4.4.2 Surviving structures

The view currently taken of Scottish medieval church buildings is inevitably conditioned by the fact that a relatively small proportion of the churches that were built has survived in a complete state. One commentator writing in 1939 calculated that only about sixty substantially medieval churches remained in use for worship by then (Anderson 1939). That was certainly a rather pessimistic view, since it is certain that many parish churches which now appear to be of eighteenth- or nineteenth-century date are in fact essentially remodelled medieval structures. Indeed, when a post-medieval church is known to occupy the site of its medieval predecessor, if it is aligned from east to west, and if it is significantly longer along that axis than from north to south, the assumption should probably be that there is underlying medieval fabric, or that it was conditioned by a medieval plan. This was clearly demonstrated in the rather over enthusiastic restoration of Fowlis Wester (Perthshire) in 1927, when a church that appeared to be entirely of 1802 was found to have incorporated the greater part of the masonry of its medieval predecessor. Perhaps more importantly for our present purposes, since the view expressed in 1939 was based purely on churches that had remained in use until that time, it ignored both the large number of churches that have been abandoned but of which there are structural remains, and those churches whichare only known through archaeological investigation.

So far as the earliest churches are concerned, there has been the tentative suggestion that a number of sub-Roman D-ended structures, including those identified at Traprain Law and The Dod could have been churches (Smith 1996). There is, however, the major problem that many sites continued in use over very long periods, with the consequence that the first buildings were overlain and largely obscured by those that succeeded them. There is the further problem that the earliest church buildings appear to have been predominantly of timber, and have left few traces. Adomnan’s description of St Columba’s monastery on Iona, founded in about 563, suggests that the oratory and associated buildings were of timber and wattle (Adomnan, 109-15), while the Venerable Bede evidently thought it remarkable that St Ninian should have built a church of stone in the fifth century, and that King Nechtan should have sought advice on how to build in stone, in the Roman manner, in about 710 (Bede, 264-65). At Whitmuirhaugh near Sprouston, Roxburgh, an E-W oriented post-built structure within an Anglian settlement thought to have been occupied between the seventh and ninth centuries may be a church (Smith 1991), while at Whithorn in the years around 800 two axially aligned buildings interpreted as oratories appear to have been joined together to form a larger church (Hill 1997, 26-48).

Nevertheless, there may have been partly or wholly stone-built churches from an earlier date than is generally appreciated. In the east of the country, at Dunbar, it has been suggested that an L-shaped length of wall could have been part of a church like that at Escomb (Alcock 2003, 214-17), while in the west, on Ardwall Isle, a timber oratory may have been replaced by a clay-bonded masonry structure as early as the eighth century (Thomas 1966). In some areas masonry was perhaps used because it was the most easily available material, and that is especially likely in the north and west. On the Argyll island site of Eileach an Naoimh is a pair of conjoined beehive cells of the Irish pattern which, if not as early as the traditionally ascribed sixth-century date, must still be relatively early in the sequence of masonry structures (RCAHMS 1982, 41-42). Perhaps the most remarkable stone-built early monastic sites are those of Sgòr nam Ban-Naomha, at the foot of the clilffs on the South coast of Canna (Dunbar and Fisher 1974) and at the Kame of Isbister in Shetland (Lamb 1974).

Over an extended period excavation has continued to locate other early stone-built churches. At Iona the so-called St Columba’s Shrine is known to have had Irish-style antae when first discovered (RCAHMS 1982, 41-42), and the same feature was found at Chapel Finian, which dates from the tenth or eleventh centuries (Radford 1951). In the east of the country, but likely to be of a similar date range, is a possible reliquary chapel found below later churches on the Isle of May in the Forth estuary (James and Yeoman 2008), and a ninth or tenth century date has been postulated for the earliest part of an estate church excavated at The Hirsel, Roxburgh (Cramp 1985). The dating of the earliest stone-built churches remains highly contentious, and is only likely to be clarified through extensive further research and investigation. Even the best preserved of such buildings have been the subject of intense controversy, as in the case of the square tower at Restenneth, and the round towers at Abernethy and Brechin. There is a growing consensus, however, that none of these contain anything earlier than the eleventh century (Fernie 1986). There may be scope however to cast new light on the interiors of pre 11^th century churches if the on-going reassessment of some of Scotland’s early medieval sculptural elements as church screens and furnishings is completed.

