What follows represents the view of Chris Henige, creator of fabricae.org. It summarizes conclusions drawn from several years of surveying nearly 300 medieval churches in the Paris Basin.
A Large and Diverse Sampling
The Creation of Gothic Architecture (COGA) Survey (see creationofgothic.org) includes nearly 1500 buildings across the Ile-de-France in which at least a part of the building dates to the period between 1060 and 1250. These buildings have been visited by John James over a thirty-year period. James and Chris Henige visited nearly 300 of the more significant of these buildings in the summers of 2015, 2016, and 2017 and surveyed each in greater detail.
The sampling includes buildings that range in size from a single bay of a small building to the giant cathedrals. Each can be critical in understanding the slow development from the very mural and heavy architecture of the 11th century to the light and airy, skeletal structures of the 13th century.
The Importance of Seeing the Stones
The history of building in the Middle Ages is not well documented in contemporary sources. In fact, it is very poorly documented. Perhaps it was so omnipresent that no one took particular notice, or felt it was worth documenting every step of what might be a decades-long process. As such, it is the stones themselves that must convey to us their own history.
John James introduced the term “toichology” as a way of discriminating between traditional archaeology, which typically involves excavation and the analysis of underground remains, from what we might call “above-ground” archaeology, the study of standing structures.
The Greek word “toichos” refers simply to “a wall”. “Toichology”, then, is the “study of walls”, in particular what those walls can tell us about the construction history of a particular building.
A great deal of study has been given to the funding of the great buildings – the huge cathedrals and abbey churches. By contrast, very little has been given to the funding of the small rural churches.
Thinking Like a Patron
Who paid for these buildings?
Assumption: The large buildings – the cathedrals and the abbey churches – were undertaken with an understanding of fairly consistent funding over long periods of time. These institutions were at the top of the economic food chain, and there was a reasonable expectation that funding could be more stable. Bishops and abbots usually pledged a portion of their incomes, the chapter did the same, and funding drives across entire dioceses often provided the necessary resources to undertake entire buildings.
Assumption: The priests of the smaller rural churches could not rely on consistent funding over long periods of time, and so they usually had to rely on the local nobility for one-time contributions that might fund a single wing of the church which might also serve as the burial locale for that noble family.
Observation: As a result of the funding situation, large buildings usually have few projects, while the small buildings have many.
Thinking Like a Builder
When a builder was brought to a site, they first had to consider what could reasonably be achieved within the budget available.
Assumption: A builder would not want to come to a site and build a partial structure, one that would be subject to damage by the elements. Similarly, the patrons would not want to be left with a half-completed building that they couldn’t use, without any understanding of where the funding would come from to complete it. It is likely that if a builder started a project, he intended to finish it, from floor to roof, even if that roof was provisional.
Stylistic Continuity and Coherence
After visiting just a handful of the smaller buildings, it will become very evident that there was very little concern for stylistic unity and coherence. These buildings almost inevitably grew piecemeal, one bit at a time, and these bits were often separated by decades. Each portion is a reflection of its age, and almost no effort was made to try to link newer portions with the older ones in any meaningful way, stylistically.
Methodology – The Onsite Survey
There are a number of strategies we take when visiting a new building in order to attempt to discern its building history.
First, a few terms that are used throughout this article and throughout the website require clear and consistent definitions.
A project is defined here as a single building effort that is uninterrupted by any significant breaks in time. It is usually a complete structure, from floor to ceiling, and would have been roofed over to protect it. The project reflects the total amount of funding that would have been available to the patrons. They would not have undertaken a project if they did not believe they had the funding to finish it.
A project would typically be completed within a decade or two. Projects are separated from one another by significant lengths of time, often decades, where there was no funding or expectation of funding. Because they were conceived as wholes, from floor to roof, the joints between projects are usually vertical and easy to discern.
If it is understood that neither a patron nor a builder would want to undertake a new part of a building that they knew could not be completed and used given the funding available, then it follows that if there are changes that occur between the floor and the roof, these were not the result of a significant break in time, but in a change of concept during the project.
A phase is a portion of a project which shares a consistent approach to general design and detailing. It was not uncommon for there to be changes in the design intentions during the course of a project. Sometimes this change was initiated by the same master, who simply came up with a new idea. Other times it was the result of a new master coming to the site and adjusting the design to suit his own vision. It might also be that the patrons changed their requirements, either because they discovered something they wanted, or perhaps they were able to secure additional funding that permitted a more ambitious scheme.
