Maxime L’Héritier (Université Paris 8, ArScAn UMR 7041) and Dorothée Chaoui-Derieux (Direction Régionale des Affaires Culturelles, Service Régional de l’Archéologie)
This paper was written in collaboration with Aurélia Azéma, Béatrice Bouet, Alexis Komenda, Catherine Lavier, Lise Leroux, Emmanuel Maurin, Florence Mousset, Olivier Puaux, Delphine Syvilay and Thierry Zimmer
At time of writing, it has been three years to the day since the catastrophic fires decimated the cathedral of Notre-Dame de Paris (France) on 15th April, 2019. In the aftermath of the flames, scientific communities from France as well as abroad gathered to try to save as much of the historical, heritage and archaeological data from this landmark building as humanly possible. Part of the building had literally gone up in smoke, while another part had collapsed. In addition to the sorting and selection of archaeological remains (the realization of which has allowed many researchers to initiate long-term research programmes), the restoration work represents a unique opportunity to observe parts of the building in detail which were previously inaccessible.
Organised clearing of the remains: the first stage of the ‘assembly line’
Beginning at the end of April 2019, the Historic Monuments Research Laboratory (LRMH), the Regional Directorate of Cultural Affairs/Regional Archaeology Service (DRAC/SRA) and the Centre for Research and Restoration of the Museums of France (C2RMF) began close collaboration to set up a protocol for sorting, removing and inventorying the elements of the grande dame that had collapsed covering the ground as well as the vaults prior to their removal. This protocol was intended to be easily applicable and to explain the criteria put in place for the selection of remains for preservation (i.e. with regards to their particular scientific interest or importance relative to cultural heritage). In April, examination of the disaster site revealed three principle concentrations of debris located in the northern transept, the transept crossing and the eastern part of the nave. In addition to these, the vaults were covered by the remains of the burnt framework and metal reinforcements of the building and spire as well as some lead elements that escaped the fire.
In the building, the clearing and sorting of the remains began a few days after the disaster and lasted until mid-November 2019. Due to the dangers of lead pollution and the hazardous nature of the site, the entirety of this work took place under very strict intervention conditions which furthermore prohibited easy access. The remains were removed by two remote-controlled robots that placed them on an improvised sorting table where they were then subjected to a rapid (but careful!) examination by scientists equipped with protective suits and breathing equipment. See Figure 7. Those blocks that were kept were then numbered and placed on a pallet bearing an inventory number: "TN (north transept)", "CR (transept crossing)" and "N (nave)". Each stage of this process has been documented photographically. Over twenty thousand shots were taken by the sorters and the C2RMF photographer. A specific orthophotographic campaign was dedicated to the pile in the eastern part of the nave where one of the cathedral's medieval double arches had collapsed. Eight photographic sensors were installed on the capitals' abacuses and were connected to a tablet, making it possible to trigger on-demand shots documenting the progress of the excavation phase almost in real time. As a result of these efforts, it was possible to reveal the true stratigraphy of the collapsed cluster. Saved to a tablet, the photos permitted live annotation and a finer numbering of the remains as they were removed from the nave.
Figure 7. Robots at work clearing the archaeological remains in the transept crossing before their sorting and inventory. Photo ©Dorothée Chaoui-Derieux / SRA.
From the summer of 2019 until March 2021, the sorting and removal of the collapsed structural elements located on the vaults were carried out. Using photographs taken by a drone just after the fire, a grid was set up on the vaults of the choir, the nave and the two arms of the transept. These operations were also monitored by means of photogrammetric acquisition using a 3D laser scanner, as well as by a device specifically designed by the National Centre for Scientific Research’s (CNRS) MAP (Models and Simulations for Architecture and Heritage) laboratory. This cable-cam system was installed along rails above the vaults and can automatically obtain hundreds of photographs as well as rapid 2D and 3D records of all the stages of removal. The large timbers were individually numbered on the photographs and then labelled in situ by steeplejack technicians, who then removed them to the marquees set up on the forecourt, where they could be observed by specialists. In addition to these individual elements of the framework, phenomenal quantities of ash and charcoal were collected in large numbered bags. These bags were filled and then emptied and sorted in the marquees by small teams from the different institutions involved (LRMH, SRA, C2RMF, Paris 8 University and CNRS), assisted by the companies present on the site, which set aside all the archaeological remains: forged nails and pegs, roof ridge elements (lead/internal iron frames), bell fragments and clock gears, etc.
Sorting and inventorying materials: a methodology forged in the field, after the fire
During these sorting operations, researchers from different institutions and backgrounds with complementary fields of expertise enabled extremely fruitful collaboration around the sorting table. Archaeologists and specialists in wood, stone, stained glass and metal shared their skills towards a single cause: so that the most information and remains as possible could be salvaged and evaluated. The objectives were manifold: to better understand the cathedral in terms of context, materials and construction site. The ultimate goal remained aiding with restoration. But, we must always contend with the possibility that it may be necessary to replace certain elements of the vaults, spire or roof. This is why, for example, the molten lead retrieved during these operations has been carefully isolated in dedicated bags, with a view to a possible later recasting and reuse, once impurities have been removed.
