Church stone vaults
In an earlier article on The Origin of the Jersey Parishes I drew attention to the fact that the 12 ancient parish churches of Jersey appear all to have been designed on a similar aisleless cruciform plan, and that their original features, where these survive, suggest that they were all built in this form at about the same date in the 11th century, when there is also the first documentary evidence for the existence of the historic parishes of which they are the centres.
I gave reasons for believing that, however some of them might be related to pre-existing Christian centres, both the parishes and the churches owed this form to the initiative of the Bishops of Coutances and their archdeacons, one of whom was Archdeacon of the Isles. These officials were engaged from about 1020 onwards in a major programme for restoring the framework of Christian life and worship throughout the diocese after its long and shattering devastation by the pagan Norsemen.
The basic uniformity of design underlying the structure of all twelve churches constitutes a most unusual aspect of that programme, difficult to parallel in any other group of neighbouring parish churches elsewhere in the dioceses of the Norman lands. There was no space in my previous article to describe in detail the early architectural features which the churches have in common. and more careful examination of their fabrics since it was written in 1977 has shown that an analysis of them on these lines could be more complete and thorough than I then thought possible.
It would now be easy, by assembling and confiating all the more significant original elements that are best preserved in each, to reconstruct a composite blueprint of the 11th century design that was evidently in the minds, and no doubt in the hands, of their builders, as the model to be followed in every case. Among the elements that could be used to build up this blueprint would be, for example, features from the naves of St Clement and Grouville, with the central part of the west front of St Peter, from the transepts of St Clement, St Lawrence and St Brelade, and the chancel of St Mary, with confirmatory details on many minor points from other churches where the main original structure is less well preserved.
It can now be shown that one of their most characteristic common features does not belong to the earliest phase of their building, but to a secondary development, a fact which raises questions of considerable historical interest. It is the purpose of this present article to examine this unexpected aspect in the evolution of the churches by considering the origin and purpose of the pointed barrel vaults of rubble stone which are such a characteristic and unusual feature of them all.
It is remarkable that all 12 churches should have these pointed barrel vaults, covering both their oldest parts and also most of the later additions to them. Such uniformity might be taken to imply that barrel vaults were intended from the beginning and formed an essential element in the original design. Their universal presence over the oldest parts of the churches might be seen as giving support to this conclusion. While it is not impossible that this was indeed the original intention, close scrutiny of the fabrics shows that even the vaults covering the oldest sections appear to be secondary. To appreciate this point it must first be noted that while all these vaults are roughly alike in their form and proportions, and appear to spring directly from the walls which carry them, a distinction must be made between those which are completely featureless and those which are divided into bays by internal pilaster arches.
In the nature of the case, the featureless ones provide no clues to their dating, except the general indication given by their pointed form. But, in the case of those with pilaster arches, it is immediately apparent that these internal arches never appear to correspond in position with the external pilaster buttresses which likewise divide into bays the walls of the surviving parts of these 11th century buildings. If the walls and vaults were integral parts of the same design, such misfits would surely have been avoided.
Structural considerations alone must have shown the desirability of setting the internal arches to correspond with the external buttresses, thus providing the arches with the additional strength which the buttresses supplied at the points where it was most needed. The misfits between the arches and the buttresses can hardly be purely accidental, for they seem to follow a uniform pattern wherever there is enough surviving evidence to show the relationship between them. In nearly all the eight churches where this is so, it would seem that the original arrangement was for three-bay vaults to he set over four-hay naves and for two-bay vaults to cover three-bay chancels.
The regularity of this pattern cannot be taken as implying that the vaults and walls, however awkwardly related, are contemporary. In one case, at least, it is certain that the vaults must belong to a secondary phase in the building. This is at St Clement, where the gable lines of a roof earlier and lower than the pilaster-arched vaults can still be seen above the nave and chancel arches of the crossing, showing that both nave and chancel roofs were raised to a higher level to accommodate the inserted vaults. There is also said to be evidence, not now very obvious on superficial examination, that at Grouville the north wall of the nave, which bears several early features which show that it belongs to the original church, has itself been heightened. If this is correct, the heightening was presumably to carry the present vault, which though it has no pilaster arches, is otherwise similar to those which have them, thus suggesting a sequence corresponding to that at St Clement.
