Open Book Publishers logo Open Access logo
  • button
  • button
  • button
GO TO...
book cover

10. Early Modern Correspondence:
A New Challenge for Digital Editions

Camille Desenclos

© Camille Desenclos, CC BY 4.0

The project of building a platform dedicated to early modern correspondence at the École Nationale des Chartes is the starting point for this contribution. Its reflections are based on the editing of two early modern corpora: the correspondence of Antoine Du Bourg, chancellor during the reign of Francis I (1536–1538),1 and the correspondence of the extraordinary embassy led by the duke of Angoulême in the Holy Roman Empire (1620–1621).2 The former encompasses approximately 1200 letters concerning every matter with which a chancellor had to deal (justice, royal finances, monitoring printed production, economic policies etc.), while the latter contains only 80 letters, which are longer and embrace the international network of French diplomacy in the early seventeenth century. Based on traditional editorial practices and the issues that we confronted while encoding, our aim was to build a documented schema that could support and highlight the specific characteristics of correspondence. Above all, this project has given us the opportunity to reconsider the issues and aims of digital scholarly editions of early modern correspondence. Through the process of highlighting its formal and historical specificities, it has become quite clear that the present form of digital editions is not well suited to correspondence. This is especially the case for early modern correspondence.

Correspondence within the digital field

Editing correspondence: aims and general issues

Letters are one of the most important sources in the field of historical research.3 They show, perhaps better than any other source, political, cultural and even commercial forces at work. They are not the result of a single act, but are, rather, part of a process, be it an exchange of information, a commercial transaction, a political negotiation, a cultural or spiritual discussion4 etc. When editing them for scholarly purposes, we must first, therefore, reconstruct this process. An isolated letter can of course be understood and used by researchers, as it may contain historical information (the description of a battle or political audience etc.). But a single letter does not allow us to understand the context of the correspondence, the very reason for its existence: the origin and evolution of the relationship between the people involved. It is only through the analysis of several letters that we can gain a deeper understanding of the nature of that relationship. The main aim of a scholarly edition should therefore be to restore the general context, the continual dialogue between two individuals: one letter may refer to a previous one and call for a response, and this dialogue may be broken due to letters having been lost. An edition cannot of course overcome such losses, but it can attempt to build bridges between letters (through links, references to other correspondence or archives etc.) in order to restore the general meaning of the correspondence and consider it not as a collection of individual letters but as an intellectual entity. As far as possible, therefore, a scholarly edition of correspondence must try to (re)present a coherent slice of life. But, if corpora are generally built around a typology of documents (e.g. royal acts), or more rarely on a theme (e.g. the legal status of religious minorities in medieval Christian and Islamic societies5), which give them a degree of consistency, epistolary documents are by their nature heterogeneous. They are produced by (at least) two people, writing in different locations, and contain a wealth of information of various kinds. The same letter might deal with information about political changes in France, military events in the Holy Roman Empire and the purchase of books. Editions of correspondence can thereby quickly appear jumbled. For editions based on the typology of documents, the aim seems quite clear: restoring the writing activity of a political or religious institution. The corpus itself is already established—the whole writing activity of one institution—and a chronological choice has to be made, but there is no need to justify the reason behind the edition. The choice is even simpler when the envisaged content of the edition comes from only one manuscript, such as for political discourses, memoirs or chronicles. For correspondence, however, the choice is more complex. The researcher must choose the sender, but must also decide upon one or more addressees. For example, he or she can only edit the active correspondence of ambassadors to their king or their active and passive correspondence with the king and/or with other ambassadors and so on. Unlike charters and documents for which the context is obvious—an institution producing documents regardless of an envisaged public—there are as many choices as there are contexts for the production of letters.

At first sight, editions of correspondence seem to be the poor relation within the family of digital editions. The collection of digital editions at the École Nationale des Chartes, Élec,6 corroborates this assertion: of the thirty editions available on the website, fourteen are scholarly editions, and only one is a scholarly edition of correspondence (that of Antoine Du Bourg). This observation is demonstrated more starkly by the Telma project, led by the Centre national de la recherche scientifique (CNRS),7 which has six digital scholarly editions, none of them of correspondence. There are around twenty projects on correspondence in Europe and North America, barely ten per cent of the editions in the world.8 Editions of manuscripts, or more generally of documentary units, seem to be preferred to those of correspondences: the text is already established and with an internal consistency.

