The Mariakerk at Utrecht, Speyer, and ltaly
By Peter Kidson

Without a doubt, by far the most interesting of the medieval churches in Utrecht is the one that is no longer there - the Mariakerk. The last vestiges were removed I50 years ago in I843. But like the Cheshire Cat, it had been in process of vanishing slowly over the previous 200 years; and were it not fot one of the most remarkable exercises in architectural recording of the seventeenth century, our knowledge of it would be confined to a handful of brief, repetitive, and certainly not very informative, notices in documentary sources from the Middle Ages and the centuries that followed.
We owe the record to the anxieties of the ecclesiastical authorities who were in charge of the church in the I630S. At that time the Mariakerk was used in a discreet but clandestine fashion by the small catholic community of Utrecht, which was tolerated, but only barely so, by the predominantly Calvinist majority of the citizens. The situation was always volatile, irequently ominous. The church had already been threatened with demolition in I58I; and 55 years later the dean and chapter, fearful for its survival, but immensely proud of it both as a building and as an institution, decided in a mood of antiquarian piety, somewhat reminiscent of our own times, that a visual memorandum ought to be made before it was too late. For this purpose they secured the services of the Haarlem artist, Pieter Saenredam, who specialized in drawing and painting buildings. Saenredam came to Utrecht in the summer of 1636, and spent, so we are told, twenty weeks there,l in the course of which time he made eleven drawings of the interior of the Mariakerk and three of the exterior. On his return to Haarlem, seven of the interiors were immediately worked up into paintings; the rest followed at a more leisurely pace. It is these drawings and paintings that now represent the building for us. Saenredam was clearly an archaeologist manqué. Neither drawings nor paintings were romanticized, and they were accompanied by diagrams and measurements that have every appearance of being mathematically accurate. What they reveal is a church that fully justified the concern of the clergy, fot it was a work of conspicuous originality, a totally unexpected maverick among the churches of Romanesque date in the city and the surrounding region. Moreover, if the chronology that can be extracted from the documents really applied to the monument that Saenredam painted, the Mariakerk was a building of European significance, which ought to be ranked with the handful of buildings that betokened a major epoch in medieval architectural history. However, the chronology bas never been firmly established, and the purpose of the present essay is to re-examine the whole question in the light of evidence not previously taken into account. The importance of the subject is my only excuse for trespassing on a well-cultivated piece of art-historical territory to which I have no prescriptive light of access.
The extant manuscript sources date from the middle of the fourteenth century,2 but these were compiled from earlier chronicles and obit rolis, which contained a basic list of dates and events to be commemorated, that are not really open to doubt. The one event which caught the attention of the early chroniclers was not the foundation of the church so much as the murder of its founder. This was noticed by Sigebert of Gembloux who put it in I098; otherwise the briefest of the local references occurs in the Annales Egmundenses, which state onder the year I099 that Bishop Conrad was struck down while celebrating mass 'in proprio domo', and was buried in the ecclesia sanctae Mariae

124 THE MARIAKERK AT UTRECHT, SPEYER, AND ITALY I

which he himself had built at great expense: The entry is in the first part of the chronicle, which up to the early twelfth century, was little more than excerpts from obit rolls. One other death that the annalist thought fit to mention was that of Emperor Henry IV in 1107 [sic], though there is nothing to indicate that this had any significance for the Mariakerk. The Mariakerk had its own annals,4 which simply recorded the killing of Bishop Conrad in 1099; but this was subsequently augmented by the scribe with the information that he was stabbed by a 'plebian'. The martyrologium adds that he was 'the founder of our church', and credits him with the establishment of the 'fratrum prebendam', i.e. the distribution of the endowments among the canons. Under 1106 the death of Henry IV was again listed without further comment. But the martyrologium reveals the reason tor his inclusion. There the emperor was referred to as 'our benefactor, who ordered the church to be consecrated and confirmed our possessions'.
Another fourteenth-century source was a list of bishops of Utrecht, of whom the last to be named by the original compiler was Johannes de Arkel (1342-64).5 This adds two further details to the medieval tradition about the origin of the Mariakerk. One is that Conrad built the church up against the walls of Utrecht in a manner reminiscent of a church at Milan which Conrad himself had destroyed. The other is that the murder was committed in the bishop's own hall by a certain Frisian who managed to escape when the deed was done.
As given in the list of bishops the story of the Milanese church is pure fantasy. Needless to say there is no evidence that a Lotharingian bishop was involved in the destruction of any Milanese church at the time when the Mariakerk was founded; and how such a curious fiction came to be fabricated is itself a substantial problem. One might easily conclude that by the middle of the fourteenth-century legends were forming around the Mariakerk. An alternative explanation, however, might be that a substantially true account has been rendered incomprehensible by abbreviation and corruption. Of this, more in due course. 
The identification of the 'plebian' of the martyrologium as a Frisian is also not beyond the need of further elucidation, though it is not in itself historically. incredible. A much fuller version is given in the chronicle of Johannes de Beka,6 where it reads like the precis of one of Chaucer's tales. Here the plebian bas become a certain Frisian by name Pleberus. The crafty bishop is said to have cheated him of the exorbitant payment he demanded for solving the problem of how to lay the foundations of the church; and in revenge the Frisian plotted the bishop's murder. De Beka does not mention the Milanese church, trom which it may be inferred that he was not familiar with the list of  bishops or its source. But the tact that he turned plebeo into the Frisian Pleberus suggests that he had access to other texts, if only to misread and embellish them.
For the fullest account of all, however, it is necessary to turn to one of Saenredam's interiors (Pl. IX). Apparently there were some not very distinguished latin hexameters displayed on the crossing piers, purporting to describe the origins of the church.7 Given
Saenredam's meticulous attention to detail, it may be assumed that these were faithfully reproduced in the painting. The translatton reads as follows:

Should any wish to know when and why
This ancient church was built, these verses will instruct you.
When Henry IV swayed sceptre over the Roman Empire And had subdued all Italy with his victorious troops, Milan alone daredc lose to him its rebellious gates.
Capturing it at last, at first the victor took his spoils,

PETER KIDSON 125

Then destroyed the town, in flames which did not spare
Even the white marble temple of the Holy Virgin Mary.
This saddened him. His soul afire with pious love He resolved to have another built Wherever else on earth he could.
At that point Conrad, bishop of Utrecht,
His former teacher and fervent comrade in arms,
Rendered the emperor a service by founding,
On an elevation in this place,
A high-towered church, built to last through the ages. In this he was helped by presents and a large donation from the emperor. So you see it, Standing on its firm columns,
The vault covering in perpetuity the lofty site.
The venerable bishop took it upon himself
To serve as primate of the church, instituting prebends
And canons to sing an everlasting song to you, 0 Virgin Mary.
Af ter the passage of three times six years
From the church's beginnings, the fatal hour arrived
When the holy bishop was killed through infamous murder,
For this reason: when the foundations for the walls Were being laid, the masons ran into a bottomless
And squelchy mire. Work came to a halt.
None of the experts could deal with the problem.
They could not believe that their highly paid staff, With unlimited funds, was unequal to the task. When all hope was given up, a boorish Frisian
Guaranteed that he would fill the pit. But when They heard his quite immonest fee, they balked, He quit, and work was stopped once more.
The man having a son he loved, the bishop called the lad
And offered him abribe. With his mother's help,
He got his father drunk, wheedled the technique out of him and,
Behind his father's back, told it to the priest.
At once the work, after long delay, was recommenced,
And all were overjoyed to know the secret trick.
The Frisian, though, enraged at being so deceived, Was driven to despair.
Beside himself,
He killed the bishop with a sharp knife
As he descended the stairs after celebrating Mass.
His death is commemorated in April by the inhabitants of Tivoli (Sic! - the latin is Tiburti). In the year 1099.

The text of these verses does not appear in the preparatory drawing in the Utrecht Municipal Archives, no doubt because the drawing was less than one third of the size of the finished picture. A pedestal and canopy on the left-hand pier, which had once been occupied by a statue, were also omitted in the drawing.To make room for the text on the left-hand pier, the picture was extended further to the left than the drawing. From all of this it can be inferred that the clean and chapter insisted that a legible account of what they rook to be the origins of the church should be preserved for posterity in the painted version.

126 THE MARIAKERK AT UTRECHT, SPEYER, AND ITALY

At first sight this narrative seems like nothing but an expanded version of the cryptic remarks in the list of bishops, making it even more of a farrago of nonsense. The tale of the Frisian's revenge was evidently based on Beka, or Beka's source. But by adding the detail that the murder took place eighteen years after the foundation, it deprives the story of whatever credibility Beka's version of it may have had. Either the problem of the foundation was insoluble for most of the eighteen years, or the Frisian nursed his grievance for a similar length of time. Neither seems likely, and more plausible explanations cao be proposed. It was said that Bishop Conrad was murdered because he was a schismatic;8 but it could just as weIl have been an episode in the on-going feud between the bishops of Utrecht and the counts of Holland, which from time to time erupted into active warfare.9 Either of these would imply that the motive for the murder was political. It is hard to believe that a mere artisan, whose identity must have been known, could have gote clean away, and eluded retribution for several years, if he was acting entirely on his own. In the end the crime was punished. In 1110, when the daughter of Henry I of England was on her war to become the Empress Matilda, the wife of Henry V, the happy pair, having met at Liège, moved on to Utrecht, where they celebrated Easter and were formally betrothed. In the course of the festivities the Emperor Henry found time to order the execution of an unnamed man who was alleged to have plotted the death of Bishop Conrad.1O No doubt a great deal about the affair will never be known, but one thing is certain: the case of a Frisian artisan would never have been referred to the Emperor. If imperial authority was needed for his execution, the plotter was a man of some social distinction. The actual assassin may have been a Frisian, and he may have been engaged in the building of the Mariakerk, but much more was involved than a personal quarrel between him and the bishop, and much of the story may be apocryphal.
The really important new item in the Saenredam text, was the prominent part assigned to the Emperor Henry IV. There was no precedent for this among the extant medieval sources apart from his designation in the martyrologium as 'our benefactor'. On the other hand, the reference to the Milanese church in the list of bishops can almost certainly be taken to imply that something akin to the Saenredam text was in circulation by the middle of the fourteenth century, for when the two are put side by side, it is clear that one is simply an abbreviated quotation from the other. With this in mind it is worth taking a closer look at the question of the Milanese church.
As history, the Saenredam text is, for all the extra detail, no great improvement on the list of bishops, but marginally it does make better sense to attribute the destruction of an Italian church to the emperor than to the bishop. At no time in his turbulent reign did Henry IV sack Milan or destroy a Milanese church. He was in Milan on precisely one recorded occasion, 14 April 1081, and then only briefly en route for Pavia and Rome.11 There was, however, one north Italian city which was besieged and captured by Henry IV. This was Mantua. In 1090 he mounted an expedition to Italy with the express intention of dispossessing the Countess Matilda, the staunchest ally of his arch- enemy the Pope, of the vast estates between the Apennines and the Po, which she had inherited at the age of nine in 1055. The siege lasted from June 1090 to April 1091. Whether it ended with a sack is not recorded, but it is just possible that a church of the Virgin was destroyed on that occasion, for there is no such dedication among the older churches of the city. Bishop Conrad of Utrecht was with him on this campaign,12 and was present when Mantua fell. Thus it transpires that a verifiable historical event was turned into a romantic fable by the corruption of the single word Mantova into Mediolanum.

