Silvia Stoyanova and Ben Johnston
Figure 1. Paragraph View from the Zibaldone Hypertext Research Platform Website
S ilvia Stoyanova holds a PhD in Italian literature from Columbia University with a dissertation on Giacomo Leopardi. She has taught Italian language, culture, cinema and literature at Columbia University and at Princeton University. She has been a postdoctoral fellow at the Center for Digital Humanities at Universität Trier in Germany and is presently a Visiting Scholar at the University of Macerata, Italy.
She is interested in exploiting the relational dimension offered by digital technologies for the dynamic mediation of modern fragmentary narratives, such as the intellectual notebook, and for the construction of knowledge sites. Since 2004 Silvia Stoyanova has been experimenting in private with hypertextual argumentative writing within the limits of Microsoft Word, i.e. with fonts, colors, diagrams, WordArt animations, hyperlinks, etc. She is optimistic that the current academic interest in technological approaches to the humanities will cultivate a fertile environment for scholarly hypertext.
B en Johnston is Senior Educational Technology Specialist, McGraw Center for Teaching and Learning, Princeton University. Johnston has worked with faculty from across the University to develop and maintain technology projects for teaching and research, and has taken a lead role in the development of several projects focusing on electronic databases for textual analysis and transcription such as the Princeton University Sefer Hasidim Database, (PUSHD) the Nationalism, Ethnicity and Self-Determination Index, (NESDi) and the Princeton University Geniza Project.
F or content | code | process (formerly Authoring Software), Silvia Stoyanova and Ben Johnston discuss the creation of their Zibaldone Hypertext Research Platform, which migrates Italian poet and scholar Giacomo Leopardi's extraordinary hypertext precursor, the Zibaldone -- created in the 19th century over the course of 15 years -- to a web-based platform.
"In its larger scope", Stoyanova notes, "the project studies the structure of this fragmentary text and its author's techniques for its semantic organization as a model for adopting hypertext to mediate the phenomenological method and relativistic thought procedure at work in migrating similar research note collections into argumentative narrative."
Based on Leopardi's hypertextual organization of his notebook, Stoyanova and Johnston's innovative project also serves as a resource for potential uses of large databases to create hypertext literature.
Silvia Stoyanova and Ben Johnston: The Zibaldone Hypertext Research Platform
T he Zibaldone Hypertext Research Platform (which is currently available at http://digitalzibaldone.net) is a digital reconstruction of the hypertextual design inherent in the research notebook (Zibaldone) of the acclaimed nineteenth century Italian poet and scholar Giacomo Leopardi. The project was undertaken at Princeton University in the Fall of 2010 by Dr. Silvia Stoyanova (French and Italian) and Ben Johnston. (Humanities Resource Center) Consultant: Dr. Clifford Wulfman, Coordinator of Library Digital Initiatives.
In its larger scope, the project studies the structure of this fragmentary text and its author's techniques for its semantic organization as a model for adopting hypertext to mediate the phenomenological method and relativistic thought procedure at work in migrating similar research note collections into argumentative narrative. This objective indeed recalls the original conception of hypertext by Ted Nelson as a medium for capturing the many possible trajectories in the course of developing an argument, instead of sacrificing them because of the limitations of the two-dimensional space of paper. It was precisely in an attempt to overcome the same limitations that Leopardi interspersed his notebook with thousands of cross-references, linking its apparently fragmented passages, creating the blueprint of a virtual hypertext.  [Figure 1]
Although its title and to some extent its contents reflect the humanist compiling of commonplace books, (in which readers and scholars extracted quotations from their readings) the Zibaldone is more akin to the modern intellectual diary and further exhibits features of the present-day academic blog, such as a date stamp and thematic tags for each entry. While Leopardi was noting down observations and commenting on books without an immediate objective in mind, at the same time he was also recording their connections to previous reflections and intended to eventually rework his material into formally cohesive discourses on a great variety of subjects -- from linguistics, to social mores, to aesthetic theory. For this purpose, he indexed his material thematically and added more cross-references between passages while re-reading.
Figure 2. 1827 Index is a detail of the beginning of the index Giacomo Leopardi wrote in 1827 for his Zibaldone. (as implemented on the Zibaldone Hypertext Research Platform) "1827 index" has about 800 themes and subthemes.
The Zibaldone Hypertext Research Platform utilizes XML and Hypertext technologies in order to render the dynamics of the associative, cumulative mode of argumentation in the text, and thereby to allow readers to engage with its fragmentariness in a qualitative manner. For example, readers can explore its temporal stratification by evidencing demarcations for marginal and interlinear comments or whether link references were added at the time of composition or subsequently.
For each paragraph there is an information window which expands right above it and lists
On the other hand, users can rearrange the fragments in semantic clusters by viewing each paragraph in the context of its incoming and outgoing hyper-links to other paragraphs along with their thematic index tags.
Figure 3. Information Window shows a detail of navigation by date entry of the text in the Zibaldone Hypertext Research Platform.
By bringing together graphically and textually a series of meditations on a subject from numerous angles and a variety of contexts, we can recreate the unified but ephemeral vision of the mental process of how an argument's many parts hang together, which Leopardi describes. In exploring these interconnections, we have assigned a value of semantic relevance to the links between fragments based on whether they were established as explicit continuation of a previous passage, in the course of composition, during re-reading, between passages belonging to the same index category, or are indefinite verbal references to previously discussed themes, which have been specified by editors.