How should we use metadata?

In addition to the editorial issues, the benefits of publishing digitally are not immediately apparent for series of correspondence. The characteristics of correspondence, such as the layout, which enables the reader to identify clearly the date, the signature and so on, can be expressed quite easily in a print edition. Yet the semantic step required by the encoding and the transfer to a digital framework turns these structural elements into editorial problems. Many issues need to be addressed: description of multiple metadata, concordance between layout and structure, semantic encoding for each part of the letter etc.

Metadata are crucial for digital scholarly editions of letters. They are the only way to identify and distinguish one letter from another. Above all, they give the letter a meaning, a formal proof of its place within the whole correspondence, by including details of the sender, addressee, date and place of writing, and sometimes even the date of receipt. As proof of the importance of metadata, almost every digital scholarly edition permits browsing of its corpus via these criteria.

If good structuring of the metadata offers a useful basis for the user, they must also be well displayed. Unfortunately, the display of the search results frequently fails to reflect the quality of the editions themselves, and does not reflect the global coherence of the correspondence. Most digital editions of correspondence, such as those of Carl-Maria von Weber9 or Thomas Bodley,10 allow access to a specific letter through these criteria, but only via their list of results. We cannot move on to another letter without going back to the list of results. By limiting the correspondence to a searchable object, the edition facilitates, in a sense, the research. Specific information can be found more easily without being lost within a number of letters. Thus a biographer can find useful information about a person under study simply by searching for his or her name within the edition.

Fig. 10.1 Screenshot from the browsing page (and main access) to the correspondence of Thomas Bodley.

This breaks the dialogue created through the correspondence, however, and prevents the researcher from examining the wider context of the writing. No continuous reading, and consequently no global apprehension of the corpus, is allowed. This hides the true nature of a correspondence: that it is a flow of information. By offering a continuous reading, an edition places the letter in its proper context as part of a dialogue, not just as a piece of information. Publishing digitally should not necessitate the loss of one of the advantages of paper: that of continuous reading. We do not deny the importance of providing several access points to facilitate navigation throughout sizeable corpora. But a scholarly edition of correspondence should, first and foremost, before providing search tools, transcribe a slice of life and make it understandable.

The main aim of producing a digital scholarly edition of a correspondence should not, then, be the creation of a database of letters—which can be a second step—but rather, editions should aim to offer a coherent object of study. The text must come first. At that point the edition can be used by researchers in various ways. An edition can, and should, be used by researchers for different purposes than those that were the initial aim of the edition. One edition can be produced, for instance, in order to recreate the intellectual network of Late Humanism, but another researcher may use it as a source of information about the Thirty Years War or about political habits. The aim therefore should be to build a solid and reusable structure with well identified metadata, for both straightforward study via the pages of the edition and more extensive use through the underlying XML files.

One of the major research aims of the study of early modern correspondence, and of correspondence in general, is the (re)construction of networks. Once perhaps considered less important, indexing, especially through traditional metadata such as sender, addressee and place of writing, has become the main value of such digital editions. We are preserving the advantages of the traditional edition (historical and prosopographical background) but with new possibilities for reconstitution of networks. By the links created first between the sender and the addressee and then between the different people mentioned within the letters, a social, commercial and even cultural network can be reconstructed. The analysis of correspondence is, for example, the first tool for studying the ‘Republic of Letters’, i.e. the long-distance intellectual community in the late seventeenth and eighteenth centuries. Different kinds of exploitation can be envisaged, from a simple prosopographical dictionary (linked to the letters) to a dynamic graph. For the time being, the first solution is the most common, through identification systems, sometimes supported by a database that provides access to the letters by date, place, sender, addressee or some other criteria. We can also find inspiration in other projects, such as the Pez correspondence.11 Although not digital, this edition has been used to create a database of every book mentioned within the letters, to recreate the theoretical library of the brothers Pez and therefore to inform us about the general cultural environment of learned people in the eighteenth century.