But if the earliest of these pointed barrel vaults thus belong to a secondary stage in the building of the aisleless cruciform churches, they were certainly all in place before the addition of north or south naves or chancels in the later medieval centuries. This is made clear by the fact that where such additions were made, as at St Lawrence, St Brelade, St Martin or St Peter, the pilasters supporting the arches of the earlier vaults are invariably curtailed or cut away by the later arcades, which normally pay no regard to their positions. It would seem certain, therefore, that the vaults of this kind over the early naves and chancels must date sometime between the erection of the buildings in the 11th/12th centuries and the main extensions made to them in the 14th/15th centuries.
This points to the 13th century as the most likely date for the earliest vaults, and this is exactly the period to which, on stylistic and architectural grounds, such pointed barrel vaults would naturally be attributed. The flat pilaster form which the pointed arches of these vaults invariably take, suggests strongly that they belong no later than the early part of the 13th century, when the Romanesque style of pilaster buttressing was already old-fashioned and on its way out. It may well be significant that many of the barrel vaults covering later medieval extensions to the churches, such as the north and south nave aisles at St Ouen, the north naves of St Brelade and St Lawrence, and the south naves of St Martin and St Peter, do not have pilaster arches at all. It may also be significant that the north chancel of St Ouen, which has what looks like a 13th century arcade of two bays, carries a two-bay vault with a central pilaster arch aligned exactly with the central pier of the arcade. It looks very much as if here the vault is contemporary with the arcade and that the north chancel as a whole is all of one 13th century date. This is the only place I have noted where such a vault seems to be an integral part of a 13th century extension of a Jersey church.
It has already been noted that the nave and chancel vaults with pilaster arches, such as St Clement, St Martin, St Mary, St Ouen, St Peter and St Brelade, follow a consistent pattern, mostly covering the nave in three bays and the chancel (invariably) in two. It looks as if they were all built to implement a policy affecting the parish churches of the island as a whole. Their close similarity suggests the use of a common blueprint, indicating to all the relevant authorities what was required. This plan did not specify exact dimensions, which would clearly have to be adjusted to suit the slightly different lengths and breadths of the various naves and chancels to be covered. The latter had themselves been built in conformity with the original blueprint of the buildings, which must also have lacked an exact scale.
This point comes out very clearly when the essential dimensions of the vaults, kindly taken for this article by Joan Stevens and Jean Arthur, are examined. In eight cases the length of the vault bays measured from centre to centre of the pilaster arches varies only between 14ft 4in at St Peter and 18ft 6in at St Martin, with three (St Mary, St Brelade and Trinity) all within a few inches of one another between 16ft and 17ft. The only exception to this module is St Lawrence, which has longer bays of 25ft 5in. There is the same basic similarity in the widths of the pilasters themselves, which all measure between 2ft 1in and 3ft 2in; both these extremes are at St Martin, all the rest being between 2ft 2in and 2ft 11in.
What circumstances can have brought about this universal adoption of a uniform and highly original, and expensive, method of reroofing the parish churches of Jersey in the early years of the 13th century? There can be only one answer. It must be related directly to the loss of Normandy by the English crown in 1204. Up to that time the islands had been very closely linked to Normandy by political, social and ecclesiastical ties. The parochial organization had been formed, and the churches built in the 11th century, when the Norman rulers were expanding their authority in a series of moves which culminated in the conquest of England in 1066. The Jersey fiefs had probably all been created by successive Dukes to reward their feudal dependents both before and after that time. The long episcopate of Geoffrey of Montbrai, who was Bishop of Coutances from 1049 to 1096, marked the culmination of a similar expansion in the ecclesiastical sphere.