Figure 4. Index Theme Dante is a Gephi visualization of one of Giacomo Leopardi's index themes in relation to additional themes belonging to the paragraphs listed under "Dante". When the cursor is pointing to a specific paragraph listed under the theme "Dante", a scholar can also see the related themes under which this paragraph is listed and thus recreate the semantic field of the main theme "Dante".
I n its encyclopedic quest for knowledge, the Zibaldone has also woven a thick intertextual network of bibliographic citations to works with which Leopardi critically engages. By connecting the contents of these references to their URLs on the web and producing chronological and spatial charts with names and places mentioned in the text, readers can gain an informative overview of the ideological environment that generated it. One aspect of the project examines the development of the modern intellectual diary from the commonplace book by distinguishing and marking textual material as quotes, bibliographic references, commentary and meditations without a defined anchor. Additional reading approaches and platform functionality, such as a social edition enabling users to add their own annotations and bibliographic networks, are under development. Thereby, the project aspires to enable a comprehensive, simultaneously close and distant reading of the text, as well as to derive a paradigm for exploring the genre of the intellectual notebook through current digital approaches to the semantic organization and the formal mediation of our research material.
Figure 5. Quotations and Bibliographic References Chart is a statistical chart, which shows a representation of the proportion of quotations and bibliographic references in the text. It can be clearly seen that towards the end of the text this proportion is quite high, (about 35%) as the text's fabric disintegrates.
T he Zibaldone Hypertext Research Platform originated as a rather simple HTML encoding of the document in which references within the work were made into clickable links. A Microsoft Word version of the text  was converted to HTML and further work was done using Adobe Dreamweaver and the open-source KompoZer HTML editors. In the process of creating this HTML encoding, a TEI encoding was suggested, which, besides being suited to literary text encoding, would contribute to the digital edition's future sustainability and encourage collaboration.
The use of XML technologies for the encoding, storage, and querying of the texts, enabled the development of more dynamic display of the text than in this project would have been possible with HTML. Hyperlinks between the author's thematic indexes and the main text, for example, allow the user of the research platform to jump to individual paragraphs or date ranges listed in the indexes, but also to generate lists of all paragraphs from the text associated with a theme. Having both the indexes and the main text encoded in TEI allows one to perform queries upon the texts and generate lists, such as these, dynamically. TEI and associated XML technologies also facilitate transformations of the text into many varying displays, which the user can call up on demand.
The oXygen XML Editor served as the primary tool for work done on the encoded files. The software provides tools for validating the correctness of work based on a pre-defined schema and provides suggestions and auto-completion. In addition to simple keyword searching of the text, the Editor allows one to search the structure of the document through XPath expressions, provides a multi-line find and replace tool, and allows for regular expressions. A server-based instance of the eXist-database was also established for this project. This open-source software not only allows one to publish XML documents to the web, but also includes a Lucene text indexer, XQuery processing, and RESTful and XML-RPC interfaces enabling the processing of requests from other web-enabled applications. Being a Java application, the software also fits nicely into the existing Princeton campus IT infrastructure.
For this project, on-demand XQuery processing was absolutely necessary. While the eXist XML database provides a full application framework, Drupal was used for the website interface. With the addition of a few plug-in modules, the versatile Drupal content management system simplifies the addition of features such as authentication, authorization, commenting, tagging, caching, and theming to the website. Custom PHP coding in Drupal pulls content from the eXist database, does any necessary processing on that content, and then displays it to the user.
Figure 6. Settings shows the settings which allow viewing the text with or without various editorial markers and additions.
T he choice of Drupal was also in keeping with the characteristics of the Zibaldone; being a project that focuses on a fragmentary style of writing, it was felt that the Zibaldone was a good match for Drupal, since the structure of Drupal is also fragmentary in nature. Content is added to a Drupal system in units of content called nodes, which can then be referenced and connected to other nodes in the system via menu structures, taxonomies, or thematic lists. In this way, units of content representing the various divisions within the manuscript can be created within the system and manipulated in various ways for display. The Gephi network graph visualization tool is used for creating visualizations of the semantic networks encoded in the text.
1. Figure 1 shows a passage on p.1155, paragraph 2, which presents several orders of semantic references in the text, which the website activates as hyperlinks. The passage begins with a reference to a previous passage on p.1121, indicating that this passage is composed as a continuation of p.1121, or in terms of the encoding, it is "subordinate" to it. In the first sentence there is a reference to a paragraph on the previous page, (1154) where Leopardi discusses a related topic and which is encoded as "parallel" semantic association. There is a marginal comment marked by a + sign as in the original manuscript. The marginal comment contains two more references to other pages: the first one refers to a previous page (1113) and is therefore also "parallel", but the second one refers to a page written subsequently, (1166) which indicates that it was added at a later time, and at p.1166 in fact we find a reference back to the margin of p.1155, thus establishing a "reciprocal" reference. The XML encoding of these references allows processing of their interrelations, while the HTML view allows quick navigation through the web-like texture of the Zibaldone.
2. The text is based on the transcription published by Project Manuzio at http://www.liberliber.it and has been partially corrected by consulting the CD-ROM edition by Ballerini-Ceragioli (2009) and the facsimile edition by Peruzzi. (1989-1994).
content | code | process Home
Silvia Stoyanova and Ben Johnston