PETER KIDSON I27

The Saenredam text concludes with another bit of mystification by introducing the Tiburti, i.e. the inhabitants of Tivoli. But here again a simple emendation is all that is needed to bring sense out of nonsense, for the correct word was obviously Traiecti, i.e. the citizens of Utrecht. How these misunderstandings came to infect the transmission is in itself an interesting question; but the fact remains that far from being the sheer make- believe that it is of ten taken to he, the Saenredam text needs remarkably little editorial attention to turn it into a potentially useful historical document.
There is abundant evidence that the Henry IV connection was a source of pride and accepted without question at the Mariakerk in the late Middle Ages. A fifteenth-century cushion cover has survived in which he is depicted standing on one side of the Virgin with Bishop Conrad on the other, in a manner that clearly assigns to them equal status as founders. There was also a statue of the emperor on the roof of the choir, almost certainly medieval in date, which the charter had restored in 1632.13 By then it was no doubt expedient for the catholics of Utrecht to draw attention to this particular imperial founder, for his life-long hostility to the Roman papacy would ipso facto make him something of a hero in the eyes of the Calvinist reformers; and it was perhaps with this in mind that, in 1641, the charter paid out quite a large som for a portrait of the emperor, whom they explicitly named as 'our founder'.14
What is curious about the whole business is not so much that the early sources were decidedly reticent on the subject of Henry IV, but that signs of real interest cannot be inferred until the story of the Milanese church began to circulate, i.e. not much before the middle of the fourteenth century, by which time would-be historians were distinctly hazy about what actually happened. It is hardly surprising that in the twelfth century ecclesiastical chroniclers should have been at pains to play down the connection between the Mariakerk and Henry IV, for papal anathemas had left him with a reputation little short of anti-Christ. But what caused the pendulum to swing the other war, 250 years later, is harder to discern. It is perhaps conceivable that the fortuitous operations of the laws of inheritance had something to do with the revival of half- forgotten memories of former imperial associations. When Count William IV died, in 1345, the County of Holland passed to his sister Margaret, who happened to be the wife of the German emperor of the day: Louis of Bavaria. If it was the social rank of founders and benefactors that gave prestige to churches, the canons of the Mariakerk may weIl have decided that the cachet of having an imperial founder more than outweighed any stigma attached to his name. But in that case the only thing that mattered was Henry's title.
Put like that, the value of the Saenredam text, even when amended, as a reliable source about the origins of the Mariakerk, remains open to question. Nevertheless, it considerably augments the other documentary material, and if there had been no pictorial evidence, the building history of the church might be summarized in something like the following terms.
Alone among the sources, the Saenredam text gives the date of the foundation. It was eighteen years before the bishop's mulder, in c.1081. The documents are unanimous that Conrad was the founder, and there is nothing to suggest that Henry IV's participation began until aftel the 'sack' of Mantua in 1091. But if the bishop acted on his own initiative in 1081, it is unlikely that he looked far beyond Utrecht itself for ideas about the sort of church that he intended to build. In 1043 one of his predecessors, Bishop Bernoldus (c. 1027-54), had founded the Pieterskerk; and like Conrad in 1099, he was buried in his own church when he died.15 The Pieterskerk lay to the east of the cathedral, and the Janskerk to the north was also ascribed to Bernoldus. As the

