Medieval settlement remains elusive throughout Scotland and the situation is no different in Argyll. Few of the strategies developed to address this situation have borne fruit. This includes concentrating on the area around later settlements, which has been the default mainstay of historic settlement studies in Scotland since the 1950s despite a spectacular lack of results (for a summary of the current state of Scottish historic rural settlement studies see the Historic Rural Settlement Group's Research Framework - Dalglish & Dixon 2008). Whilst this may partially be a product of the lack of readily identifiable medieval material culture and buildings constructed from highly perishable materials leaving little archaeological trace, this - more probably - suggests some form of discontinuity and break in the nature, location and/or form of settlement. Work in the Outer Hebrides (Raven 2005) indicates that a dramatic realignment of settlement occurred somewhere between the fourteenth and sixteenth century and that the post-medieval settlement pattern is largely a product of the eighteenth century and not the product of continuity.
With the exception of the fewbuildings associated with duns and castles very little has been identified that can be dated with any certainty from the twelfth to fifteenth centuries. A number of circular post built structures associated with charcoal burning activity at Dunloskin Wood, Cowal, have been dated to between the eleventh and fourteenth centuries (Rennie 1984, 1997, nd). Whilst some have recently suggested the domestic use has been misidentified, recent excavations in Lochaber have confirmed this interpretation (Ellis forthcoming). Their shape and location suggests they may not be representative of normal houses, and may be seasonal charcoal burners' huts, similar to the shielings on Jura that accompanied Pennant's eighteenth-century itinerary (1774). The only other domestic building dating to the thirteenth century is the recently excavated sub-rectangular building at Kilchoman (CANMORE ID 187906) (Ellis 2015). Of similar date is the kiln-barn at Baliscate (CANMORE ID 294740) (Ellis 2017), suggesting grain processing might have been more intensified than would otherwise be indicated by what appears to be an absence of contemporary mills. Post-built houses appear to have been common in the twelfth century in the Scandinavian world and eastern Highlands: the post-built structure within Dun Fhinn (CANMORE ID 38467) perhaps indicates this tradition may have spread into Argyll but, currently, no other examples have been identified.
Those sixteenth-century houses that have been closely dated are often unusual in their siting: Gunna (CANMORE ID 141630) (James 1998a, 1998b) sits on a hard to reach tidal island while those at Finlaggan (CANMORE ID 37708) (Caldwell and Ewart 1993) relate to what is effectively a closing off and obliteration of the MacDonald lordly use of the site. They are, however, similar to contemporary houses excavated in mid-Argyll by James (2004, 2009) and in Lewis, North Uist, Skye and Perthshire (Raven 2005). Characteristically, these houses are relatively small and insubstantial, sub-rectangular in plan, mostly with a central hearth (although some have end hearths) and often largely turf built or wicker walled.
There is often an assumption crucks were used throughout the medieval period, although some would argue against this (Dixon 2002). However, despite some limited evidence suggesting the possibility of cruck construction in neighbouring districts (eg Dixon 1998, Walker 2008), the evidence is not persuasive. Cruck building elsewhere only seems to have become common from the fifteenth centuries onwards and there is little evidence that this technique was imported into Argyll any earlier. More evidence may redress this picture.
It is curious that similarly formed buildings identified through survey in Islay (Caldwell et al. 2000) and mid-Argyll (James 2009 and Regan) tend to be more upland in setting than associated with later low-lying arable fields. Although these late medieval houses often sit within or alongside small areas of rig and furrow, their ephemeral form and more upland distribution may suggest a number of possibilities, none of which are necessarily mutually exclusive: a greater concern with pastoral resources, displacement through war (recent work has certainly suggested this is the case in Colonsay in the early seventeenth century - Breen and Raven 2012) or a reaction to environmental stress (Raven 2005). A widespread programme of excavation would test the date, function and environment of similarly shaped buildings. Environmental work may also help to answer if any of these scenarios has any validity and compliment larger scale survey and excavation. This would also help answer questions about the extent of woodland and the importance of woodland exploitation for charcoal manufacture and galley building to the Argyll economy (see below).
New models for identifying medieval settlement are required.
