and can vary from circular in plan to oval or flattened circles (Gibson 2004, 71). For example, both Weston (Canmore ID 48914) and Normangill (Canmore ID 47386), both in South Lanarkshire, are elongated ovals in plan, while Cairnpapple, West Lothian (Canmore ID 47919), Broomend of Crichie, Aberdeenshire (Canmore ID 18621), and Broadlea, Dumfries and Galloway (Canmore ID 67148) are also all broadly oval (Figure 6). Therefore, the sub-oval form cannot exclude a possible henge classification. Nevertheless, the form of the Inchnadamph enclosure would be unusual (R. Younger, personal communication 2018), though the considerable variety within the henge classification means this cannot conclusively remove it from consideration as a henge or similar form of monument.

Figure 6. Comparative plan of henge monuments, including plan of Inchnadamph for comparison (Weston and Broomend of Crichie SC343705, SC2109769 © Crown copyright: HES; Broadlea AT003727 © Historic Environment Scotland; Cairnpapple after Piggott [1950], redrawn by Kirsty Millican).

Henge monuments are characterized as having one or two opposed entrances (Harding and Lee 1987, 34). At Inchnadamph, breaks in the bank have been noted on the north, east, and southeast. The eastern break is probably original, while the origin of the remaining two is uncertain. Henges with single entrances are generally broadly circular in form, while those with two opposed entrances are often elongated in plan, with the entrances on the shorter sides. Inchnadamph does not fit either of these patterns, regardless of which breaks are original, suggesting that a henge interpretation is less likely.

Cavers and Douglas (2013, 23) cite the profile of the ditch, which is steep sided on the outer face but shallower sided on the inside, as further evidence that the enclosure cannot be interpreted as a henge. Again, this is problematic, as the considerable variety within the henge classification means that this is not a defining feature. Indeed, the henge at Pullyhour, Caithness (Canmore ID 8366) was shown to have a similar ditch profile to the Inchnadamph enclosure on excavation (Bradley 2011, 122). While considerably smaller than Inchnadamph, this clearly demonstrates that ditch profile on its own cannot be used to discount interpretation as a henge.

While the dates recovered from the site and presence of metalworking could indicate that an early prehistoric date should be discounted, many henge monuments show evidence of re-use. Some, such as Pict’s Knowe, Dumfries and Galloway (Canmore ID 65681), Moncrieffe House (Canmore ID 28012), and henge 1 at Forteviot (Canmore ID 26565), both in Perth and Kinross, have been shown to have been re-used as craft-working sites in later prehistory and the early medieval period (Younger 2015, 124–125; Brophy and Noble 2020). At Inchnadamph, none of the radiocarbon dates or the finds recovered appear to relate to the construction of the enclosure. Therefore, evidence of Medieval or post-Medieval activity cannot be taken as evidence that this site did not originate as a henge monument.

Can the location of this monument aid our interpretation? The low-lying location of the enclosure, overlooked by higher ground, would certainly not be unusual for a henge and neither would its watery and low-lying situation (Harding and Lee 1987, 34; Watson 2004, 85) on the floodplain of the River Traligill. The enclosure’s position within northwestern Scotland, however, places it outside the generally known distribution of henge monuments (Figure 7). This could suggest that a henge interpretation is unlikely. This distribution, though, likely reflects our current level of knowledge rather than the distribution of henges in prehistory. Many of the henge monuments recorded in Scotland have been identified as cropmarks on aerial photographs. This means that the known distribution of henge monuments is largely biased towards the crop-growing south and east of the country and, at least in part, reflects methods of recording as much as prehistoric distribution. So even this does not permit the wholesale exclusion of a henge interpretation. Indeed, the fact that the monument is outside the normal distribution of moated sites has not prevented that argument from being regularly presented.

Figure 7. Location map of henge monuments in Scotland recorded in the NRHE (contains OS data © Crown copyright and database rights 2022).

Therefore, it is clear that some of the arguments previously presented to argue against interpretation as a henge do not stand up to scrutiny. However, other characteristics, such as the configuration of the entrances, indicate that a henge interpretation cannot be fully supported. Consequently, other interpretations must be sought.

Prehistoric settlement

What of an interpretation as a prehistoric settlement or defensive enclosure (Cavers and Douglas 2013, 7)? The configuration of the bank and ditch would certainly be very unusual for a settlement or defensive enclosure, though not entirely unknown. Examples from the Highland region include The Ord (Canmore ID 5008), Dalmor (Canmore ID 6215), Carn Nam Buth (Canmore ID 6700), and Ach An Fhionnfhuraidh (Canmore ID 7170), the latter two interpreted as brochs with outer enclosures. All these examples, though, are smaller than the enclosure at Inchnadamph. The Ord is one of the largest with an internal diameter of 38 m, while Dalmor measures 12 × 9 m within a wall spread up to 3 m and a ditch 6.5–8 m wide. Indeed, prehistoric settlements in this region are generally small in comparison to the enclosure at Inchnadamph and often circular in form, although some are broadly oval.