128 THE MARIAKERK AT UTRECHT, SPEYER, AND ITALY

Mariakerk lay due west of the cathedral, it has been inferred that Bernoldus planned to build a cross of churches with the cathedral at the centre, to be named after the four great basilicas of Rome, and that Conrad's church was part of the implementation of this scheme. If so, there is astrong probability that it would have been conceived as a basilica of the kind that can still be geen in the Pieterskerk or the Janskerk.
The annals of Egmond state that 'the great expense' of the church was incurred by the bishop alone. The martyrologium called Henry 'our benefactor', but then specified that he 'authorised the dedication of altars and confirmed our possessions' which, rather carefully perhaps, stops weIl short of presenting him as co-founder. The author of the annals of St Mary's thought it important to add as an afterthought to entries reporting the murder of the bishop and the coronation of Henry V at Aachen, that the altar of the Virgin, which was later removed to the west end, had been consecrated by Conrad as the high altar - as though this proved that the church was indeed Conrad's. And though the omission of any mention of Henry IV in the reference to the Milanese church in the fourteenth-century list of bishops is confusing, it does at least preserve intact the article of faith that Conrad alone had been the founder of the Mariakerk.
That might be described as the early house tradition. But a quite different picture could be constructed trom sources outside Utrecht. This starts from what monern historians would probably think to be the most important thing about Conrad, namely that for twenty-three years, between 1076 and 1099 - the period when the power struggle between the pope and the emperor for control of the German church was at its height, he never once wavered in his support for the emperor. His origins are unknown, but he was said to have come from Swabia, and all the early steps in his career were taken in German churches such as Hildesheim and Goslar.16 He owed his elevation in 1076 entirely to the emperor and his loyalty was as valuable tothe imperial cause as it was rare among German ecclesiastics. The backing which he gave to Henry's attempt to depose Gregory VII, and his support for the anti-pope, Clement III, had the inevitable consequence of making him persona non grata at Rome, where technically he was in a state of schism. In 1087 he was put in charge of the education of Henry's son, the future Henry V, and in this capacity he must have spent much time at court. It is inconceivable that he was not familiar with the emperor's own architectural project at Speyer. Conrad was at Speyer with Henry IV in 1090, and was his companion in Italy for the first part of the campaign which culminated in the siege of Mantua. His murder in 1099 deprived Henry of one of his few personal friends, and the loss was all the more grievous in the unhappy circumstances of the last years of his reign. It would have been entirely appropriate if he had adopted the Mariakerk as a gesture of gratitude for his dead friend, whether or not he was suffering pangs of remorse for deeds of violence done at Mantua.
It is obvious that the Saenredam text is the only document in the case that does anything like justice to the close relationship between the emperor and the bishop. It is also quite explicit that it was 'this church', i.e. the church in which the verses were displayed, that received the benefit of Henry's generosity, and that there was only one church in which the emperor and the bishop were involved. It should therefore follow that the house tradition and the Saenredam text were referring to the same building. But were they, and if not, which of the two was mendacious?
This is where the visual record enters the argument. Saenredam's pictures should, indeed in the last resort they must, dovetail with the documentary evidence to sustain a single, historically coherent interpretation of the chronology of the building. But it is precisely here that the complications begin, for architectural historians have been

PETER KIDSON 129

virtually unanimous that the church depicted by Saenredam could not have been built in the time of Bishop Conrad or Henry IV. The consensus of recent opinion is that it belonged to the second quarter of the twelfth century.17 A previous generation even brought it forward into the second half of the twelfth century. The fact that the documents might lead to a different conclusion was regrettable, but, as a distinguished art historian once said 'if style tells me one thing and documents another, I will always trust my sense of style'; and the Mariakerk is a case in point. To escape from the impasse, architectural historians have postulated two Mariakerks: the first of which was founded in 1081 by Bishop Conrad; while the second was a complete rebuilding from same date in the twelfth century. The former, having satisfied the documents, then conveniently disappeared without trace; the latter was still standing in the seventeenth century. It so happens that the annals of St Mary's record under the year 1131 the occurrence of a fire in which a large part of the city of Utrecht was said to have been burnt; and the Mariakerk was one of several notable casualties singled out for specific mention.18 At first sight this is just what appears to be needed to justify the theory of the second church; but from that point of view it is almost certainly a snare and delusion. Chronicles abound with remarks of that sort, and unless there is a follow- up, they do not always have to be taken at their face value. In the next fire to afflict the city, which happened in 1148, both the Pieterskerk and the Janskerk were said to have been burnt,19 but it is doubtful whether many architectural historians would be prepared to redate either of them on the strength of this evidence. The year af ter the fire of 1131, a reconsecration service was performed by the Archbishop of Cologne and four bishops.20 As it was not thought necessary to specify the church which received these ministrations, the obvious inference is that it was the annalist's own church, i.e. the Mariakerk. If so, the fire damage can only have been very superficial. Certainly two or three years later the church was in full commission when Floris of Swarte sought sanctuary there for three months before he was murdered (1134-35)23 It has been claimed that the removal of Conrad's high altar to the west end is evidence that rebuilding was in progress. A date of c. 1138 has been proposed for this change of location, though why a rearrangement of the altars could not have been the reason for the re-consecration service performed by the Archbishop of Cologne in 1132 is not clear.
If there was a rebuilding in the twelfth century, the clues in the annals are curiously muted and oblique. No one reading them would instantly leap to the conclusion that there was a second church, unless they were under pressure to find one. The idea was perhaps permissible as long as the narrative in the Hamburg painting could be treated as a romantic invention; but once it is recognized that there was a sub stratum of sound history behind the embroideries, the theory of the second church becomes an expedient of last resort. Before going to such lengths it is worth asking whether the stylistic arguments are so watertight as to exclude any other possibility. Not much more can be done about the documents, but art-historical inferences are always open to revision, and in this case they are positively crying out for re-examination.
The reluctance of architectural historians to subscribe to the view that the Mariakerk of the documents and the church painted by Saenredam were one and the same, sterns from the current orthodoxy about the evolution of Romanesque architecture in the lower Rhine-Meuse area. This has come to be organized around a single critical event, namely the inception or reception of what the Germans call the 'gebundenes system' (gebonden stelsel in Dutch,3 système Lombardo Rhenan in French). Before this arrived most of the great churches of north-west Europe still fell into the category of the

