Reassembling the Republic of Letters in the Digital Age

Tommaso Elli and Beatrice Gobbo DensityDesign Research Lab, Milan Edited by Howard Hotson and Thomas Wallnig Between 1500 and 1800, the rapid evolution of postal communication allowed ordinary men and women to scatter letters across Europe like never before. This exchange helped knit together what contemporaries called the ‘respublica litteraria’, a knowledge-based civil society, crucial to that era’s intellectual breakthroughs, formative of many modern values and institutions, and a potential cornerstone of a transnational level of European identity.

For, as Nevalainen rightly notes, 'letter writing has always been situated activity' and therefore, in order to understand the concept of a letter in the early modern period, we must first look at early modern epistolary theory to see how contemporaries understood the term. 3 'To expect all letters to conform to a single type, or to teach that they should, as I notice even learned men sometimes do, is in my view at least to impose a narrow and inflexible definition on what is by nature diverse and capable of almost infinite variation'. 4 So wrote Erasmus of Rotterdam in his Opus de conscribendis epistolis (Basle: Johann Froben, 1522), a text on letter writing which did much to shape understanding of epistolary theory in the early modern period. When pushed to produce a concise definition, Erasmus provided two formulations, each based on ancient and medieval commentators: a letter is 'a mutual conversation between absent friends', and 'a letter is a conversation between two absent persons'. 5 The first formulation owed much to ancient definitions: for example, the Ancient Greek rhetorician Isocrates' letters were invariably addressed to known associates and written in a familiar style. 6 The even more influential Demetrius of Phalerum (c. 350-c. 280 BC), looked back to the definition of Artemon, the editor of Aristotle's correspondence, who stated that a letter was a conversation halved. 7 The idea of a letter as a continuation of a conversation by other means (in this case written), was further popularized by Cicero (106-43 BC), who spoke of 'Amicorum colloquia absentium'. 8 Thus in ancient epistolary theory there were four essential elements in the definition of a letter: it was (1) a written conversation; (2) written in a familiar style; (3) between people who were known to each other; (4) who were absent from each other. Absence was the necessary pre-condition for any letter. 9 Medieval commentators of the ars dictaminis, such as the early thirteenthcentury author Guido Faba, added more elements to the definition of a letter. In his Summa dictaminis, written c. 1228-9, which Camargo notes was 'probably the single most influential treatise of the ars dictaminis', he included the following definition of a letter: An epistle is a booklet sent to one or several absent persons, and it is called 'epistle' from epi, which is 'beyond', and stola or stolon, which is 'sending', because it makes the sender's desire clear 'beyond' a messenger's capacity to expound it. For on account of the mind's forgetfulness and the multiplicity of affairs and the distances of journeys, many things would be omitted, which an epistle represents like a mirror.
The epistle was invented for two reasons. The first was so that the secrets of friends might be concealed through it, whence it is named from epistolo, that is, 'I conceal'. The second reason was so that it might express better than a messenger what is sent. 10 Faba's definition thus added four additional elements to that of the Greek and Roman commentators: (5) a letter could be a 'booklet'; (6) it was something that was 'sent'; (7) it provided secrecy; (8) it allowed for greater accuracy. Later thirteenth-century commentators, such as Conrad of Mure, added a fifth: (9) to enable conversation between persons unable to communicate directly with one another. 11 This moved the definition beyond a communication between friends to one between potential strangers, a definition later mirrored by Erasmus's second formulation, that a letter might be defined as 'a conversation between two absent persons'.
Erasmus's Opus de conscribendis epistolis spawned a host of imitators. As the sixteenth century progressed, writers not only sought to give advice on writing letters in Latin but increasingly identified the vernacular market as a growth area. 12 In general, such treatises on letter writing focused less on definition of letters and more on the division of letters into various genres. However, one area relevant to definition, which both medieval and early modern epistolary theory considered, was the question of the parts of a letter. Medieval treatises on letter writing agreed that there should be five parts to a letter: 1. the salutatio: the address, usually including the name of the author and the name of the addressee; 2. the exordium: a short statement or epigram outlining the qualities of both author and addressee and the circumstances of writing the letter; 3. the narratio: the main subject of the letter; 4. the petitio: the request; 5. the conclusio: the subscription, which might also include a brief summary of the argument. 13 Of these the most important were the address and farewell, for it was these which identified the document as a letter. 14 Additional medieval requirements -that a letter be brief and that it should keep to a single subject were largely ignored in the early modern period. However, as Burton notes, the fact that the five parts were modelled on the six rhetorical parts of Ciceronian orations ensured that they were retained in one form or another by humanist scholars in their reform of the ars dictaminis. 15 The salutatio and farewell continued to be given pride of place: Erasmus spends some part of his Opus de conscribendis epistolis on the formula for greetings and farewells which should be employed.