Diocesan officials, and notably the Archdeacons of Bauptois and the Isles, and of the Val de Vire, the latter holding the advowson of St Saviour's as the focus of their administration in Jersey, must have provided the plans and supervised the building of the churches. In due course they would also have encouraged the transfer of their patronage from the local seigneurs who had originally held most of them, to the various Norman abbeys into whose hands all except St Saviour had passed before Normandy had been lost to the English Crown in 1204.
But the Norman empire had begun to crumble almost as soon as it had reached its zenith in the reign of Henry I. The death of his only son in the loss of the White Ship left a disputed succession on his own death in 1135, and the much less tightly knit, if larger, Angevin Empire which inherited the Norman lands was itself soon challenged by the rising power of the French monarchy under Philippe Auguste. In due course he wrested Normandy from King John in 1204.
These events created a new political situation, and must have had a traumatic effect on social life in the islands. The seigneurs belonged to families whose members held lands on both sides of the water. They were immediately faced with acute problems of loyalty when their possessions in Normandy became subject, directly or indirectly, to the French king, while the island fiefs continued to owe fealty to the English. Above all, there was suddenly a risk of destructive raids on island properties as both French and English factions endeavoured to assert their authority against one another. The old relationship of peaceful collaboration between the life of the islands and Normandy came to a sudden and disastrous end.
It was immediately followed by a ding-dong struggle for the possession of the islands, which lasted for at least 13 years. Pierre de Préaux, who had been Lord of the Isles under John, appears to have surrendered them to the French as part of the settlement which he was forced to make on the capitulation of Rouen on 1 June 1204. But by September 1205 they were once again in English hands, following a successful expedition led by the pirate chieftain Eustace the Monk, who had entered John's service after losing his previous employment as seneschal to the Comte de Boulogne.
In the poetical Roman, written in celebration of Eustace's buccaneering exploits not much more than a generation after the events, we are specifically told that, after his recapture of the islands, they were subjected to great devastation by the Conqueror, who burnt the contents of castles and manors. There is no direct reference to churches, but Eustace, in spite of his monastic background, was not the sort of character to spare them, although the Crown did apparently order him, or some other commander, to maintain peace in the islands and cause no damage there.
According to later records, however, the French recovered the islands at least once after Eustace's initial success in 1205, and may have held them for some years, perhaps with his collaboration. But after his death in 1216 his brothers were ordered under the Treaty of Kingston between Prince Louis of France and the young King Henry III, in September 1217, to return the islands to the English. Probably they had been in French hands at least since September 1215, for it is known that Eustace had deserted John by then and had joined the rebellious English barons in league with the French in the year of Magna Carta.
It would appear that an uneasy peace followed the Treaty of Kingston, leaving the islands under precarious English control from 1217 to 1259. In that year, their final separation from France was confirmed by treaty, when Henry III renounced his claim to Normandy, while the French king recognized English suzerainty over the islands. Both John and Henry III seem to have attached great importance to their hold on the islands, no doubt because of their strategic position as stepping stones on the sea route from England to their possessions in Gascony. Great pains were taken during the 13th century to strengthen the defences of the castles and other fortifications and to maintain garrisons in them against the ever recurring risk of French onslaughts.
At the same time every effort was made to conciliate the local seigneurs and the ecclesiastical authority by permitting as much freedom of intercourse with the mainland as was consistent with political security. Before the century was out the most destructive raid of which there is any detailed record occurred in 1294, and was followed by complaints from the islanders that the invaders had cut down crucifixes and statues, dragged women and girls from the churches, killed 1,500 men and women, burnt houses and corn, and cut up church vestments to make trappings for their horses. The contents of the churches evidently suffered severely at this time, but it would perhaps be unwise to conclude from the absence of reference to structural damage that none of them was actuallv burnt.