130 THE MARIAKERK AT UTRECHT, SPEYER, AND ITALY

unvaulted basilica, which had descended with very little change from prototypes which belonged to the early Christian period. These were essentially aggregates of large blocks of interior space shaped by walls or arcades, i.e. the nave, aisles and transepts of the familiar church plan. The gebundenes system may be described as the articulation of these primary spaces by the device of inserting into them smaller cells framed by wall- shafts, arches, and in theory if not always in practice, some species of cross vault. The great questions for architectural historians are when did this ncw type of church make its appearance in north-western Europe, and where did it originate?
The One thing about which there is no doubt, is that the Mariakerk in Saenredam's pictures was a church of this kind. It was, however, by no means a typical example. The original cast end was extensively Gothicized in the fourteenth century, and for some reason Saenredam paid very little attention to the choir. But even in the parts on which he concentrated,i.e. the nave and transepts, there were three unusual features. Of thirty-three churches cited by Kubach and Verbeek as in stances of the gebundenes system in the region,23 all but three were two-storeyed, and only two had proper galleries above their side aisles. Both were in Utrecht. One was the Mariakerk, and the other, the Nicolaaskerk, was patently monelled on the Mariakerk.
A second peculiarity was the presenceof a dwarf transept in the middle of the nave. Here again the sample in Kubach and Verbeek included no more than three instances. The second was at Kerkrade (Rolduc), near Maastricht;24 while the third was once a feature of the Onze Lieve Vrouwekerk at Maastricht. These nave transepts were clearly related to One another, and as they were an intrusive element, the question of which ca me first is likely to be settled only on the basis of contact with sources outside the area.
The third rarity was the rib bed vaults of the nave and transepts. Here it was very much a question of date, for ribs became common among the later instances of the gebundenes system. More. than anything el se it was the presence of ribs which persuaded Gall that the Mariakerk belonged to the second half of the twelfth century. Firm dates were hard to fiod, but there was One documented date which seemed to provide a plausible peg on which the whole chronology of ribs in the Rhineland could be hung. This was 1159, when there was a fire at Speyer which provided a convenient terminus post quem for the ribs in the transepts there.25 As the nave of Speyer had groined vaults and the choir was barrel-vaulted, the sense of evolution was difficult to resist. During the inter-war years and for long afterwards, it was axiomatic that ribbed vaults were later than groined vaults. It was also widely accepted that rib bed vaults originated in France, from where, by a process akin to osmosis, they gradually percolated into the surrounding areas, though not to any great extent before the second half of the twelfth century. Speyer appeared to confirm this theory to perfection, and it became the pivot around which undated or controversial ribs, like those of Worms, Murbach or Utrecht could be grouped.
In recent year~, this edifice bas begun to crumble. Dendrochronology produced dates between 1132-36 for the timberwork at Worms, which made a date in the 1160s for the vaults untenable.26 Then, in the aftermath of the restoration of Speyer (1958-71), it emerged that the ribs of the transepts could not have been inserted into a structure that was not already standingf7 Accordingly, they ceased to have any connection with the fire of 1159. If the ribs were integrated into the upper stages of the transept, they had to be part of the remoneUing of Speyer that was set in motion by Henry IV; though it could still be argued that they belonged to the latest phase of that operation. This permitted a cautious revision of the chronology. Rather reluctantly, One feeIs, Kubach