Surviving evidence of internal features, generally in the form of earth or stone structures such as hut circles or stone enclosures, is another characteristic of many prehistoric settlements in the region. Indeed, some have been classified as brochs or duns based on surviving stone structures. One exception to this is the defensive settlement excavated at Achiemore, Lairg (Canmore ID 97112; McCullach and Tipping 1998, 69–72), at which, prior to excavation, no internal features were identified. Excavation, however, revealed an internal palisade set 1.8 m from the inner lip of the ditch, as well as revetting to the inner face of the ditch (McCullach and Tipping 1998, 69–72). Measuring only 22 m in overall diameter, this settlement is small in comparison to Inchnadamph. Inchnadamph, at about 52 × 32 m, is generally much larger than any of the examples in the region and has no clear evidence for internal structures and certainly no stonework to match the hut circles, enclosures, or possible brochs identified elsewhere. The small bank or wall on the inner edge of the ditch may represent such an internal feature, though the nature of this feature is unclear.

Most prehistoric settlements in the north of Scotland tend to be positioned in defensive or at least defensible locations. For example, Dalmor is positioned on a small knoll on the steep western slopes of Strathnaver, with the natural slope forming its northeastern defense, while the settlement excavated at Acheimore is located on a spur of better drained land on the eastern slopes of Achany Glen. The non-defensive location of Inchnadamph, on low-lying ground overlooked by higher ground to the north and east, also contrasts with the positioning of prehistoric settlements elsewhere in the region. Altogether this suggests that interpretation as a defensive enclosure of prehistoric date is not supported. It may still be possible to entertain interpretation as a non-defensive settlement, but the larger dimension and lack of internal features, as well as the unusual bank and ditch configuration, all suggest otherwise. Therefore, while we cannot completely rule out this interpretation, it is considered here to be poorly supported, and parallels should be sought elsewhere.

However, a note of caution is required when drawing comparisons. In the absence of excavation, Medieval interpretations have also been suggested for some sites identified as possible prehistoric settlements (see, for example, Borgie Bridge [Canmore ID 5739] or Skail [Canmore ID 6186]). This is a group of sites that have also proven difficult to interpret.

Early Medieval ecclesiastical site

Cavers and Douglas (2013) also raise the possibility that the Inchnadamph enclosure may represent an early Christian ecclesiastical settlement, the ditch representing the vallum of an early monastic establishment. Proximity to the church and burial ground at Inchnadamph, coupled with discovery of two fragments of an early Medieval free-standing cross within the churchyard, one of which was being used as a marker at the head of a gravestone (Weeks 2001, 51), are cited as indications that the location may have been of ecclesiastical importance prior to the construction of the current church. Parallels are also drawn with monastic enclosures elsewhere in western Scotland, in particular, with the enclosures on Eileach An Naiomh, Gravellachs (Canmore ID 22361) and Nave Island, Islay, both in Argyll and Bute (Canmore ID 37472; Cavers and Douglas 2013, 8).

Recorded use of this site for a church dates to sometime between a.d. 1436 and 1443, when a church was established by Angus Macleod, laird of Assynt (Innes 1851–1855). This church stood at Kirktown, Balnaheglise, or Achnahiglash at the eastern end of Loch Assynt. Interestingly, all three of these placenames incorporate words that mean “church” and are suggestive of an earlier, unrecorded Christian foundation here. The only surviving remnant of the early church may be the Macleod burial vault which stands to the southeast of the present church. The current church at Inchnadamph dates from the late 18th century a.d. and stands in the northwestern corner of the present churchyard. It is located on slightly higher ground than the neighboring enclosure.

Assessing the interpretation of the enclosure as an ecclesiastical site is difficult, as our knowledge of early Medieval ecclesiastical structures is limited (Foster 2011). Archaeological research has tended to focus on larger, higher status sites such as Iona (Canmore ID 21649), Portmahomack (Canmore ID 15662), and Whithorn (Canmore ID 63298). Early rural ecclesiastical sites are harder to identify and, consequently, less is known about these. However, excavations and geophysical survey have increased our knowledge of early church sites such as Tullich (Canmore ID 174125) and Kineddar, both in Aberdeenshire (Canmore ID 16470), and Govan, Glasgow (Canmore ID 44077), as well as other early Christian burial enclosures such as Montfode, North Ayrshire (Canmore ID 269034) and Kirk Hill, South Ayrshire (Canmore ID 40881).

