The Correspondence of Pierre Bayle, Seventeenth Century Philosopher

The Correspondence of Pierre Bayle, Seventeenth-Century Philosopher

La Correspondencia de Pierre Bayle, filosofo del siglo XVII

Éric-Olivier Lochard et Antony McKenna

Traduction de Michelle Arriss, Delaina Haslam et Séverine Sofio

Traduit de :
La Correspondance de Pierre Bayle, philosophe du xviie siècle

Citer cet article

Référence électronique

Éric-Olivier Lochard et Antony McKenna, « The Correspondence of Pierre Bayle, Seventeenth Century Philosopher », Biens symboliques / Symbolic Goods [En ligne], 2 | 2018, mis en ligne le 12 avril 2018, consulté le 18 août 2018. URL :

How did the database project begin?

The Correspondance de Pierre Bayle is a publication produced from the electronic work entitled Le Monde de la correspondance Bayle, devised in 1999 and since developed by our team with the help of the Arcane software program in accordance with the operating instrumental paradigm “Le monde selon Arcane” (addressed more fully below). The Arcane project was launched in the early 1990s at the CNRS Centre of Eighteenth-Century Studies, a Humanities and Social Science laboratory in Montpellier.

How was the corpus defined?

The corpus of the active and passive correspondence of Bayle (Labrousse & McKenna 1999-20171) consists of some 10,000 manuscript pages, housed for the most part among major libraries (Paris, London, New York, Leyde). Our point of departure was the collection of more than 1,800 letters, a critical inventory of this data (set up by Elisabeth Labrousse), the transcription of the texts, the translation of Latin and Italian letters, and finally the creation of a databank of digital images. These images accompany the annotated edition of the letters’ content on the Bayle correspondence website. The correspondence’s metadata is copyright-exempt and has been integrated into the EMLO database at Oxford. The online edition currently (as of February 2017) consists of 1,099 letters. The entire correspondence will be freely available to the public in 2022. It is also integrated into the corpus of Lumières électroniques (Electronic Enlightenment) of the Voltaire Foundation as well as in XML format in the ePistolarium database (Utrecht University).

Is the design of the database supported by a specific theory and/or discipline?

The Arcane Project aimed to build on the advances in computer science for the research on the history of ideas and sociability in the eighteenth century; these advances were used to produce critical editions of correspondences, and an Atlas des correspondances à l'époque moderne.

This research is structured around an epistemological reflection concerning the academic editing process and, in a more general sense, the scientific creation of knowledge in the humanities and social science in a new economic and technological environment. Numerous issues were examined in this reflection in line with the requirement to suggest operational solutions.

Which software programs were used to build the database infrastructure and, as the case may be, to treat the data statistically?

This work resulted at the same time in the development of the Arcane software program, a prototypical electronic writing instrument resembling the word processing that researchers are used to, and designed to test responses to current critical editing challenges. The characteristics of this program can be found in three articles in particular, the first written in 2001 by the author of the program (Lochard & Taurisson 2001), the second in 2008 (Lochard 2008), and the third in 2014 (Dufour 2014) by a doctoral student who had used Arcane for the critical edition of the correspondence of Marceline Desbordes-Valmore.

Among the solutions proposed, some are commonly accepted today, although they appeared original in nature at the time.2 Others constitute the originality of the Arcane instrument. We detail some of them briefly by specifying their application in the Correspondance de Pierre Bayle.

One requirement of the Arcane project was to situate the instrument in the “anthropological” environment of the production of knowledge in the humanities and social science by taking into account the habits of authors.3  One thorny issue was that of suggesting, for the editing work, another mental image than that of the book as the author’s horizon of work. This image had to be natural, functional, and independent of the form of the publication: a website was thus excluded as a satisfactory solution. Hence, we proposed considering electronic scholarly editing more abstractly as the intentional and scientific, potentially infinite, composition of a world of multimedia knowledge

• that the reader can read and explore
• from which one can extract information for publication in various forms
• and which can dialogue with other worlds in order to compile electronic libraries.