This reassessment will include the St Andrews sarcophagus, (see Blackwell, forthcoming) where  the discovery of the sarcophagus elements as a buried sculptural group (see Foster 1998, 45-47) may have influenced its interpretation as a burial monument (albeit the specialised form of accessible, burial monument that is a shrine). It is possible that this phenomenon is a reverencing, through careful burial of that which has gone before. Similar concerns are evidenced, for example, by the recent discovery of the buried angel sculpture in Lichfield cathedral (Rodwell et al. 2008, esp. 58-60,) which, in turn, recalls several other examples of carefully buried architectural furnishings in major churches. The fragments of wall paintings carefully buried in the church of St Susanna, Rome (Basile 2004) may reflect similar sentiments. Re-interpretation of the sarcophagus parts as screen and altar elements would thus not only allow a closer understanding of the lay-out of early churches in Scotland but demonstrate how subsequent re-modelling reverenced what it was replacing.

Moving on to the better documented provision of churches from the twelfth to the mid-sixteenth centuries, a higher proportion has survived, and, although it is certainly true that more has been lost than has survived. Here some basic statistics about the numbers of church buildings that are likely to have existed at the time of the Reformation will be offered (the numbers of foundations are based largely on Cowan and Easson 1976, and Cowan 1967). Complete accuracy cannot be claimed for all of these figures. In particular, account must be taken of the following points:

1. the numbers of parish churches fluctuated to some extent in the course of the middle ages, as new parishes were formed and others suppressed;
2. there is no idea of the numbers of parochial or private chapels that supplemented the cure of souls provided by the parish churches;
3. there is only a very inadequate notion of the numbers of the hospitals.

the figures do, however, at least give some idea of the ecclesiastical buildings that existed across late medieval Scotland.

Of the cathedral, monastic and collegiate churches, only those in Table 2 remain wholly or partly in use for parochial worship.

The greatest areas of uncertainty are those associated with the parish churches and with the chapels of various kinds located within those parishes. So far as the parishes are concerned, Ian Cowan’s work on the documentation has at least established with a reasonable degree of accuracy the numbers that existed (Cowan 1967), and those parishes are illustrated cartographically in The Atlas of Scottish History (McNeill and MacQueen 1996, 347-60). In the majority of cases the location of the churches that served those parishes can be ascertained. In an attempt to assess the scale of survival and loss, Richard Fawcett, Richard Oram and Julian Luxford have been looking at the parishes in the dioceses of Dunblane and Dunkeld, an area that has not so far been intensively researched. A principal aim of this AHRC-funded project has been to determine which churches are likely to embody medieval fabric, even if that is not immediately apparent, and to put together a short account of the medieval history of each parish. The detailed results of this project have now been disseminated online http://arts.st-andrews.ac.uk/corpusofscottishchurches. Of the 105 parishes that existed in the two dioceses in the course of the middle ages, the provisional findings suggest that: 14 churches that remain in use have retained a significant and identifiable element of their medieval appearance; 22 ruined churches have retained a significant and identifiable element of their medieval appearance; 17 churches appear to occupy the footprint of their medieval predecessors and probably incorporate some medieval fabric; and 26 churches are thought to be wholly or partly on the site of their medieval predecessor.

It must be stressed that these figures are as yet provisional, and it cannot be known how far they would be reflected in the other dioceses. A second phase of the project, again funded by AHRC and led by Fawcett, Oram and Luxford, is to investigate the dioceses of St Andrews and Brechin, and it is hoped that it might be possible to extend the study to the other nine dioceses in due course. There has so far been very little work on either the documentation or the architectural survival of chapels, of which there were probably very large numbers. Perhaps the most important architectural and archaeological contribution in this area so far has been made by the Royal Commission of Ancient Monuments in its work across Argyll (RCAHMS 1971-92), an area where the small numbers of very large parishes necessitated the provision of many chapels, and where limited development pressures have resulted in relatively high survival rates. But it is also known that some of the more densely distributed Aberdeenshire parishes also had significant numbers of dependent chapels, and it is certain that close study of the documentation would be very fruitful in that area. The possibilities of unconsidered aspects of the early medieval church in Scotland are raised by innovative new studies of church form and usage before 1100 in both Ireland (Ó Carragáin 2009) and Anglo-Saxon England (Hare 2009).

Dalmeny Parish Church (West Lothian), mid-12th century. One of the finest of the churches built to meet the needs of the network of parishes that was beginning to take shape across Scotland from the earlier twelfth century. The architectural details suggest it may have been built by a master mason initially brought to Scotland from England to work on David I’s major new monastic foundation at Dunfermline. The church will be one of those investigated in the course of the second phase of the Corpus of Scottish Medieval Parish Churches project, based at the Universities of St Andrews and Stirling. The architectural sculpture will be analysed for the Corpus of Romanesque Sculpture.©R. Fawcett