Introduction

§ 1 This article will provide a brief introduction to the Brown University Virtual Humanities Lab (VHL)'s major new project: the electronic Esposizioni.^[1] Like other online texts available on the Decameron Web (http://www.brown.edu/decameron/), the Esposizioni sopra la Comedia di Dante is a major vernacular work by Giovanni Boccaccio; however, unlike these other texts, the Esposizioni is a vast, non-fictional work, being the text of Boccaccio's unfinished lecture series on Dante's Commedia. The task of developing an electronic version of this commentary has allowed us to address, define, and finally visualize the relationship between the primary (commented) text and the secondary (commentary) text. We hope that our work thus far will be of use to the wider scholarly community. We are at an early stage of this project and would welcome any comments from interested parties as we begin to implement it.

§ 2 The article is divided into three sections. First, we provide a brief introduction to Boccaccio's commentary on Dante and present some of the features of the text that are highlighted in the electronic edition; we then provide a short history of the project and some technical background. The third part of the paper is devoted to a presentation of the electronic edition of the Esposizioni, comprising a technical description of the commented and commentary texts, a discussion of the structural and semantic encoding principles, and a practical demonstration of the encoded accessus. Finally, we hope to show how Boccaccio's Esposizioni is a true medieval hypertext, as Theodore Nelson originally defined it in 1965: a body of written or pictorial material interconnected in such a complex way that it could not conveniently be presented or represented on paper (Nelson 1965).

§ 3 The Esposizioni is a motley text. It is highly digressive and thus appears disorganized from the modern point of view. However, we should seek to understand it in its historical context, and its polychromatic nature is best explained by its origin as a text written in response to another text (the Commedia). As hypertext, the Esposizioni surrounds and contextualizes Dante's poem, providing a point of access into and multiple routes out through its many references to and citations of other authors and works. The Esposizioni is an extraordinary—and unique—book, at once a record of one author's intertextual engagement with another and an encyclopaedia of literature. As such, it is particularly suitable to the electronic milieu.

§ 4 Historically, Boccaccio's Esposizioni has been considered almost exclusively in terms of its status as a commentary and, as a result, there has sometimes been an assumption that this work is secondary and in some way subordinate to Dante's Commedia. However, in recent years, Boccaccio scholars have begun to move away from this Dante-centred response to the text by relating it to other works in Boccaccio's corpus. One of the principal aims of the VHL edition of the Esposizioni is to return the text to Boccaccio and present the text as a key example of his expository and encyclopedic writing, rather than merely as an adjunct to Dante's poem.

Introduction to the Esposizioni

§ 5 Before describing our electronic edition, it is worth highlighting some of the unusual characteristics of Boccaccio's commentary. In 1373, some of the citizens of Florence petitioned the comune (city government) for a public reading and explanation of Dante's Commedia. As the acknowledged expert on Dante in Florence, Boccaccio was appointed to the task for a period of one year and began his readings on Sunday 23 October 1373 in the church of Santo Stefano di Badia (Padoan 1965, vii). Although records are scarce, it seems from the manuscript evidence that Boccaccio gave about 55 lectures in total before he fell ill and suspended the series at canto 17 of the Inferno. He was never well enough again to resume his lecturae and died on December 21, 1375.

§ 6 The keen interest of the Florentine people in Dante's poem was not a new phenomenon: demand for exposition and clarification of the Commedia can be seen even in the years immediately after Dante's death in 1321.^[2] The combination of an eschatological thematic, accessible language, and easily memorizable structure made the Commedia an instant success, and thus the text penetrated all sectors of society very quickly.^[3] For the semi-literates and vernacular literates (i.e. those people who could read and write to a greater or lesser degree in the vernacular), the demand for exposition was met by public readings and lectures; the vernacular literates and the litterati (i.e. those who could also read and write in Latin) could also access learned commentaries written in Latin or Italian.

§ 7 The written commentaries to the Commedia that precede Boccaccio's lectures are thus all essentially directed towards a fairly restricted reading public made up of those people who occupied the upper end of the literacy spectrum. Of the eleven commentaries written before 1355, six are in Latin and five are in the vernacular (Caesar 1989, pp. 6-7); and although quite different in form and emphasis, all of them are at least concerned in some way with a literal clarification of Dante's poem, often also including information on literary allusions and scientific knowledge.