Characteristics of early modern correspondence

TEI and early modern correspondence

A letter is identified first through its form: salutation, date, signature and address. These elements must be considered along with the edition and highlighted in order to facilitate a general understanding of the letter. Requirements differ, however, between early modern correspondence and that from nearer our own time. In the former, for example, there is no envelope,12 as the address was normally written directly on the verso of the last folio. The editor does not need to describe an object such as the envelope, but rather a new block of text,13 which has to be differentiated from the main text (the body of the letter), which was written on the same material. The form of the letter is also much more strictly structured in early modern correspondence. The form of early modern letters was not free and had to follow various rules, especially within the area of political correspondence, as part of a wider protocol: salutation, sometimes a formula of politeness, a courtesy formula, time and place, signatures.

This form should be re-transcribed clearly and accurately when preparing the digital part of the edition. These elements could be structured as simple sections of text, but this option would not capture their original meaning. Structures such as salutations, courtesy formulas or signatures could appear as equivalent to diplomatic structuring for charters. For correspondence, these elements are not optional, and furthermore a letter cannot be formally identified if they are not distinguished from the main body of its text. These structures are also objects of study for Diplomatics, which focuses on the critical analysis of the form of documents. However, the form is as important as the content, since it allows immediate identification of the nature of the text (a letter), its aims (who is writing to whom) and its context (through the choice of the politeness and courtesy formulas). Only the encoding of each specific element can accurately render the text, its coherence, legibility and initial form.

Fig. 10.2 Diplomatic formulas of early modern correspondence.

The encoding guidelines developed and maintained by the Text Encoding Initiative (TEI) are the most common standard used in the preparation of digital scholarly editions today.14 Like many other projects, we have chosen these guidelines for the implementation of ours. The different structural features of the diplomatic letter are not sufficiently well distinguished by the TEI, however. These features can be encoded separately as various segments of text, but the TEI does not provide enough specific elements in order to identify clearly the meaning of each segment. Created to encompass a large variety of documents, the TEI is too general to cover the specifics of each particular kind of text. That is one of its major advantages. But correspondence is too specific to fit into this general standard. In fact, instead of making the encoding easier through the Guidelines and years of user experience, TEI complicates the encoding of early modern correspondence. It might indeed be possible to use some general TEI elements such as <seg> (segment), with a type attribute, to convey the features specific to this type of document. However, in order to achieve the required degree of granularity, it is necessary to use one or more attributes for each element: one can, for example, mark later additions with the help of a @type attribute, but it is then also necessary to add @subtype in order to state whether such the addition represents a date of reception, a summary etc., as one can see from the example below:


<seg type=”salutation”>Monsieur,</seg>


<p xml:id=”l6-p1” n=”1”>L’avis que l’empereur a eu et vous aussy […]</p>


<seg type=”courtesy”>Vous asseurant de nostre affection, laquelle vous sera plus cognue lors que l’occasion s’en offrira, et que nous vous verrons pour dire de vive voix que sommes, Monsieur, </seg>

<seg type=”courtesy”>Vos plus affectionnez à vous faire service</seg>

<seg type=”signature”><name>Charles de Valois.</name> <name>Bethune</name><name>Preaux</name></seg>

<seg type=”mention” subtype=”adresse”>À monsieur, monsieur de Puisieux, conseiller du roi en ses conseils d’Estat et privé et premier secretaire des commandemens de Sa Majesté</seg>


A solution such as this, while TEI-conformant, results in verbose encoding which is open to mistakes. Some lighter semantic encoding is, however, possible with the TEI: in fact, the TEI offers a limited number of elements that can be used for this purpose, such as the element <opener> for opening formulas, <signed> for signatures and <closer> for closing formulas. But because of the general nature of the TEI, which offers support for every kind of text, these elements have a broader meaning than is required. As defined in the TEI Guidelines, the <opener> element ‘groups together dateline, byline, salutation, and similar phrases appearing as a preliminary group at the start of a division, especially of a letter’.15 Although clearly designed with modern letters in mind, these can be made to work with older material, but the Guidelines should offer better documentation and further examples to help researchers working with such material to make the appropriate choices.