If it is right to interpret the architectural evidence as showing that the parish churches, which had been built in conditions of comparative security during the 11th century, were still roofed in 1204 with timber and thatch, as the surviving early gable lines of the nave and chancel at St Clement suggest, they would have presented a massive fire risk in the troubles that followed the loss of Normandy. They stood wide open for many years to the attentions of raiders from both sides, intent on spreading maximum destruction in the island possessions of each other. It is not known what damage the churches actually suffered in the half century before the treaty of 1259, but it is more than likely that many of them lost their roofs by fire.
The diocesan authorities at Coutances were clearly in an embarrassing position in coping with a situation in which their own friends may have done major damage to churches whose revenues were still earmarked for the support of Norman monasteries, and in the case of St Saviour, for their own diocesan administration. But the English officials, apparently endeavouring during much of this time to pursue a policy of conciliation between the islands and their Norman neighbours, are not likely to have objected to any measures which both their duty and their interest led diocesan officials to take in restoring damage to the churches, and insuring, so far as possible, against further peril. It looks as if they may well have decided that all wooden roofs, whether already destroyed or still surviving, should be replaced by massive vaults in stone. There would have been plenty of time during the years of comparative peace that followed the Treaty of Kingston in 1217, or the more definitive treaty of 1259, for such major building operations to be put in hand.
If this is really what happened, the policy was certainly successful, for these drastic fire precautions seem to have ensured the survival of the main fabrics of the churches, not only in the aftermath of the earlier disasters and in 1294, but still more effectively perhaps during the later medieval centuries when the islands suffered terribly from French incursions throughout the Hundred Years War. That the effectiveness of the policy was appreciated at the time, is well shown by the fact that nearly all the many additions made to the churches in the 14th and 15th centuries were roofed from the beginning in substantially the same fashion. Certainly these massive stone vaults made the main fabrics of the churches as nearly incombustible as could be, whatever damage later raiders, fanatic intolerance, or accidental disaster might inflict on their internal furniture and fittings.
Inspiration from Coutances
How safe is it to conclude that this remarkably successful long-term achievement was inspired by the initiative of the bishops and archdeacons of Coutances? There is unfortunately no direct documentary evidence on the matter, and no indirect allusions in contemporary chronicles have been noticed to throw any light upon it. Had the vaults on the early parts of the churches been demonstrably of different dates, or in different architectural styles, there would be no grounds for attributing them to any common initiative. It is their similarity of design and construction that makes it virtually certain that, at any rate those with pilaster arches, were all inspired from a single source and erected at roughly the same time.
One is bound to ask whether any other authority but the diocese can have been responsible. Could the patrons of the churches have done it? Their patronage at that time was in the hands of a number of independent monastic houses scattered across the countryside of northern France. It is in the highest degree unlikely that they can ever have been induced - and by whom? - to collaborate in producing a common design for reroofing all the churches in their several patronages in the same way.
Could any lay authority have undertaken it? The local seigneurs were a heterogeneous group intent rather on promoting their individual fortunes than safeguarding the common interest of the island churches. Moreover, those of them who had been concerned with establishing the original parishes, as we know some of them had been, had all divested themselves of the responsibilities and profits of their patronage in favour of religious houses in Normandy; they would not be eager to undertake the organisation of major repairs to buildings that were no longer in their jurisdiction. Might the representatives of the English crown, whether custodes or the so-called Lords of the Isles, have taken the necessary initiative? Many of them held office for short periods only, were often non-resident, and generally ineffective in action unless stimulated by direct Royal pressure.
While they were certainly at the receiving end of a stream of Royal Writs throughout this period, these were mainly concerned with the King's interest in the building and repair of castles and other fortifications, and with the provision of men and materials for their garrisoning and maintenance. Had the Crown taken a direct interest in the repair of churches, none of which were at that time in its gift, this would most likely have left some traces in the Patent or Close Rolls, or other collections of official correspondence. The only hint of Royal interest that has been noticed is in a writ of 1224, addressed in general terms by Henry III to ecclesiastics holding property in the islands, who were urged to contribute to the cost of defence, 'as you will want aid from us when you need it'. But this indirect hint of possible assistance is most unlikely to be intended to refer specifically to necessary church repairs.