§ 8 Boccaccio's commentary text differs from the texts of the previous Dantean commentary tradition in several crucial ways. First, and most importantly, the text of the Esposizioni was not written with the intention of becoming a published book; the Esposizioni is simply the notes for Boccaccio's lecture series, and as such resists characterization as a typical medieval written text. The oral intention and nature of this text will thus be one of the elements upon which we will be focussing in the online edition. (In fact, one of the major benefits of the online format is that we are able to distance this text from its material format as a bound book, which fixes and preserves the oral text as a canonical and immutable object.)

§ 9 Secondly, in terms of its manuscript history, the text itself is uncertain: Boccaccio's notebooks containing the lecture notes have been lost and the text has been transmitted via only five manuscripts. Furthermore, only two of these manuscripts were produced in the period immediately following Boccaccio's lectures, having been tentatively dated to the late fourteenth or early fifteenth century (Padoan 1965, 713-15). The text that we possess is thus doubly distanced from Boccaccio's lectures: what we have is a (possibly) incomplete or inaccurate copy of Boccaccio's autograph, a version that may have served only as a cue for the complete text, which would have been delivered orally at the lecture. The Esposizioni is thus rather a peculiar artifact in the world of textual scholarship, and this may be one of the reasons that critics have often found fault with it.^[4] However, the unstable and unusually fluid characteristics of this text are precisely why it is so suited to the digital environment.

§ 10 Although it is unusual among Dante commentaries, Boccaccio's text nonetheless demonstrates contemporary medieval commentary techniques. Even though the text is designed for oral delivery to an audience of various levels of literacy, it is as highly structured as any other medieval commentary. The strutture portanti (support structures) behind Boccaccio's lecture notes are the architectural structures of Dante's own poem, as he organizes his lectures around a systematic and methodical line-by-line (and sometimes word-by-word) progression through the Commedia. Like all his other works, the Esposizioni is a highly organized work, which is a great advantage for us as architects of the electronic edition. Within this systematic progression, Boccaccio also employs typical rhetorical structures derived from the dictates of the ars dictaminis.^[5] After the general accessus, which serves as introduction to the entire work, Boccaccio then proceeds to examine each canto in turn, first according to the literal interpretation and secondly to the allegorical interpretation. Being Boccaccio, and thus with an almost uncontrollable tendency to digress and diverge from even self-imposed restrictions, this standard structure soon breaks down.^[6] For example, according to his notes, Canto 10 has no allegorical content whatsoever (Questo canto non ha allegoria alcuna, Esp., X, 110); likewise, he states that he is suspending the allegorical discussion of cantos 15 and 16 until later (a promise that he is unable to keep due to his illness). It is clear from even a cursory examination of the text that he devotes much more attention and time to his literal exposition than to the allegorical. This may well be due to the fact that the literal exposition is the place where he displays one of the most characteristic traits of his writing, the synthesis of many different literary sources in the service of his narrative.

§ 11 Although he attempts to follow the established principles of rhetorical divisio (probably most successfully at the beginning in the accessus), in fact, as a text-type, Boccaccio's commentary resembles an oral sermon much more than it does previously written Dante commentaries. Boccaccio himself refers to his lectures as sermoni, demonstrating that he himself envisaged his task as primarily a didactic one;^[7] and it should not be forgotten that the lectures themselves took place in a Florentine church, nor that Boccaccio himself had taken minor holy orders. The presence of homilectic devices within the text is thus the final element of our analysis and is the key to understanding how the Esposizioni is able to bridge so many dichotomies: as a document it is both oral and written, secular and religious; it looks back to scholastic techniques and forward to humanist learning.