The enclosure at Inchnadamph does bear some similarities to some of these early ecclesiastical sites. It is a sub-oval enclosure formed by a bank and ditch, which could be a small vallum, and early Medieval carved stones have been found in close proximity. Archaeological work, though, suggests that there is some variety among early ecclesiastical sites. Some have both inner and outer enclosures, as evidenced at Tullich (Geddes, Murray, and Murray 2016, 163) and on a larger scale at Iona. There is also an increasing body of information suggesting that some of the larger Scottish early Christian sites were laid out within sub-rectangular enclosures, for example Portmahomack, Kineddar, and Iona. Other early ecclesiastical sites, such as the sub-circular burial enclosure at Montfode (Hatherley, Franklin, and Henderson 2010), appear to be defined by a single enclosure ditch. Therefore, the variety of shape and complexity of early ecclesiastical sites means that it is not possible to entirely rule out an early Christian origin for the enclosure at Inchnadamph based upon morphology nor is it possible to definitively place the enclosure within this class of site.

What then of the proximity of the enclosure to Inchnadamph church? The presence of carved stone fragments at Inchnadamph could be seen as stronger evidence of an early ecclesiastical presence on the site. Early Medieval carved stones are found at numerous later church sites, often hinting at an earlier foundation, such as Nevay, Angus (Canmore ID 32155), Navidale, St Ninian’s chapel, Sutherland (Canmore ID 7438), and St Peter’s Chapel, Invergowrie, Perth and Kinross (Canmore ID 31926). Interestingly, these are all sites of probable early Medieval chapels (MacDonald and Laing 1968, 1970), and all have oval or sub-oval enclosures associated with them.

The carved stone fragments are part of the head and top arm of a freestanding cross, for which there are other Medieval parallels in the West Highlands (Fisher 2001). The cross may be of the Anglo-Scandinavian hammer-head type, for which one of the closest parallels is from Kilmory Knap Chapel, Argyll and Bute (Canmore ID 39050). While the cross fragments were found in the present kirkyard, it is unclear where exactly they came from. Beyond proximity, there is no clear evidence to connect the cross fragments with the enclosure.

The recent excavation and geophysical survey have provided no evidence for ecclesiastical use. It is unlikely, though, that geophysical survey alone could provide such clarity, and the excavation was of such a small scale that any evidence could well have been missed. Taken together, the evidence for any connection with the Medieval kirk or for an early Medieval origin is tenuous and as yet unproven. Therefore, while we cannot rule out the possibility that Inchnadamph was an ecclesiastical site, neither is there any evidence that would support such an interpretation.

Moated site

Moated sites are a distinctive monument type of the medieval period, found throughout Britain, Ireland, and northwestern Europe (Coleman and Perry 1997, 176; Coleman et al. 2002). Although English moated sites have received a fair amount of academic interest since the 1970s, Scottish sites have not received the same attention. Around 90 are noted in the RCAHMS Inventories or within the National Record of the Historic Environment (NRHE). However, only a very few sites have been surveyed and excavated in recent times, for instance Bishop’s Loch, Glasgow (Canmore ID 44990) and Dunrod, Dumfries and Galloway (Canmore ID 63927).

Moated sites in Scotland are predominantly found in the areas of Anglo-Norman occupation (McNeill and MacQueen 1996, 431): in the southwest, the eastern Borders, the central valley, and the eastern coastal plain north of the Forth (Figure 8). These sites are a subset of a larger group of defensive earthworks including motte and bailey and ring-works which appear as part of the introduction of feudal systems during the 12th and 13th centuries a.d. There has been insufficient research at present to determine the exact timespan within which these sites were constructed, but most will fall within this period. Two examples are recorded in the northwestern highlands, one of which is the enclosure at Inchnadamph. The other is “Tigh Dige” Flowerdale House (Canmore ID 11961) near Gairloch. However, this was not a moated site but rather a house enclosed by a moat, perhaps dating to the earlier 15th century a.d. Herein lies an issue with classifications that we will explore later. Therefore, the enclosure at Inchnadamph is even more notable in that it is located outside the area of Anglo-Norman influence and is firmly within the territory of Norse overlordship.

Figure 8. Location map of moated sites in Scotland recorded in the NRHE (contains OS data © Crown copyright and database rights 2022).