It is thus difficult to see who else but the dioceasan authorities of Coutances, and in particular the archdeacons, who had direct responsibility for seeing that the fabric of parish churches was properly maintained, could have been in a position to act effectively in this matter. And, if they did so, it was almost certainly not for the first time. I pointed out in my earlier article the compelling reasons which led one to believe that the aisleless cruciform plan on which all the churches had originally been built was in all likelihood also provided by the Coutances archdeacons.
They alone, therefore, would have had in their hands, or readily accessible in their office records, the necessary plans, specifications and contractual details that would be required as background information for this fresh task of providing for all the naves and chancels of the churches, a new standard design for pointed barrel vaults. If the design was drawn with bays between the arches longer than those between the old wall buttresses, that is most easily explained as a simple error in the Coutances drawing office, due perhaps to the flat buttresses being inadequately shown on the old plans. But it might be due to some building convention traditional among the craftsmen used to construct these vaults, which were no doubt put up in sections demarcated by the pilaster arches that divided them into bays.
The main interest of this enquiry thus lies in its intriguing suggestion that the characteristic form of the Jersey churches with their barrel vaults was determined not just by a single initiative from Coutances, but by two. The historical and architectural evidence seems to dovetail neatly together to demonstrate the decisive part played by the diocesan officials, and surely notably the archdeacons, first in creating, and then, some century and a half later, in fireproofing a more or less uniform set of twelve parish churches required in Jersey for the maintenance, in exceptionally difficult conditions, of the Christian way of life.
The realization that the earliest stone barrel vaults of the Jersey churches must represent a uniform, but secondary, phase in their architectural development makes it much easier to understand that development as a whole. As noted above, the first phase of these aisleless cruciform churches is best illustrated by the comparable 11th century features surviving in the naves of St Clement and Grouville, the west front of St Peter, the transepts of St Lawrence, St Clement and St Brelade, and the chancels of St Mary and St Peter. The picture which these features combine to present can be filled out by confirmatory detail from several other churches, such as the stubs of central pilaster buttresses surviving under the later east windows in the chancels of St Martin and St Brelade, or the signs of similar early buttresses incorporated in the angles of the west end of the nave of St Martin. That the west front of St Mary was originally similar to that surviving at St Peter is suggested by the drawing of 1815 which shows the head of a central pilaster buttress projecting above the gable of the later addition.
But it has never been easy to understand what happened at the crossings, where the naves, transepts and chancels of these 11th century churches met. One would expect crossings of this date to comprise low and narrow round-headed arches, set on piers sufficiently robust to support a low tower above. But none of the existing crossings are like that at all. All have pointed arches, which are unlikely to be earlier than the 13th century, although at St Lawrence and St Brelade these now carry what were evidently Romanesque towers with saddle-back roofs, and, as explained above, there must have been something similar above the early crossing at St Clement. The tower at St Lawrence owes its present height to the addition of an extra storey in the 19th century, and the low, stumpy, tower at St Brelade, though possibly incomplete, is probably more typical of their original dimensions.
One is bound to ask what effect the raising of the nave and chancel walls to carry the high stone vaults of the 13th century would have had on these early crossings with their low towers, and perhaps in some cases with no towers at all. It would surely have made some alteration desirable, even if not actually necessary, since, otherwise, whatever had surmounted the old crossings might hardly have stood proud of the new roof lines of nave and chancel. Moreover, the new high, pointed vaults would have made the old semi-circular crossing arches look inconveniently low and cramped, and it may well have been thought that to replace them with taller pointed arches would both match the new spaciousness afforded by the pointed barrel vaults over naves and chancels, and provide a sounder structural basis for taller towers, especially if it was already planned to surmount these with the heavy rubble stone spires which so many of them still retain.