The breaks between phases are much more subtle than those between projects. A small change in a molding, or the abandonment of one vaulting scheme in favor of another are visible in the details, but only on careful examination.
Because a project is defined as a continuous construction effort, and a phase is a portion of a project, there would be no break in time between phases unless something very unusual happened.
In exceptional cases, it is possible and useful to look even more closely at the sequence of construction. Sometimes the stones tell us enough to allow us to break down a phase of a building project to campaigns, defined here as the work of a single building season or less. If the same master and masons were present at a site over a number of years, it is almost impossible to discern exactly where they left off between seasons.
Initial Assessment – Gross Building Changes
It has proven very effective to start by trying to “chunk” the building – breaking it into discrete parts that are clearly different from one another in terms of general style or even subtle details. The first go around in this process is to try to identify each of the different projects.
Although there was little regard for the work of previous masons when it came to the next project at a particular building, the level of consistency within a project tended to be very high.
Differences in Overall Height
One of the most obvious visible attributes of different parts of buildings is their overall height. We do not generally consider the roofs in this analysis, because they are so often replaced, and just as often adjusted in their pitch. Sometimes there is evidence in adjacent walls as to the intended or actual line of the original roof, which can help.
More useful is the height of the walls themselves. Again, one must be somewhat careful here because the tops of the walls are an area that is subject to damage over time, from water, but also from the racking of the roof timbers in the wind. When roofs are replaced, the tops of the walls are often consolidated and solidified to provide adequate seating for the new framing.
One of the observations that has come out of the Survey is that it is very common for adjacent wings of smaller buildings to be different heights, even if that difference is small.
Changes in Plinth Design
As we move around a building, we look for the presence of any plinth setbacks near the base of the wall, and we look for discontinuities in the slope, the shape, and the size of that setback.
Changes in Window Design
Usually windows within a single project share the same character, but window design evolved over time, and it would be very unusual to see the same window designs used in two different projects decades apart.
Changes in Corbel Design
Although one has to be cautious when one observes the tops of the walls, when two projects of equal height come together, there is also usually a change in the design of the corbels at the tops of the walls. This too can indicate a change in project or phase.
Confirmation – Finding Joints
Having looked at these easily visible general characteristics of the buildings, we have formed an opinion about the nature and location of the chunks of the building in question. As noted above, because a project is defined as a floor-to-roof effort, breaks between projects tend to be vertical in nature.
Discontinuities in Horizontal Coursing
Although there may be differences in the heights of the courses, differences in height between stones within a course is highly undesirable. Because it is necessary to stagger the joints between stones from one course to the next, (think of a typical brick wall,) having stones of different heights would necessitate extra work to notch stones around one another.
Assumption: The heights of the stones in a single course will be equal, and when they are not, this probably signals a joint between projects or phases. However, in unusual cases a later mason may try to match the course heights of new work with those of existing work, making it more difficult to discern the joint. Even so, the match is rarely perfect, and close examination will usually expose the joint.
It is quite common for a new part of the building to simply butt against the old, creating a continuous butt joint running up the entire wall.
Buildings move, and it is likely that most butt joints are also pinned using iron pins set in lead. This keeps the joint from sliding in any direction.
Much less common is the stitching together of the old with the new using a “keyed” joint. Here an effort would be made to intersect one wall with another, and because the coursing of these two walls is almost never the same, it involved notching stones and considerable labor.
Logical Stopping Points
Because it is assumed that a project constitutes work from floor to roof, and separate projects are built against one another in some manner, one of the key places to look for joints on the exterior of a building is in the interior corners. Simply follow the courses from one wall around the corner to the adjacent wall. If they are consistent, they were built at the same time. If they are not consistent, there’s probably a break in the work.
Assumption: Masons have no interest in matching course heights of earlier work. They use their own measure, and make any necessary adjustments at the joint.
Sometimes a building might not have lateral wings, or perhaps a couple of bays of the nave were completed, and others were appended later. Here the discontinuity in course heights will usually occur at an exterior wall buttress. Sometimes the wall coursing is continuous on either side of the buttress, but the buttress itself has different coursing. This indicates that the buttress was added later.
Working vertically, one of the logical stopping points for erection is at the sills of the windows. This would occur at the top of the first course of carved stones of the window sills, or at the bottom of the window opening.