The salutatio provides us with a useful starting point in our search for the essential characteristics of an early modern letter. If we concentrate less on the function of a letter and the circumstances of its creation, and more on its form, we may isolate the following essential elements whose combination enable us to recognize a text as a letter: • A letter text includes the name of the sender/s.
• A letter text includes the name of the intended recipient/s.
• A letter text includes the place name of the letter's origin.
• A letter text includes the place name of the letter's destination.
• A letter text includes the date the letter was written.
True, there are examples where the address of the recipient may be missing, or the date of the letter might have been lost (or, in the case of a partial draft, not included), but in the main, the combination of these five elements indicate that a text is a letter. Ultimately, then, a letter is a written communication (in manuscript and/or 13 Rita Costa Gomes, 'Letters and Letter-writing in Fifteenth-Century Portugal', in Schulte and Von Tippelskirch, eds., Reading, Interpreting and Historicising, 24. 14 Constable, Letters and Letter-collections, 17. As Constable notes, not all letters had a subscription, for writers of secret letters necessarily wished to conceal their identity (18). The salutation and valediction were considered vital areas of a letter by ancient writers also: Sullivan, 'Classical Epistolary Theory and the Letters of Isocrates', in Poster and Mitchell, eds., Letter-writing Manuals, 7-20, esp. 9. 15 Gideon Burton, 'From Ars dictaminis to Ars conscribendi epistolis', in Poster and Mitchell, eds., Letterwriting Manuals, 92-3. print form), between two different entities (sender/s and recipient/s), which is located in space and time by the addition of addresses and a date.

Elizabethanne Boran with contributions from Antonio Dávila Pérez
At its most basic level, a letter is both text and material object and any definition of a letter must take into account the interaction of both in order to understand it fully. 16 Any definition of a letter must consider the fact that early modern letters often survived in more than one state. The arrival of printing further complicated what was already a very complicated process. In the early modern period a letter might survive in the following multiple states: 1. as a draft -this might exist in various stages of composition; 2. a holograph letter (sometimes called the 'original'), written by a scribe and signed by the author; 3. a holograph letter, written wholly by the author; 4. a copy of the letter, written by a scribe; 5. a copy of the letter, written by the author; 6. a copy of the letter, written by the recipient/s; 7. a printed copy of the letter; 8. a revision of the letter; 9. a summary/abstract of the letter in manuscript form; 10. a summary of the letter in print form; 11. a translation of the letter by either the recipient or someone in her/his circle. . 18 Ibid., vol. 3, 1143. An example of a letter that was substantially revised during its composition is Gerardus Joannes Vossius's letter to Ussher of 1/11 January 1632 from Amsterdam, now in the ence was drawn up by scribes and simply signed by him. 19 Nonetheless, many of his holograph letters survive, sometimes in manuscript volumes devoted to letters. 20 Copies written by others were also kept in miscellaneous volumes. Some letters were evidently more privileged than others: for example, Ussher kept copies of his outgoing letters to many of his new continental correspondents, presumably because he was afraid that his initial letters to them might be lost in the post -and also to keep track of his correspondence with them which, unlike that of correspondence closer to home, might take more time to elicit a reply. 21 Equally, some of Ussher's letters were reprinted by contemporary scholars keen to link themselves to him. 22 Indeed, many exist only in print form, primarily because his seventeenth-century editor, Richard Parr, on printing an edition of Ussher's life and letters in 1686, evidently decided to destroy the holographs, which Ussher (unlike so many other authors), had thoughtfully archived for posterity. 23 Some later editors silently revised the text (though in Ussher's case such textual changes were minor). 24 Some of Ussher's letters to William Laud, archbishop of Canterbury, were kept as manuscript abstracts and are now in the Bodleian Library, Oxford. 25 One letter exists only in translation: Ussher's letter to James Frey, 17 September 1635, was translated into German, presumably by Frey (or by a member of his circle), who was professor of Greek at the University of Basle, so that he might send on the import of the original to a different, vernacular-speaking readership. 26 To these eleven states we might, to further complicate matters, add a ghostly twelfth -the all too frequent missing letter, whose existence may be deduced but which is no longer extant. 27 As these examples demonstrate, letter states are chiefly categorised by using the following criteria: The contents of the letter (nos. 1, 8, 9, and 10). Here the existence of a holograph letter greatly aids identification of other states, such as drafts, revision, and abstracts in other formats, but, as Ussher's letter to Rothe demonstrates, it is also possible to identify drafts in the absence of holographs. In this case the subject matter indicates that the letter is part of a known ongoing communication process. As Vossius's letter to Ussher of 1/11 January 1632 shows, another useful indicator of draft status is the frequency of deletions in the exemplar.