Defining and identifying moated sites is equally difficult. In Scotland, this has been largely done through field or aerial survey and is predicated on both the morphology of the enclosure and the topographic location. Moated sites are found in a variety of configurations from known examples (Figure 9). An added complexity is that many survive only as cropmarked archaeological sites and are difficult to differentiate from rectilinear prehistoric enclosures, particularly south of the Forth where the latter are a feature of the cropmark record. All moated sites are enclosed by a wide ditch, sometimes water-filled, within which would be an area of ground sufficient for a dwelling and perhaps associated structures. Many of these are rectangular, square, or trapezoidal, such as Peel of Gartfarran (Canmore ID 44640) and Fortingall (Canmore ID 24963), both in Perth, and Kinross or David’s Fort, Highland (Canmore ID 18266). However, circular, oval, and D-shaped moated sites are also known, such as Hall of Logy Rothwayne, Aberdeenshire (Canmore ID 17070) or Peel of Claggans, Stirling (Canmore ID 44638). The enclosing ditch does not always completely surround the interior; in some cases, natural features such as ravines form one or more sides, such as at Ardargie, Perth and Kinross (Canmore ID 26671).

Figure 9. Comparative plan of moated sites, including plan of Inchnadamph for comparison (Ardargie, Fortingall, Kinbattoch, Fordoun House, and Caskieben SC2213711, DP234470, DP148830 © Crown copyright: HES).

As noted above, the ditches of many of these sites may have been water-filled. This would have been achieved in several ways—directly from a watercourse, through channels and sluices, or even by locating the moated site within an area of low-lying marshy ground. Morphologically, the enclosure at Inchnadamph could be a sub-oval moated enclosure; it is located on ground that regularly floods and is likely to have done so in the past, and it is close to the River Traligill. Either of these could have provided a water source, although no visible evidence survives of a connection between the moated site and the river. There are three breaks in the bank of the enclosure, to the north, east, and southeast, although only the east appears to be original. The latter provides no obvious connection to a water source and in fact functions as a means of access to the causeway across the ditch rather than channeling water. Therefore, if it is the only original break, there is no obvious method for feeding water into the ditch.

The upcast from the creation of the ditch was often used to create or raise an interior platform upon which there would have been a dwelling and associated structures. The upcast was also utilized to create earthen banks on the exterior and interior of the ditch. As with the outer ditch, there are a variety of shapes for internal platforms, ranging from rectangular or square, circular, oval, or D-shaped. There may even have been more than one platform within the outer ditch, as is evidenced at Muirhouselaw, Scottish Borders (Canmore ID 56968); however, these are rare. Sub-oval shaped, the Inchnadamph enclosure fits into the overall pattern of these sites; however, the internal platform, although raised, is not well defined and may be a result of the creation of the surrounding bank. Characterizing it, though, is difficult, as only a small area in the interior was excavated.

The excavations by Cavers and Douglas provided four radiocarbon dates for the enclosure, all falling within the 15th–17th centuries a.d. If these dates are to be taken as indications of use of the enclosure, this would suggest that if this is a Medieval moated site, it was clearly re-used and potentially altered at a later date. Although there is a limited body of information on moated sites in Scotland, we would expect a date range for the enclosure to be within the 12th–14th centuries a.d. However, as previously noted, the dates retrieved from the excavations do not necessarily reflect the initial date of creation of the enclosure and may have come from secondary deposits.

Morphologically, the enclosure at Inchnadamph does appear to fall broadly within the group of monuments defined as moated sites; however, it cannot be definitively placed in this categorization. Its location outside the area of Anglo-Norman influence, the range of dates recovered from excavation, and limited evidence for water management suggest that a different interpretation is possible.

Late Medieval or early post-Medieval high-status enclosure

If we take the dates recovered at face value, could this enclosure have another explanation based upon these dates? A fifth interpretation is possible. Cavers and Douglas (2013, 23) conclude that Inchnadamph may represent one of the principle high-status residences of the early post-medieval period in this region. However, as outlined above, while these dates reflect use of the monument, it is possible that the monument is earlier in date.

Additionally, if Inchnadamph were a post-Medieval high-status settlement, we might reasonably expect some documentary evidence to corroborate this. However, the enclosure does not appear on any maps before the 1967 edition of the Ordnance Survey 1:10000 map (NC22SW), and there is very little documented beyond the almost throw away reference to the site being a place where the chief of the Macleods slept in times of danger with his followers (RCAHMS 1911, 6). This particular assertion would appear unlikely, as the site is overlooked by higher ground to the north and east. There are documentary references to Assynt dating back to the mid-14th century a.d. when Torkile Macleode was granted 4 davachs of land in Assynt together with the “fortalice of the island of the same” (Innes 1851–1855). There are further references to Assynt and the fortification of the Isle of Assynt. Although this could refer to the enclosure at Inchnadamph, it seems more likely to be a reference to Ardvreck Castle, which is located on Loch Assynt and accessed by a narrow causeway, giving the castle the feel and appearance of being on an island. There is another possible site, Eilann Assynt, a small island to the northwest of Ardvreck Castle, which has been suggested as the location of a Medieval castle (Canmore ID 4534). There is, however, no specific mention of Inchnadamph. The minister in the 18th century a.d., writing in the First Statistical Account, makes specific reference to the Macleod family visiting Rome to get permission to build what is now the burial vault beside the parish church (Sinclair 1795, 209). No mention is made of a high-status dwelling beside the church, which one would expect to be mentioned if such a site did indeed exist. Moreover, Ardvreck Castle and its successor, Calda House (Canmore ID 4661), would have been known to the minister and are more likely candidates for a place of refuge. Therefore, lack of mentions in the existing documentary evidence combined with the uncertainties surrounding the radiocarbon dates means that interpretation as a high-status residence is unproven. While it is clear the enclosure was used during the post-Medieval period, we cannot definitively state that this was the primary period of use of the site nor what the enclosure may represent if it is of post-Medieval date. Consequently, an earlier date remains possible.