Nor does the TEI offer elements designed for the semantic encoding of the final folio. The addressee often wrote on the back of the last folio the name of the sender, the date of the letter and sometimes the date of receipt, or even a brief summary of the letter. This practice is not yet documented or even envisaged by the TEI (or by the TEI Special Interest Group for correspondence, on which see below). At this point, two solutions can be envisaged: using a <seg> element—with the issues about the verbosity of encoding already mentioned—or creating a new element which could be added to the TEI schema. The use of <seg> or <div> (division) is therefore possible, but it is not really convincing as a solution, as these elements are too general in their scope. For scholarly purposes it would be much better to have a specific element that possesses the right semantics. This would permit, for instance, distinctions to be made between the different parts of the letter as hierarchical units. In this way the text on the back of the last folio would not be required to be part of the body of the letter, but could stand as a separate structure in its own right, which would enable the exploitation of the information it contains. Even if the last folio is not considered a priority for many scholars, the information contained in it should not be forgotten, as these folios form part of our understanding of the context of the correspondence.

Semantic elements would bring the encoding closer to the original source and above all make it more understandable for the researcher, and perhaps even easier for him or her, as it forces the researcher to think of the structure of the letter no longer in terms of simple text divisions but rather as units of meaning. Thus the process of encoding becomes scholarly work in itself. This could encourage researchers to produce digital editions, or at least be part of the process and therefore link at last both the scholarly and technical parts of a digital edition. The definition of new elements at the local level should not be considered lightly, however, as adding new elements can present an obstacle for interoperability. It can be considered as a temporary solution, and indeed, the TEI itself is aware of the issues created by its generality and encourages the creation of new elements as the first step of a formal proposal to the TEI Council toward their integration into the general schema. Following this cautious and responsible approach, we have created at the École Nationale des Chartes a new element called <nota> based on a <seg>-like model in order to ensure a proper dialogue with the TEI.16 The main purpose of this element is to make the encoding more precise by using a @type attribute to indicate the nature of the added text, such as address, date of receipt etc.

<enc:nota type=”addresse” place=”back”>À monsieur le chancellier.</enc:nota>

We could even envisage, in future, the creation of even more specific elements for each kind of information, such as:

<enc:address place=”back”>À monsieur le chancellier.</enc:address>

<enc:reception place=”back”>L’empereur du XXII avril 1620. Receue le XIII may 1620</enc:reception>

The former would be dedicated to each address, which appeared either on the last folio (for early modern correspondence) or on postcards or envelopes. The TEI Guidelines already offer an <address> element, which can be used within an <opener> element, for instance. But for the time being, the Guidelines restrain the use to postal addresses, so that plain text, or even a <p> (paragraph) element, cannot be inserted directly within this <address> element. Or if structured addresses exist for contemporary letters, for early modern ones they are mostly names without any precise address as ‘To the chancellor, at the Court’. Ideally it would be good to find a solution that would fit with both contemporary correspondence (with envelopes) and that of the early modern period (with the address written directly on the last folio), by allowing both structured and unstructured encoding of the address. The latter would encode each addition made after the writing of the letter. Most of the time, these additions—the dates when the letter was sent and received, as well as sometimes a short abstract—were made after receiving the letter by the office of the secretary of state in order to manage more easily state papers and affairs. Other name elements or means of encoding could be envisaged. We have much to learn from the work currently being done on genetic editing for the encoding of different hands and other further additions.17 The purpose of these examples is not to suggest new encoding elements, but more to highlight the need for further encoding specifications for correspondence. This work would require more than two corpora and would need to include various types of correspondence (commercial, personal etc.).

Documentation as a solution?

The main requirements for early modern correspondence are, however, not really new elements, but rather a more constrained encoding, to fit its specific structure. Writing, especially to a king or a secretary of state, is subject to various protocols or ceremonies, and the encoding should reflect this. This structure can also be considered as an advantage, since it forces the editor to think more carefully about the encoding requirements. For the two editions of correspondence at the École Nationale des Chartes, the first step has been to write a schema, based on TEI, but with a few additions, accompanied by a documentation that has been written specifically for early modern correspondence. The aim of this schema was to constrain the use of TEI and to document it in order better to adapt the elements to the particular nature of early modern correspondence, and therefore to make the use of TEI more straightforward for the researchers. The documentation is a result of a combination of both digital and traditional editorial practices. In fact, writing specific documentation provided an opportunity to link editorial norms to those of encoding. The École Nationale des Chartes has an established editorial practice through the publication of manuals such as Conseils pour l’édition de textes médiévaux18 for medieval texts, L’édition des textes anciens for early modern ones19 and very recently L’édition critique des textes contemporains.20 Our documentation for early modern correspondence continues in this tradition, bringing it into the digital age.