The handwriting of the letter and signature (nos. 2, 3, 4, 5, and 6). To distinguish whether an item was written/copied by its author, a scribe, or a recipient, it is necessary to compare the handwriting of the letter with exemplars from both sender and recipient in order to clarify correct identification.
The format of the letter (no. 7). Early modern letters exist in both manuscript and/or printed form. In the early modern period it was relatively common to destroy manuscript holographs once they had been published in print.
The language of the letter (no. 11).

The Letter as Object
Signed, Sealed, and Undelivered (Rebekah Ahrendt, Nadine Akkerman, Jana Dambrogio, Daniel Starza Smith, and David van der Linden) Letters do not simply bear the words of authors to their recipients, they can also be interpreted as carefully crafted composites of substrate and writing substance. The reading of a letter begins long before it is opened, as its material features communicate a series of silent cultural assumptions. In the last decade and a half, scholars have increasingly turned their focus to the material features of letters, particularly in the early modern period. 28 Digital resources for the study of letter collections have also begun to factor materiality into their research remits, raising 28  new possibilities for archival access and data-driven analysis of epistolary materiality. 29 This section summarizes the essential concerns of the study of letters as objects, and explains how two current interrelated projects -Signed, Sealed, and Undelivered (brienne.org) and Letterlocking (letterlocking.org) -are seeking to theorize them in new ways and implement tools for their further study. For much of history, a letter could not simply be rushed off: even a short, informal note would require some degree of planning and preparation. As objects, letters should be considered in relation to a series of other objects on a letterwriter's desk, which might include an inkwell, standish (inkstand), candle, feather quill (carved with a pen-knife) or pen, seal matrix, dust-box, blotter, scissors, whetstones, wax jack, and rulers.
The letter proper begins with a substrate: paper, papyrus, and parchment are most familiar to us now, but in other traditions clay, wax tablets, bark, etc. were also used. Scholars of material letters ask how thick this substrate is, whether (if it is paper) it is hand-made, if its chain and wire lines (also known as laid lines) are visible, and if its watermark enables us to identify its source. All these details enable us to understand the document's make-up. We also need to ask if it has survived largely intact or whether damage (such as mould and ink corrosion) or interventions over the years might have destroyed or altered some of its material evidence. Before writing, the substrate may need to be trimmed for neatness, prepared to ensure it did not absorb too much ink, then folded into a suitable shape for writing, often a bifolium; a crease running parallel to one edge can serve as a writing margin.
Early modern letter-writers usually made their own ink, and the quality of ink can drastically affect a letter's afterlife -too much acidity, and it will eventually eat through the paper. Invisible inks can more subtly alter the physical state of the paper, if made visible by the recipient or interceptor: chemicals need a reagent such as water, which might leave the paper crinkled; fluids such as milk or the juice of citrus fruit need the heat of a flame to oxidize, which might leave the paper scorched. Many letters may survive with hidden writing which has never been made visible.
Once written, letters' contents have their own materiality, for example, where signatures or marginalia are placed, if cross-writing is employed, or how much blank space is left around the writing. 30 Letter-writers may employ cryptologywhich can take the form of cryptography (ciphers) or steganography (codes, riddles, invisible inks) -to disguise their message, or to embed a hidden communica- tion within the overt one. These, too, have their own material conventions and histories. Other material features sometimes found on historical correspondence include postal marks, seals, ribbons, and sealing wax. The borders and edges of a writing substrate can sometimes be decorated: letters announcing a death, or written during a period of mourning, might be edged in black; gilt edging is common; blue-and green-edged letters also exist. Edging is a feature which can easily be overlooked on a digital surrogate.
Early modern letters were composed in an age before the mass-produced gummed envelope had been invented. This usually meant that, after writing, the writing surface itself had to be folded up to become its own sending device. This process is called 'letterlocking', 'the act of manipulating and securing an epistolary writing substrate (such as papyrus, parchment, or paper) to function as its own envelope. 31 Letterlocking is a subcategory of a 10,000-year information security tradition, pertaining to epistolary materials, and its study encompasses the materially engineered security and privacy of letters, both as a technology and a historically evolving tradition. Letterlocking demonstrates that letters were for centuries folded and otherwise manipulated to become their own envelopes, and that this process has a rich history. Archival letters can today seem like flat, fossilized, twodimensional artefacts, but letterlocking reminds us that they were once dynamic, three-dimensional objects which travelled through space and worked as engineered objects, often including sophisticated anti-tamper mechanisms.