The electronic Esposizioni: background

§ 12 In 2003, the electronic Esposizioni project was submitted as part of a broader grant proposal to the National Endowment for the Humanities. Following the success of this application, the Esposizioni project is now able to go ahead as part of the Brown University Virtual Humanities Lab (VHL). The electronic Esposizioni project is at the centre of the VHL, both in terms of its content and in the opportunities it offers in developing new online tools for textual editing and collaboration.^[8]

§ 13 The encoded text of Boccaccio's Decameron always has been the centrepiece of the Decameron Web (http://www.brown.edu/decameron/) since its inception in 1995. Since then we have added a number of further electronic texts to that site. Now, the Decameron Web contains five full texts: the Italian-language Decameron; the first anonymous English translation of 1620 and the 1903 English translation by James Macmullen Rigg (Anonymous 1620, Rigg 1903); and two further important vernacular works of Boccaccio (the Corbaccio and the Elegia di Madonna Fiammetta). These texts are encoded structurally and semantically, with varying degrees of complexity. The three Italian texts are semantically encoded in XML; the 1620 English translation is included for historical purposes and therefore presented in simple HTML; finally, the Rigg translation is XML-encoded with structural cross-references that enable the user to move between the main Italian text and parallel English translation.

§ 14 In early 2002, when the editors made the decision to add another text to the five already online, the Esposizioni rapidly became the obvious choice. The text is of comparable size and textual complexity to the Decameron, running to some 709 densely printed pages in the critical edition, and the Decameron Web has much technical experience of handling works of this size. Notwithstanding the richness of the content, the Esposizioni's primary interest is due to its genre. Three of Boccaccio's fictional works have already been included on the site, and thus it seemed a propitious time to highlight another aspect of Boccaccio's literary endeavours that is often overlooked—his compendious encyclopedic writings. Boccaccio wrote a considerable number of this type of work during his lifetime (e.g., the De mulieribus claris, the De viris illustribus, the Genealogia deorum gentilium). The Esposizioni stands out among these by virtue both of the language in which it was written (Italian rather than Latin) and by the fact that it is a learned commentary on the work of a near-contemporary vernacular author.

§ 15 As editors of a widely used, open access online educational resource (over the last year the site has been accessed an average of 793 times per month, 774 of those from outside Brown), we felt that it was imperative to concentrate on Boccaccio volgare rather than Boccaccio latino. Thus we chose to focus on the Esposizioni rather than on, for example, a text of considerable interest but potentially forbidding language such as the De mulieribus claris. In this way, we will be able to reach as large an audience as possible, including scholars of Italian language and literature at undergraduate, graduate, and the highest level, as well as the Italian audience itself. Since one of VHL's primary goals is to create a virtual editing house—that is, a place where scholars can collaboratively annotate and edit texts—the Esposizioni seemed a natural fit, being an immensely rich and as yet largely unexplored text.

§ 16 Of course, the Dantean subject matter of the Esposizioni also attracts a wider audience than Boccaccio students and specialists. Dante studies remains one of the largest and most popular areas of study in Italian literature worldwide. The Esposizioni is already available for searching as part of the Dartmouth Dante Project, which is concerned with commentaries on the Commedia and forms part of the Princeton Dante Project.^[9] The Dartmouth Project currently contains seventy-one different commentaries to the Commedia, from the earliest Trecento commentaries to Nicola Fosca's 2003 commentary. The searchable database can be used to perform searches and cross-comparisons within the entire corpus of Dante commentaries, but differs from our project in several significant ways.

§ 17 The most obvious difference between the two projects can be seen in their interfaces. The Dartmouth Dante Project is essentially a searchable text database that is accessed and navigated via a telnet connection or web interface.^[10] In theory, the telnet delivery system is perfectly serviceable, but the text-only interface has to be navigated by option keys in a series of menus, and is slow and counter-intuitive for users who are used to accessing material through a web interface. Although the project's web interface improves usability, the commentary search engine is primitive at best. The search facility is forbidding enough to deter all but the most determined users, and requires careful study of the information pages that accompany the search interface. Even with the instructions, a certain amount of trial and error is required to display the text, and it is only possible to access this text through the search menus. Visually, the interface is very plain, with a small black typeface on a white background. The only way to interact with the Esposizioni within the site is via Dante's poem: the text is navigable through the location references of the Commedia and is not free-standing. Of course, this is understandable, given the fact that Boccaccio's text is merely one of seventy-one which make up this corpus of commentaries.