Interpretations

So, what light, if any, do these interpretations shed on the enclosure at Inchnadamph? Firstly, it is clear that the Inchnadamph enclosure shares some superficial characteristics with all of these site types. For example, the configuration of the bank and ditch is a characteristic shared with henge monuments, the sub-oval form is similar to some moated sites, and the simple fact that it is an enclosure defined by a ditch and bank is a feature Inchnadamph shares with all of these site types. However, closer examination indicates factors suggesting it does not quite fit neatly within these classifications. For example, the position of the site in the northwest of Scotland and the generally accepted distribution of moated sites or the poorly defensible location and the generally defensive location of prehistoric settlements in the region. Overall, close examination of each site type and comparison with the Inchnadamph enclosure raises many questions, particularly when taken together with the results of the geophysical survey and excavation.

Here, our attempts at interpretation are not aided by the small-scale nature of the excavation and its focus on the earthworks and ditch. This is a normal excavation practice but does not always provide information about how the interior space was used, relying instead on indirect information from material that has found its way into the ditch. Also, as we have seen, the dating of a site from deposits from an open ditch can be difficult. Therefore, excavation strategies focusing on the location within which activities took place during the use of the site, the interior of enclosures, as opposed to ditches and earthworks, may provide more direct information about the way in which sites were used and therefore better help us make sense of and characterize both Inchnadamph and similarly difficult sites elsewhere.

At Inchnadamph, therefore, it is clear that at present we do not have enough evidence to determine which, if any, of the suggested interpretations fit best. Clearly, some are less likely than others, such as prehistoric settlement, but that leaves us with a range of interpretations to consider. None of the evidence presented, either through geophysical survey, excavation, documentary sources, or morphological parallels, provides us with a definitive or even relatively certain interpretation. Indeed, other interpretations are also possible: for example, an assembly site, an agricultural structure, or a metalworking site. So where does this leave us? Herein lies the problem. We have a site that does not fit readily into existing classification schemes and archaeological interventions that, far from providing clarity, complicate the picture further.

Classifications

Classification schemes and typologies are how archaeologists place order on the diverse remains of the human past (Adams 1988; Banning 2020, 23–41; Neusius and Neusius 2020). They are indispensable for summarizing, ordering, understanding, and communicating the archaeological record and are central to archaeological practice. We use them every day in all aspects of archaeological work (e.g. Edis, Macleod, and Bewley 1989; Bortolini 2017; Lyman 2021). This paper is no different; the discussion above is littered with classifications and types. Yet, while we could not do without them, classification schemes can at times also unhelpfully structure our understanding of sites and simplify what is a very diverse and sometimes poorly understood archaeological record. Sites that do not readily fit within these classification schemes become difficult to explain and are often ignored, viewed as a “problem” or placed within very broad classifications such as simply “enclosure.”

Classifications schemes are artificial constructs created and used by archaeologists within which certain characteristics, often based on morphology or dimensions (Brophy 2005, 8), are selected in the present to define classes of sites constructed in the past. While it would be overly negative to suggest that none of the elements selected to create a classification had any relevance in the past, what is clear is that classes are brought into being within the practices, structures, and conceptual schemes of contemporary archaeology (Baines and Brophy 2006b; Pollock and Bernbeck 2010). In other words, they are embedded within and reflect the concerns and values of the prevailing archaeological thought. As such, they are not necessarily “natural” or even inevitable groupings nor are the terms used to describe them neutral (Baines and Brophy 2006b, 213; Boozer 2014, 96). It is perhaps inevitable, then, that not all sites will fit neatly within our classification systems.