The material features of letters -in particular the letterlocking aspects -have hitherto rarely been captured in epistolary databases, which have largely focused on standard content-related metadata such as date, place, author/sender, and keywords. One notable exception is Bess of Hardwick's Letters, for which Wiggins et al. recorded a number of material features in the metadata, making this 234-item corpus searchable by fifteen standards, including 'Letters with seals', 'Letters with significant space', 'Endorsements', 'Subscriptions', and 'Sewn'. 32 These metadata capture standards usefully group letters that exhibit common physical features, enabling them to be studied and compared more easily.
The main repository which records materiality among its metadata is the epistolary union catalogue EMLO, which enables contributors to note postage marks, endorsements, enclosures (both letters with enclosures and letters that are enclosures), seals, paper type, paper size, and handling instructions. 33  already allows for a rich variety of data to be captured, the incorporation into EM-LO of the Brienne Collection, an extremely well-preserved archive of 2,600 undelivered letters sent from all around late seventeenth-century Europe, 34 has prompted the development of new metadata standards to record different kinds of evidence, both material and ephemeral. 35 From a material standpoint, the Brienne Collection presents a series of opportunities and challenges related to the state in which the letters have been preserved: all are archived in their folded state, and some 600 of them have never been opened (even by their original addressee). These features have inspired two new metadata fields capturing (1) whether letters are still unopened and (2) if they have been stored folded. The Signed, Sealed, and Undelivered project is also developing metadata standards which more overtly define and capture evidence of letterlocking. In particular we seek to record letterlocking formats and categories. Formats refer to the shape the letter takes when folded into a packet (e.g. 3 = triangle, 4 = quadrilateral, 5 = pentagon). Categories are distinguished by the number and combination of steps required to make a packet, including (for example) folds, slits, and locks. The Unlocking History research team, led by Dambrogio and Smith, is working to refine format and category information into metadata standards that can be globally adopted. 36 The material features of the letterlocking data -and thus, by extension, of the letters -open up new and exciting avenues for scientific analysis, allowing scholars to relate letters' content to their material features, and to explore technological trends and innovation across centuries, borders, and cultures.
From a more immaterial perspective, but still pertaining to letters as objects, the Brienne letters challenge commonly accepted notions about the nature of correspondence routes which may necessitate further revision of EMLO's metadata fields. The letter as an object is the product not just of one 'author', but of an entire system. EMLO, like most correspondence databases, had operated on the assumption that the origin of a letter is one fixed geographical location where it was physically written, while the destination is the location of the addressee. But what if there are multiple hands writing from various locations to a destination that is never achieved? Many of the Brienne letters came to The Hague, after all, not according to the will of their senders but by accident or omission: a good number of the letters now in 'La Haye en Hollande' were intended for 'La Haye en Touraine', a small village in France. Furthermore, The Hague was not the only place where these letters stopped along their originally intended routes; many bear the marks of other post offices along the way, sometimes including the dates on which they were there. In a sense, all of the letters in the Brienne Collection arrived at the 'wrong' destination, whether due to incomplete or indecipherable addresses or to the absence, death, or non-acceptance of their addressees. The reasons for non-delivery were carefully recorded by the post office in The Hague on nearly every letter; notes such as 'in England', 'departed', 'refused', attest to the many hands and voices implicated in the traffic of correspondence, which deserve themselves to be acknowledged and recorded. Thus, Signed, Sealed, and Undelivered and EMLO have developed metadata fields to record the address as intended, the route taken by the letter across time and space (including special handling instructions), and the reasons why it was never delivered.

Marie Isabel Matthews-Schlinzig
Descriptive metadata records on letters should ideally also include information on epistolary subgenre/s. In order to identify an appropriate way of recording this information, it is necessary to reflect on four aspects: (1) how epistolary subgenres are conceptualized; (2) how they were defined in the past; (3) how they are studied in the present; and (4) how the current state of knowledge on letter genres can best be integrated into data sets.

Conceptualizing Letter Genres
Letter genres (i.e. epistolary subgenres) 37 are commonly conceptualized in relation to a particularly distinct feature: this includes (1) a function, theme, or purpose (e.g. farewell letter, love letter, blackmail letter); (2) institutional or social contexts (e.g. chancery letter, children's letter); (3) a particular linguistic quality, style, and/or textual form (e.g. gallant letter, epistolary treatise); or (4) prominent material qualities (e.g. illustrated letter). Some letter genres are characterized by a combination of such defining features -think, for instance, of the use of black wax to seal letters of mourning. Letter genres emerge from and develop through epistolary cultures and practices, and in particular due to specific social, educational, economic, and institu-tional needs. 38 Hence, many subgenre designations (e.g. condolence letter) traditionally reflect specific types of social and cultural interaction. This social -as well as cultural -'grounding' (Charles Bazerman) of letters accounts for the fact that the existence, lifespan, age, characteristics, and names of epistolary subgenres can vary across geographical space and time, i.e. between cultures and languages, and within societies. Furthermore, developments in printing and publishing, as well as (modern) scholarly interests have played key roles in subgenre creation and change.
