Publishing Theoretical Apparatus
This all raises an important question: what kind of digital format should be used? There are, obviously, many options for how to publish a text online. Beyond the more aesthetic considerations of typeface, visual representations of navigation (e.g. the common webpage “scroll” format vs. the skeuomorphic “flipbook” style popularized by Kindle), and file type (e.g. PDF / OCR-analyzed PDF, TEI-encoded XML rendered through XSLT, HTML with or without CSS), a text edited and published electronically for academic and pedagogical purposes also must consider the kinds of editorial apparatus which will be embedded within the coded text. How these apparatus function and what they contain clearly have a foundational effect on how the text is read (or at least how it can be read) and thereby how meaning is built through the text.
In attempt to determine the most effective method of hypertext reading for students, I will construct four separate versions of the “Aeolus” chapter from James Joyce’s Ulysses. The first will be a sort of control edition – a simple HTML version presenting merely the text with the navigational and aesthetic qualities outlined below in the enumerated “Bibliographic codes and paratexts” and “Technologies of textual navigation.” The second and third versions will be authoritatively-edited and editorially-controlled annotated editions, wherein reference materials and editorial apparatus will be constructed and edited solely by the editorial team (currently: me). The fourth version will be an annotation-enabled edition, but in which students are given full control over adding annotations to the text via a layered Genius.com interface. More explanation of these three versions can be found in the section “Methodology of Construction.”
But first, I feel it prudent to briefly justify hypertextual editing and to explain my ethos behind digital (re)production of text. As I argue above, the particular value in hypertextuality for literature is its ability to ease accessibility, but without simplifying or alleviating difficulty. Shillingsburg, with a touch of polemic flare, establishes his similar ethos for hypertextual publication:
The challenge here is to develop, with the aid of the new medium of computers, sites of textual complexity from which the beauty of complex coherence shines: where text and counter-text, annotation and image, singularity and multiplicity of perspectives can serve readers upon whom nothing is lost. Let us leave to the simple-minded the creation of dumbed-down editions designed for the simple-minded; the world of knowledge, of scholarship, and of research demands editions that clarify without simplifying the textual condition. (23-4)
Hence Shillingsburg reiterates my argument for hypertextuality: it can “clarify” (or “increase accessibility” in my terminology) while simultaneously preserving or even increasing “complexity” through its intertextual structuring, multi-modality, and plurality of authorship in constructing meaning in the collective textual discourse of academic commentary and criticism.
Furthermore, hypertextual editions have the added benefit of being highly malleable. Unlike a codex publication, which can only be edited post-printing by the reader through palimpsestual amendments or by editors and printers through a full edit and reprinting, hypertextual editions can undergo radical transformations repeatedly at any given moment. Jerome McGann’s arguments for the potential virtues of hypertext over print – that a digital edition is able to “pass forward through time infinitely”  and “evolve and change” as it does, and that it can “gather new bodies of material” in the process (71) – demonstrate not only potentialities for hypertexts, but, I would argue, necessary considerations for the publication of text in a digital format.
The idea of evolution and mutation for a published text is, for many, troubling. John Updike fears that the digitization of texts, particularly in terms of the increased intertextual movements between texts, robs them of their value as “the site of an encounter, in silence, of two minds, one following in the other’s steps but invited to imagine, to argue, to concur on a level of reflection beyond that of personal encounter” (2). Katherine Hayles further notes that, for many scholarly editors, “The specter haunting textual criticism is the nightmare that one cannot then define a ‘text’ at all, for every manifestation will qualify as a different text” (My Mother 103).
To some extent, I agree with these textual concerns. As I have previously argued in my work on Graham R. Thomson’s “Vespertilia,” some logical consideration and respect for the semiotic materials of the text (and, perhaps, that text’s author) is needed. As Shillingsburg argues, “nothing [an editor] does is neutral. It all has meaning” (19). Changes to a text need logical justification and reasoning if we are to consider such changes “academic” or “scholarly” at all, and if we are to consider the text as being the “work of _______” with its originary title. For an extreme example, if one takes all the words of Moby Dick and manipulates them to the extent that it becomes a story about gardening, it becomes difficult to call it Moby Dick anymore or attribute any part of its composition to Herman Melville.