By no means all the churches can be shown to have experienced the architectural transformation here suggested, for later rebuilding and reconstruction have removed or obscured the evidence in many cases. St Helier and St Saviour can tell us nothing, for both their crossings were rebuilt from ground level in the late Middle Ages to carry the existing tall square towers, which as Bois has pointed out, were probably the work of the same craftsmen. Neither Trinity nor St John show any sign of original south transepts, though the rebuilding of the tower at St John in 1804 may have eliminated significant evidence there. The tower of St Martin has suffered from repeated storm damage, and it would probably be unwise to argue from its present condition what the original crossing was like. At St Mary only the pointed arch between the crossing and the former south transept, which was absorbed into the new south chancel in the 14th century, survives to suggest that the crossing once had the typical 13th century form.
Thus the earliest evidence of the reconstruction of the original crossings at that time is that provided by St Lawrence, St Brelade and St Clement. In all three, enough remains of what had surmounted the earlier crossing to show that it was evidently possible to reconstruct the main arches in pointed form without dismantling the Romanesque towers above. At St Clement, the drawing of 1815 confirms the evidence of the early roof lines in nave and chancel that the old structure remains with the new roofs rising above it. But in the other churches that have rubble stone spires, no visible evidence has been noted to show the extent to which the towers which carry them may incorporate work earlier than the 13th century. It may, however, be worth noting that, whether or not any of these spires were actually built before l300, the idea of capping the central space with stone may well have arisen at the same time and had the same purpose as the new stone vaults, to make the churches as fire-proof as possible. Certainly stone spires were far less vulnerable to fire than saddle-back wooden roofs like those on the towers of St Lawrence and St Brelade.
There is some indication that in redesigning the main crossing arches in the 13th century a similar specification was used in several cases. Thus St Lawrence, St Brelade and St Peter all show nave and chancel arches of a single order, while the narrower transept arches are of two orders carried down to ground level without capitals or bases. That this similarity is not accidental is suggested by the fact that it occurs in the two churches which retain Romanesque towers. It may of course also have occurred originally in other churches, such as St Ouen, where the crossing piers and arches now show signs of having been modified or rebuilt in later centuries.
A remarkable parallel to the early architectural development of the Jersey churches is provided by the church of Bampton-in-the-Bush, Oxfordshire. This fine cruciform church with central spire almost certainly owes its origin to Leofric, first bishop of Exeter 1050-1072, who made Bampton part of the endowment of his new bishopric. Its first phase must therefore have been nearly contemporary with the building of the early cruciform churches of Jersey, and it may well have been closely similar to them in appearance.
While the later development of its nave and chancel followed an English rather than a Jersey model, the crossing at Bampton shows exactly what must have happened in the conversion of the original Romanesque crossings of the Jersey churches to their present 13th century or later forms. At Bampton the nave and both transepts now have at the crossing tall pointed arches, carried on simple piers very similar to those of many of the Jersey crossings. But for some reason the conversion of the chancel arch to this form was never completed, and although a pointed arch corresponding to the others was inserted in the wall above it, the Romanesque arch and its abutments were left in position below, where they still remain. It is particularly interesting that this partial transformation of the crossing at Bampton was evidently part of a plan which included the addition of a stone spire, exactly as happened in so many Jersey churches from the 13th century onwards.
Dr A J Taylor has suggested to me that the fact that the interior pilaster arches of the churches do not correspond in position with the exterior pilaster buttresses of the naves and chancels which they cover may be due to the circumstances in which the vaults were built to replace the previous wooden roofs. The latter would have been carried on trusses placed to receive maximum support from the external pilaster buttresses. It would have been desirable to retain what remained of the wooden roofs while the stone vaults were being built above them, not only to keep out the weather but to serve as a basis for the centering on which the vaults were being laid. Thus the one position in which the pilaster arches for the new vaults could not be placed was that occupied by the trusses of the wooden roofs which preceded them. If this was so, it not only helps to explain why the pilaster arches of the vaults, while similar in number and position in nearly all the naves and chancels which have them, never correspond with the external buttresses: it also explains why the stone vaults had to be set so much higher than the wooden roofs which they replaced. I am extremely grateful to Dr Taylor for this ingenious suggestion which so happily reinforces my argument in a manner which had not occurred to me.