The ARTL@S project offers researchers a digital database of catalogues of exhibitions held all over the world during the nineteenth and twentieth centuries. The addresses contained in these catalogues are geo-referenced, regardless of era and geography. The tool is designed to enable monographic research as well as a study of the circulation of works of art, and the construction of careers and tastes in this process. Through its capacity to centralize and cross-reference sources that are often difficult to access, the ARTL@S database aims to breathe new life into the international history of exhibitions and the culture associated with them. It has just been opened to the public and it is accessible from the project's central online portal. ARTL@S is financed by the École normale supérieure (ENS) Paris, l’Agence Nationale de la Recherche (ANR, Projet Jeunes-Chercheurs 2011-2016), PSL Research University Paris, l’Institut d’histoire moderne et contemporaine (UMR 8066, CNRS/ENS), and the laboratoire d’excellence TransferS (programme Investissements d’avenir ANR-10-IDEX-0001-02 PSL* et ANR-10-LABX-0099). It is headed up by Béatrice Joyeux-Prunel (ENS Paris) in collaboration with Catherine Dossin (Purdue University, USA) and Léa Saint-Raymond (Université de Nanterre).
The project has its roots in the early 2000s, with a team of doctoral students and young scholars in history, history of art and literature. We were searching for a social, quantitative, comparative, and transnational social history inspired by Fernand Braudel and the first École des Annales, and we were also nurtured by the legacy of Pierre Bourdieu and wanted to take into account issues of cultural transfer. We wanted, therefore, to carry out our studies on long periods of time and various scales, while working on both the levels of micro-history and of macro-history. We used a quantitative approach to better pinpoint our monographic spot-checks. The development of data processing tools enabled us, among other things, to take up the best of the Annales legacy with the certain knowledge that by not abandoning detailed and timely analysis, nor lived history (contemporary or past accounts of an event), nor, finally, the artworks themselves, we could move towards a more contextualized and explanatory history. That way, we thought that the forces at work in the history of art and literature could be seen more clearly than under the partial light of decontextualized histories of creators and movements, too often presented as exceptions.
For my PhD thesis, I created a database of exhibition catalogues in order to study the international circulation of Parisian avant-garde works from 1850 to 1914.1 Thanks to this database I was able to study the internationalization of trajectories, reputations, and aesthetics, before undertaking more traditional studies, supported by a preliminary reconstitution of structural logics. At the same time, several other colleagues were also working on databases or wanted to learn about quantitative approaches, but everyone was feeling somewhat isolated with his or her spreadsheets. Consequently, when I arrived at the ENS in 2007, I decided to organize a seminar on “Art and Measurement” in which Claire Lemercier, who is passionate about quantitative methods, helped us to create a better framework for our approach. During the subsequent conference I organized on the same subject in December 2008, I observed how many researchers were processing exhibition catalogues’ databases. Why not gather our data? It was especially clear that a digital, online database of exhibition catalogues on a global scale and over a long period would be most useful for the history of art.
My own database of exhibitions and catalogues was a good place to start. I had given a part of it to the Musée d’Orsay’s Salons database project, but the museum was only interested in exhibitions held by French institutions, and therefore of an academic nature. The Institut national d’histoire de l’art (INHA) which I also contacted in 2006, was interested, but no progress was made. We therefore began to work on our own with funding from the ENS. We quickly hit upon the idea of creating a Geographical Information System (GIS) from the database – thus making it possible to map the many addresses contained in the exhibition catalogues. We had to quickly find more ways of financing the technical creation of the database. We were able to recruit researchers who were specialists of geographies and of periods very different from ours in order to test the database with other spaces and eras, and link the database to actual research activity – a seminar, conferences, and publications. The stakes were twofold: to create the database with an easy-to-use interface in order to assist art historians who are traditionally unfamiliar with quantitative and global methods (“global” in the sense of both the study of cultural transfer and of comparatism) and help them experiment new approaches that would complement their traditional methods; and to encourage the global and social history of art, which was at that time not very developed. This led to a funding application to the ANR. The project was accepted with funding for four years beginning in 2011, which was extended to May 2016.