Classifications can also become more than a means to order archaeological information. Some have come to represent the cultural concepts, functions, and purpose behind the sites in question (Read 1982, 57; Brophy 2005). Equally, the names used to define some classes can also carry with them implicit assumptions of function, use, or date. For example, a henge monument is usually understood to represent a particular form of ritual monument built during the Late Neolithic or Early Bronze Age, bound up in a particular form of social organization (Harding 2003). Similarly, the term “moated site” implies a type of Medieval monument surrounded by ditches which may have held water, constructed within a particular political system (Coleman and Perry 1997). It is clear, however, that not all sites classified as moated fit this presumption. Indeed, this classification is at times utilized as a catch-all term for any site that is enclosed by a moat. Examples where this appears to have happened include Tigh Dige Flowerdale House, Highland noted previously or Balamani Castle, Perth and Kinross (Canmore ID 28043), a 16th century a.d. L-plan fortified residence. Without critical examination, assumptions of date and function are applied to those mis-classified sites, and we end up with a potentially skewed view of the distribution of such sites, as well as sites that are not Medieval moated sites in our records and in our discussions. Without critically examining each site in turn, this can be difficult to break this apart.

Therefore, once such sites are placed within classification systems, they can acquire pre-understandings and expectations. In turn, these can influence the way that a site or monument is approached or understood (Hodder 1992, 213–240; 2004, 34–39; Baines and Brophy 2006a, 79–80), including the research questions posed and resulting survey and research strategies. For example, different questions will be asked of a monument classified as a moated site as opposed to a henge, resulting in different survey and excavation strategies to answer those differing questions. Such pre-understandings can lead to interpretations through the lens of these expectations, potentially constraining and to an extent controlling what can be said about the sites being studied as attempts are made to fit the data into classifications and the preconceived notions of what they represent. This is something that seems to have occurred at Inchnadamph as seen, for example, in the interpretation of features identified in geophysical survey as feeder channels for a moat but also in the research questions the excavation was designed to address, which were primarily focused on determining whether the site was a moated or monastic site (Cavers and Douglas 2013, 10).

Such classification schemes, based simply upon morphological features, also tend to remove sites from their landscape setting (Watson 2004, 84–85). This is an important consideration, as sites were not placed at random within their landscape. Instead, they are intimately connected with their locations, built through the specific activities of people in a place (Gow 1995, 47; Ingold 2000). Therefore, drawing landscape, setting, and context into interpretation of sites is important and has the potential to alter the way in which sites can be understood and interpreted. At Inchnadamph, for example, the non-defensive situation of the enclosure indicates that some interpretations are less likely than others. Proximity to the church has also led to the suggestion the site could represent an ecclesiastical enclosure, something that is unlikely to have been considered were the church not located nearby. Similarly, the enclosure’s low-lying location in an area prone to inundation and marsh clearly influences the interpretation of the site as a Medieval moated site. The connection made between a site and its landscape is often done instinctually but has a lasting effect when this then leads to the classification of a site.

Once a site is placed within a classification, it can be difficult to challenge that interpretation. This is partly because of the way in which classifications allow filtering of data. So, when a site is placed, for example, within a broadly Medieval classification, it is largely removed from the view of researchers interested in prehistoric sites. This can mean the interpretation of that site remains unchallenged or that similarities to sites of different date or interpretation are not noticed or emphasized.

Classifications also tend to assume that archaeological sites are static entities and that the characteristics we see today are representative of the site throughout its life. We know, however, that many sites were adapted, reinterpreted, and re-used over time. In some cases, a single site may have been re-used and transformed on multiple occasions, thereby effectively transitioning through more than one classification. This is well illustrated by henge 2 at the prehistoric ceremonial complex at Forteviot, Perth and Kinross, eastern Scotland (Brophy and Noble 2020, 209–241). This site was first recorded as cropmarking on oblique aerial photographs as an oval enclosure defined by a broad ditch, widening distinctly on the north but with no obvious entrance gap. In the interior, a large elongate “blob” and series of small pits were recorded. Within the NRHE, it was classified simply as “enclosure.” Excavation demonstrated that this site began as a timber setting or settings, which was then enclosed within a henge ditch and later transformed into a barrow with the closure of the henge entrance, deposition of burials in the interior, and probably the throwing up of a barrow mound or, less likely, a cairn. The site was re-used again in later prehistory and the early Medieval period with the deposition of objects of Iron Age and later date in the ditch, a later Iron Age fire setting beside a paved surface hinting at an industrial process, such as metalworking, and the digging of the large central pit, possibly within the 5th–7th centuries a.d. Altogether, this is a site that was transformed from one classification to another more than once and within which activity took place for which we would not assign a monument classification. Forteviot may be a particularly complex example, but such re-use and transformation is not unusual. Other examples that could be cited include the stone circle later re-used as a roundhouse at Strichen, Aberdeenshire (Phillips, Hampshire-Monk, and Abramson 2006) or the complex site at Cairnpapple, West Lothian (Piggott 1950; Barclay 1999; Younger 2017). The chronology of this site remains uncertain, but it comprised at various times the site of fire-lighting, a cremation cemetery, a timber or stone setting, a henge monument, cairns, and beaker burials.