However, this concern is ultimately outdated from a post-structuralist understanding of textuality per se. Without even considering digitization, we must note that texts do not build meaning individually, but intertextually. As Roland Barthes (and every other intertextual theorist thereafter) argues, “a text is not a line of words releasing a single ‘theological’ meaning (the ‘message’ of the Author-God) but a multidimensional space in which a variety of writings […] blend and clash” (146); and the semiotic building of meaning in a given text relies on how its words relate to similar uses and associations of those words in the mind of the reader (and, for Barthes, for the textually-created “scriptor”), not the author (147). Famously, Jorge Luis Borges’s “Don Quixote of Pierre Menard” demonstrates this by arguing that, by transcribing Miguel de Cervantes’s novel verbatim in the present time, one would still be creating a new work because the contexts (or, in terms of Kristevan intertextuality, the semiotic field) of the present moment are fundamentally different from those of Cervantes, and thus the idea that Cervantes is even the author or that there is such a thing as an essence or purity to the text is undermined. As Hayles notes, Borges “draws into question the very idea of an ‘original,’ for temporal priority does not signify ontological priority when the original is regarded as simply one draft among many” (My Mother 114). The idea of an essence behind a given text, thus, is an inappropriate way of imagining textuality; and, as Willard McCarty suggests (conveniently with a quote from Ulysses), the (re)production of digitized texts should therefore be considered less of a metempsychosis and more of a metamorphosis (111). Hayles, like Marx before her, appropriately welcomes her haunting specter, and so should we.
This is not to say, again, that there is no responsibility toward preserving anything about the text being (re)produced – metempsychosis is an inappropriate metaphor in that it assumes an essential “truth” or identity behind the text, but a metamorphosis still requires that there is a text of some sort being altered rather than newly created. As Shillingsburg argues, “What is being ‘electronified,’” and therefore altered, “in an electronic scholarly edition is not the [lexical] texts but the access to texts and textual scholarship. The potential effects are profoundly textual, both in the sense of changing readers’ relationships to the text and changing their interpretations and uses of text” (85). If the text being (re)produced were treated with complete disregard, then the title and authorship of that text logically should be that of the editor (me) rather than the author (Joyce). In this sense, the text itself (coined the “lexical codes” by Shillingsburg and McGann) should be maintained as closely as possible with respect to the versioning decisions between multiple previous editions and manuscripts. The editorial alterations should primarily come in terms of various levels of bibliographic codes and paratextual elements, technologies of textual navigation, and editorial apparatus.
The first two of these concerns are fairly simple to theorize:
Bibliographic codes and paratexts.
Typeface ideally should retain a close semblance to the original (TNR for my purposes); but as browsers do not necessarily have all typefaces built in, the generic font family name (here “serif”) also needs to be permitted such that one respects the individual browser’s role as part-creator in a unique rendering of the coded text. Margins should be preserved on all edges to “frame” the document, but also because, as Christian Vandendorpe notes, they “allow the eye to regenerate from the tension produced by the innumerable saccadic movements of the eye during the reading process.” Vandendorpe furthermore points out that full-screen edge-to-edge text is incredibly difficult for the human eye to follow: “As the eye proceeds [along a line of text] in a jerky fashion, the longer a line of characters, the greater the risk that the eye loses track of the line on which it is fixed.” He thus suggests the same general length as lines in codices: 70 characters per line. For editions in which annotations appear on the same screen (as my versions 2-4), the text should be left-justified with a smaller margin (one inch or so) to allow annotations to appear along the right side of the screen. If no annotations are included (as my version 1), centered equal justification or enlarged margins which center the text logically make the text occupy the same screen space as most websites (which often have side bars) or word processing documents. Coloration should be left default – normally black text on white background, but if a user prefers a different coloration (for purposes of preference or Mearles-Irlen syndrome) the editor should not force the user into a specific aesthetic rendering. Generally, bibliographic codes should otherwise be left default to best match the most common visuals of hypertexts which students commonly read.