Several people in our group wished to store diverse sources in the database: exhibition catalogues; sales catalogues; journals and reviews; museum catalogues; biographical indexes – and people still ask me today why the ARTL@S database does not also include journals' tables of content, biographical notes, and even photographs of artworks. The potential of a database is always a reason to dream – one tells oneself that one will detect networks, track circulations, link several databases to one another, increase sources, etc. However, it is important to think of concrete uses and of the durability of the database: a meta-database where one can store many things is all well and good, but as soon as one enters a search query, the diversity of the data found by the system makes the use of that data particularly complex. One ends up not being able to see anything with it. Moreover, it is impossible to compare heterogeneous data with any sort of coherence. Moreover, technically it is not easy to create cartographic systems that distinguish data according to its original identity; the structure of the database becomes highly complex. It was safer and better to create a database using one sole type of source, knowing that the associated conceptual model is already highly complex because of the multiplicity of tables needed to map all available addresses, and also because of the scientific demands that are essential for a research database: all information is dated, sourced, associated with a contributor, and traceable. In addition, and this is important, a database containing multiple objects would have no end. It would never be complete! I had no desire to embark on a research project around a database that was never-ending and therefore impossible to be made available for other researchers, and thus the public at large, to use. Too much money had been spent, regardless of the country, and is still being spent today, in order to create databases that, in the end, cannot be brought to fruition due to a lack of time and means, or that are not ready to be put online due to a lack of clear limits defined from the start. In short, I had to impose a scope for the project – to choose is to eliminate: the ARTL@S database contains only exhibition catalogues, and it will never contain anything else. Even within this vast category, a distinction needs to be made between complete collections and isolated catalogues. An average user always believes that a database is complete. In fact, it offers only what it contains! Only complete collections should be made available online – but it is also necessary to explain exactly what is being made available to the public. We have worked hard, therefore, to imagine and realize public interfaces for searches and visualisation that would clarify this message and not give the impression that ARTL@S offers more than it can. For a simple search, if you wish to know, for example, the titles of the works shown by Claude Monet at the first impressionist exhibition in 1874, the database will provide reliable and clear information with the source indicated. For advanced quantitative searches, however, without a thorough reflexion beforehand on sources and their representativeness, research interfaces should only allow searches on complete collections, so that one exactly knows what this percentage means, in these precise geography and era. But this is technically possible only if the user defines by him·herself the corpus s·he wants to search in his·her request.
It made sense, at the beginning of the 2000s, to go against the trend of a “soft” history of mentalities (which was interesting but rarely theorized), and even more so against the empire of monographs, which still play such a strong role in art history today. There are still so many monographic articles and books explaining how “their” artist invented, deconstructed, understood everything, and that he or she renewed modernity. After reading twenty-five of them, it can be tiring – after reading 250 of them, it becomes funny. That is the first – I would say sociological – point of ARTL@S, for which I take responsibility and that I defend. Serial processing has never been an automatic reflex for a discipline that traditionally focuses and emphasizes individual artists – or at best a few groups – following the logics and methods of connoisseurship. At best, serial processing is used in iconological analyses (which study a motif in a particular artform – a method that is rarely used on exhaustive collections) and stylistic analyses (but here again, exhaustive research is rarely a purpose, and the approach is often critical [Joyeux-Prunel 2008]). In short, the results are disappointing. This realization, as harsh as it is, is often poorly received, especially when you arrive at a conference in art history with conclusions drawn from exhaustive surveys that put into question the conclusions of each individual search. Certainly, there is symbolic violence in the figures, and in all pretence at objectivity, but the discipline in which I work should make more room for the issue of representativeness – running counter to the ideology of uniqueness being a major component of the modernist value system. Things are beginning to change. There is currently a sort of armed truce with the quantitative and socio-historical approach. We have yet to make clear all that serial processing has to offer, nor the fact that it does not, by the way, necessarily call into question established values in the history of art. Processing catalogues in series and encoding them in the form of a geographical database facilitates the cross referencing of sources, and is all the more useful when other sources are rare. Certain gaps in archival work are filled by new information produced by connecting catalogue series. Considering the information on a quantitative level, the unique character of exhibitions and the hapax of a period are submerged in the mass – they are no longer representative, and long tendencies can be revealed. The visualization stage also helps to find exceptions in the mass, or to identify new ones. In one sense, the quantitative approach is therefore not contrary to qualitative questioning, and even less so to the study of exceptions, forms and genius to which it is sometimes said art history should be limited. We also still have to make clear that, on “our side” – if we must accept that there are camps, according to the reproach that has been made to me of being “on the side of the sociologists” and of “quantitative cartographers” – we do more than statistics and cartography; we are interested in forms and material; we read the works of others (which is not that reciprocal); and we do go into archives and museums! There is major informative work to be done in art history in France, especially among the over-fifty generation, still very distrustful of sociology and digital humanities.