The architecture of these worlds must be at the same time generic enough to apply to a wide range of editing projects, and simple enough to make it easy for authors and readers to locate and explore the information. A solution to this dual constraint is to consider the editorial work as comprised of three types of object, dynamically interconnected, both necessary and sufficient to compose the semantic combination of a world of multimedia knowledge, and which can function as portals into the electronic publication:

subjects of intentional interest of the editorial project (“what he is talking about”), categorized according to type (in the Bayle database: people, places, books, letters), and which make up the materiality of the world and its inhabitants, by generalizing the cumulative indices;

enhanced documents (“what has been said about them”) multimedia information streams to describe, illustrate, annotate, and interpret the world (in the Bayle database: letters that thus serve as subjects and as enhanced documents, notes, appendices, manuscripts, etc.);

relationships (“what is known about them”) in order to organize the subjects, describe the factual relationships or formally locate predicative information (in the Bayle database: relationships among correspondents and the individuals mentioned in letters).

Within the world of the Correspondance de Pierre Bayle, there are essentially five types of subject: the inventory consists of 2,200 subject-letters which describe the letters (their content is a document); 11,000 individuals (senders, addressees, people cited in the letters); 2,700 modern critical publications, 150 geographical locations and 75 (literary and theatrical) characters. Each subject possesses a metadata form composed according to its type: the subject-letters form, for example, consists in particular of the author, the addressee, the places the letter was sent from and to, as well as the date it was sent.

An Arcane-document is a link comprised of a multimedia support (sequence of signs for a text, of pixels for an image, of units of time for sounds and videos), and of the combination of its enhancements – objective or subjective values that the author explicitly and intentionally attaches to a sequence of the medium in order to qualify, represent, structure, index, or annotate it, etc. In the world of the Correspondance de Pierre Bayle, manuscripts (1,793) and established texts (2,040) are Arcane documents associated with the corresponding subject-letter, thus facilitating movement from one to the other.

An enhancement is a logical piece of information attached to the document medium by an anchor (a sequence, a rectangular part of an image, a sequence in a sound or a video) with the help of an enhancer; the enhancements (title, italics, emphasis, foreign language, etc.) are freely created and stipulated by the editor according to the editing project, by setting various automatic operations such as fixing the anchors’ typographical form, or choosing the format to use when exporting the document – for example, using a TeX or XML markup. In Arcane, enhancements are recorded separately from the media; this unique characteristic has several advantages. In particular, it enables enhancements to be ports of entry into the world, and allows the author to suggest different views of a document by making only part of its enhancements operate on the medium.

The Correspondance features two types of annotation, enhancements which link the annotation (an Arcane document) to an anchor: the explicative annotation (22,500 separate notations regarding the identification of the people, events, questions, addressed in the letters) and the critical annotation (13,000 separate notations about the condition of the manuscript: crossings-out, ink stains, faded ink, illegible text, etc).

The indexation, whether simple or relational, is an enhancement which attaches a subject (publication or individual) to the anchor: the 70,000 anchors of the Correspondance have enabled the automatic compilation of cumulative indexes and bibliographies for each volume, and general indexes for the collection of volumes as a whole. The indexing also allows access via its form to all occurrences of a subject in the documents.

The “Glossary” enhancements (2,000) link a term from the dictionary of the world to the anchor.

The other enhancements (185,000 in total) consist of adding labels to the anchor in order to describe it: title, foreign language, verse, citation, list, etc., or to create links with various targets.

Relationships are formal combinations that serve to establish semantic and dynamic connections between objects in the world, in particular among subjects, or to formally interpret the content of a document’s anchor. They are produced by the application of a relator (hyperlink). These are freely created and prescribed by the editor according to the editing project, thus determining the list of the combination arguments as well as their type: one could, for example, introduce relationships of all types among all the basic subjects (arborescence, inclusion, family, dependence, domination, protection, friendship, or more, etc.).

The Correspondance de Bayle uses two relators4: ProduitPar (ProducedBy – 2,650 occurrences) which links an author of the type subject-person, and a work of the type subject-work; and EtreEnRelation (BeRelated – 150 occurrences) which links the person who initiated the relationship with another (whose name is mentioned in the letter). These 150 relationships (which we have established for experimental purposes for a given period in order to investigate their pertinence) are anchored in the letters by an ad-hoc enhancement.

Dynamic documents are defined by a script composed by the user and applied to elements of the world. They are effectively calculated and produced in response to an instruction by the user; for example, the document composed of the anchors of an index by a given subject or the anchors of an enhancement such as “passage to be re-examined,” or even the cartographical representation of subject-letters on a map of Europe using coordinates of the geographical subject-places acting as place of expediting and of destination in the subject-letters file.

Publications, such as exchanges with other information systems, are produced by automatic exportation of a document, led by a script specifying the objects to be exported and a style sheet indicating the operation to be carried out on the enhancement anchors – for example, on markup.

