EARLY RIB VAULTS IN ITALY
For a long time it has been thought that rib vaults were invented simultaneously but separately at Durham and in Lombardy, around 1100. The five buildings from which stems the dating for all the other ribbed vaults in Lombardy are San Ambrogio and San Nazaro in Milan, San Michele in Pavia, San Savino in Piacenza and Rivolta d'Adda.
As published in the Avista Forum
Dr Jane McKinne's doctoral thesis on Rivolta suggests a date around 1130 for its rib vaults. ["The church of S. Maria e S. Sigismondo in Rivolta d'Adda and the double-bay system in northern Italy in the late eleventh and early twelfth century", dissertation Berkeley 1986.] She hypothesises a date for San Nazaro of about 1110-12, and a similar date for the San Ambrogio narthex, with the nave vaults there after 1128. However, there are no documents for Pavia, Piacenza or San Nazaro that would securely date their rib vaults. The documents for San Ambrogio are open to various interpretations.
Also, most of these vaults have been restored or replaced. The cells of the easternmost bay at San Nazaro were replaced, though the most important part for this hypothesis, the ribs, were preserved. The San Savino vaults were partially reconstructed by Martini in 1902-3. The Rivolta vaults were reported as being in poor condition by its restorer, but we have no details of what work he did other than capping them with portland cement.
In four of these five buildings it is unlikely that ribs were intended from the beginning, for their pier bases and the capitals under the ribs are set square to the walls. The awkward junctions between the capitals and the ribs suggest that in all four the decision to have ribs was not made until after the walls had reached the height of the capitals. At Rivolta a Burgundian-type barrel vault had already been constructed over the two eastern bays before ribs were installed over the next two.
It is only in San Ambrogio that the capitals were set at 45o to the walls, in both the high vaults and the narthex. There is a document which states that the tower in the north-west corner was "largely built" (in maxima parte edificatum) by 1128. As the southern wall of the tower forms the northern wall of both the church and the narthex the narthex ribs would have been in place by then, and those to the high vault planned if not completed.
However, when I examined the capitals in the gallery I recognised forms that were familiar to me from capitals in northern France that we would date there to 1130 or later, but would not under any circumstances date a generation earlier (I am aware that some of these may have been in part or in whole restored - an Italian student is studying this at the moment). For this reason I became dubious of the usual date of 1100 for the San Ambrogio vaults, and spent part of this summer visiting these five churches in Italy. For this discussion I want to suggest:
1. that it was the earthquake of 1117 that stimulated the Italian builders to reinforce their big vaults with ribs, and
2. that the technique for these first ribs was introduced into Italy by an Anglo-Norman builder, possibly from Lessay around 1120.
The first point became more logical the more I thought about it. Early Lombard ribs occur only over large spans and never in aisles, suggesting they had a utilitarian purpose rather than an aesthetic one. If we could devise an experiment that would show whether earth tremors caused the stones in the groins of ribless vaults to fall out first, then we may have a good explanation for their decision to use ribs to relieve a post-earthquake anxiety. Ribs seem to have gone out of fashion by the middle of the century (until re-introduced in Gothic mode later on) suggesting that once theimmediate memory of the earthquake had passed builders no longer felt they had to use them. Lastly, even when restoration is taken into account, I saw few signs of the dislocation we should expect in ribs shaken by the earthquake as we would if they had been erected before 1117.
The second point needs a short introduction:
The earliest rib vaults in England and Normandy were constructed in the manner of groin vaults in which the role of the rib was more decorative than structural. This can be seen in the shape of the boss. Where the junctions between the boss and the ribs are not at 90o to the ribs, they are non-structural because these ribs could not have supported themselves if the cells had not been erected with them. Though some ribs had a provable structural purpose in the 1120s (Gloucester crypt) it took more than thirty years for the builders to understand that the rib could be constructed independently of the cells (Saint Denis ambulatory). It follows that until the masters realised this it is doubtful whether they would have perceived that the rib had more than a marginal effect on the structural integrity of the vault.
The three signs of early ribs are discussed in my Gesta article on Durham [xxii/2 1984, 135-145.]. They are: the shape of the boss, the ribs being offset at the crown and/or being twisted in plan. It follows that any ribs displaying these technical imperfections are likely to be earlier than ribs that could have been built to support themselves independently of the cells.
In Lombardy bent and offset ribs and non-structural bosses are to be found in the easternmost bays of all these five churches, while most of the western bays are technically perfect. This shows that the western bays are later than the eastern.
It is often said that the typical Italian rib vault is very domical, with pointed ribs and transverse arches. They are compared with the typical English and Norman vault with level crowns and round arches. To keep the crown level Anglo-Norman ribs were usually set out from the segment of a circle, in which the centre of the arc was located well below the capitals. It was for this reason that Bony thought that English and Italian rib vaults had been invented independently of one another.
However, in the earliest Lombard vaults the crowns of the vaults are not domical but level: in the eastern bays of San Nazaro, San Michele and San Savino. Further, the ribs are neither round nor pointed, but are segmental. Their spacial arrangement is therefore Anglo-Norman.
Where nearly all Lombard vault profiles, and all those later vaults with true bosses, are rectangular, those in the eastern bays of San Nazaro, San Michele and Brebbia, are circular. This circular profile is also a French and English characteristic.
This combination of non-structural bosses, level spatial organisation and circular rib sections is so typical of northern vaults that it is hard to believe that the concept of the rib and the mode of construction was not imported from there.
It has been argued that circular ribs are later than rectangular ribs as the latter, being simpler, must be earlier. We only have to compare the Durham ribs of 1100 with the square ones at Gloucester from c.1125 to raise doubts about this argument.
Some have wanted to date the vaults with circular profiles to the 1160s. As they all have non-structural bosses and bent ribs I believe this is much too late. The non-structural boss does not exist in northern Europe after 1140, or 1145 at the latest. By then every builder understood how the rib could be best used to support the vault, and the formwork techniques had been refined so they aided the workmen in the accurate laying of the stones. Once the Italians had understood this (as can be seen in most western bays) it is hard to believe they would revert to the more primitive technique two generations later. I believe we should consider dating these ribs to the 1120s.
If the earthquake was the trigger that brought the rib to Lombardy, I would postulate that they sought out some Anglo-Norman builder or someone who had worked in the north to advise them on how to reinforce vaults over the larger spans. This person brought the techniques and methods he had learned to Italy. Once the Italians had absorbed what he had to offer they improved on it, using heavier square sections for the ribs and making the vaults even safer with domical cells.
Some of the domical vaults in Lombardy with rectangular ribs have the same non-structural bosses and other technical imperfections as the Anglo-Norman vaults. Those that are structurally true tend, on the whole, to be those over the most western bays, and hence, in most cases, the last to be erected. In the north the true structural boss over straight ribs was rarely found before the mid 1130s, so it may be reasonable to suggest that these last vaults should be dated to the 1130s or later. The others would, by this argument, have been constructed between 1117 and the mid-'30s.
It should be remembered that the rib was not the only concept the Italians learnt from the north at this time - consider also alternation, square bays, barrel vaulting and, later, the Gothic style itself.
I hope these thoughts may help future discussion on this very important topic.
Click here for my other articles.