The second point of ARTL@S is to offer an approach that could be called decolonial. In his call for “a new ecology of knowledge,” one of the proponents of this new current of thought, Boaventura De Sousa Santos, suggests that researchers cross their discipline methods with other modes of thinking, in particular those of “non-western” cultures and spiritualities (2009; Boidin & Hurtado López 2009). Supposing there is a “western” culture, which I do not much believe is the case, it is still very difficult to step outside one’s own culture, especially when we come from a perception of the world in which it is my culture that offers salvation, freedom, etc. For art history, there is an intense reflection, particularly on the modernist “canon” which the majority of museums of modern art expose and impose, and which excludes so many geographical, social and stylistic as well as ethnical, gendered, spiritual, and religious categories. How can we make room for everyone? It seems to me that even when one introduces artists considered peripheral in the canon, one does not modify it: the new arrivals are made to fit the Procrustean bed of modernist criteria (originality, critical capacity, deconstruction of the social, gendered, or scopic systems of its era, etc.). Basically, they become little clones of the major artists of the canon, making the latter all the more admirable. In order to decolonize this canon, it is first necessary to understand how it was constructed, hence the importance of a social and collective history of art, which requires an alternative narrative. So, how can we write this alternative narrative? It first requires access to sources other than those produced by the “centres.” By increasing the possibilities of accessing catalogues of exhibitions that took place in all European countries and not just France or Germany, in all America, including Latin America, and not just in New York or Chicago, and as far as Asia, Africa, the Middle East, and Australia, ARTL@S already contributes to giving another idea of the history of modern art or the history of fine arts – and to no longer automatically thinking that an exhibition is ultimately a European exercise, and that it has nothing to say about the cultural history of peripheral countries.
One could also reflect on ways to decolonize our methods, despite the difficulty of decolonizing our minds and our sources. The main methodological premises of the modernist canon contribute not only to the dis-symmetries denounced by the postcolonial approach, but also to their naturalization: methodological nationalism (assigning an art, an artist, a work, to a nationality), a corollary of a focus on a few capital cities (Paris and New York, for the contemporary period); monography (focus on a single individual who of course is a genius, who invented everything, understood everything, “deconstructed” everything, revolutionised everything); formalism, and its corollary, the denial of social and economic logic. One could choose to adopt different methods. Certain historians of the “Peripheries” invite a new type of history of the avant-gardes, a horizontal geography rather than a vertical narrative (Piotrowski 2009).2 And as a matter of fact, serial processing approach puts everything on the same level. Tracking the circulation of artists and artworks can also help there (DaCosta Kaufmann, Dossin & Joyeux-Prunel 2015: 1-22). Consequently, ARTL@S does not limit itself to the “basics:” we introduce researchers to transnational approaches (connected history, cultural transfers, the study of artistic circulations) and to decolonized readings, which helps them to emancipate from the national myth as self-sufficient culture. We also introduce them to quantitative methods, cartography, and sociological problematics.
Which software programs were used to build the database infrastructure and, as the case may be, to treat the data statistically?
Could you offer one or two examples of scientific (whether consensual or surprising) results obtained with the help of the database?
