Electronic Textual Editing: Digital Facsimiles in Editing [Kevin Kiernan]
According to the Guidelines for Editors of Scholarly Editions and Digital Facsimiles of the Committee on Scholarly Editions (CSE), the first question an editor should consider in choosing the medium of publication is: ‘Is the source material itself printed, electronic, or both?’ For a great many texts, including medieval manuscripts, the answer is Neither. One of the few certainties we have about authorial intentions from the Middle Ages is that neither authors nor scribes (who may sometimes be the same person) ever intended their work for print or PC. These writers none the less supplied their texts with meaningful text-encoding, which modern editors, because of the limitations of print and the different conventions of modern literacy, have routinely ignored or unavoidably misrepresented in modern editions. Editors today can most effectively recover these significant lost features through image-based scholarly editions.
It is worth distinguishing between plain old digital facsimiles and image-based scholarly editions. The primary purpose of a facsimile in editing is to provide editors and textual scholars with a reliable version of the source document or documents. Editors of all theoretical dispensations must ultimately base their editions, whether they want to or not, on source documents, which for unique manuscripts are seldom readily available for sustained reference. High-resolution, full-color, digital images, particularly when acquired directly from the manuscripts themselves with sophisticated electronic cameras, are demonstrably more revealing, easier to access, and far more productive for further image-processing, than microfilm or printed facsimiles.
Like digital facsimiles, image-based scholarly editions of Old English manuscripts provide high-quality facsimiles and digital restorations for editors and textual scholars, but they also serve the broader purpose of helping modern readers negotiate a culturally strange text in its original form with all its foreign formats, including its codicology, paleography, scribal abbreviations, shifting orthography, unexpected word-divisions, unknown word-hoard, missing metrical layout, and lack of any helpful punctuation or capitalization. Modern print editions of medieval texts have always approached these problems by quietly modernizing the text, replacing unfamiliar letters and ligatures with rough equivalents from the Roman alphabet, silently expanding scribal abbreviations, normalizing spellings, imposing word-boundaries, emending unusual or unexpected words, laying out explicit, numbered, lines of verse, and adding Modern English punctuation.
This editorial process, excluding the major editorial interventions of emendations and conjectural restorations, may itself easily involve thousands of editorial changes in a long text. Although editors do not treat them as real editorial emendations, these pervasive changes from script to print should in fact be viewed as radical modernizing translations of source documents, not the usual stuff of scholarly editions. It is comparable to an editor of a modern text removing all punctuation, dividing words into syllables, providing ancient spellings, and displaying poetry without lineation. Before the advent of digital technology, editors of medieval texts were virtually forced to make these translations of their sources, if they wanted modern readers to understand the texts. The role of the editor of medieval texts has been that of a solitary scholar, imbued with great authority by rare knowledge of primary sources, transferring a text from unique script to generic print for the less erudite masses. In the course of this work, modern editors have obscured manuscript records to an extraordinary degree. The resurgence in interest in medieval manuscripts derives from the growing recognition that modern print editions have not accurately represented them.
These editors of course included codicology and paleography of the manuscripts as part of the apparatus of their editions. But print is an extremely inefficient and inadequate means for representing manuscript texts. Codicological overviews and paleographical details in support of textual notes are all poorly portrayed without illustrations, even though these aspects are almost always treated without them. Codicological information is usually presented in extremely austere ways. Ff. ii, 1-38, 4-66, 7-108, 10 wants 1 for example, means that the editor discovered by collating the leaves of a manuscript that there were ten gatherings and two prefixed leaves; the first three gatherings were made up of four sheets of eight leaves; the next two gathering had three sheets or six leaves; and gatherings 7 through 10 again were 4-sheet quires of eight leaves, but the first leaf was missing from the tenth quire. With a complete digital facsimile of the source material there would be no reason to present the collation in shorthand so schematized as to be useless to all but experts. Print editions are increasingly providing diagrams of gatherings, but even with diagrams it is difficult to conceptualize the make-up of the book and impossible to analyze and assess the possible significance of any structural peculiarities. With access to all the images, an editor could present the collations in diagrams, indicating hair and flesh arrangements, sheet and quire signatures, and information on rulings, and easily link the diagrams to comprehensive image collations, from thumbnails to full-resolution views of conjugate folios.