PETER KIDSON 131

settled for a date in the 1130s,28 and it was in the wake of Speyer that the hypothetical second version of the Mariakerk at Utrecht came to be located in the same decade.
These reappraisals have been made almost entirely on the basis of archaeological evidence. As steps in the right direction they are weltome, but an historian might think that they do not go nearly far enough. To an extent that makes it unique among European cathedrais, Speyer was deeply embroiled in the momentous events of the years when it was being updated for Henry IV, and the building cannot be properly understood without reference to them. When the purpose that inspired the work is taken into account, it becomes very hard indeed to resist the conclusion that the ribs were an essential part of Henry's project, and almost impossible to sec how they could have been added to the design at any time af ter Henry's death in 1106. To justify these claims a digression is required into certain aspects of the emperor's quarrel with the Roman papacy.
In its first incarnation Speyer was started by Conrad 11 in 1030, continued by Henry 111, and completed in 1061 during the minority of Henry IV. In 1030 Conrad's intention was to make Speyer his own burial church, and in this respect it was not fundamentally different from ütto the Great's Magdeburg or Henry lI's Bamberg. Although the building cannot have been far advanced when he died in 1039 Conrad was in fact buried there. It was Henry 111 who broke with precedent. Instead of adopting or founding a church of his own for the purpose, he toa elected to be buried at Speyer; and it was this decision that launched Speyer on its career as the mausoleum of the Salian dynasty. Henry IV rook the idea One stage further. The inspiration of the vast campaign of works which commenced shortly af ter 1081 was nothing less than the wish to convert the cathedral into a vast imperia I mortuary chapel, or collection of such chapels. The physical evidence of this intention is the presence of six specially prepared tomb chambers let into the thickness of the walls - two in the choir, One on each side of the high altar, and two in ead1 of the end walls of the transept. It may be conjectured that the two choir niches were designated for Conrad and Henry 111, and that the transepts were reserved for Henry IV, his son, and the royalladies, though as the whole scheme was aborted before it was implemented, it is pointless to speculate about the details.
So far as I am aware, it has never previously been suggested that the recesses were intended for tombs; but no other satisfactory function has ever been found for them, and the circumstantial case is astrong anc. The transepts were the principal beneficiaries of Henry IV's attention.29 In Speyer I they were little more than bare boxes of masonry, devoid of both altars and articulation. The immense envelope of cladding that in stage 11 allowed the niches to be inserted in the end walis, also made provision on the cast side for shallow arses. In other words, niches and altars arrived together.
I The transformation of the transepts into veritable chapels was further enhanced by the
cross of arches overhead. Whatever the structural advantages of rib bed vaults, in medieval eyes they would almost certainly convey the sense of a ciborium over a consecrated space of especial holiness, or at least express an aspiration to be understood in that way. The ribs may have been an afterthought in so far as they were not foreseen in 1081, but they added an extra nare of solemnity which distinguished the transepts from the nave, and which would have been meaningless unIe ss they were introduced while it was still the purpose of the transepts to house the imperia I tombs..
Externally, it was planned that the windows of the transepts should receive elaborate frames of carved ornament, and if all of them had been embellished in this war no One would have been in any doubt that the transepts were the most important part of the

132 THE MARIAKERK AT UTRECHT, SPEYER, AND ITALY

church. But the programme was abandoned rather precipitously in mid-window}o This is perhaps the most important clue of all, for the abrupt departure of the sculptors signified not jOgt the death of Henry IV in IIO6, but the premature col lapse of all his burial plans.
Henry died at Liège. His wish to be buried at Speyer was not at first respected; but by order of the papallegate; the body was taken to Speyer, where it lar for five years in an unconsecrated chapel, which Godfrey of Viterbo identified as the Afrakapelle}l This was because he died in a state of excommunication. When in the end the Church gave war, the place of burial was not in the transept but at the east end of the nave. None of the niches was ever occupied by a tomb; and whatever soul masses were celebrated on behalf of the defunct Salians rook place at the nave altar, or in the chorus regum which was subsequently set up over the royal tombs as a kind of collegiate chantry.
There is nothing in the documents to indicate that the withholding of burial rites was used by the Church as a bargaining ploy to confound Henry's grand design for the exaltation of the Salians. But given the architectural evidence that there was such a scheme, and the historical evidence that it came to nothing, it is reasonable to suppose that the matter was raised in the course of negotiations between Henry V and Paschal II at Rome in 1110, when the papal ban was finally lifted.
It would be a mistake to underestimate the seriousness of the tomb issue. For both sides what was at stake was a matter of principle, indeed nothing less than the issue at the heart of the quarrel between pope and emperor. Henry IV was heir to an exceedingly lofty notion of his status as a consecrated ruler, by virtue of which he considered himself entitled to exercise far-reaching powers over the affairs of the Church. His father, Henry III, consistently invoked the sacral kingship as the foundation of his authority without compunction or challenge. If necessary, precedents could be cited among his predecessors extending back through Ottonians and Carolingians to Byzantine emperors and Constantine himself. Then, in the middle of the eleventh century this whole tradition was suddenly thrown into question by a radically different conception of the light order of power in the world, emanating from Rome, and based on the fundamental distinction between secular and sacerdotal. From this point of view kings and emperors were no more than laymen, and however great their temporal privileges, these did not, or should not, allow them to cross the lines of demarcation which separated the spheres of the laity and the spiritual hierarchy. These extended to where and how laymen should be buried in church.
What Henry IV appeared to be doing at Speyer was expropriating strictly ecclesiastical symbols, like the croisée d'ogives, to proclaim that consecrated kings deserved the sort of veneration normally reserved for saints or relics. The gesture was all the more provocative for being made at the moment when the conflict between Henry and Gregory VII reached its climax, shortly after they had solemnly deposed one another in 1080, and when Henry was about to set out for Rome in order to be crowned emperor by his anti-pope.. The Church party could not fail to construe the plan as a flagrant act of calculated defiance; it could not afford to turn a blind ere, and in the prevailing mood of intransigence had no option but to veto it. It was still a sticking point in 1110. By then, however, Henry V was no langer fighting for the sacral kingship as this had been understood by his father or grandfather. He had shifted his ground, and when at last he was in a position to wring concessions from the pope, it probably cast him little to give way on the question of where the late emperor should be buried.
It follows from this interpretation that to all intents and purposes the ribs of Speyer were already in position when Henry IV died in 11O6. Unless they had passed the point