Figure 1: Dartmouth Dante Project search interface (50% actual size)

§ 18 In comparison to this, the electronic Esposizioni explicitly benefits from the access, usability, and presentation advantages offered by the web interface. In our project, Boccaccio's commentary will be presented as a primary, stand-alone text, readable in its entirety. The text's relation (but not necessarily subordination) to the Commedia is expressed through the option of viewing the text of the Commedia in a separate and adjacent window at the same time. The search engine we are building will be based not on the relevant canto or line of the Commedia (that is easily enough accessible through indexes, which we are also building) but rather on sophisticated semantic encoding centered around aspects such as Boccaccio's literary sources, the complex rhetorical structures of the text, themes recurring in the sermoni, etc. Finally, the annotation engine will allow a further scholarly commentary to be built around the text, as users add their own comments and references.

Figure 2: Prototype annotation interface (50% actual size)

§ 19 Overall, then, the difference between the Dartmouth Project and the VHL's electronic Esposizioni is both one of orientation (the one directed towards Dante, the other Boccaccio) and magnification (one offering a broad view of all the commentaries, the other a close-up of one text). We emphasize the text's status within Boccaccio's oeuvre and allow for a highly detailed examination of its literary sources and structures independently of—or at least on a par with—its renowned subject. The Dartmouth Project is necessarily directed towards Dante and allows (among other things) the user to compare how different authors have responded to the Commedia. The two projects are fundamentally complementary and, taken together, allow different approaches to the relation between commentary and commented text. Finally, in addition to boccaccisti and dantisti, we hope that the XML-encoded Esposizioni will also be of interest to scholars of medieval literature and rhetoric in general, since it is a prime example of the genre of medieval commentary.

A more detailed look at the electronic Esposizioni

§ 20 In keeping with academic good practice, we have taken the decision to semantically encode the electronic Esposizioni in XML. At the moment we are not following a pre-created Document Type Definition (DTD) and have opted instead to create our own as an interim measure; this gives us the flexibility we may need for such an idiosyncratic text. In a second phase, we will translate the final version of our DTD into one compatible with the Text Encoding Initiative (TEI) Guidelines, so that our project might be more easily cross-referenced.

§ 21 Our electronic Esposizioni is actually made up of two encoded texts: the Esposizioni itself and the relevant sections of Dante's Commedia, linked together for cross-referencing and navigation purposes. Given the lack of an autograph manuscript for either of these two texts, problems immediately arose with regard to textual discrepancies. Within the Esposizioni, Boccaccio cites the Commedia from the manuscripts in his possession; in fact, the manuscript evidence shows that he worked with several different copies of the Commedia to hand in an attempt to create the most authoritative version means of by textual comparison. Accordingly, for our electronic edition of the Esposizioni we reproduce the text established by Giorgio Padoan in the knowledge that this is the best version of an imperfect text and one that preserves Boccaccio's Dante without any paradoxical modification to modern critical norms (Padoan 1965). Likewise, the text of Dante's Commedia that we use reproduces the critical text established by Giorgio Petrocchi in his Commedia secondo l'antica vulgata, which has become the standard reference edition (Petrocchi 1966-1967). We use this text in the knowledge that there is an incongruity between Petrocchi's Commedia and the text used by Boccaccio. Boccaccio's work as editor of Dante was an attempt to establish a definitive text; as a result, the manuscript tradition of the Commedia is considered to be contaminated by Boccaccio's editing. Petrocchi's edition thus uses the pre-Boccaccio manuscript tradition in order to best recreate the language of Dante himself.^[11] After some discussion, we decided that it was best to use the standard critical editions for both of these works, acknowledging the variants that would arise. Differences between the two texts are generally minor and, in the case of more serious divergences, the textual variants are extremely interesting in their own right.

§ 22 Elements of the work that are of particular interest to us are evident in what we have chosen to encode:

• documentation about the project, including entities responsible for its progress and funding, the principles upon which this particular text encoding is based, as well as details of the print critical editions used as the basis for the electronic text files, details of editorial practice, and a description of the history of the project;
• basic structural elements such as chapter divisions, the divisions between literal and allegorical exposition in each chapter, paragraphs and milestones (following the numeration established in the critical edition);
• proper names, along with contextual information about the people, places, and entities they represent;
• citations from other authors quoted by Boccaccio and notes as to the erroneous nature of some of these citations (with information about their authors and the works in which they appear—in some cases, the quotation is anonymous and we have found that these tend to be either proverbial sayings or Boccaccian constructions intended to give personal statements a more general import);
• Greek and Latin terms and their meanings, to be indexed;
• the many terms, Italian as well as foreign, that Boccaccio treats as lemmata, explaining their meaning in detail to his audience, and the definitions themselves; these lemmata will also be indexed and the glosses cross-referenced to them—particularly useful in cases where Boccaccio defines terms more than once;
• rhetorical devices explicitly and implicitly used by Boccaccio;
• the complex rhetorical structure that Boccaccio lays out every so often, and divergences from it, if any;
• digressions in which Boccaccio stops addressing Dante's text directly and spends at times entire pages addressing a broad topic of interest to him (for example, poetry).