While some people continue to think of electronic texts as exclusive of images, the fact is that digital images of manuscripts are electronic texts, as well. The most compelling scholarly editions of the future will make full use of markup schemes such as XML (or its TEI manifestation), but not without extensive integration of images. The TEI has begun moving in this direction, as its latest guidelines indicate: ‘Work on areas still not satisfactorily covered in this manual will continue, and resulting recommendations will be issued as supplements to the published Guidelines. Work is expected to continue in ... [among other areas] manuscript analysis and physical description of text.’ Text-encoding can include coordinates in images to link text and images, so that search facilities will not only return results of text searches but also hypertextual links to corresponding images. It seems likely, however, that computer science will contribute powerful additional methods for searching manuscript images, even though, with our current experience, optical character recognition (OCR) does not appear to be a fruitful approach with scribal handwriting.
The CSE Guidelines assert that editors establish reliability by ‘explicitness and consistency with respect to methods, accuracy with respect to texts, adequacy and appropriateness with respect to documenting editorial principles and practice.’ Any editorial theory should be able to justify its methods in the face of its source, no matter how explicit or consistent the editorial departures from its sources might be. Textual notes should accordingly incorporate thumbnails of the documentary sources of error and expand to the full manuscript context of the error. In addition to searchable tagging for editorial emendations, conjectural restorations by the editor, and restorations from independent manuscripts, transcripts, and collations, image-based electronic editions should also provide searchable encoding and a facility for linking to corresponding images for such manuscript features as scribal abbreviations, accented letters, additions, alterations, letters covered or partly covered by bindings, damaged letters, deletions and erasures, readings enhanced by special lighting, faded readings, and technological restorations. For some editing projects high-resolution, digital facsimiles, acquired from a range of special lighting techniques, are indispensable for what the CSE calls the ‘basic task’ of scholarly editing, ‘to present a reliable text.’ The reliability of editors of the future will not depend so much on historical or interpretive introductions bound to a single edition, but on their ability to show the relationship between the edition and its documentary sources and to bring their readers to an ever-expanding digital library of supporting ancillary material.
Image-based electronic editions with thoroughly integrated text and images will thus have an advantage over traditional printed editions and facsimiles by providing the reader greatly facilitated access to an edited, readable text, including easy structured search of the apparatus, as well as full, linked, visual representations of all the myriad uses of the source documents. Manuscripts often unabashedly highlight features that are completely ignored or deliberately submerged in print editions. One of the most exciting and interesting aspects of image-based electronic editions is that unique resources once restricted to the solitary scholar are opened to the scrutiny of all. The best electronic editions will be those that for one reason or another improve on traditional print editions. A major advantage of image-based electronic editions over print editions is that the display of the text is transparent. The accuracy of the materials is either apparent or correctable. An image-based scholarly edition provides a very effective means of proof-reading and an ever-present failsafe for locating and correcting any residual errors.
Editors of image-based electronic editions must become newly sensitive to the power of sources to reveal the different state of literacy a medieval manuscript evinces. We must presume, for example, that the communities that made and used our surviving Old English manuscripts shared a very practical, non-magical, knowledge of runes as both letters and as shorthand for words, and that they had the extraordinary ability to read verse without the kind of formatting modern readers require even for accentual-syllabic rhyming verse. By making manuscript images always available, editors can help transmit insights into paleography, another aspect of medieval literacy, which the reduction to typescript strips away. Focused, comprehensive access to scribal letterforms might be mediated through the glossary, by linking all head-letters to salient examples in the manuscript. However it is accomplished, examples of all letterforms should be described and illustrated, and any letter used to support an editorial emendation (for instance, typical confusion between insular c and t) should be specifically linked to its manuscript context. XML markup is good at distinguishing different letterforms, such as insular, caroline, and uncial s, for searching of text, but to be of real value, the editor and the researcher should be able to link any search results to the specific instances in the manuscript images.
If the manuscript is present through images it will both demand and enable more explicit justification than is often given for significantly different representations in the edition. The most effective way to describe the authoritative or significant texts, as well as to present the punctuation, capitalization, and spelling is by displaying them and providing the means to understand them. For example, the strange word-division of manuscripts will not be difficult to understand if the word-elements are mapped together, glossed, and clearly linked as informative illustrations of the normal word-divisions of the edition. In the same way the image-based scholarly edition subsumes the purpose of a diplomatic edition and removes the fruitless frustration of trying to preserve the exact layout, illumination, and physical appearance of a manuscript in print form.