The multiple factors contributing to the development of epistolary subgenres are not the only reason why it is not always straightforward to delineate letter genres clearly or assign an individual piece of correspondence to a specific subgenre. Such difficulties are exacerbated further by basic characteristics of the epistolary form as well as of letter cultures as such: key factors include the polymorphism and malleability of letters, the epistolary form's inherent poetic qualities, and the often complicated transmission history of letters.
Letters, like 'texts' more generally, 'may', as Michael Sinding points out, '"coincide with" multiple genres in various degrees'. 39 Genre mixes -both in a specific as well as a broader sense -are common: one letter can be associated with several epistolary subgenres. Furthermore, just as letter writing in general can integrate several ways of recording and representing information as well as adapt other genres of writing (e.g. poetry, calculations, drawings), so epistolary subgenres can encompass miscellaneous types of text in several styles and material forms. The subgenre of scholars' letters, for instance, comprises such different forms as familiar correspondence, epistolary treatises, and a range of paratexts, including epistles dedicatory. 40 Ultimately, this multiplicity of forms relates back to the question raised earlier in this chapter: what actually constitutes a letter and how -if at all -can one differentiate it from other types of writing? While its perhaps most defining characteristic nowadays, the 'epistolary bracket' (i.e. the presence of an address and a salutation plus a signature and, potentially, a leave-taking formula) persists in genres of writing that have emerged from the epistolary form (e.g. legal documents such as letters patent), 41 it is not always fully present in letters, let alone a defining feature of some epistolary subgenres (e.g. billets, published letter series). 38  The fundamentally rhetorical nature of letters adds a further layer of complexity to the picture: while some epistolary subgenres are primarily non-fictional (e.g. papal bull/brief), others (e.g. letter poems) tend to encompass a large proportion of fictional or semi-fictional texts. Although this chapter centres on non-fictional writing, it should nevertheless be noted that letter writing takes place on a continuum of literary and factual writing activities. 42 Not least because of this, fiction can play a key role in the emergence and/or development of letter genres; after all, fiction does not only respond to, and document, but can also modify or set new standards for contemporary genre conventions (e.g. through epistolary novels).
Last but not least, a number of processes that take place during the transmission of a letter (e.g. publication) can also influence its association with one or more epistolary subgenre/s. The transformations that letters can undergo from being first drafted to -potentially -being circulated in print (e.g. fictionalization), can be manifold.
From these observations it is evident that letter genres are fundamentally as dynamic as the social and cultural processes they both manifest and shape. If we wish to identify epistolary subgenres and assign them to individual letters we must therefore be cognizant of two interrelated factors: firstly, the culture-historical contexts for the classification as well as use of letter genres in the past and, secondly, our own theoretical premises.

Early Modern Letter Classes
As Gabriella Del Lungo Camiciotti has remarked, '[i]n the early modern period […] correspondence acquired the characteristics and uses that were to become typical of the genre in the whole modern period'. 43 That this is also true of many epistolary subgenres can be gleaned from a range of sources (e.g. manuscript letters, letter collections, collections of model letters, letter manuals, and treatises on the epistolary form). However, while many a letter genre, especially from the later early modern period, might seem familiar to twenty-first-century readers, the subgenre's historical conceptualization can deviate considerably from present understanding. A comparative study of source material both from different parts of the period and from different language contexts will reveal significant variation both in classification methods and definitions of letter genres; there are also discrepancies between letter theory and letter writing practices.
Already before the sixteenth century, letter writing manuals had included definitions of a number of letter 'genres': e. tiva to the epistola domestica)'. 44 Initially, classical oration alongside classical letter writing served as the key frameworks for describing and categorizing epistolary subgenres; this was especially pertinent to letters written in Latin. 45 New thoughts on classifying epistolary subgenres evolved gradually.
Two different versions of Erasmus's Opus de conscribendis epistolis (1522) illustrate such a process. In an early draft (c. 1499-1500), Erasmus 'classifie[d] subcategories of letters under the categories of the oration (deliberative, demonstrative, and judicial)'; letters he could not describe in relation to these categories he called 'extraordinary'. 46 The 1522 version of the Opus refined this model. Not only did Erasmus differentiate between 'mixed' and 'unmixed' letters, i.e. those that cover one or many topics, before subdividing the latter again in relation to the rhetorical categories mentioned. He further replaced the term 'extraordinary' with 'familiar', and added, almost as an afterthought, a fifth category of letters that dispute, investigate, and teach. 47 This and other developments of letter genre classifications were in part due to changing conceptualizations and uses of the epistolary genre: from the Renaissance onward, letter theoreticians and practitioners began to consider the letter as a more flexible form than it had been to their medieval predecessors and contemporaries who continued to use the ars dictaminis. 48 Other causes driving epistolary subgenre development, which cannot be discussed in detail here, include increases in the variety of social groups using letters (e.g. a growing amount of non-official correspondence and increasing numbers of middle-class, as well as, later, lower-class letter-writers and -readers), and more frequent use of vernacular languages in letter writing.