Fig. 10.3 Screenshot from the documentation for the general structure of a scholarly edition of early modern texts.

The documentation is now composed both of editorial requirements and encoding recommendations, in order to provide an adapted and easily useable handbook for digital scholarly editions of early modern texts. This confers added scholarly value onto the schema, which is not only a computing tool but the result of an effective dialogue between researchers and developers, based on the meaning of the various segments. In fact, this makes its use easier for the researchers, since it not only talks about XML elements but also about editorial norms. Overall, correspondence projects, because of the large sizes of their corpora, are often collaborative projects. With such documentation, we can guarantee a common and consistent encoding which will permit interoperability between projects.

In order to make the encoding more meaningful for the researchers, we have made several choices which differ somewhat from the traditional recommendations of the TEI. The data required to identify the letter—title (i.e. sender and addressee), place and time of writing, witnesses, summary—are metadata and could be inserted into the <teiHeader>, which is probably the most common practice. This kind of encoding would require that each letter be encoded as a separate XML document. At the École Nationale des Chartes, the choice was made to create one single XML document for the entire corpus, in order to facilitate the creation of internal links and to match the structure of other editions, mostly those of charters. In this case, the <teiHeader> is dedicated to the general metadata of the corpus and the information specific to each letter is inserted within a <front> element. Moreover, the structure of that element is less constraining than that of the <teiHeader>. We were thus able to build, within the <front> element, a structure which duplicates the traditional structure of scholarly editions: addressee and sender, which are considered as the title of the documents in order to match to the structure of editions (<head>), date (<docDate>), witnesses (<div type=“tradition”>) and summary (<argument>). We would probably benefit from modifying this structure in further editions, especially for addressees and senders. Certainly we could simply encode sender, addressee and place name with <name> elements. But this is not as satisfactory a method for letters as it is for charters. In fact this practice requires the use of the same heavy encoding mentioned earlier in regard to the structure of the letter. Indeed, if we want to identify the sender and address with precision, we would need to use @role attributes as follows:

<head><name role=”sender” ref=”#FrançoisI”>François Ier</name> à <name role=”addressee” ref=”#DuBourg”>Antoine Du Bourg</name></head>

In this way, we should use the elements <sender> and <addressee>, as was suggested by the TEI — Special Interest Group on Correspondence within its proposal for a whole <correspDesc> element. We would only need to enlarge this use by allowing it both within the text and the <front> element.

The TEI-SIG on Correspondence was formed in 2008 and illustrates well the real encoding issues concerning these texts. To solve them, the work of the SIG relies mostly on the DALF schema and its documentation.21 This is a specific schema for correspondence, particularly for features such as the main metadata (addressee, sender—who is considered as an author—etc.), envelopes and postscripts.22 A task force for Correspondence Description was created in October 2013 during the TEI Members’ Meeting in Rome and has led, very recently, to the addition to the TEI Guidelines of a new <correspDesc> element.23 Contained by the <profileDesc> element, <correspDesc> allows the description of ‘the actions related to one act of correspondence’, including sender, addressee, place and time of writing, date of reception, type of correspondence etc. Only a few new specific elements have been introduced, however. For the sender and addressee for instance, the Guidelines still recommend general elements such as <name> or <persName> with a type attribute. If the introduction of new elements is an admirable first step towards recognising the specific characters of correspondence, things need to go further in order to provide a lighter, but still precise, encoding for correspondence, especially where the early modern period is concerned.

In our model, we have constrained the body of the letter in order to follow the strict structure of early modern correspondence: <opener> for the opening formulas (salutation), <p> for the text of the letter, <closer> for closing formulas (courtesy and politeness formulas, mention of the date and signatures), <postscript> for postscripts and <nota> for additional material (address, back notes etc.).