One further illustration: the wood specialists working on site decided to keep all wood with a cross-section larger than five centimetres. See Figure 8. Even pieces several meters in length have been carefully set aside, regardless of whether they date from the medieval period or the 19th century or whether they originally came from the spire or the framework of the nave and choir. Each sample is attentively photodocumented beforehand, so that no information about context is lost. Once sorted or extracted from the heaps, the timbers are labelled, numbered and subjected to initial observations in situ, before being stored in the marquees.
Figure 8. Numbering of wooden beams from the framework collected on the vaults before taking them down. Photo ©C2RMF / Alexis Komenda, with permission.
For the stone material, numerous sorting criteria were defined for preservation. These include, e.g., having at least three sides preserved, the presence of polychromy or sculpted decoration, having stone marks, and/or the presence of mortar, etc. Blocks that were too fragmentary were not inventoried, but were nonetheless stored in large scrap bags which are kept in warehouses under the responsibility of the Etablissement Public in charge of the conservation and Restoration of Notre-Dame de Paris cathedral (EPRNDP). It is hoped that this will eventually enable studies to be carried out on the cathedral’s supplies and construction methods, as well as its economic context, and/or to furthermore diagnose damage to the vaults that have not collapsed as well as the upper portions of the cathedral that were also subjected to the fire. The first identifications and assessments were made in the field during the sorting process, which provided an opportunity for careful and statistically representative repeated observations. Here again, each block is photographed prior to removal, then placed on pallets bearing an inventory number before transfer to the marquees.
In terms of metal, distinctions were made between molten lead (which was generally isolated, but also sometimes sampled to assess the nature of the mixtures generated by the fire), iron from the spire or the framework, nails of different sizes and shapes and copper from the gutters. Finally, many small fragments of charcoal remain, most of which end up in large bags like those already mentioned.
Once this sorting period was over, there followed a phase in which researchers made a more detailed inventory of the pallets kept in the marquees. In addition to their location on the ground, the typology of each type of vestige (e.g. stone wedges, iron pegs or lead decorations) has been fundamental for their identification. A specific database was created which allowed for an initial listing of those remains which required urgent rescue. Due to the very restrictive physical working conditions (mask, motor, gloves, etc.) and the changing composition of the small team on the site, field inventory sheets were set up to avoid data entry errors. The database was updated using these collected sheets and the numbered pallet shots.
At the end of the operations, nearly 10,000 pieces of wood, over 650 pallets of stone and nearly 350 pallets of metal were removed from the site. All of the remains are now kept in warehouses specifically rented by the EPRNDP in the north of Paris, and placed under the responsibility of the DRAC. Though they do not at present form the body of the cathedral, these building parts are nonetheless classified as historic monuments (they have in fact become classified as “movable objects” by law) and further benefit from having the status of being "movable archaeological property". An agreement signed between the DRAC and the CNRS regulates all requests for access to the remains and sampling (destructive or not) subject to prior agreement from the DRAC. Without this heavy sorting work, carried out under difficult conditions and under duress while still in alignment with the most rigorous methodology possible given the context, the removed remains would no longer have any scientific value. This was, therefore, the first stage in the scientific process, which will now provide input for a large number of forward-looking research projects.
Setting up a scientific project
In addition to the security worksite itself, researchers from France and around the globe gathered under the auspices of a scientific association (Scientifiques au service de la restauration de Notre-Dame de Paris) in order to put their knowledge and expertise at the service of the restoration of Notre-Dame. Shortly thereafter, this initiative was followed by national research and heritage institutions. From June 2019, eight working groups were set up as part of the scientific program coordinated by the Ministry of Culture and the CNRS. The programme brings together nearly 175 researchers on wood and framework, acoustics, stone, metal, glass, structure and civil engineering, emotions/commitment, digital data and monumental decoration (the latter added in 2021) from all over the world. The groups specifically working on materials needed access to the remains, not only to answer certain questions from the architects in charge of the building conservation/restoration project or from the contracting authority, but also to launch research programmes. Access to the site only became possible from the beginning of 2020. In addition to the sorting and selection of the remains, the restoration work represents a unique opportunity to observe in detail parts of the building that were previously inaccessible. For example, the collapse of the roof structure has revealed iron staples on the gutter walls, a detail that had been unknown prior to the fire. After a phase of in situ survey and measurements (which required the use of lifelines and/or suspension under the temporary roof; see Figure 9) some of the staples were removed. As some of these will not be reused as part of reconstruction efforts they have now become the subject of a series of studies. The analyses provide further reflection on the use of such material in the cathedral structure from the medieval period onwards, as well as on the metal’s provenance. This has allowed scholars to reconstruct the supplies funnelled into the building site over time, and also sheds light on the building's construction campaigns through radiocarbon dating of ferrous alloys. The evaluation of the quality and mechanical behaviour of these ferrous alloys highlights craftspeople’s knowledge of ancient materials as well as also providing information for their further reuse in the view of the building’s restoration.