Field walking in the more fertile fields of Kintyre noted, but did not record, concentrations and spreads of imported medieval wares (Cummings pers. comm.). Revisiting these fields along with analysis of the context of middens containing medieval pottery in Coll (Crawford 1997) may provide opportunities for investigations into lower lying settlement. Concentrating on one, defined settlement unit, such as a pennyland, through a combination of survey, test pitting and chemical analysis from lowland to upland may also prove highly productive. Another possibility may be to excavate around the vicinity of the earlier, smaller chapels, as, throughout the North Atlantic, from Orkney to Greenland, there is a strong correlation of chapels with important farmsteads from the Norse period onwards. Closer analysis of the Historic Land-Use Assessment may also bear fruit by providing areas of survival of medieval settlement to focus upon. Work in Ireland has shown how difficult creel houses, or creachts, can be to identify in the archaeological record, often only a scatter of stake holes remain, occasionally accompanied by a hearth. The grey literature contains odd reference to 'isolated hearths' and other un-identified stake holes that may, just, on further inspection, prove to be indicators of medieval settlement. Excavators, especially when prompted by developer funded work, should be encouraged to look for possible indications of ephemeral structures, many of which may only survive within the top soil.
The upland location and ephemeral form of many of the houses that have been identified as medieval could also suggest that many sites which have been identified as shielings, simply because they are upland sites, may, upon closer inspection, be discovered to represent medieval houses. The similarity of their remains could say something about the transient, transhumant and/or seasonal nature of medieval settlement and land use. Do they, for instance, as noted above, simply reveal a reaction to the most effective method for achieving subsistence farming in an arable poor, upland marginal environment; a reaction to environmental collapse; a shift towards cattle and pastoral resources; a defence mechanism at a time of increased raiding and decreased stability, and/or; the adoption of mass-migratory systems, with whole communities permanently shifting from pasture to pasture, similar to that Tudor English commentators witnessed in Ireland (see Raven 2005, 2012b, Dixon forthcoming for different perspectives). Shielings are poorly understood as a rule, as archaeologists use the classification to cover a diverse range of structures, almost any upland building, and land uses, and the subject tends to be dismissed. Much further excavation and analysis is required. Further, in-depth work, incorporating excavation may allow for more nuanced identification and expand the understanding of the extent and nature of medieval upland land use.
Further work on the development of dykes and field systems is needed in order to understand their date and development. An initial place to start could be the Historic Land-use Assessment which has identified small upland spreads of rig and furrow and distinctive globular field-systems in Mull and Islay which could represent a different form or date of agriculture (RCAHMS 2002). Further investigation of medieval and localised cultivation practices (see Dixon 2016) may also help elucidate this further.
Recent excavations at Baliscate (CANMORE ID 294740), Mull, may also help rewrite the settlement history of Argyll. Baliscate also continued to be used into this period, the early medieval chapel was reused by a building similar to but slightly larger than those buildings outlined above. It has a large quantity of pottery associated with it (Ellis 2017)), which may suggest a domestic use rather than a later chapel. However, the post-excavation analysis of Baliscate has the potential to re-write Hebridean settlement history. The dating of the hand-made pottery produced throughout the Kingdom of the Isles and potentially Ulster and Man has proved particularly hard to date and interpret. Diagnostic features rely on vessel form and decorations which are often all but undecipherable but on a small number of sherds within an assemblage, all of which are poorly phased and dated (Raven 2005). Many pottery specialists had even begun to lazily attribute all historic hand-made Hebridean wares from the tenth to nineteenth centuries to the meaningless cover-all term: 'craggan-ware'. Some form on consensus was, nevertheless, beginning to appear in the way of typologies and very broad contextual dating (see Campbell 2003, Raven 2005). However, carbon dating analysis at Baliscate has revealed that pottery of a form that has previously been dated to the later Middle Ages dates, instead, to the thirteenth century (Hall et al 2017). This calls for a substantial reanalysis of medieval hand-made pottery throughout the western seaboard. Added to this should be studies of why so many of these sites contain imported continental wares and why there seems to have been little exchange or trade of Hebridean wares: for instance, whilst there are some morphological similarities between Manx and Hebridean medieval pottery, no Manx pottery has been identified in the Hebrides, or vice versa.
There is an assumption that the medieval landscape of Argyll was largely deforested, with some areas of woodland that was sufficiently mature and well managed to provide materials for castle, church and galley building. The likelihood is that the high medieval period was sufficiently warm (see Section 10.2) and dry enough to support wide scale arable farming, especially in flatter, lower lying areas. Limestone areas, like Lismore, and a few pockets of land, such as around Campbeltown in Kintyre, appear to have been particularly desirable in this period, which seems to be largely a result of their arable potential. Using data from elsewhere, it appears that the Little Ice Age is likely to have significantly reduced this capability. This is most evident around coastal zones, where storms, sand blow and soil destabilisation appears to have been increasingly common. Many medieval remains in machair environments, such as Coll and Tiree, were subsumed in the sand. Oram (2014b) has argued that Highland and Hebridean Scotland is unlikely to have been immune from the waves of cattle disease that spread across Europe in this period, and James (2009) has identified several recorded examples relating directly to Argyll. Oram has also identified one reference that may suggest that the Hebrides were not isolated from the Black Plague (also see Oram and Adderley 2010). This environmental collapse, especially if complimented by pestilence, is likely to have had a significant impact on the Argyll economy, adding to, and perhaps even being a significant cause of, the political collapse that characterises Highland and Hebridean history from the 1400s onwards. This is likely to have resulted in less dependency on arable and contributed to a shift in settlement patterns, with the relocation of settlement away from fields which were no longer as agriculturally productive as they had been. However, very little work has been done by environmental archaeologists or soils scientists to reveal the character of medieval landscapes in Argyll and assess the nature and scale of change.