<div type=”transcription”>

<opener><salute>Mons<hi rend=”sup”>r</hi> le Chancellier,</salute></opener>

<p xml:id=”l6-p1” n=”1”>Pour ce que je desire que le partaige de mon cousin le marquis de Rothelin […] Et vous me ferez service en ce faisant.</p>


<salute>Priant Dieu, Mons<hi rend=”sup”>r</hi> le Chancellier, qu’il vous ayt en sa garde.</salute>

<dateline>Escript à Valence le <hi rend=”sc”>ix</hi><hi rend=”sup”>e</hi> jour d’aoust mil <hi rend=”sc”>v<hi rend=”sup”>c</hi> xxxvi</hi>.</dateline>



<enc:nota type=”note” desc=”note dorsale d’une écriture contemporaine”>Le roy pour le faict du partaige de Mons<hi rend=”sup”>r</hi> le marquis de Rothelin.</enc:nota>

<enc:nota type=”adresse” desc=”adresse au dos”>À mons<hi rend=”sup”>r</hi> le chancellier.</enc:nota>



Only the elements required for the structure of correspondence are permitted here. This is not a new standard but, rather, a proposal to stimulate discussion. Matters are not fixed and must be adapted according to each experience and practice of encoding. Our proposals solve some specific issues, but they do not cover the whole field of early modern correspondence.

Towards a general documentation

The TEI-SIG on Correspondence can be considered as a first important step towards much-needed collaboration. First of all, it tries to link different projects dealing with correspondence and encourages a general and free discussion through a mailing list and annual meetings.24 Its work focuses on documenting and finding solutions for encoding the specific features of correspondence, which in turn enables researchers to identify and list the main issues they encounter when editing correspondence. The SIG does not offer immediate solutions, even if one of its eventual aims is to make some encoding proposals to the TEI Council. It is however a space in which to exchange and discuss proposals about topics such as

  • the description of metadata (sender, addressee etc.);
  • the content model of the postscript;25 the issue of enclosures or attachments which are part of the letter but written on another leaf and need consequently to be distinguished from the letter but linked to it at the same time;
  • the encoding of addresses.

In order to succeed, it is important that any proposal that will become part of the standard must take into consideration the largest dataset possible, covering material from many periods, countries and contexts. This is why we have to call for greater cooperation between DALF, WeGa,26 our own project and other projects in order to build a model that is as strong and efficient as possible.

The type of collaboration put in place by the EpiDoc community, which has built recommendations for the encoding of epigraphic documents, could inspire us. EpiDoc is indeed a specialisation of TEI for ancient documents, inscriptions and papyri, and all EpiDoc files are also TEI-conformant. However, its Guidelines27 only contain those TEI elements which are required for this type of source material and have been constrained in order to fit to the epigraphic context. EpiDoc is therefore, as the Guidelines themselves assert, a supplement to, and not a replacement for the TEI Guidelines. We could envisage similar guidelines and an associated schema for correspondence.

Early modern correspondence requires, perhaps more than many other types of sources, a semantic and specialised encoding. Digital scholarly editions of such sources should not be a transposition of a text into a digital medium but a genuine restitution of the meaning of a text. Only by correctly labelling structures such as names or formulas can we retain the flow of the letter and consequently produce a virtual recreation of the relationship(s) inherent in a correspondence. In order to give to its specific structure sense and utility, a documented set of guidelines has to be created. The schema and documentation, created at the École Nationale des Chartes, should not be considered as anything other than a proposal for further development, especially by merging the documentation from our project with others such as DALF. We therefore call for the creation of a dynamic community around correspondence, so that we might build a general documentation that is suitable for correspondence of all kinds and from all periods.

1 La correspondance du chancelier Antoine du Bourg, a project at the École Nationale des Chartes, At this time, there are around 100 letters published, and the edition is still in progress.

2 L’ambassade extraordinaire des ducs d’Angoulême, comte de Béthune et abbé de Préaux, a project at the École Nationale des Chartes, This is only a beta version.

3 Les correspondances: Problématiques et économie d’un ‘genre littéraire’: Actes du colloque international ‘Les correspondances’, Nantes, 4–7 octobre 1982, ed. by Jean-Louis Bonnat, Mireille Bossis and Hélène Girard (Nantes: Université de Nantes, 1983).

4 Letters contributed for example greatly to the establishment and spread of late humanism, as the principal way in which the members shared and discussed their opinions. See Axel E. Walter, Späthumanismus und Konfessionspolitik im konfessionellen Zeitalter: Die europäische Gelehrtenrepublik um 1600 im Spiegel der Korrespondenzen Georg Michael Lingelsheim (Tübingen: Max Niemeyer Verlag, 2004).