Technologies of textual navigation.
Though the page-turning skeuomorphic navigations are more akin to the codex version and thus preserve at least the aesthetic quality of previous codex forms, hypertext editions should ideally follow the navigational technologies of common webpages to more accurately mirror students’ common navigation of digital texts. Navigation between web pages or sections of web pages should be conducted through imbedded hyperlinks within the text (for specific references) or sidebar “Contents”-style hyperlinks (for general site navigation) to make such navigation as similar to common hypertext layouts as possible.
The editorial apparatus is a bit more complicated, as this is the site where the hypertext most differs from the codex. Hypertext can demonstrate similarly self-contained apparatus to a codex edition – using hyperlinks to move the document to footnoted information or draw up marginal information, presenting site-contained introductory contextual commentary or critical materials, etc. However, hypertext can also link to other extra-site hypertexts. The primary theoretical interest here is on drawing attention to an intertextual view of textuality in general, and on the rather overt intertextuality of, fittingly titled, “the Web.” McGann notes,
Unlike a traditional edition, a hypertext is not organized to focus attention on one particular text or set of texts. It is ordered to disperse attention as broadly as possible. […] The point is that the hypertext, unlike the book, encourages greater decentralization of design. Hypertext provides the means for establishing an indefinite number of ‘centers’ and for expanding their number as well as altering their relationships. One is encouraged not so much to find as to make order – and then to make it again and again, as established orderings expose their limits. (71)
But, I would add, this “making order” rather than “finding order” and this explosion of textual “centers” is not a unique quality of hypertext, but of text broadly. As Hayles argues, both echoing Barthesian intertextuality and giving a nod toward Deleuze and Guattari, all texts are really “assemblages” of text, in which the “text in an assemblage intermediate one another without necessarily bestowing on any one text the privileged status of the ‘original.’ Everything is simultaneously a translation of everything else, each united to the others in a rhizomatic network without a clear beginning or end” (My Mother 115). The difference between hypertext and codex text – and McGann’s language is very careful in demonstrating this – is the organization and enabling of textual engagement established through the medium.
Building off of Marshall McLuhan’s theory, Hayles notes that “for information to exist, it must always be instantiated in a medium” (How We Became 13). It is, quite essentially, impossible to separate the medium from the message, the technology of engagement from the information it displays, as one cannot have information without a means of its organization, representation, and access. The medium inherently and necessarily shapes the method of reading. Thus, even if the same information is represented through a series of different media, that information, owing to its different mediations, operates differently in terms of “speed, scale, and forms of human association” with it (Gordon iv). Through the speed of the hyperlink and the palimpsestual layering of windows, the spatial divisions between separate texts blur, allowing for a conception of the text as being pluralized or correlative – or, for Hayles, in assemblage. Through overtly encouraging the reader (as my version 4) or a disembodied hyperspace (as my version 3) to make rather than find order through dynamic linking movements within such assemblages of text, the digital medium draws attention to the way the reader’s intertextual movements and associations are making order (or meaning) rather than discovering it.
This, I would argue, is one of the clearest benefits of hypertextuality: through its unique qualities as a medium, it is capable of drawing attention to qualities of textuality which have always already been there. As such, hypertextual editions should allow the medium to work with these ideas as fully as possible, and the editor of such an edition should construct all editorial apparatus with an eye ever-directed toward decentralization of the text into “text in assemblage,” drawing out the relationships between texts, and encouraging the user to “make” rather than “find” order.
 Shillingsburg reiterates this sentiment, noting that digital publishing projects are more valuable specifically because they “are open-ended and can be added to and manipulated after their original editors have retired” (82).