A look at classifications of epistolary subgenres from both the earlier and later parts of the period indicates, also, that the classical rhetorical tradition continued to influence subgenre concepts arising in the seventeenth and eighteenth centuries. In Empire of Letters (2005), Eve Tavor Bannet briefly contrasts letter 'classes' present in eighteenth-century English-language manuals with those found in some sixteenth-century manuals. 49 In the latter, she writes:  (1599). Compare e.g. for an introduction to epistolary subgenres in German-language letter manuals in the seventeenth and early eighteenth centuries, Stylistically, letters were grave, learned or jocular; functionally, they were 'exhortatorie, accusatorie, commendatorie, excusatorie, congratulatorie, resonsorie, consolatorie, jocatorie, nunciatore, criminatorie, lamentatorie, mandatorie, dehortatorie, objurgatorie, petitorie, comminatorie, expostulatorie, ematorie, conciliatorie, laudatorie, or intercessorie'. 50 [In the eighteenth century] recognized classes of familiar letter […] included: letters of business; letters of advice; letters of praise or commendation; letters of recommendation; letters of remonstrance; letters of exhortation; letters of entreaty, request or petition; letters of counsel; letters of complaint; letters of reproof; letters of excuse; letters of thanks; letters of invitation; letters of congratulation; letters of consolation, comfort, or condolence; letters of visit; letters of compliment; letters proffering assistance; letters of merriment or raillery; and letters mixing two or more of the above. 51 Irrespective of potential similarities and differences between older and newer letter genres (which in many cases are still awaiting analysis), one of the more significant links between them is that their names reflect 'speech-act functions', i.e. they 'name possible verbal actions'. 52 Other, more recent forms of classification, found for instance in German-language letter manuals, identify letter subgenres in relation to the letter authors' 'intent', or a particular social group they belong to (e.g. merchants, 'gentlewomen', chancery officials, students, and soldiers). 53 As Bannet reminds us, conceptualizations of subgenres emerging in the later part of the period should also be approached with care, since some '[e]ighteenthcentury letter classes […] contained types of letter that we do not now think of together'. 54 'Letters of Business', for instance, 'dealt with court, political, administrative or government business, as well as with commercial affairs. […] [T]hey could include letters of advice, counsel, remonstrance, command, request, recommendation, offering assistance, complaint, reproach and excuse'. 55 The hierarchical relationship between letter genres observable in this last example is another feature that is subject to historical change and cultural variation (as already indicated, e.g. by Erasmus's categories of 'mixed' and 'unmixed letters', etc. Bannet's description highlights that letter genre mixes could be an integral part of historical conceptualizations of epistolary subgenres.
Finally, when studying historical subgenre concepts, we should remind ourselves that historical letter theory does not represent a complete record of epistolary subgenres in use in the early modern period. Letter treatises and manuals mostly had a clear agenda: they were used to codify practices that were already present, establish best practice guidelines, and/or to instigate change. Manuals also tended to have educational purposes, were influenced by changing ideological concerns, and often addressed a particular readership. As a consequence, some letter genres that were commonly used and even present in other types of publication either did not find their way into letter treatises and manuals, or did so only comparatively late.
An example of this is the epistolary subgenre of last letters written before death. The English language offers a number of names for letters in this subgenre, including 'farewell letter', 'farewell note', 'last/final letter' as well as 'last note', 'death-bed letter', 'suicide note', 'death note', and 'suicide letter'. 57 This -in comparison to other European languages -very long list of designations suggests a particularly differentiated cultural presence of the subgenre. Hence, in the eighteenth century, at least some examples of such letters found their way into Englishlanguage letter manuals. 58 This was, however, not the case in other languages and cultures, such as, for instance, in German: in fictional and non-fictional epistolary practice, the subgenre was well known and designated mostly either with letzter Brief ('last letter') or Abschiedsbrief ('farewell letter'). Yet, letter manuals did not include 'last letters', and defined Abschiedsbrief (or Abschiedsschreiben), only as polite notes which it was customary to write 'to good friends and patrons' when one parted from the place where the latter lived. 59 Given such gaps in historical letter theory, the widest possible range of relevant sources and manifestations of letters have to be taken into account in order to reconstruct historical knowledge about a letter genre.