Could you offer one or two examples of scientific (whether consensual or surprising) results obtained with the help of the database?

These relationships enable the establishment, for example, of the pertinence of lost letters and indirect contacts with the people mentioned in the known letters. If we were to rely only on the representation of the network designed by the known letters, we could be seriously mistaken about the world of the principal subject (in this case, Pierre Bayle) because his correspondents appear to have no relationship to one another. By revealing letters that have been lost, the correspondence is far more complete. If we add to that the total number of the relationships (established by relators), we have a much better insight into Bayle’s own representation of this world of correspondence and of more or less direct relationships; we can then better understand the objective nature of these networks.

In another example: relationships enable the hierarchical representation of direct and indirect networks in such a way that we have a better understanding of Bayle’s situation and intentions. For the purpose of compiling his Dictionnaire, he was able to call upon a small number of friends who acted as “Secretaries of the Republic of Letters.” They would pass his questions (bibliographical, historical, etc.) on to specialists of various fields (bibliography, history, numismatics, genealogy, etc.). In this way, we see that Bayle received, directly or indirectly, the information for his Dictionnaire from a great number of subjects who, considered together, make up his representation of the Republic of Letters and the objective structure of this community.

PDF versions of the fifteen volumes of the paper edition were produced by exporting the document-letters to LaTeX format, then composed with the help of a style sheet interpreting the page layout instructions of the commercial editor, the Voltaire Foundation.

The web edition of the Correspondance is served by a version of the free publishing system SPIP developed by Pierre Mounier and, from then on by Cindy Tessier (DSI, Université Jean-Monnet de Saint-Etienne), and is supplied by the exportation of an XML document re-imported into the server.

The integration of the Correspondance inventory into the EMLO database at Oxford was done through the exportation of the subject-letters metadata, and that of the complete Correspondance into the ePistolarium database (Utrecht University) through the exportation of the document-letters in TEI format.

1 With a team consisting of Wiep van Bunge, Edward James, Bruno Roche, Fabienne Vial-Bonacci. With the collaboration of Éric-Olivier Lochard. Over the

2 Such was the case, for example, for the proposal of using a database, which was poorly received at the launch of the program; it was also the case

3 This requirement led to the choice of a laboratory in humanities and social science for the production of the software instead of a laboratory

4 Other works, those dealing with sociability, for example, consist of dozens of them in order to distinguish the different types of relationships

Labrousse Elisabeth & McKenna Anthony (eds.) (1999-2017). Correspondance de Pierre Bayle. Oxford, La Fondation Voltaire, 15 vol.

Lochard Éric-Olivier & Taurisson Dominique (2001). “‘Le monde selon Arcane’: un paradigme instrumental pour l’édition électronique.” Cahiers GUTenberg, 39-40: 89-105. [On line] [accessed on 15 january 2018].

Lochard Éric-Olivier (2008). “L’édition de manuscrits au prisme du paradigme instrumental Arcane.” Recherches & Travaux, 72: 83-95. [On line] [accessed on 15 january 2018].

Dufour Delphine (2014). “L’enjeu numérique dans l’édition de la correspondance de Marceline Desbordes-Valmore.” PhiN-Beiheft, 7: 64-75. [On line] [accessed on 15 january 2018].

1 With a team consisting of Wiep van Bunge, Edward James, Bruno Roche, Fabienne Vial-Bonacci. With the collaboration of Éric-Olivier Lochard. Over the years, the following people also collaborated on this publication: Laurence Bergon, Hubert Bost, Maria-Cristina Pitassi, Ruth Whelan, as well as Annie Leroux, Dominique Taurisson, and Caroline Verdier.

2 Such was the case, for example, for the proposal of using a database, which was poorly received at the launch of the program; it was also the case for the formal markup of information to ensure its perpetuity at a time when the SGML (Standard Generalized Markup Language) had not yet given birth to its simplified version, the XML (Extensible Markup Language), and when the web was still confidential.

3 This requirement led to the choice of a laboratory in humanities and social science for the production of the software instead of a laboratory specialized in science and technique in which its development would have been more “at home.”

4 Other works, those dealing with sociability, for example, consist of dozens of them in order to distinguish the different types of relationships among individuals in a more nuanced manner.

Éric-Olivier Lochard

Antony McKenna

Université de Saint-Étienne – IHRIM

© Presses Universitaires de Vincennes 2018