Ask people how medieval books were written, and they’ll likely describe a scenario something like this: a lone monk sits in the library—the scriptorium—of his monastery, laboriously copying a text onto parchment. Every now and then he’ll stop to warm his fingers over the candle burning on top of his desk. That’s the romantic image of how these beautiful texts took shape. The truth is more dynamic. To perform the painstaking steps needed to make a manuscript, teams of people actually worked together in the cramped quarters of a stationer’s shop.

First, there was the master scribe. This was the owner or head of the shop, who “designed” the manuscript—the codex—by laying out the pages or leaves, called folios. His was no easy task. In addition to deciding whether to create two columns or three of text on each page, the designer also had to determine where to place the miniature paintings or “illuminations.” He had to fit each of these with their caption, and then to place the red text headings called “rubrics” (from Latin rubrica, from the red ochre material in the ink used for these interventions) throughout the manuscript. Finally, the master scribe had to choose marginal decorations for every page. These most often took the form of vines or plants climbing up the side of the folio, then twining luxuriantly across the bottom of the page and on up between the columns of text.

Once the page layout was planned, it was next the master scribe’s job to supervise the team of scribes and artists that would actually carry out the design. These scribes would set to work copying the text—they weren’t shy about adding their own touches or cutting the boring bits. At the same time, artists specializing in manuscript painting would begin executing the miniatures and marginal decorations programmed by the master scribe. Rubricaters would soon come along to fill the blanks left for captions and headings. Finally, it would be time for the binders to fold the parchment into “gatherings” that could be sewn and bound into the manuscript codex or book.

When we look at a medieval manuscript, we, of course, see little of how it was fashioned. Instead, we behold a handsome artifact that reveals no hint of the complex teamwork that brought it into being. So intent on studying the text and images are we that we generally give no more than a passing thought to how it was made. With careful study, however, there invariably comes a moment when we begin to sense the manuscript’s personality. We recognize something new, something different about how its text differs from others telling the same story: we notice variations in decoration, in rubrics and paintings. We see how these decorative characteristics bespeak an originality that we’ve not seen elsewhere. Such moments remind us that the manuscript is an important interpretive agent in its own right. We recognize, in fact, that every manuscript is not simply a passive recipient for text and image, but a collaborator that offers a new way of presenting the work.

What could a scenario like that possibly have in common with digital technology in the 21st century? Ten years ago, I would have said without hesitation: “absolutely nothing.” That’s when a French colleague and I had the idea of creating a digital library of manuscripts of the Romance of the Rose, the most popular vernacular French romance of the Middle Ages. In its time, the Rose was better known than Dante or Chaucer, which explains why there are over 250 surviving manuscripts today, widely scattered in libraries and museums in Europe and North America for the most part. Given the number and geographic dispersion of the manuscripts, few if any scholars had ever seen them all.

That was the problem we hoped to correct: if a medieval work existed in many different versions, then it is historically incorrect to pretend that this literature, like modern fiction, circulated in a single, “authoritative” fixed text. In short, we wanted students and scholars to be able to experience these works much as medieval people could have—in a rich variety of versions and styles. Even though a medieval reader would have been unlikely to own more than one copy himself, he would see the different formats the stationer offered for sale when he went to order his copy. Price could be reduced for less expensively illuminated copies (gouache or wash drawings), with or without abridgements to the text. In any case, the stationer displayed examples of the different treatments.

Why shouldn’t students today have the same opportunity as their medieval forebears to experience the same work in a variety of treatments? Just to see what these manuscripts looked like in “living color,” we thought, would put the zing back into reading romance. Instead of limiting student’s experience to one dimension of the work, a black and white printed text, we wanted to offer an historical encounter with medieval codices. Faced with images of the real thing, scholars and students would be able to see the colorful pictures depicting scenes recounted in the poetry, the headings reminding readers of the subjects discussed in different sections, and the colorful marginal decorations, some of them fanciful or even downright racy. To achieve our goal, it would not do to have simply a handful of examples. We needed as many manuscripts as we could gather. The idea involved creating a digital library that would show the evolution of the Romance of the Rose codices from the late 13th century to the early 16th, when manuscripts still had the prestige of elegance and luxury that printed books could not hope to emulate.