In the field, we are usually presented with the final incarnation of a site. In some cases, as with the cropmarked Forteviot henge 2, there may be hints of complexity and change, but at others this may be more difficult or indeed impossible to discern without excavation. The Forteviot example also illustrates that some re-use may alter but may not fully transform a site into another recognizable classification and even, in the case of the large pit dug into the center of the barrow, major transformations may occur for which we do not have a classification to apply. Therefore, the characteristics we use to define the classification of a site may represent alteration and modification, rather than necessarily a defining characteristic of a specific “type” of site, while some sites could be classified differently at different periods of their biography. Overall, such re-use and transformation can add to the difficulty of understanding and interpreting a site from field remains or an aerial photograph. That the site at Forteviot was originally classified simply and non-committally as an “enclosure” ably illustrates the interpretative difficulty presented by this site before excavation. Therefore, if the dated activity at Inchnadamph is indeed secondary, then this site may well have been adapted and re-used over time and, like the site at Forteviot, it may have fitted into more than one classification during its use life. This makes the assignment of a single classification difficult and indeed perhaps not even appropriate. In fact, a single classification may prevent us from giving due consideration to other possibilities. If the site has been re-used and adapted, this may also be a contributing factor to the difficulties of interpretating this site. This does not mean that classification schemes are of no value; they are a vital part of all archaeological work and are necessary to order, interpret, and make sense of the rich archaeological record. Nevertheless, there are issues associated with their use, and some critical awareness is necessary.

All these issues are illustrated by the Inchnadamph enclosure. It is clear that the early placing of the enclosure within a Medieval moated site class and rejection of a prehistoric interpretation has affected the way this site has been approached, discussed, and interpreted. Most fundamentally, classification as a moated site has meant that there has been little or no full assessment of the enclosure within a henge or broadly prehistoric enclosure context, at least until the interventions outlined above. Though, as argued, this seems largely based on the assertion by the OS that the monument is too irregular for a henge rather than a full assessment against the henge classification as a whole. It also seems to have affected the way in which some of the evidence has been interpreted. For example, the interpretation of some paleochannels as feeder channels for a moat is predicated on a particular interpretation and assumption of function. Although interpretations other than a moated site are considered by both Hodgson and Cavers and Douglas, a prehistoric interpretation was rejected early on (Cavers 2012, 3–4; Cavers and Douglas 2013, 7–8). Therefore, whether intentionally or not, we suggest that this interpretation and broad dating likely played a part in the way this site was approached, the expectations of the survey and excavation, and the interpretation of the features found. Without challenge, this type of sequence of events can create a self-perpetuating loop.

This form of interpretative loop is difficult to remove oneself from, and we intend no criticism of the previous investigators of this site. This is likely something all archaeologists have found themselves within at times, and it can require considerable critical reflection, as well as more definitive evidence than was found at Inchnadamph, to break free from. Indeed, we ourselves, in writing this paper, found ourselves drawn into a similar loop, reflecting personal preference and individual period expertise in our writing and interpretations rather than a fully balanced consideration of the evidence, and it required critical observations by colleagues to enable us to see this. None of us can approach the data we study in an entirely objective or neutral manner. As human beings, we all inevitably bring elements of our own biases, viewpoints, and preconceptions to our work. There is nothing wrong with this, and indeed, varied viewpoints and perceptions can serve to challenge accepted interpretations and ways of thinking. However, this does highlight the need for critical reflection and the need to critically appraise previous interpretations.

We consider the uncritical acceptance of a rejection of a henge interpretation at Inchnadamph as problematic, while the general acceptance that the enclosure is a moated site is equally problematic. The available evidence does not permit us at present to offer a definitive interpretation or even conclusively date the primary construction and use of this site. One possibility that we must consider is that it could represent a new type of site not considered or encountered before, something that perhaps requires a new classification. But with our limited knowledge of this site, it would be difficult to develop a meaningful type, nor would it be wise to construct a classification around a single site. At present, the most we can say, based on the available evidence, is that the site is an enclosure in use during the late medieval to early post-medieval periods. This then raises one of the issues highlighted here, as there are more than 2000 “enclosures” recorded in the NRHE in the Highland Council area alone. Choosing to classify the site simply as an enclosure may well have the effect of making this site disappear from the view of both Medieval and prehistoric researchers.

Conclusions

Having explored the interpretative and classificatory challenges through the case study of Inchnadamph, we can draw some more general conclusions. Firstly, Inchnadamph is not unique in being a site that is difficult to interpret. Much of the evidence we deal with as archaeologists can be ambiguous and “difficult.” This should not prevent attempting to examine and investigate such sites and archaeological evidence. It is through tackling “difficult” sites that issues such as the need for critical awareness come to the fore, and it is through attempting to deal with difficult issues that our knowledge and archaeological approaches develop. Choosing to acknowledge and engage with the difficulties and ambiguities, rather than uncritically forcing sites into limited categories, can push us to think differently and to consider new possibilities and interpretations.