PETER KIDSON 133

of no return, they would have been jettisoned once the prime mover was no langer there. If the argument is sound, the case for postulating a second Mariakerk in the 1130s evaporates, and with it go the archaeological reasons for disregarding the account of the origin of the church (suitably corrected) given in the Saenredam text. There was only one Mariakerk. Bishop Conrad was the founder in the sense of setting up the chapter in 1081, but in all probability the church itself was not designed until the late 1090s. Given Henry IV's friendship with Bishop Conrad, and his recorded interest in the foundation, it makes good sense to suppose that the man who was responsible for the vaults at Utrecht, was seconded from Speyer when he had completed the task of putting ribs over the transepts there. Thequestion then arises whether it was the same man who gave the Mariakerk the galleries and nave transepts which, together with the ribs, set it apart from the main body of gebundenes system churches in the lower Rhine-Meuse area. If so, there is a very strong likelihood that the man was Italian.
The Mariakerk bas often been called an Italian church in the north, but how it got there is a question that bas usually been side-stepped, as though it were a trap for the unwary, and answers have tended to be very cautious. This is another consequence of the reluctance to take the Saenredam text seriously, for the story of the Italian church seems to offer the rudiments of an explanation. If Bishop Conrad was present when Mantua was captured in 1091, and the Mariakerk became an expiatory offering for a church destroyed on that occasion, it seems rather perverse not to connect these circumstances with the Italianate appearance of the building. At least the hypothesis ought to be explored further. The problems do not entirely vanish, but they are not perhaps as intractable as sometimes supposed.
It may be accepted for the sake of argument that the Mariakerk was not the replica of a destroyed Mantuan church, which, unless it were brand new, wouldprobably have been a simple basilica. The Italian churches that resembied the Mariakerk were not to be found in the lands of the Countess Matilda, but further west in Lombardy. There were some in Milan, but the places par excellence were Pavia and Novara. Oddly enough, the Romanesque cathedrals of both these cities, which are known to have had nave transepts, have disappeared: Pavia in the sixteenth century, Novara as lately as 1863. The evidence for the double cathedral of Pavia is meagre; but there are drawings of Novara which show it to have possessed all three of the features which made the Mariakerk exceptional: galleries, including one across the west front, combined with an internal porch, which was very rare; dwarf nave transepts, and rib bed vaults. The ribs rested on corbels, and in the nineteenth century it was inferred that they were additions,32 but the grouping of the clerestory windows of the nave implied that vaults were always intended. It is dangerous to compare two no langer extant buildings on the basis of drawings alone, but ostensibly the cathedralof Novara and the Mariakerk were too close for them not to be related. Is it possible that Novara was the model for Utrecht?
The trouble with almost all the relevant Lombard buildings is that none of them is firmly dated. On stylistic grounds Novara bas usually been put weIl into the twelfth century. Kingsley Porter offered a date c. 1125,33 which suited the theory of a second Mariakerk weIl enoUgh. The date could be disputed, but it is perhaps more to the point that Novara cannot have been the first church of its kind. It was toa provincial, and toa close to both Milan and Pavia to escape the influence of the architectural styles that were fashionable in those cities. Pavia in particular, as capitalof Lombardy, was still able, at the turn of the century, to compete vigorously with its ancient riyal in the matter of great churches, and it was probably there, rather than at Novara or Milan,

134 THE MARIAKERK AT UTRECHT, SPEYER, AND ITALY

that Henry IV recruited the expert who constructed the vaults of Speyer for him. Henry seems to have avoided Milan, but he was in Pavia on several occasions: in April I08I on his war to Rome, when he presumably taak the opportunity to have himself crowned king of Lombardy; again a year later; and he was there for several weeks at the end of I092 when the Mantuan campaign was over.34
As far as the vaults are concerned the argument is circumstantial. There is no general agreement that large-scale rib bed vaults were to be geen anywhere in Lombardy when Henry IV was there; although the reaction which set in after the failure of Kingsley Porter's crusade to instate Italy as the truc home of ribbed vaulting al most certainly went toa far in the opposite direction. The crucial thing about the Speyer ribs is that they were the work of someone who was familiar with Roman vaults. That is to sar there is no proper keystone. Instead One arch was built before the other, and appears to pass through the crown of the second arch, which takes the farm of two arcs abutting it. This was how the brick ribs embedded in the concrete of the Boths of Diocletian were built; and it may be surmised that part of the attraction for Henry IV lar in the knowledge that the croisée d'ogives was Roman as weIl as ecclesiological- a perfect symbol for a half Roman emperor. ThoUgh it does not followas a matter of course that the Speyer man was Italian, the bulk of the classical evidence was to be found in Italy, and the intuition which led Kingsley Porter to the view that the first moves in adapting what was in effect a Roman practice, for use in Romanesque churches, were made in Italy rather than in France, was essentially sound. One by one, his putative list of eleventh-century examples was whittled away; but so far as I am aware one case still remains outstanding: S. Nazaro Maggiore in Milan. This church was one of the casualties of the religious fiats which wrought havoc in the city in I075, and its restoration was the work of Archbishop Anselm 111, who was buried there when he died in I093. The diagonal ribs over the nef-unique of S Nazaro Maggiore have a span of over 20 m; which though not quite on the sc ale of those at Speyer (c. 25 m), at least proves that the necessary technology and experience were available in Italy.
With the sculpture we are on firmer ground. The idea of framing windows with braad bands of carved ornament was nowhere common, and apart Erom Speyer the most conspicuous instances are to be found in the choir of S Abbondio at Coma, for which there was apapal consecration in I096. One of the Coma window patterns was copied, more or less move for move, in the window of a chapel in the south transept of Quedlinburg, so there were sculptors with access to the Coma pattern hook active in Germany at about the right time. The men who carved the Speyer windows almost certainly knew S Abbondio, but an even closer parallel can be found in the pulpit of the church on the Isola di S Giulio in the lake of Orta just north of Novara. Here the similarities extend beyond motifs and formal arrangement to the war the shapes were actually cut; and it caD be inferred with confidence that the Orta sculptor worked at Speyer.
If the sculptors of Speyer came Erom Italy in the I090S it becomes very difficult to postulate a different date or provenance for the author of the vaults. Speyer was a remonelling job, and references to Italy were confined to details; but given a tree hand to design a complete church for his northern patrons, such a man would almost certainly come up with something out of the Pavian constellation - especially if his brief encouraged him to do so, which is precisely what, according to the Saenredam text, happened at Utrecht. AII in all that seemingly fanciful document turns out to have drifted no great distance Erom the verifiable facts of history, and the time has come for it to be rehabilitated among the primary sources of the Mariakerk.

