Rachel Kathryn Onuf

Hypercontextualizing the Work of William Blake

The work of William Blake (1757-1827), particularly the illuminated manuscripts, has raised interesting issues of interpretation for both art historians and literary critics since its creation. The illuminated manuscripts demand interdisciplinary study and yet have rarely received any, due to cultural biases and technological limitations. In many ways the capabilities of new technologies and William Blake's "composite art" are well-suited to each other, as numerous web sites devoted to Blake attest (Chayes, p.442). Hypertext and digitization combine with current trends toward interdisciplinary approaches in academia to create a climate uniquely qualified to exhibit and interpret Blake's work.

Literate people tend to privilege the written over the visual and in the past, Blake's work has been read and studied primarily as text. During the Blake revival of this century many illuminated manuscripts, including the Songs of Innocence and Experience, were known only as texts. Visual elements were completely ignored and published editions provided his handwritten text in standardized typeface. Thus Blake was reduced to a book, which is how/what we are accustomed to reading. In addition to cutting off any meaning contained in the images and in the interrelationship of word and image, this nullified the differences among editions and set up one as the "standard", obscuring the uniqueness of each manuscript. I refer to these illuminated works as "manuscripts" rather than the more commonly used "books" in order to emphasize this uniqueness, a quality crucial to issues of interpretation.

Next came facsimile editions, most notably those of the William Blake Trust and Trianon Press, that "recreated" the work. Although a vast improvement over the text-only versions, these were also dangerous, for they did not necessarily identify their relationship to the original or what printing processes they used, making it easy for readers to forget that they were seeing a copy. Pre-digitization, a copy was never an exact copy, but without access to the original, the reader had no way of identifying difference. Improvements in color reproduction and growing awareness of the importance of making process and relationship to original explicit have informed the most current edition of Blake's illuminated manuscripts, a six volume set published collaboratively by Princeton University Press and the William Blake Trust .

These volumes, two of which were published in 1995, "reproduce full size in color what the series editors regard as the finest extant copies of each of Blake's now-scarce illuminated books" (Baker, p.1). And yet they are limited. Not to mention expensive. They are still discrete volumes; choices were made, compromises reached, the cover closed. Not all extant editions of the various books can feasibly be reproduced. More than one scholar has realized the impossibility of a "definitive" edition of Blake and has suggested an alternative; an electronic Blake archive. A physical archive is not possible, nor is it what Stephen Carr and Jerome McGann are suggesting. Instead, they envision an electronic archive published on the Internet, allowing for maximum accessibility and useability of the material. Morris Eaves, Robert Essick and Joseph Viscomi are currently working on just such a Blake Archive, with the support of the Getty Grant Fund and The Institute for Advanced Technology in the Humanities at The University of Virginia. There are several other more modest sites aimed toward intellectual study as well, including The Blake Multimedia Project, The Digital Blake Project, and The William Blake Page.

These smaller sites all include scanned images of varying degrees of quality. Two of the sites are hypermedia archives, allowing users to "interact" with the material. People who browse the World Wide Web recreationally are accustomed to the "click and go" method of reading and may not realize the conceptual impact hypertext is having on scholarly research practices. Changing research methods in turn affect how the source material is "read", coloring the interpretation. Before, a Blake researcher had to either limit her source material to a particular archive's holdings, travel long distances and rely on her memory or use reproductions. Trends in academic scholarship make the last unacceptable for most research purposes. Examinations of Blake's methods and applications of post-structuralist theory, reader response theory and cognitive/perceptive models now proliferate in Blake studies. These all require close textual and intratextual readings of word and image and the bits in between. Comparative readings rather than studies of isolated texts are now more common.

These trends in diversifying the subject matter to be studied and the interpretive models with which they are studied mirror the ever-increasing capabilites of technology. By 1998 The Blake Archive will contain 3,000 images, 2/3 of those from 55 editions of 19 illuminated manuscripts. The grant ends there but that does not mean the archive is completed. Unlike the recent "definitive" six volume set, material and links to new relevant sites could be added to the Archive. Even when work on what constitutes the Blake Archive levels off, work with the archive is potentially limitless. The same could be said of a book, but the hypermedia archive includes much more supplementary information and claims to come closer to the quality of the originals, thereby contextualizing Blake and his work and encouraging "a deeper, more responsible understanding of his aims and methods, which have been regularly misunderstood and misrepresented" (The Blake Archive).