By building our digital library, we reasoned, we could let students and scholars all over the world “take out” as many volumes of the Rose as they wanted at any one time. They could then compare the different versions, marvel at the changes in artistic style of the painted miniatures from the stylized and somewhat stilted figures of the early 14th century to the flowing, gracefully modeled figures of the late 15th and early 16th centuries. Classes in medieval literature would access the Rose library online to compare different versions of the text, witness the evolution of the language, and observe the “editorial” remarks scribes inserted in the text. So many versions of the same work, involving input from so many scribes, artists, and manuscript architects over more than 200 years, we thought, would also provoke reflection on just what it meant to be an author in the Middle Ages. What better a means of demonstrating that the modern concept of "intellectual property" simply had no traction in this period?

This scholarly rationale for the digital library project proved the easy part. What we never guessed were the thousands of technical steps, the multitude of procedures, that would be required before our grandiose conception for a digital manuscript library would begin to take shape. In our naïveté and ignorance, we thought all we would need to do was obtain permission to have the manuscripts digitized—a matter of scanning or photographing them, we believed. Then we would only have to put them up on a Web site. What could be simpler? Although none of us would have dreamed of underestimating the effort involved in producing one of the 250 surviving copies of the Rose, somehow we never imagined a comparable effort for the digital surrogate. Indeed, our view of how digital objects reach our computer screen turned out to be just as witless and romantic as the image of the medieval monk hunched over his parchment with his stylus. We imagined that some “techies” would simply feed images into servers, which would in due course pop up on our laptop screens ready to do our bidding. Painful as it is to admit this level of ignorance on our parts, the confession reflects the huge divide between scholarly users and library providers that existed a decade ago. Today, I am happy to report that this gap has slowly narrowed as digital literacy has taken hold throughout the academic community.

Once my colleague and I took our project to the Digital Research and Curation Center at the Sheridan Libraries at Johns Hopkins, we quickly gathered the magnitude of our task. The discovery has a wonderful historical irony. What became apparent was that with digital technology we were re-creating the medieval stationer’s shop. Just as a team of diversely trained artisans produced each original manuscript, so must a group of specialists, each with a different competence, produce each new digital version. And as with each medieval manuscript, each digital surrogate is a unique object.

Like the master scribe who designed the medieval codex, a digital architect heads the team of programmers and metadata specialists who produce the digital surrogate. As the architect determines the parameters of the surrogate, he refers to an intellectual data model that the group has constructed from information provided by today’s scholarly patrons, the eventual end users of the digital objects. This plan plots the kinds of programming and data models that will make possible the functions that scholars have said they would like as they study the digital surrogate on the computer screen. Some of these functions are necessary because there’s no physical object. Navigation through the manuscript therefore becomes a prime concern. The codex is a book with pages; to go from one to the next, you must turn the page. Easy enough when you have the actual manuscript in your hands, but to turn the pages of a digital object requires a program that allows for single or multiple folios (pages) to be turned forward and backward. Since Rose manuscripts can have 250 or more folios, navigating one folio at a time can be both time-consuming and frustrating.

But page turning is not the only reason why it’s necessary to move through surrogate manuscripts. Scholars today want to compare treatments of lines, passages, scenes, or the sequence of illuminations found in different manuscripts of a work. That practice alone illustrates how this new access to manuscripts made possible by digital surrogates has changed the way they work. Only one generation ago, literary scholars carried out their research and interpretation using a modern critical edition of a given work. Rather than having multiple manuscript versions at their fingertips, they studied a printed text prepared by an editor who chose one particular manuscript to represent the whole corpus. In cases where the editor felt the selected manuscript to be defective, he would substitute words, lines, or sometimes even entire passages from other manuscripts. And, naturally, all of the elements of the medieval artifact not strictly pertinent to the text—particularly the images—would be omitted.