We in the ARTL@S team often come from research fields that are antipodal and diachronic, from the internationalisation of modern art in the nineteenth and twentieth centuries to that of African art. We value the catalogues, however, as commensurable sources that allow the tracking of artistic globalization over the long term, from its spatial, social, and economic dimension to the circulation of art objects and the cultural transfers that this circulation implies (Joyeux-Prunel & Marcel 2015). We are not the only ones to look for comparative elements in order to discuss cultural internationalization. With the development of Global Art Studies, more and more researchers feel the need for common ground to compare different contexts, events, or collective trajectories of artists and works of art. The success of the database is related to this context. It is possible from now on to implement studies that are larger, longer, and perhaps also more ambitious.
The sources and digital tools provided by ARTL@S also assist in the study of what Arjun Appadurai calls the “social life of things” (1986), and in better understanding the successive contexts of a transcultural process (Joyeux-Prunel 2017b). The database helps with the comparison of internationalization schemas, with detecting the presence of certain mediators, of specific networks, localizing the trajectory of an artefact in a given economic, social, or geopolitical context. Case studies such as the internationalization of Picasso’s career which was based on a total absence of exhibitions in Paris prior to 1914; or the calling into question of the supposed world domination of US modernism after 1945, can from now on be addressed faster and more efficiently than with the manual methods that we had to use at the start of the 2000s.3
The ARTL@S database also gives its contributors a very new idea of the possible chronologies and geographies of art over nearly two hundred years. In particular, we have set in motion a complete collection of exhibitions in North Africa and the Middle East from 1900 to 1990 – close to 1,400 exhibitions indexed and referenced, before eventually accessing their catalogues. This work illustrates the importance of genuine modern activity in the cultural history of a wide region reaching from Morocco to Iraq, despite the lack of memory of this recent past in those regions.
Another significant result is the GeoMap project, explained in several publications by Léa Saint-Raymond, Félicie de Maupeou, and Julien Cavero (2016). This reconstitution of the geography of exhibitions in Paris between 1815 and 1955, which began from a homogenous source, the Bottin du Commerce, shows that the culture of Parisian exhibitions was spread among very diverse spaces: clubs, bookshops, journals, cafés, shops selling artists’ materials, halls rented for the occasion, cinema hallways, ateliers, photographers’ studios, etc. It demonstrates that the arrival of galleries was not a phenomenon of the 1860s as was believed, but of the beginning of the twentieth century. I was delighted with these results. They confirmed what I myself had observed, that is to say, the inexistence of the so-called “dealer-critic system” described by Harrison and Cynthia White in the 1960s, based on biased and insufficient sources, extrapolated case studies, and without a profound understanding of the period. It will take some time before the Whites cease to be cited in discussion of the art market at the end of the nineteenth century, but it will happen! By cross-referencing the data in the ARTL@S database with dealer and artistic archives and sales catalogues, knowledge about certain lesser-known phenomena could also be deepened, particularly concerning the circulation of works of art between dealers, or the connections between the social and artistic geography, and commercial spaces of art.
Over time, we have become aware of the subversive significance of ARTL@S. The merging of sources from regions considered peripheral in the world in the global history of art and artistic modernity, challenges the favourable consideration accorded by historians and museums to a few centres-Paris prior to 1945 in particular, and New York from then on. This has become a way of provincializing the symbolic centres of the artistic canon – a new interpretation that many call for (Joyeux-Prunel 2015b). ARTL@S offers a useful, and not exclusive, methodological and theoretical tool for a re-reading of the global history and geography of art since the nineteenth century (Dossin 2015; Joyeux-Prunel 2016c and 2017a4). With a transnational history of art that challenges the artistic geographies imposed by the canon, an openness to other geographies, the taking into account of geopolitics in the history of art, and the refusal to ignore social contexts, cultural transfers, and market issues, I dare to believe that we could attempt to write another history of artistic modernities – and that it is worth spending a great deal of time doing so.