§ 23 To illustrate some of the above points, we will consider two passages from Boccaccio's literal exposition (esposizione litterale) of Inferno I. The first passage shows examples of Boccaccio's citation of Dante and other authors:

Dice adunque così: <quote work="inferno" linebegin="I_1" lineend="I_1">Nel mezzo del cammin di nostra vita</quote>. Ove, ad evidenzia di questo principio, è da sapere: la vita de' mortali è, massimamente di quegli li quali a quel termine divengono il quale pare che per convenevole ne sia posto, settanta anni, quantunque alquanti e pochi più ne vivano e infinita moltitudine meno, sì come per lo <name nameid="salmista" nametype="person" myth="yes" biblical="yes" authored="bible">Salmista</name> si comprende nel <name nameid="salmi" nametype="work" partof="bible" author="salmista" rend="italics">Salmo</name> LXXXVIIII, dove dice: <quote work="salmi" linebegin="89_9" lineend="89_10" rend="italics">Anni nostri sicut aranea meditabuntur; dies annorum nostrorum septuaginta anni. Si autem in potentatibus circa octoginta anni; et amplius eorum labor et dolor</quote>: e perciò colui, il quale perviene a trentacinque anni, si può dire essere nel mezzo della nostra vita. (Esp.,I (I), 3)^[12]

§ 24 The opening phrase in quotation marks is a citation of the opening line of the Commedia and would be marked up as such (Inf., canto I, line 1). The tagged citation itself acts as a hyperlink to the appropriate point in the commented text. Users will thus be able to orient themselves within Dante's poem via the metadata structure of the encoded commentary text. At line 4 in the above extract, Salmista would be encoded as an individual author (il Salmista / the Psalmist) and as one of the authors of the Bible; the name of the literary work is also encoded (Salmi / Psalms) and it is marked as part of the Bible; finally, the quoted passage is tagged as a citation with its full reference (Psalm 89, lines 9-10). In addition, a note will be added within the encoding to the modern version of the text (which reads nostrorum in ipsis septuaginta anni... in potentatibus octoginta) and a further reference to Padoan's footnote in which this information is found (Padoan 1965, 775 n. 4). A note to the critical edition used for this reference (in this case, the Vulgate Bible) will also be included. The encoding of the above elements will therefore allow the generation of indices showing any of these elements: e.g. an index of citations from the Psalms or, more broadly, from the Bible; a list showing which lines of Dante are commented (and, perhaps more interestingly, which lines are not commented); or a list of incorrect references.

§ 25 The second passage shows some examples of our standard informational encoding:

<quote work="inferno" linebegin="I_71" lineend="I_71">E vissi a Roma</quote>. Certa cosa è che <name nameid="virgilio" nametype="person">Virgilio</name>, avendo l'ingegno disposto e acuto agli studi, primieramente studiò a <name nameid="cremona" reg="Cremona" nametype="place">Cremona</name> e di quindi n'andò a <name nameid="melano" reg="Milano" nametype="place">Melano</name>, là dov'egli studiò in medicina; e, avendo lo 'ngegno pronto alla poesia e vedendo i poeti essere nel cospetto d'<name nameid="ottaviano" reg="ottavianoaugusto" nametype="person">Ottaviano</name> acetti, se ne andò a <name nameid="napoli" reg="Napoli" nametype="place">Napoli</name> e quivi si crede sotto <name nameid="cornuto" seealso="LuciusAnnaeusCornutus" nametype="person">Cornuto</name> poeta udisse alquanto tempo. (Esp., I (I), 61).^[13]

§ 26 Again, the first encoded element is the reference link to the relevant line in Dante's text (Inf., I, 71), so that a user may view the commentary alongside its original context if she wishes. The proper names are encoded as such and will later be indexed; and a regularized spelling of Milano is encoded into this occurrence of Melano so that the city represented by these words may be found regardless of how it is spelled by Boccaccio. A note will be added to identify Cornuto as the Stoic philosopher Lucius Annaeus Cornutus.