Therefore, it is wise to remain open to the possibility that some sites may not fit our constructed categories. It would be arrogant to presume that we understand every action carried out by our predecessors, that we recognize the function of every site that was constructed, or even that it is possible to fit the whole span of the varied actions of humans in the past into our constructed categories. We will, inevitably, always encounter new forms of sites, even for activities that we have already defined. There is merit and, indeed, a need to remain open to new possibilities, something that surely serves only to enrich our understanding of the archaeological record.

At times, then, we need to think beyond the traditional categories available to us. As explored above, this is not always easy. However, the interpretation of archaeological sites should be grounded in the available data and information and not in the classifications themselves. If classifications and typologies are the starting point for our interpretations, we will reproduce what we already know, run the risk of replicating and reinforcing historical bias, and surely serve only to “confirm” the existence of certain site types. We also do a disservice to the complex and varied sites we study and those who created and used them. We are not suggesting that we should dispense with our classifications. This would be both unrealistic and undesirable. But, as Sørensen (2015, 90) has pointed out, while the fundamentals of classification have been revised within other disciplines, they have remained static within archaeology, and so we should perhaps develop a better critical awareness of the issues surrounding classifications and their use so that we can use them with greater care and attention. In so doing, we may also be able to offer better and more nuanced interpretations of “difficult” sites such as that at Inchnadamph.

How we deal with the issues raised in this paper within the more formal settings with which this discussion began is more difficult to determine and not easy to solve. National and regional records and heritage management systems are very firmly wedded to established classifications, often for good reasons; they must structure their data within databases adhering to international standards and clearly disseminate this data to a varied audience of historic environment professionals and the general public. Baines and Brophy (2006b) have offered some pertinent suggestions, and we do not intend to explore these in any detail here. However, developing the means to record and index sites against differing and sometimes competing classifications must surely be one possible solution. While often unavoidable (and used in the past by both authors, so no criticism is intended here), careful consideration of the use of generic terms such as “enclosure” must also play a part. Whatever the setting, critical and nuanced use of classifications seems appropriate.

The case study presented here, therefore, raises many questions and issues that lie central to the practice of all those who deal with site classification, whether that be within formal indexing within historic environment records or in less formal settings. It demonstrates both the value and the challenges of critical examination of our applied classifications and the evidence as presented to us. As a site that defies ready classification, close examination of the Inchnadamph enclosure has challenged some of our preconceptions and shed light on the limits of our knowledge and the limits of our classification systems. It has illuminated wider issues that require careful consideration. This is not a negative. It can spur us on to think differently, to challenge interpretations and preconceived notions, and to tackle other “difficult” sites elsewhere. This is not confined to the Scottish context of the example selected nor to earthwork enclosures, though simple enclosures can be particularly difficult to deal with. Instead, the issues and challenges discussed here are relevant to all those who apply classifications to archaeological sites, which includes all of us who deal with such sites.

Acknowledgements

We would like to thank Graeme Cavers, Rebecca Jones, Anne MacSween, Dara Parsons, and Allan Rutherford for their comments on various drafts of this paper. Thank you to Nick Hannon for his help with the interpretation of the geophysical survey results and to Georgina Brown for preparing Figure 1. We would like to thank Highland Council for giving permission to reproduce the resistivity results and AOC Archaeology for giving us permission to publish the radiocarbon dates. Thank you to Historic Environment Scotland for providing funding to make this paper open access.

Disclosure Statement

The authors report there are no competing interests to declare.

Additional information

Notes on contributors

Rory McDonald

Rory McDonald (M.A. [Hons] 1992, University of Edinburgh) is an independent researcher and archaeologist employed in the public sector. He is a heritage professional of 30 years’ experience in heritage protection, database management, and heritage interpretation. His research interests include Scottish Medieval archaeology, aerial archaeology, and how nature and archaeology interface. ORCID ID: 0000-0002-4344-6614.

Kirsty Millican

Kirsty Millican (Ph.D. 2009, University of Glasgow) is an independent researcher and archaeologist working in the public sector. Her research interests focus on Neolithic and Early Bronze Age Scotland, aerial and landscape archaeology, archaeological interpretation, and the relationship between humans and the environment in prehistory. ORCID ID: 0000-0001-5260-2204.

Notes

1 This paper uses Scotland as its geographical focus, and throughout sites will be referenced by their Canmore ID, their unique reference number within the National Record of the Historic Environment (NRHE) of Scotland. This can be accessed at https://canmore.org.uk/.