This means that no text of the Rose in any critical edition corresponds exactly to any single actual manuscript. At a time when manuscripts were all but inaccessible, the availability of those editions satisfied scholars’ needs. Today, it is ironic that the inaccessible versions—the original manuscripts themselves—have become more available than printed editions ever were. Still, working with multiple versions of manuscripts poses a problem of referencing lines and passages that critical editions solved by the use of line numbers.

Scribes did not number lines in manuscripts. As a result, cross-referencing lines, passages, or scenes between manuscripts is exceedingly laborious. Since comparative study of Rose manuscripts is one of the main scholarly rationales for the digital library, the Rose team needed to devise a mechanism for mapping scenes and then incorporating this protocol into the intellectual data model. That meant they needed to formulate a way of referring to episodes in the work without using the line numbers of critical editions. What emerged is an entirely new convention for segmental analysis of the romance. The protocol for scene mapping assigns symbols to each section and sub-section. This mapping of the work allows a grid to be established for each manuscript that shows all the scenes contained in that manuscript. It is thus possible to superimpose the grids for the manuscripts in the corpus in order to plot variances between them. Not only is it possible to determine scenes or sections that have been omitted (or added) in given manuscripts, but we can now determine variations in scene ordering. Finally, the scene-mapping protocol allows readers to locate with accuracy lines, passages, and episodes they want to compare among the 150-odd manuscripts planned for the site.

To appreciate the complexity of creating a digital surrogate of a Rose manuscript, we need to recall the elements of the original codex design: the rubrics, the paintings and their captions, the decoration and other marginal annotations. All of these have potentially important information for modern readers, so the intellectual data model must find ways of including them. This might seem negligible until we realize that rubrics and captions often take the form of what we might call editorial comments. They fall into the category of what we call metadata today. As such, they play an important role in informing us of the attitudes of the medieval scribes about the poem they were transcribing. Scribes did not hesitate to take sides with characters at times, or even to disagree with the original author. At such moments, scribes might not simply insert a heading-as-comment, but add some lines to the poem itself. These interpolations need to be marked for ease of reference.

These are but a few of the functions that must be incorporated into the intellectual data model for the digital surrogate of a manuscript. Even so cursory a list demonstrates the complexity of the digital architect’s job. Besides handling the overall design, today’s architect must also incorporate requests from scholarly users whose input determines the significant parameters of the data model. Once formulated, of course, the paradigm must be implemented by the rest of the team. Just as artists choreographed the virtual space of the parchment into a vibrant microcosm of competing media—images, text, captions, rubrics, marginal decorations, etc.—so the graphic designer determines how pages will appear on the screen, programmers write the code and convert different elements of the raw data into “objects” attached to the platform (in this case, Fedora), and translators convert data from different formats into the one currently used.

This labor-intensive work must be performed separately and often differently for each manuscript. This is particularly true in the case of adding searchable functionality to the surrogate. Every item that scholars have said they need to search requires a "tag" at the appropriate place in the image data. Some of the items, such as the captions of painted miniatures, can be tagged fairly quickly, since they do not occur on every folio. Other items, such as proper names, place names, rubrics, scene mapping, and distinctive items of historical or artistic interest in a given manuscript, are painstaking and time-consuming to tag.

That observation evokes a paradox. Producing digital surrogates of medieval codices is even more labor-intensive than the production of the original manuscript. For all the analogies that we can draw between the medieval and the digital scriptorium, it turns out that there are also real differences of scale and outcome. To make a digital library of Rose manuscripts requires more people and far greater material resources than medieval scribes would ever have imagined. Imagine the awe they might feel upon hearing that people on continents not yet discovered in their day and in countries then unknown would transform their work into a format that could be accessed every hour of every day by anyone with Internet access (whatever that is). What’s especially moving is the fact that the digital experts whose devotion and enthusiasm make this modern-day project a reality not only do not know medieval French, but had never heard of the Romance of the Rose before the project began.