Another logical stopping point, particularly between seasons, is at the arches over windows or in the interior. A wall would be raised to a few courses above the springs of the arch, including the bottom stones of the arches themselves. This provided solid lateral support for the rest of the arch. Scaffolding was installed under the arch, the remainder of the arch was put in place, and the scaffolding was removed. Because of the nature of the mortar, the arch stones would settle slightly, and allowing this to happen before the rest of the wall was installed around the arch insured there would be no later separation of the arch from the wall. This was a good place to stop between seasons.
The coursing of things like piers often does not match the coursing of adjacent walls, or even adjacent piers. But it seems obvious that any shafts and capitals on either must be consistent, so when one gets up to the level of the bottom of the capitals, this height needs to be consistent, and therefore we consider this a “target” height.
Assumption: The capitals and imposts above them for any particular piece of work are all intended to be the same height. Therefore, the height of the top of the course immediately beneath them must be consistent.
Once the piers or wall shafts and their capitals and imposts are in place, it was not unusual to run timbers between them in order to stabilize them until the work above them could be undertaken. This is another logical “leaving-off” point.
Another logical target height was the bottom of the window sills, in particular the top of the plain course of the walls just below the first carved course of the sills. Insuring that this height was consistent insured that the windows would be consistent.
Changes in Stone Type and Quality
One of the first things that happens at the beginning of a project is the contracting of a quarry as a source of stone. This might not always be the same quarry for a particular building, or the same part of the quarry, or even the same stratum of a quarry. As a result, the stone will vary in quality and appearance. This is usually readily apparent in the walls of the finished building. Differences in color, texture, and wear are good indicators of a change.
Changes in Ashlar
The way the stone is cut and set also affect its appearance, and can signal change. The thickness of mortar joints and the sizes of the stones help guide us. In some cases, course heights vary in one part of the work, but are consistent and regular – almost modular – in other parts.
Reflecting on the Gross Observations
All of the considerations addressed above are useful in the initial analysis of a building. They are the obvious signals of change, which suggest changes in time or master. They are not subtle, and neither are the differences between projects.
This simple process of looking for these indicators should result in a reasonable partitioning of the building into its projects.
Refining the Assessment – Looking for Anomalies
Although the signs of change between projects are usually quite obvious and easy to discern, the indicators of change within a project are not always so obvious. In this next section we will learn to look for these more subtle indicators, and try to determine what they tell us about the building process, and about the history of the building in question.
Changes in Window Design
In the earlier section, the sills were considered a logical stopping point for construction, and the course below the sills a logical “target” height. Where these vary, even slightly, we should look for other evidence of change.
Sill and Jamb Profiles
The slopes on either side of the sill, interior and exterior, are generally consistent within the work of one crew. The same is true of the profile of the jambs running up each side of the window. When these change, from one window to the next or even occasionally within a window, this is an indication of a change.
Tracery refers to the stonework design within the general opening of the window. Its design tends to be a rather personal thing, and changes in it likely signal a change in master, or a change in concept.
Changes in Capitals
Who carved capitals and who carved imposts?
After having visited a great many buildings, it became clear that when the imposts are not carved with the capitals themselves, they are generally more consistent across a portion of a building than the capitals. This indicates that while the capitals are probably carved by a number of different individuals, each in their own idiom, the imposts constitute less “skilled” work, and were likely done by less skilled masons working to a common template.
Change in Carver
A change in carver is not a very reliable way of separating phases in a building. There are typically several to many different hands carving the capitals, and there is likely to be considerable variation among them.
Change in Impost Profile
Impost profiles tend to be far more consistent than the styles of the capitals, and the same impost profile will often be found above the work of numerous different carvers. Therefore, although the style of the capitals is not necessarily indicative of a change in phase, changes in the profile of the imposts above the capitals may be.
Change in Heights
Another good indicator of change is a difference in height between adjacent capitals and/or imposts.
Changes in Piers
Once a pier design is set out, it is not likely to be altered as it rises. Changes do not tend to occur within a pier, but from pier to pier.
Changes in Shaft Diameters
Even when the pier plan seems to be the same from one pier to the next, it is often the case that the diameters of the shafts on those piers vary enough to indicate that a different mason using a different geometry or system of measure has taken over. Mapping the shaft diameters in a building can be a very useful indicator of the changes in designers.
Changes in Bases
The template for a particular set of piers or wall shafts had to be the same for the top of the base above the torus and the underside of the capital, because the piers and shafts that connect the two are of course perfectly vertical and must remain consistent all the way up.