PETER KIDSON I35

The conclusion that the ribbed vaults of Speyer and Utrecht belonged to the end of the eleventh century rather than a generation later is the limited objective of the present paper. It would make them almost exactly contemporary with the ribs of the choir of Durham, and open fresh vistas of speculation about the relation between Anglo-Norman architecture and the gebundenes system. In a more general sense it would provide prima facie grounds for reopening Kingsley Porter's argument that ribbed vaulting began in Italy; and it would do nothing but good if it helped to clarify the distinction between Romanesque and Gothic ribs. But these are all matters to be explored elsewhere.

ACKNOWLEDGEMENT
This is an amended version of the lecture delivered at the BAA conference at Utrecht in July I993. I wish to record my particular thanks to Koos Wynia-Gils, to whom I am indebted for information about Dutch sources.

REFERENCES
I. G.Schwartz and M. Jan Bok, Pieter Saenredam, the Painter and His Time (London I990), I3I.
2.. S. Muller, Fzn., 'Drie Utrechtsche Kroniekjes vóór Beka's Tijd', Bijdragen en Mededeelingen (Utrecht I888),465-508.
3. 'Annales Egmundenses' in Fontes Egmundenses, ed. o. Oppermann, Werken uitgegeven door het Historisch Genootshap, 3rd ser, 6I (I933), I34.
4. S. Muller, op. cit., fol. 465. 
5. Ibid., fol. 482..
6. Chronographia Iohannis de Beke, ed. H. Bruch ('s Gravenhage I973), 94-95.
7. The picture is in the Kunsthalle, Hamburg. See G. Schwartz and M. Jan Bok, op. cit., catalogue no. IS8. Latin text, 2.83-84, English translation, I4I-44.
8. G. Meyer VOD Knonau, jahrbücher des Deutschen Reiches unter Heinrich IV und HeinrichV (Leipzig I89O-I909), v, 68-69.
9. S. Muller, op. cit., fol. 497.
10. Annales Patherbrunnenses, ed. P. Scheffer-Borchorst (Innsbruck I870), I2.2.. 1
11. E. Kilian, Itinerar Kaiser Heinrichs IV, (I886).
I2.. G. Meyer VOD Knonau, op. cit., v, 335.
I3. G. Schwartz and M. Jan Bok, op. cit., I47.
I4. Loc. cito
I5.. S. Muller, op. cit, 49I.
I6. Series Episcoporum Ecclesiae Catholicae Occidentalis. Ser V, vol I, Archiepiscopatus Coloniensis (Stuttgart I982.), I93-94.
I7. H. E. Kubach and A. Verbeek, Romanische Baukunst an Rhein und Maas, 4 (Berlin I989), 2.3°.
I8. S. Muller, op. cit., 476.
I9. Ibid., 479-80.
20. Ibid., 477.
2I. Loc. cito
22.. H. E. Kubach and A. Verbeek, op. cit., 2.3°.
23. Ibid., figs I87-88.
24. Ibid., 2.47.
25. R. Kautzsch, 'Der Dom zu Speyer', Städel jahrbuch, I (I923), 7S-I08.
26. H. E. Kubach and A. Verbeek, op. cit., 2.32., n. 2.6A.
27. D. VOD Winterfeld, 'Die Rippengewolbe des Doms zu Speyer', jahrbuch des Vereins (ür Christliche Kunst in München, XVII (I988), IOI-I2.
28. H. E. Kubach and A. Verbeek, op, cit., 232.
29. H.E. Kubach and W. Haas, ed., Der Dom zu Speyer. Die Kunstdenkmäler von Rheinland-Pfalz (Berlin
1972), textband 673-77, 722-44, tafelband 9,19.
30. Ibid., bildband pl. 760.
31. Godfrey ofViterbo, Monumenta Germaniae Historica, Script XXII, 253.
32. A. Kingsley Porter, Lombard Architecture (New Haven 1915-17), lOS-IS. 33. Ibid., 115.
34. E. Kilian, op. cit., 89, 118, 147.

terug naar boven