And finally, what about the question of finances? Given the resources required for the project, skeptics might well ask: “Is it worth it?” That is a question we had to address when we approached funders for the project. It is best answered, I believe, by looking at the historical context. From the beginning Rose manuscripts were expensive. The work is long and frequently lavishly illuminated. In the Middle Ages, miniature paintings could be costly, especially if artists employed stone-ground pigments and gold leaf. When produced for aristocrats or rulers, Rose manuscripts were sumptuous objects. As time passed, these manuscripts continued to appeal to collectors who paid ever greater prices for them. Today, even a relatively ordinary Rose codex will fetch close to $1,000,000. A truly elegant, richly illuminated manuscript can be valued at three or four times that figure.

But besides being precious objects, these codices are also historical artifacts of great value for teaching and research about the Middle Ages. Often, however, the interests of repositories concerned with the preservation of their manuscripts conflict with users seeking to consult them. Private owners can simply refuse scholars access to their treasures. Happily, in our century, the vast majority of Rose codices reside in museums and libraries. Those public institutions of course also have an obligation to preserve their artifacts. And with the growth of scholarly interest in manuscripts they face a difficult dilemma: how to reconcile the need of scholars for intensive access to manuscripts while preserving these fragile artifacts from the deterioration caused by frequent use?

The problem is hardly a new one. As long ago as the early 20th century, with the advent of microfilm, repositories set about recording their manuscripts in that medium. The quality, however, could be poor to begin with, and became worse with frequent copying. Moreover, bad black and white copies of richly illuminated folios did not allow scholars to analyze text and image appropriately, still less convey to their classes the beauty of these objects. Digitization resolves many of these problems. It also offers the advantage of color and of text and image manipulation through the zoom, pan, and rotate function.

Digitization also reconciles the conflicting obligations of manuscript preservation with the need for access to vulnerable artifacts. For the first time in history, we have a means for preserving objects that simultaneously diffuses them to new and larger audiences. More intriguingly, this new world of digitization accomplishes all this by creating a collaborative community among libraries and museums, on the one hand, and the scholarly and public sectors, on the other.

So is digitizing manuscripts worth the cost? Let’s look at the numbers. As of this year the entire expense of creating the digital library of the Romance of the Rose—some 150 manuscripts to date—has yet to exceed the price a single luxury manuscript would command at auction. What’s more, the digital library will enable repositories to realize significant economies in preservation costs while allowing them to bring in revenue from the sale of images on the site. Meanwhile, at the same time that scholars conceive new ways of thinking about medieval works, thousands of their students around the world will be able to see and use beautiful books once reserved for kings and queens, dukes and duchesses.

And this is only the beginning. The scholarly advances and new insights now possible with the digital library of over 150 original manuscripts of the Romance of the Rose are unfathomable. What’s clear is that the deeper understanding of the past this resource makes possible opens exciting vistas for the future of medieval scholarship.

Stephen G. Nichols, James M. Beall Professor of French and Humanities at Johns Hopkins University, specializes in medieval literature, art, and history. One of his books, Romanesque Signs: Early Medieval Narrative and Iconography, received the Modern Language Association's James Russell Lowell Prize for an outstanding work by an MLA author in 1984. In 1991, The New Philology, conceived and edited by Nichols for the Medieval Academy of America, was honored by the Council of Editors of Learned Journals. In 1992, the University of Geneva conferred on him the title of Docteur ès Lettres, honoris causa, while the French Minister of Culture made him Chevalier de l'Ordre des Arts et Lettres in 1999, and Officier in 2007.

Chair of The Johns Hopkins Department of German and Romance Languages and Literatures, he is also a Fellow of the Medieval Academy of America, and a Senior Fellow of the School of Criticism and Theory, which he also directed from 1995-2001. Recent publications include: Re-Thinking the Medieval Senses (2008), L'Alterité du Moyen Age (2003), Medievalism and the Modernist Temper (1996), The New Medievalism (1991), and Mimesis: From Mirror to Method, Augustine to Descartes (1983, reprinted, 2004). Current projects include: The Enigma and Exasperation of Laughter, Seeing Voices: On Reading Troubadour Lyric, and Building History: the Politics of Medievalism in Restoration France.