Of course images on the web are reproductions and purists may certainly object to using them for research purposes without consulting the originals. The people constructing The Blake Archive are aware of the critical importance of capturing high-quality images: "all materials will be scanned either from new transparencies, or directly from originals whenever possible, and monitored carefully for accuracy" (ibid). The images are currently stored on a hyperdense Kodak Photo-CD and the project heads are optimistic about Kodak's continued commitment to advancing Photo-CD technology.

The many added benefits of working with digitized images make the issue of using digital copies for research purposes fade, if not disappear. Simply put, digitized images can be manipulated in ways that the originals cannot. Enlargement, juxaposition and computer enhancement are all possible. If the most important concern for an image database of this kind is the quality of the image, the second is, arguably, the search engine. The 3,000 images are useless without a sophisticated search engine. The Blake Archive claims they will have "an ingenious image annotation tool" called Inote 3.0 that will

construct a kind of visual concordance by creating pathways within and across different media. The user may highlight details within an image, attach hypertextual annotations, link the details to related details in other images (and texts) for comparison, and present the annotated images in an open-systems, networked environment (ibid).

The specifications for Inote 3.0 state, using slightly "harder" language, that it "will permit multiple overlays, three types of free-hand annotation (point, circle, rectangle), and two types of automatic overlay generation (including automatic line-numbering of page-images)" The Blake Archive).

Initially surprised that the project would not be implementing query-by-image-content techniques, I consulted a colleague. "Image content searching just doesn't work very well" he said with finality (Mark Handel, conversation, 10/25/95). After reading Vlad Wielbut's paper I agree that The Blake Archive may not be the type of database that would benefit from the ability to perform query-by-image content searches, whether they work well or not. Users would probably know what they are looking for or at least know enough about Blake and his work to be able to use Inote with success.

Typical searches would not be for those elements that Vlad Wielbut identifies; i.e color and shape. The one query-by-image-content element that might be useful would be line. In addition to focusing on the relationships between images and texts, Blake scholars scrutinize his use of line to blur the distinction between the two. Erza Pound identified three levels of expression in texts - language visible, auditional, and intellectual. Visible language operates through gesture and (type)script. In Blake's work, this line often escapes from the text - and from the image as well, twirling around between the lines, marking and opening up other relationships and meanings within the page. As Carol Bigwood puts it, the nuances of Blake's handwriting make us notice the words themselves: "[m]oreover, because his words often burst into flowers, leaves, tendrils and various squiggles, the words are seen as quasi-drawings that clearly show up their visual presence through the gesturally drawn lines" (Bigwood, p.310). This description celebrates the decentered-ness of a Blakean "text" as a fruitful site for critical intervention.

Blake's unique practices and the resulting illuminated manuscripts could be thought of as a prototype of hypertext. His gestural line disrupts attempts to read a poem straight through, instead leading the reader on a "sinuous perceptual pathway which will be formed uniquely by every person and at every reading" (ibid, p.309). In my understanding, Inote's free-hand annotation would allow the entire illuminated page to be searchable, which alleviates my concern that the material between the traditional lines of text would be untaggable. If that is true, William Blake's work and our study of it can only be enriched by the creation and utilization of a hypermedia archive published on the Internet.


Baker, Kenneth. William Blake's Ingenuity as a Poet and Artist.

Besser, Howard & Jennifer Trant. Introduction to Imaging: Issues in Constructing an Image Database.

Bigwood, Carol. Seeing Blake's Illuminated Texts. Journal of Aesthetics and Art Criticism, v.49 (Fall 1991); pp.307-15.

Carr, Stephen L. Blake's Illuminated Printing in the Age of Post-Mechanical Reproduction: Variation/Intention/Textual Editing. Unpublished manuscript of talk given November, 1994.

Chayes, Irene. Picture and Page, Reader and Viewer in Blake's 'Night Thoughts' Illustrations. Studies in Romanticism, v.30 (Fall 1991); pp.439-71.

McGann, Jerome. The Rationale of Hypertext.

Wielbut, Vlad. Computer, give me [inserted image] and [inserted image]!

Wright, John. William Blake's Urizen Plate Designs: A graphic essay in facsimile of his copper plates for 'The First Book of Urizen'. Ann Arbor: 1980.