The catalogues database was published online in January 2018. We develop and expand its query and visualization interface gradually. We did not make much publicity about this publication, but the news circulated immediately on the social networks. The database, however still to be augmented, has a great success, not only among art historians but also among many players in the art market – merchants, experts, inheritors, etc. – as well as museums and other art institutions in search of new digital and interactive content to enhance their events, particularly their exhibitions.
In market terms, for a long time I thought it would make sense to have only certain levels of the database available to the public, that is to say those which concern the exhibitions, their chronology and geography, the artists and their circulations. The idea was that information which would help to increase the prestige of the works (and therefore increase prices) would only be accessible for free to the contributors and scholars associated with the project, or to others in exchange for a financial contribution – modest in reality when compared with the amount our research institutions have invested in ARTL@S. After all, what is an annual subscription for a major gallery in comparison with the information that it can find in the ARTL@S database to increase the value of the works it sells? However, arranging the legal and technical elements of such a project was, in fact, much too complex, and would have required me to build my own start-up company. I have no desire to do that! I realised, above all, as things progressed that it is better to spend my time moving ARTL@S forward and finetuning our interfaces – if they are attractive and convincing they will be funded again; I have confidence in the project. Up to this point, we have always been able to find the money we need for it.
I also resolved to work on partnerships with museums, art institutions and other research projects that help us to extend the collections. These teams have helped us for several years to see all that the database has to offer. We have thus been able to position our work on visualization and search interfaces in the most logical manner possible. Several museum teams have sought to collaborate with us – with ARTL@S having to provide technical solutions and sources to facilitate the cartographic approach for digital and animation sites, as well as for terminals available in future exhibitions designed for the public. The idea is to provide our partners with “widgets,” personalize-able interface sections for quantitative, spatial, and textual analysis of the data available in the database that they can display on their own sites and according to their own issues. We work with the RKD (Royal Institute for Art History, Leyden) for an exhibition at the Petit Palais in Paris and the Van Gogh Museum on Dutch artists in Paris in the nineteenth century; the Picasso museum for a project on Picasso and the Mediterranean; the Asian/Pacific/American Institute and its “Virtual Asian American Art Museum” project; a project based at the Open University in the UK, on exhibitions in British India between 1888 and 1914; a team from the University of São Paulo on the São Paulo Biennales; the University of Barcelona’s MoDe(s) project; and for our own teams whose entire research we need to put online, in particular that of Léa Saint-Raymond, Félicie de Maupeou, Julien Cavero, and Tatiana Debroux on the Parisian artistic geography of the nineteenth and twentieth centuries, and the works of Annie and Gabriel Verger on the students of the Académie de France in Rome.
However, the durability of the project will depend most of all on the manner in which we improve the data collection process, which is still too manual, time-consuming, and error-prone, and which requires us to leave out certain information, in particular the illustrations contained in the exhibition catalogues. We already have many partners, especially from Arab countries, but also in Israel, Latin America, and Spain, who would like to contribute to the database but can do no more than send us photographs of catalogues, and sometimes re-read the data once it has been processed. After eighteen months of work with a team from the University of São Paulo, we understand the difficulty of training contributors from a remote location, on a really complex French database. However, Ana Paula Simioni and Eduardo Dimitrov’s team did not give up! It is with them in mind that I am currently working on a partnership with a team of Computer Vision specialists. The aim is to automatically process the scans sent by our partners in order to add them into the database after recognition and systematic encoding of the catalogues’ texts and images. From this data, it should be possible to conduct a fairly powerful study of artistic globalisation over a century. We are working on an exhaustive series of biennial catalogues – a form of exhibition that after 1950 has been going global and has multiplied in an exciting way – and over a long period, 1890-2010, if we achieve it technically. We want to draw from these incredible collections information on artistic globalisation and circulation in all of their dimensions (economic, geographic, sociological, aesthetic, stylistic, political, gendered, etc.). We will do this not only from the textual information contained in catalogues, but also by examining reproductions, photographic content, and page format. The team is in the process of being assembled, from Brazil to Japan. It is the start of a promising and fascinating adventure in research and in friendship!