From our present, scholarly point of view, both the lack of a historical theorization of specific epistolary subgenres, as well as the variety of different conceptualizations and designations that have developed over time/across cultures for what can essentially be described as one letter genre (e.g. letters consolatorie and letters of consolation), make it necessary to formulate new, 'historical-critical' definitions of epistolary subgenres. Although such definitions should reflect historical concepts and practice, they will, at least in part, also be constructs that are retroactively applied.

Modern Concepts
In the more recent past, the number of research projects and studies focusing on individual letter genres has been increasing. 60 Nevertheless, more comparative and interdisciplinary research on the theory, history, and relationships of epistolary subgenres is needed. There is, for instance, still no consolidated typology of letter genres that are frequent in the early modern period. Ideally, such a catalogue would also include definitions and examples, and be flexible enough to serve as a basis for work on correspondence across different disciplines and cultures. As a step toward such a typology, the end of this section offers a (not yet comprehensive) list of epistolary subgenres that makes use of 'historical-critical' letter genre designations.
Conceptually, this list (and indeed this contribution as a whole) is informed by work currently under way for the interdisciplinary project Handbuch Brief ('Letters -A Handbook'). 61 In order to throw into greater relief the relative merits and limitations of the handbook and thematize alternative approaches, it will be discussed here alongside two other, pertinent pieces of scholarship that define models for describing epistolary 'subtypes', genres, and/or subgenres. Both have a specific historical focus and are based on different disciplinary points of view; in what follows, they are used to highlight aspects of letter genres that current 'historicalcritical' conceptualizations do not necessarily capture.
In his article 'Letters: A New Approach to Text Typology', linguist Alexander T. Bergs defines letters as a '"surface" or "super" text type' which, he suggests, 'can be subdivided into […] socio-pragmatic subtypes on the basis of sociopsychological and pragmatic dimensions and factors, including speech act and accommodation theory'. 62 In defining his subtypes, Bergs reflects on (1) a particular set of primary source material, the late Middle English Paston letters; (2) the social positions of letter-writer and addressee; and, adapting primarily Karl Bühler, (3) the function that language fulfils in each case. 63 Accordingly, Bergs distinguishes five subtypes: • report (= descriptive, neutral) • request (= appellative, socially inferior to superior) • orders (= appellative, socially superior to inferior) • counsel (= descriptive) • phatic (= phatic-descriptive-expressive). 64 From his analysis of the Paston letters, Bergs concludes that '[i]n late Middle English, these socio-pragmatic subtypes do not necessarily differ in their overall structure and make-up, though they do show interesting and significant differences in their use of formulaic language, speech acts, and their functional elaboration of certain linguistic variables'. 65 '[S]ocio-pragmatic subtypes […]', he also acknowledges, 'may show a great deal of overlap and should thus be treated as non-discrete constructs'. 66 It is not only in this respect that Bergs's subtypes relate to historical letter genres as discussed above: as he himself points out, historical letter theorists such as Erasmus also adopted a classification of letters 'based on functional properties'. 67 Indeed, some of Bergs's subtypes appear very close to epistolary subgenres as defined above (see 4.1): e.g. his 'request' is similar to 'letters of petition', and Bergs himself identifies the 'Petition [i.e. 'request'] as belonging to the text type "letter"'. 68 As Bergs acknowledges, one of the main limitations of his list of subtypes is that it is not comprehensive. Yet, the subtypes' simplicity and flexibility, which is partly due to their 'reflect[ing] universal language functions', as well as their inherent focus on social relationships also makes them a potentially useful tool. 69 In this context, Bergs seems to suggest that his typology could be transferable to other periods, languages, and cultures, and that Bühler's model could, for instance, be used to identify clusters of epistolary subgenres that share specific language functions. 70 The typology of letter writing Claudio Guillén developed in his seminal Notes toward the Study of the Renaissance Letter also foregrounds certain formal linguistic and literary qualities as well as publication contexts of the epistolary form. 71 When surveying the field of the 'Renaissance epistle', Guillén writes, 'it becomes neces-sary to distinguish at least seven kinds of writing, each of which can be seen as following its own career and rhythm of development'. 72 His typology includes: • the neo-Latin prose letter In relation to our previous discussion, this list could be understood as a typology of epistolary subgenres that is alternative to the classifications introduced so far (e.g. in Erasmus, or by Bannet). However, it can also be seen as a typology of epistolary genres that, while subordinate to the 'letter as genre' in general, are superordinate to epistolary subgenres as defined earlier (see 4.1). Two features support the second understanding: firstly, Guillén, in his discussion of 'the prose letter in the vulgar tongue', mentions the 'Latin letter' as well as the 'Humanist and philological letters' as kinds (or, in our terminology 'subgenres') of prose letters. Secondly, many of the letter genres mentioned earlier (e.g. scholars' letters) could be associated with several or all of Guillén's categories. In this situation, it is also helpful to remind oneself that, as Guillén says: 'A piece of writing can be a hybrid; and to the question of its generic definition the answer need not be […] either yes or no'. 73 His typology has many merits: it is broad yet includes specific foci on language and form that allow us to capture both similarities between epistolary subgenres and their manifold incarnations in manuscript and print. The inbuilt attention to the trans-or, as Guillén terms it, 'supranational' existence of epistolary forms as well as to the close interrelationship of epistolarity and literature, is a particularly attractive feature. 74 However, although in theory this typology could be extended to include later material, it would not be finely grained enough to capture individual letter genres and their development across the whole of the early modern period in great detail.