The apparatus, a Help function always a click away, should be comprehensive, incorporating such concepts as contents, preface, acknowledgements, introduction, footnotes and endnotes (whether explanatory or textual), glossary, appendices, index, and bibliography. Editors of image-based scholarly editions must ask, ‘What kind of editorial mediation is required for a modern reader to use the manuscript images with understanding?’ To take full advantage of encoding strategies to inform the source with meaning, image and text must be thoroughly integrated. The most successful ones will always present image and text, manuscript and transcript or edition, in tandem. A menu bar is itself a digital image to help furnish comprehensive use of the images of the manuscript. On any given page, the menu bar containing the full apparatus should link to any topic relating to problems on that page. If there is evidence of corruption, or for that matter, evidence of reliability, such as intelligent scribal corrections and erasures, the editor should make it easy to link to it. A good principle is to link any and all tagged text to the source of information. Just as print editions gmove from manuscript to typescript, an image-based electronic edition must move from typescript back to manuscript.
Providing an effective, easily understood, user interface (or GUI) is always an important issue in an electronic edition. Users of print editions generally know where in the edition to find introductory discussions, including descriptions of manuscripts, textual and explanatory notes, glossaries, and so on, although these important resources are not always easily or quickly accessible in print editions. In many respects, the best possible user interface is the image of the manuscript page, provided it is integrated with the encoded text. What better illustration can there be than the physical text that led to editorial intervention? By mapping links on the image, editors enable users to click on areas of interest to display textual or explanatory notes. For example, by clicking a quire signature on an image of a folio, the user could have access to the quire it organizes, from a diagram giving the hair and flesh arrangement of the vellum leaves, to thumbnail gatherings, to full resolution images of the folios in question. To investigate an emendation, the user should be able to click the editorial change and get to both the textual note, including a thumbnail image illustrating the source of the change, and a link to the full image providing the context for the presumed error. The images thus become continuing resources for deeper investigation and possibly new theories.
There is no compelling reason to place a Preface or an Introduction at the beginning of an electronic edition, although links to them should be always available for users to examine whenever they wish. As their names indicate, prefaces and introductions, endnotes and appendices, are all bound by the restrictive sequential order of print editions. Editors of image-based electronic editions can effectively integrate these resources into the fabric of the edition. In an electronic edition, it is much more useful to have an always available guide or Help or Apparatus that incorporates any prefatory or introductory matter, because their important nature will require continuing reference while using an edition. An electronic apparatus should not pointlessly mimic a print format, forcing a reader to scroll through it screen by screen. It is annoying and unnecessary to scroll screen after screen of dense, scholarly discourse. Editors must intelligently employ the hypertextual environment and the supporting images they have at their disposal to guide and engage readers in efficient ways through the pertinent scholarship. There is no reason not to include bibliographical items to the links for both textual and explanatory notes, and the bibliographical items might well be linked also to online articles and image archives. An electronic edition should not simply reproduce a traditional print edition, which normally does not have free and wide access to supporting illustrations.
In the past several years it has become imperative to bring the research agendas of the humanities and computer science together to advance our abilities for representing, editing, and accessing our cultural records. What is urgently needed to take full advantage of encoding schemes of scholarly editions is the seamless integration of text and image, which will ultimately serve the interests of not only image-based scholarly editions but also the broader cultural world. The Computer Science research community is very interested in image search and techniques for developing efficient and accurate searching, while the humanities computing research community is similarly absorbed in textual encoding. Neither area is likely to make real progress, however, until their complementary research agendas converge. There is an attendant need to develop interactive tools combining search strategies for images and encoded text, which would permit viewers to find an image instantly and then query any part of it to bring up informing analyses by specialists as well as technological breakthroughs, such as 3D imaging, the precise understanding of size, scale, coloration, and texture, texture-analysis, and other complementary insights made available by the research of computer and imaging scientists. Through cooperative cross-disciplinary research we can revolutionize the editing of texts, providing editors with robust tools to create, maintain, upgrade, and preserve image-based electronic editions, and readers to see what we read with new insight and enjoyment.