§ 27 Realizing that this text, and therefore also its annotation, are full of uncertainties, we do not presume to publish a definitive edition of the Esposizioni online. Rather—and this is another principal aspect of the VHL under development—we are constructing a virtual editing house. Anyone with internet access will be able to read the annotated Esposizioni, but in addition scholars with privileged access will have the opportunity to add their own annotations to the work and to comment on others' annotations. Inevitable disagreements about aspects of the semantic encoding will be discussed and resolved through variant encodings, and it will be possible to discuss any annotation or section of text in a discussion forum.

Figure 3: Prototype Virtual Editing House interface showing demonstration annotation (80% actual size)

Summary and conclusion

§ 28 We believe that our electronic edition of the Esposizioni offers much to the scholar of medieval literature. Our simple encoding scheme offers readers a number of ways into and through the text of both these major literary works. The most obvious benefit of the electronic medium is the fact that, for the first time, it is possible to view the full text of both works together and to move through them both with equal facility. The interface of the electronic Esposizioni links the commentary and commented text together in a new and original way: through the common system of canonical references, it is possible to navigate both texts with reference to the other. The texts can be displayed side by side and the user can call up the corresponding passage in the other text through a simple click. The presentation of these texts within a highly usable interface designed specifically for this purpose is a major step forward from both the traditional delivery method (the printed page) and other digital versions of the texts. The benefits of such convenience are not negligible, as anyone who has ever wrestled with several hefty commentaries open together on their desk will know.

§ 29 A particular benefit of our electronic edition is that it allows the scholarly user to reconstruct Boccaccio's own Commedia, not so much in terms of the textual history (which has been amply covered elsewhere) but rather in relation to his view of what was important and worth explicating within Dante's text. Since all the citations from Dante are identified according to the referencing system, it is possible to build up the actual text commentated by Boccaccio that itself well may differ from the well-known, canonical version of the Commedia. Favoured passages have proportionally more exposition than others while some lines go entirely without comment; it will thus be possible to draw conclusions from the Esposizioni about Boccaccio's use of Dante in his previous works.

§ 30 In addition, the computerized indices we are building based on the semantic encoding will allow the user to find out much more about the text than before. It will be a simple matter to call up a list of all the literary authorities that Boccaccio cites in his work; and from there to generate a list of all the particular citations from one author, for example. We have no doubt that this facility will suggest further associations and intertextual references that have been previously overlooked. The word search feature will also work as a simple concordance, so that it will be possible to investigate Boccaccio's use of certain references that have not been formally included in the encoding.

§ 31 The aforementioned glosses are another major feature of this electronic edition. The list of terms glossed by Boccaccio will be of great use to Boccaccio scholars and medievalists alike, in that it is primary evidence of the type of literary debates occurring in intellectual circles of the time. For example, the example included in the accessus shows Boccaccio's contribution to that never-ending debate of the meaning of Comedìa (Esp., Acc., 17-26). The index of glossed terms will thus work almost as a dictionary of Boccaccio's personal usage. The eventual situation of this index on the Decameron Web will mean that Boccaccio's usage can be compared in the Esposizioni with the texts of three of his earlier literary works, the Decameron, Fiammetta, and Corbaccio.

§ 32 The Esposizioni is a remarkable example of the commentary genre and a fitting conclusion to Boccaccio's work. For this reason, we feel that the electronic Esposizioni project will allow us to re-present and re-dimension this highly significant text, and we believe that the very encoding process will cast light on numerous aspects of the book that have been hitherto overlooked. We hope that our bringing this work to your attention here will not only generate early feedback on the project, but also spark more interest in electronic scholarly resources, and critical semantic encoding in particular. Readers who would like to contribute thoughts are encouraged to do so either through e-mail (vhl@wordsend.org) or at our weblog: http://www.brown.edu/Departments/Italian_Studies/vhl/.