Refining Our Notions of What (Digital) Images Really Are

Matthew G. Kirschenbaum, Organizer and Chair
Department of English
Institute for Advanced Technology in the Humanities
University of Virginia

Johanna Drucker
Jerome J. McGann
Joseph Viscomi
Worthy Martin


This panel brings together authorities from the fields of art history, literary studies, textual editing, and computer science -- all of whom also command significant applied knowledge of printmaking and the graphic arts, or electronic production and editing, or both -- to discuss the aesthetic, ontological, and computational nature of digital images (within the context of much broader traditions of visual representation). Panelists will address both technical issues pertaining to the creation, manipulation, and dissemination of image data, as well as the current state of the critical/theoretical models undergirding our understanding of digital images, and do so with particular reference to the role of images in electronic libraries, editions, and archives.

Full Rationale

The panel has two related objectives:

  1. To advance its audience's thinking about digital images as structured data by reporting on recent technical developments in such areas as file formats and data standards (PNG, JPEG 2000), image editing, image annotation, computer visualization and modeling, and image searching and retrieval based on pattern-recognition technologies;

  2. To advance the humanities computing community's current understanding of digital images as modes of visual representation, and, in conjunction with the topics listed in item number one above, to scrutinize the prevailing theoretical and philosophical assumptions behind our thinking about digital images.

As may be obvious already, the panel is motivated by its participants' shared desire to subject digital images to the same degree of intellectual scrutiny that has been brought to bear in recent discussions of electronic text encoding and text-based editing -- exemplified by the frequently-referenced series of papers by Steve DeRose, Allen Renear, et al. from which we derive the session's title. The panel is predicated on the assumption that the humanities computing community at large does not yet engage with digital images and imaging technologies at anywhere near the same level of sophistication as electronic text, not least because the underlying computational structure of images is so radically different from that of machine-readable text. These computational differences in turn reflect and recapitulate certain elemental differences in the epistemology and ontology of images and text. (Consider markup languages in this regard: markup languages allow one to describe text with more text, but there is no comparable procedure for describing images -- one cannot "markup" an image with another image.) We do not propose to address the manifold epistemological differences between word and image in any comprehensive way, but rather to suggest a variety of computational approaches to image data while discriminating among a range of editorial strategies for incorporating images into digital library collections, electronic editions, and electronic archives.

In a recent article Julia Flanders maintains that "[t]he role that images currently play in electronic editions seems to move between decoration, scholarly substantiation, and bravura display" (309). This has been largely true thus far, but new tools and techniques have lately begun to allow us to treat images with far greater subtlety. Moreover, even the seemingly straightforward notion of a digital "facsimile" involves basic editorial decisions which too often go unrecognized: should the digital images be color-corrected, and if so what measure of chromatic fidelity should be employed? Should the facsimile be presented on the screen at its true physical dimensions, a prime consideration in a field such as art history? (The William Blake Archive uses Java technology to resize images on the fly.)

But the heart of the issue, as Flanders and others have recognized, is the need to begin treating images as structured data rather than as simple supplements to machine-readable text. Various strategies to accomplish this, ranging from descriptive SGML-based image annotation to search technologies employing sophisticated pattern-matching logics are gradually coming available, and humanists now need to begin thinking through the implications of images as structured data, in the same way that recent years have produced vigorous discussions of text-based electronic editing and text-encoding. The panel will suggest starting points for such a discussion, as well as provide examples of projects that have taken early steps in this direction.

The panel will also consider forms of visual knowledge that lie beyond the familiar horizon of facsimile reproduction. As Jerome McGann has recently demonstrated, for example, the algorithms and filters available in common desktop packages such as Adobe Photoshop have the potential to expose formal elements of an image's structural composition through processes of computational deformation. Or else consider modeling and visualization tools, which have a long history of applications in the hard sciences: a visualization of search results or the narrative structure of a text is no less an "image" than a facsimile of a manuscript page, and the increasing ease with which digital tools allow us to generate such visualizations suggests that our thinking about visual materials in electronic editions and digital libraries will soon evolve to encompass this new, non-referential class of images.