The encyclopedic Handbuch Brief: Von der Frühen Neuzeit bis zur Gegenwart takes a more generalized approach: it aims at presenting an overview of current scholarly 72 Ibid., 71. 73 Ibid., 82. 74 Ibid., 91. tion/recommendation; letter of invitation; letter of merriment or raillery; letter of praise or commendation; letter of remonstrance/reproof; letter of thanks; letter of visit; letters close; letters of credit; letters patent; letters proffering assistance; letter to the reader; moral letter; missionary letter; papal bull/brief; polemical/satirical letter; report; royal letter (including e.g. lettres de cachet); secret letter.

Letter Genres in Metadata
The previous discussion of three scholarly approaches to classifying letter writing and epistolary subgenres or types has reminded us not only of the differences between disciplinary approaches to the epistolary form. The relative merits as well as limitations of these classifications also illustrate that capturing -in one neat model -all the many parameters which we can refer to in conceptualizing letter genres is difficult. Consequently, the inclusion of information on letter genres in descriptive metadata is also a challenge.
Yet, not least in view of the evident lack of scholarship on letter genres (including, for instance, in the field of genre theory), a systematic collection of information on epistolary subgenres in digital projects (e.g. databases, digital editions of correspondence, archives, etc.) appears particularly necessary and would unlock great research potential. 77 Including data on letter genres in digital projects would, to name just one example, facilitate locating source material (which is particularly valuable in cases of epistolary subgenres for which only few examples have survived). Altogether, it is fair to assume that the use of digital tools to collect as well as analyse information on letter genres would lift research in this field -both nationally and internationally as well as within and across disciplines -to new levels.
Mindful of these chances but also the challenges mentioned, I would like to suggest that information on letter genres should be tagged not in relation to a typology that is specific to a discipline, language, or culture, but instead with the help of a limited number of keywords. Transcending fixed letter genre definitions, these keywords would reflect the categories most commonly used both in historical and present-day conceptualizations: i.e. (1) speech acts and social actions (e.g. 'advice'); (2) major themes (e.g. 'love'); (3) social and/or institutional identity of author or context of letter (e.g. 'merchant', 'church'); (4) key linguistic, stylistic, formal features (e.g. 'expressive', 'gallant', 'verse'); and (5) miscellaneous aspects (e.g. 'illustrated'). Ideally, keywords such as these should be derived from as wide a range of letter material as possible and consolidated over time.
Using these five dimensions, the large and varied genre of 'scholars' letters' (a more recent relative of a key early modern letter class -the 'learned letter') could, for instance, be tagged and recognized with keywords such as those below: 78 1. 'inform', 'recommend', … 2. 'patronage', 'philosophy', 'science', … 3. 'university', 'courtly service', 'emigré', … 4. 'familial', … 5. 'containing bibliography/scientific objects', 'illustrated', … Various types of scholars' letters -such as the 'epistle dedicatory' (or the 'letter to the reader', etc.) -could be differentiated with the help of other, sometimes additional keywords: e.g. (1) 'dedicate', (4) 'verse', 'epistle'. By the same token, more detailed information on publication contexts or text types could also be recorded, e.g. in dimension (5) 'poetry collection', 'spiritual'. Specific combinations of keywords would then allow users of databases to look for scholars' letters in general as well as, for instance, for epistles dedicatory only, or, even more specifically, for epistles dedicatory published in a poetry collection and/or a piece of spiritual literature.
To allow comparability across different databases, the sets of keywords to be used for recording genre information should be standardized and publicized to users of the database. The keywords should always be chosen with an eye to optimizing flexibility, and expandability. Of course, including this information would require additional labour and potentially create considerable sets of data. However, given the potential benefits of making this data available, which include a better understanding of letter genres and their interrelatedness with various types of texts, it seems definitely worth the effort. All this, of course, must be preceded by a consideration of how we model letters as metadata, which is the subject of chapter II.7 of this book.