Only a few medieval middens have been identified, never mind sieved and analysed. There is therefore little to inform us about diets, subsistence and changes in farming. What has been investigated is predominantly high status, associated with castles, or so denuded to the point where we can only say cereals and animals were the main foodstuffs. The middens at Finlaggan (CANMORE ID 37708) (Caldwell pers. comm.) and Gunna (CANMORE ID 141630) (James pers. comm.) appear to corroborate documentary evidence and research in the Outer Hebrides and Northern Isles regarding marine exploitation. At Finlaggan cod was being extensively consumed in the thirteenth century, with limpets becoming increasingly common in the sixteenth century. It is unclear whether the predominance of limpets relates to a crisis subsistence diet or its increasing use as bait, but its use at the higher status and non-coastal Finlaggan perhaps suggests a more cultural predilection. Considerably more research is needed.
The paucity of potential mill sites probably suggests that, as elsewhere in the Gaelic world, there was little attempt by Gaelic lords to coordinate centralised control of grain processing and that this was primarily seen as a subsistence and domestic activity. However, discounting references to mills in the sundry items listed in rentals as these are likely to be formulaic and not a direct record of realities on the ground, there are hints that milling may have been more widespread than the archaeological record immediately suggests. Oral tales occasionally refer to events said to be held at mills, sixteenth and seventeenth-century maps contain the odd reference to muilinn place-names, and there is evidence in the form of a millstone from Finlaggan (CANMORE ID 37708) (Caldwell pers. comm.). More landscape survey may identify early mill sites, lades, etc on the ground.
Poetic sources and carved stone imagery make it clear that hunting was central to Gaelic lordly life and status. However, with the exception of deer traps in Jura, few trappings of hunting activity and hunting forests have been identified in Argyll. More in depth study is likely bear good results and help demonstrate how these shaped the medieval landscape and contributed to lifestyles and economies.
Whilst the contribution of off shore fishery exploitation to the Hebridean economy and trade has been discussed above, much less is known regarding other resources. Sixteenth century records indicate cattle droving was taking place, with cattle being shifted from Islay to Knapdale in 1505. There are also references to Argyll cattle located much further south. Hides were also being exported in large numbers. There are hints that cloth, in the form of mantles, was also a major export (Caldwell pers. comm.). There is a high likelihood of medieval mining of silver and copper resources in Islay (Cressey 1995; Caldwell et al. 2000) which would have contributed to the production of high status goods for export if not exported as raw materials. Charcoal burning and bloomeries (Atkinson 2003) further highlight small scale industrial processes for domestic production and export and help substantiate the claims of some families to be hereditary armourers, boat builders and craftsmen. Unfortunately, though, few production sites have been identified or excavated, nor have the schools for carved stone creation. How these families were housed, the patronage they received, what status they had, the details of craft production are therefore highly obscure.
Equally obscure is how and where trade and exchange took place. Droving indicates that much was taken direct to markets outwith Argyll. However, it seems highly unlikely that trade did not occur within the various territories within it. The chance find of Viking Age ring silver on Dunstaffnage (CANMORE ID 23036) strand hints at open air beach markets in the preceding period and it is highly likely that the practice of holding less formalised markets at easy landing places continued, both on the occasion that a merchant vessel might turn up and on regular annual dates in well-known and traditional places: a pattern well known throughout contemporary Scandinavia and Ireland. Regular markets dates are also likely to have been accompanied by fairs and it seems probable that the small number of known rural fair sites that are recorded from the eighteenth and nineteenth centuries (Black 1999) had their origins in the medieval period. On recorded fairground is Cnoc nan Dubh Leitre (CANMORE ID 22043) in Mull which is only readable on the ground by a large area of open ground surrounded by the footings of small bothies and tents.
Other market activities are likely to have occurred at harbour sites and despite the valiant efforts of Colin and Paula Martin (Martin and Martin , Martin 2008, Martin 2009) much more work is required, especially to test and date their findings. This should be coupled by research into historic coastal change to help identify and understand maritime use, maritime routeways and marine resources exploitation better.