Changes in Plinths
Sometimes this change is manifested in a difference in the dimensions of the base and plinth at the floor level.
Changes in Torus Profile
In other cases, the profile of the torus or the presence or absence of griffes at the corners can indicate a change in designer.
Changes in Height of Torus above Floor
It is quite typical for the bases of the piers of different phases of a project to have different overall heights, another indicator of change in designer.
Changes in String Course
In buildings that have a string course below the window sills, or in the main elevation above and below the triforium and/or gallery, changes in their design are good indicators of change in crew.
Changes in Profile
This can be subtle, especially if the height of the molding above the floor remains the same.
Changes in Height above Floor
This is a more obvious and visible change.
Changes in Arches
Change in Profile
Except for the earliest period, arches are typically framed by some kind of molding. These change over time, and also with designer. If this profile changes, it is likely that the master mason changed.
Change in Shape
Whether an arch is round or pointed, and the degree to which it is pointed can be an indicator of change.
One of the more remarkable aspects of medieval building is a fairly consistent resistance to total replacement when parts of buildings are repaired or updated. It is quite common to see the filled-in remnants of earlier windows or other fragments of earlier work preserved in walls that otherwise have a later character.
Starts Without Finishes…and Vice-Versa
The architecture of this period is quite logical – a base leads to a shaft, a capital, and impost and an arch. Sometimes a pier design has shafts for one particular scheme of vaulting, and then when the time comes, one or more of these shafts ends up not supporting any arch above it. This can signal a change in master, but it also can simply indicate that the master who set out the piers, changed his mind when he got above the imposts.
It is also not uncommon for there to be no shaft at the floor level, but suddenly, partway up, a corbel supports a new shaft destined to support a new and different vaulting scheme.
Developing the Relative Chronology
The survey of nearly 1500 buildings and their building features and capitals has provided a massive amount of data about the development and evolution of medieval architecture. By correlating this data and calibrating the timelines of various kinds of evolution, it should be possible to develop a much more refined timeline for medieval architecture between 1060 and 1250.
Using Capitals Based on Stylistic Evolution
There is a clear general development of the design of capitals throughout the period, but using only the design of capitals as a means of dating buildings is subject to multiple possible pitfalls. A particular carver could work in the same mode for his whole career, at the end of which it would be considered “outdated”, while another might be particularly innovative and flexible. Capitals might be carved at one time and placed at another, capitals might even be reused.
In general using capitals can give a pretty good indication of date, but it must be combined with and validated by other toichological factors.
Recognizing the Currency of Features
There are a number of features that emerged during the period that can give an indication of date, or can mislead.
Flying buttresses emerged as a means of dealing with taller and thinner buildings, but it is often the case that they were added retroactively to existing buildings that were either showing signs of spreading at the tops of the walls, or just needed an aesthetic “update”.
During the period, there was a shift over time from simple lancets to combined groupings of openings to plate tracery to bar tracery. Knowing when these were current helps narrow the dating of various parts of a building. Again, there is a caveat. One of the simplest and quickest ways to update a building is to update the windows. Window design alone should not be used as the only indicator of date.
Diagonal Corner Buttresses
Throughout our period, from 1060 to 1250, buttresses at the corners consist of pairs set perpendicular to the adjacent walls. It is not until much later that the diagonal corner buttress made its appearance. In parts of buildings that otherwise have few or no dating indicators, the corner buttress can indicate that these simple additions were built much later as a way to complete an incomplete building.
Sadly, medieval commentators took remarkably little interest in all of the building activity around them. At this time, masons were probably not literate, and didn’t keep their own records, and the clergy was more interested in other pursuits. As such, very few accounts of building activity survive, and even these are sketchy at best.
The best sources are contemporary to the actual work, and not retrospective in nature. With the exception of Gervase at Canterbury, no one has left a year-by-year accounting of what was built, and even Gervase’s account is somewhat vague. Suger’s account of the rebuilding of the abbey church at Saint-Denis is retrospective and probably somewhat hagiographic and does not necessarily indicate direct involvement on his part. He simply took credit for it.
One of the more recent technologies that has helped to confirm or alter the dating of buildings is the study of tree-ring dates, called dendrochronology. This can be useful in dating fragments of timbers left behind from the building process, or in roofs. Care must be used in discerning whether the timbers in question are in fact contemporary with the buildings around and under them. But where the original roof framing actually survives, they provide a latest possible date for building beneath them.