5 Relmin: le statut légal des minorités religieuses dans l’espace euro-méditerranée (Ve–XVe siècle), project led by the University of Nantes and hosted at the L’Institut de recherche et d’histoire des textes (IRHT),

6 Élec (Éditions en ligne de l’École des chartes), To this collection must be added another platform which contains beta versions of editions ( with three other digital scholarly editions. Only one of them is an edition of correspondence.

7 Telma (Traitement électronique des manuscrits et des archives), a project hosted by the Institut de recherche et d’histoire des textes,

8 There is no complete list of editions of correspondences. In order to find some quantitative information, we have compared various lists:

9 Carl-Maria-von-Weber-Gesamtausgabe, a project at the Akademie der Wissenschaften und der Literatur, Mainz,

10 The Diplomatic Correspondence of Thomas Bodley, a project at the Centre for Editing Lives and Letters at University College London,

11 Monastic enlightenment and the Benedictine Republic of letters: The correspondence of the brothers Pez, a project at the Institut für Österreichische Geschichte,

12 The envelope is a part of the letter as general object. Without the envelope, which contains the address, the letter cannot be identified. But the text written on the envelope cannot be considered as part of the content of the letter or even be considered purely as metadata. An appropriate form of encoding has to be found. A similar issue is represented by attachments that can be found within early modern correspondence. They complete the content of the letter but are written on a different piece of paper; they must be separated but linked at the same time. At present, no satisfactory solution has been found.

13 Most editors do not record addresses and similar text found on the last folio. This omission is unfortunate, as these often contain useful information, such as the date of reception or the exact rank of the addressee. At the very least, the possibility of encoding such text should be available to the editor.

14 TEI P5: Guidelines for Electronic Text Encoding and Interchange, 2016,

16 This <nota> element could appear to overlap with the existing TEI element <note>. In our view, however, the content of the <nota> element is not, in a semantic way, an annotation, but rather a part of the letter which has been written either during the writing or the reading of the letter and should therefore not be considered as an addition but rather as a step in the writing process.

17 See Elena Pierazzo, ‘Un nouveau module de la TEI pour l’encodage des manuscrits modernes et les éditions génétiques’, in La génétique des textes et des formes: L’oeuvre comme processus, ed. by Pierre-Marc De Biasi and Anne Pierrot Herschberg (forthcoming in 2016).

18 Groupe de recherches La civilisation de l‘écrit au Moyen Âge, ed., Conseils pour l’édition de textes médiévaux (Paris: CTHS, 2001).

19 Bernard Barbiche and Monique Chatenet, Conseils pour l’édition des textes de l’époque moderne (XVIe–XVIIIe siècle) (Paris: Inventaire général, 1990).

20 Christine Nougaret and Élisabeth Parinet, L‘édition critique des textes contemporains (XIXe–XXIe siècle) (Paris: École Nationale des Chartes, 2015).

21 Digital Archives and Letters in Flanders, a project at the Centre for Scholarly Editing and Document Studies in Gent,

22 It is important however to note that the SIG Correspondence itself recognises the orientation of its current work and seeks to enlarge its perspectives. There are other projects of early modern correspondence, such as the projects developed by the CELL (Centre for Editing Lives and Letters, University College London), which do not offer any open documentation or access to their TEI files, therefore making it is also impossible to compare or to learn from their practices.

23 Release 2.8.0 (6 April 2015).

24 The aim of the TEI-SIGs is indeed to provide an opportunity for TEI users with similar interests to meet and exchange ideas. Eleven SIGs have been created: computer-mediated communication, correspondence, education, libraries, manuscripts, music, ontologies, scholarly publishing, TEI for linguists, text and graphics and tools. Each of these has a wiki space, a mailing-list and the opportunity to meet during the annual TEI Members’ Meeting.

25 Until recently, the <postscript> element could not be inserted before the <closer> element, although in reality, it does not have a fixed place. It can be written after the signatures, on the same level within the letter or even before them. Since release 2.1.0 this element can be inserted anywhere within the body, which allows a better semantic encoding.

26 Carl-Maria-von-Weber-Gesamtausgabe, mentioned above.

27 Tom Elliott et al., EpiDoc Guidelines: Ancient Documents in TEI XML (Version 8, 2007–2013),