Panelists (In Order of Presentation)

Johanna Drucker has been printing artists' books using letterpress and offset production techniques since the 1970s. She holds a doctorate from the Berkeley Visual Studies program and has published widely as an art historian and graphic design critic. She has held appointments at Columbia and Yale, and currently directs a visual design program at SUNY Purchase. She will become the Robertson Professor of Media Studies at the University of Virginia in the fall of 1999. Drucker has recently been conducting extensive investigative work on "the ontology of the digital image," which will form the basis for her remarks at this panel. Email: <>

Jerome J. McGann, University Professor at the University of Virginia and a Fellow of the Institute for Advanced Technology in the Humanities, has long been an advocate of "image-based" electronic editing, and his work on the Rossetti Archive has involved extensive investigation of methods for annotating and encoding images as structured data. More recently, McGann has been engaged in a series of "deformative" exercises, using the filters in common image editing software to expose the formal properties of digital images. Email: <>.

Joseph Viscomi is Professor of English at the University of North Carolina, Chapel Hill and a Fellow of the Institute for Advanced Technology in the Humanities. He worked as a curator and graphic artist before reconstructing and reproducing -- accurately for the first time -- Blake's illuminated printing techniques. As an editor of the electronic William Blake Archive, Viscomi's work focuses on color correction and image editing. He will address the twin chimeras of radical subjectivity and false positivism when engaged in image editing, and will also discuss image editing vis-a-vis traditional editorial theory. Email: <>.

Worthy Martin is Associate Professor of Computer Science at the University of Virginia and Technical Director of the Institute for Advanced Technology in the Humanities. His fields of expertise include machine vision, pattern recognition, and image databases, and he has published widely in these areas. Martin will address the question of digital images from a computational perspective, focusing on how computers "see" and process pictorial content. Email: <>.

Matthew G. Kirschenbaum (Organizer and Chair), is completing his dissertation in the Department of English at the University of Virginia. He is also the Managing Editor of the William Blake Archive. Kirschenbaum will become Assistant Professor of English at the University of Kentucky in the fall of 1999. In this session, he will briefly assess the current range of software for comparative and analytical operations on structured image data as well as outlining provisional specs for a new image-based software tool entitled LOOKSEE. Email: <>.


The panel is allotted 90 minutes. After a short set of introductory remarks from the panel chair (5 minutes), each member of the panel will be asked to speak for 10-12 minutes, developing one or two specific talking points from their particular background and expertise (total time for five presentations: about 55 minutes). With these initial presentations concluded, panelists will be asked to briefly respond to one another's remarks (clearly time will not permit every member of the panel to respond to every other member, and so the objective here will be to foreground a few major points of agreement and dissent). The remaining 20-25 minutes of time will be given over to open discussion with members of the audience.


DeRose, Steve, et al. "What is Text Really?" Journal of Computing in Higher Education, 1(2), 3-26.

Flanders, Julia. "Trusting the Electronic Edition." Computers and the Humanities 31.4 (1998): 301-310.

Kirschenbaum, Matthew G. "Documenting Digital Images: Textual Meta-Data at the Blake Archive." The Electronic Library 16.4 (August 1998): 239-41.

McGann, Jerome. "Imagining What You Don't Know: The Theoretical Goals of the Rossetti Archive." General Publications, Institute for Advanced Technology in the Humanities <>.

---. "The Rossetti Archive and Image-Based Electronic Editing." The Literary Text in the Digital Age. Ed. Richard Finneran. Ann Arbor: The University of Michigan Press, 1996.

Renear, Allen. "Out of Praxis: Three (Meta)Theories of Text." Electronic Text: Investigations in Method and Theory. Ed. Kathryn Sutherland. Oxford: Clarendon Press, 1997.

Renear, Allen, et al. "Refining Our Notion of What Text Really Is: The Problem of Overlapping Hierarchies." In Research in Humanities Computing. Ed. Nancy Ide and Susan Hockey. Oxford: Oxford University Press, 1996.

Note: This short list is intended to portray neither the broad range of technical literature on digital images and imaging technologies, nor critical/theoretical work on visual representation, but rather to suggest the panel's positioning in relation to an ongoing and parallel discussion of text-encoding and text-based electronic editing.