The two World Wars caused a great deal of damage to many of the churches of the region. Pre-war and post-war photographs of these buildings tell us about the damage done, and the faithfulness of subsequent restorations.
Restoration and Maintenance Records
The French government maintains an archive of documentation of restoration of buildings that have been classified as historic monuments. Although it is generally quite easy to discern modern stonework from that of the medieval period, such documents can confirm where work may have been done in buildings.
One of the latest trends, one that actually complicates the analysis of the stonework, is the restoration of interior painting. At Chartres Cathedral, this has literally transformed the building from a dark interior where the windows dominated, to a light and airy space where interior painting and glazing complement one another. It has changed people’s perceptions about the nature of these spaces in the middle ages. But it also hides the actual stonework, and makes it impossible to read the interior joints. We hope that all such projects are accompanied by systematic photography of the “before” condition of the walls for future reference by interested scholars.
Calibrating Relative Chronologies
The final step in this project will be to take the relative chronologies of carvers, of building features, of the documentation and calibrate them all in a way that results in a credible and highly-refined chronology for building in the period.
Some Conclusions So Far
What Have We Learned About Patrons?
The small churches could not rely on consistent funding, and the projects seen in those buildings are often confined to single wings or chapels. It was even quite common for the central vessel of the nave to be constructed, with an arcade anticipating aisles, but for those aisles to either never be constructed, or to be constructed much, much later. Funding came in small chunks, and resulted in many projects resulting in small chunks of buildings.
In the larger buildings, cathedrals and abbey churches, the chunks of funding were much larger and more reliably consistent. They resulted in much more consistent buildings, typically built within continuous windows of time – few projects resulting in larger sections of buildings, or even complete buildings.
What Have We Learned About Builders?
The builders came to a site and were faced with a simple question: What can you build for us with the money we have on hand?
Masters versus Masons versus Carvers versus Erectors
The survey has revealed some hints about how builders worked, and the division of labor on site.
The Master of the Works was a position that probably emerged out of necessity at the largest projects. There was no such thing as a “supercontractor” in the medieval period, and there were no massive crews such as might be required on a large building site. It is likely that there were multiple crews on these sites, each with their own master and each working in their own idiom. What was needed was for someone to take ownership of the overall design concept and convey – not impose – that on the individual crew masters, to coordinate the work on the site, and to act as intermediary with the patrons.
At the smaller projects, Masters had a vision of the whole of the project at hand, and directed the building activity. They may have provided templates to the masons under them for particular details. They directed their own crews and workshops, even on a larger site where they were guided by the vision of the Master of the Works.
Carvers seem to have been much more transient, moving from building to building as their skills were needed – perhaps several in a single season. They had the skills to carve capitals and figure sculpture. It seems to be no coincidence that the simpler crocket capitals became prevalent at the same time as more emphasis was being placed on portal sculpture. Perhaps the carvers gravitated toward these portal projects at the expense of the fancier capitals, and the masons with less skill took over the responsibility for the capitals. They were less ambitious, and well suited to the rather formulaic crocket capitals. The work of the carvers was probably left in the sheds until such time as they were ready to be placed.
Masons dealt with simpler carving such as continuous moldings.
Erectors were the unskilled laborers who put everything in place. In more complicated construction, they required guidance, and where this guidance was absent, we frequently see errors in placement. This is particularly true with the finer sculpture left behind by carvers who had long since moved on.
Strategies for Building over Time
Another of the revelations of the Survey has been exposing the strategies of the builders, depending on the nature of the funding available.
Small Buildings versus Large
For the small buildings, where a particular donation was probably all that was available or expected, builders limited their work to whatever they could cover with a roof. This simple formula is evident everywhere, and it explains why no building looks like any other building. They are all amalgamations of different chunks constructed at different times by different teams under the direction of different masters. Consistency and unity were not of interest to these men, because they could not be achieved.
With the large buildings, there was a long-term plan, because there was an expectation of continuous funding. But it was also evident that the duration of a project was often longer than the career of a Master of the Works, so different strategies were employed to maintain some level of consistency, and to insure that a new master could take over the project from another without significant disruption to the overall scheme. Buildings tended to be stratified rather than wing-by-wing, because if the same master and his crew set all the bases, and a different master and his crew set all of the capitals, who would be able to tell? All the bases would look the same, all of the capitals would be similar, and unity would be achieved even if the approaches of the two masters were completely different in character. This is why the large buildings appear so much more consistent than the smaller ones, all because of far more reliable and consistent funding.