Figure 9. Survey and sampling of metallic armatures (staples) on the top walls. Photo ©Maxime L'Héritier.
Similarly, the presence of scaffolding inside the building allows stone specialists to investigate the vault and sample stones and mortars. The stones that have fallen to the ground are also under examination, particularly in the collapsed double arch of the nave (which could be reassembled on the ground). See Figure 10. Close examination (i.e. under each face) of the stones which made up the arch allowed for hitherto impossible observations which have provided information on the location of the stone quarries, the direction of stone-laying (concluded by observation of stone marks; an ‘x’ is visible in the third block from the right in Figure 10) and construction date (some hints in this regard were given by notches in the arches used to lay the vaults). In addition, the stone can be physically and mechanically characterised, leading to structural modelling calculations or the evaluation of the degradation of collapsed blocks, as well as of the blocks still in place. An exhaustive inventory of the stone marks is also being carried out.
Figure 10. Archaeologists and scientists at work on the stones of the fallen arch of the nave. Photo ©C2RMF / Alexis Komenda, with permission.
The Casimodo project, financed by the French National Research Agency (ANR) is directly concerned with the timber of Notre-Dame Cathedral. The project brings together the themes of "framework", "forest" and "climate" from the 11th to the 13th centuries, all contextualised by historical, palaeoenvironmental and bioarchaeological archives. It aims to address the response of medieval society to the availability of wood resources during this period of climatic optimum through several research axes. On the one hand, it includes the characterisation of the medieval Parisian climate by means of a dendrological study of the framework oaks. A second objective is to study the wooden framework from an archaeological point of view in order to characterise the construction techniques and the supply procedures on the site. Finally, it focuses on the forests exploited between the 11th and 13th centuries in the Parisian environment and the possible forest management systems put in place for adequate wood production, as well as to reconstruct wood supply networks.
The lion’s share of the work took two years of effort within a dangerous and physically-demanding environment. Our task was urgent and we often had limited means to hand. However, from out of the ashes of the disaster, a human adventure emerged. Through inter-service and inter-structure collaborations, a renewed study of the cathedral will engage several generations of researchers in the years to come.
- Baillieul, E., Chaoui-Derieux D. and Leroux, L. (in press), ‘Reconstructing the fallen arch of Notre-Dame. New insights into Gothic vault-building’, Journal of Cultural Heritage, Special issue: Notre-Dame.
- Chaoui-Derieux D., Bouet B. and Puaux O. (2021), ‘Les opérations d’archéologie menées dans le cadre du chantier de sécurisation de la cathédrale Notre-Dame de Paris’, Patrimoines, La revue de l’Institut national du patrimoine, 16, Paris: Éditions de l’Épure, pp. 122-129.
- Collective (2021), ‘Laboratoire/Recherches, Le LRMH à Notre-Dame de Paris’, Monumental, Revue scientifique et technique des monuments historiques, Semestriel 1, Éditions du patrimoine, pp. 98-117.
- L’Héritier, M. (2021), ‘Archaeology in the Ashes of Notre Dame’, Sapiens, https://www.sapiens.org/archaeology/archaeology-notre-dame-cathedral/
- L’Héritier, M. et al (in press), ‘Revealing the iron skeleton of Notre-Dame de Paris. Methodologies and issues of archaeological and analytical investigations following the blaze’, Journal of Cultural Heritage, Special issue: Notre-Dame.
- Moulis, C., Phalip, B. and Ybert, A. (in press), ‘Archaeological analysis of the voussoirs of the collapsed transverse arch of the nave. First results and new hypotheses’, Journal of Cultural Heritage, Special issue: Notre-Dame.
- Penagos, C. et al. (in press), ‘Naming, relocating and dating the woods of Notre-Dame “forest”, first results based on collated data and archaeological surveys of the remains’. Journal of Cultural Heritage, Special issue: Notre-Dame.
- Zimmer, T. (2020), ‘Appraising and Responding to Emergencies’, in: 1970-2020 The Historical Monuments Research Laboratory. Half a Century serving French Heritage, Paris: BeauxArts Éditions, pp. 22-25.
- Zimmer, T. et al. (in press), ‘From debris to remains, an experimental protocol under emergency conditions’. Journal of Cultural Heritage, Special issue: Notre-Dame
Go back to top