Document structure. A basic difference is between documents that are prestructured (for example, with a table of contents), in contrast to those that are self-navigating. A prestructured hypertext imitates the design of a book with chapters, perhaps with subdivisions within each chapter, etc., and could be seen graphically as a tree diagram. The self-navigating, once you have moved past the opening, depends on links embedded within the text. Since there may be more than one link, the choice about where to go next will depend on your current interest or need. This kind of hypertext can be seen as a web. Some critics would argue that only the second kind is true hypertext.

Prestructured: see my essay on computers and literature.

Self-navigating: see Kaplan's essay on hypertext and ethics.

Thus in a self-navigating hypertext, the order in which you visit the nodes or lexia (terms for the separate sections of text) is determined by you, the reader. There is no overall logical order; links can be circular, recursive, or multiply related. This is the postmodern image of intertextuality. It may generate the anxiety of completeness: i.e., have I seen everything? Some authors help to allay this by providing a list of nodes somewhere, which as you visit them are given a different colour by the browser.

In the self-navigating hypertext, text is presented as short sections, or nodes, not in the form of a continuous, linear essay. Thus the nodes (as I will call them) must have some degree of self-sufficiency: each should make a coherent point or raise an intelligible issue.

(I am discussing only text here: I don't refer to other forms of presentation such as an image map, table, diagram, chronology, etc., each of which has its virtues in supporting self-navigating hypertext. Incidentally, the body of this document is presented in a table with three columns.)

This distinction may be illegitimate, if you accept recent arguments about the nature of text. In practice the distinction is often ignored, e.g., by Landow, where critical commentary and primary text are treated as if they were the same thing in a hypertext system. But it seems likely that different texts elicit different strategies in reading: critical or expository prose calls for model building (Johnson-Laird); in contrast, a literary text such as a novel or a poem invites an experiential and affective response. A part of this is an anticipatory sense for the text that helps to guide and shape the response to it (see Olson, et al., 1981). In this outline I assume we are considering non-literary documents: i.e., the contents are descriptive, analytical, expository, etc. But, as an author, how critical is it to you that the elements of your writing should be read in a given sequence? Do you mind where your reader starts? If you do, then hypertext may not be your preferred medium. (See also the comment on spatiality below.)

recent arguments: see our paper (Miall and Kuiken) on forms of reading.

Landow: consider the implications of the writerly text in his account.

P. N. Johnson-Laird, Mental Models (Cambridge, 1983).

G. M. Olson, et al., "Cognitive Aspects of Genre," Poetics 10 (1981), 283-315.

The sections described below are standard parts of a traditional essay; but they can also have an important function within a hypertext, whether in framing the discourse that it offers, or conveying the argument.

Types of node:

• introduction [places the forthcoming hypertext within its context: what is it? why was it written?]
• overview [introduces main topics and subdivisions of the argument; may provide several links to these, allowing the reader to jump first to whichever part of the argument attracts interest]
• conclusion [sums up and tells you where you have been; may suggest further implications beyond those that have been dealt with]
• argument [presents a specific issue within the larger topic]
• example [illustrates an argument with evidence, an illustration, etc.]
• context [points to the larger picture implied by the immediate issue]

Location:

• internal (self-authored: a link to another node within your hypertext)
• external (other-authored: when you jump outside to a remote node elsewhere on the internet, e.g., at #1 and #2 above: Kaplan, or Dyck )

External links are analogous to footnotes in an academic paper: you signal the authors you have been reading and/or have quoted from. If the other authors are available online, the reader can go and read their writing in full for themselves (whereas with a conventional print essay you are less likely to go to the trouble of following up a footnote). This is similar to the "docuverse" concept of Paul Delany.

• introduction
• overview
• conclusion
• argument
• example
• context

Delany: see Paul Delany, "From the Scholar's Library to the Personal Docuverse," in The Digital Word, ed. George P. Landow and Paul Delany (Cambridge, MA: MIT Press, 1993), pp. 189-199.

Types of link:

• structural: e.g., exit; contents; nodes-map [provides overall orientation and a place to get started]
• internal, within current document [for an internal link, see "spatial" above]
• internal, to a separate node [all the links to the right are to other nodes, i.e., to text or graphics that are not part of the present document]
• graphic [two kinds: a graphic that is called up, or an inline graphic. An inline graphic is provided in section 1 above; for a call to a graphic, see the link on the right]
• reference [to author cited, etc.: embedding links from an author's name within a text is now a common way of referencing; the link jumps you to a specific place in a bibliography]
• external node [a jump to a remote location on the internet]

• structural
• internal: to "spatial" above (click on "back" button to return)
• graphic (a small view of Vesuvius from Naples, c. 1820)
• reference (see the links to authors embedded in this text)
• external (this links to a valuable resource for hypertext fiction and theory on the internet, by Michael Shumate)

Forward and Backward links

-- in fact, a system such as the browser you are now using treats these as equivalent: all is reversible in the space-time of hypertext. In this respect hypertext doesn't correspond to actual physical or psychological systems (the entropy, or non-reversible change, of most natural systems, etc.), including the act of literary reading -- but no forward-only systems exist, as far as I know (and these would be against the spirit of the new rhetoric). In reading literary texts, non-reversible transformations take place across the course of the text; a feeling or a realization made during reading cannot be unmade. One cannot begin reading the text over again and repeat the same experience. The linear nature of reading thus cannot be jettisoned without certain consequences.

Another way of thinking about the directionality of links, but now in explicitly spatial terms:

• vertical (microscope analogy): zoom in / zoom out
• horizontal: associative; logical

Placement:

• implicit: a word or phrase
  -- as when I write: we can categorize link placement into two types
• explicit: "see xxx" / "click here for xxx" / etc.
  -- but now I write: for the two types of link, see placement

Rhetoric of departure / of arrival

Will the destination make sense to readers when they click the link and find themselves somewhere else? How much advance warning can you provide of what to expect at the other end of the link? (Netscape doesn't even distinguish between local and remote links.) Or do you want to surprise the reader? Stuart Moulthrop does. Postmodern writers celebrate the gap, the jump, the bricolage of hypertext, its unexpectedness.

Some hypertext programs facilitate understanding of a link. Hyperties, for example, allows a one sentence advance organizer to appear on screen before you click to activate the link; HyperWriter and Netscape don't.

For a more elaborate account of the "rules" of hypertext see Landow (1991). He introduces the essential issues in this way:

authors of hypertext and hypermedia materials confront three related problems: First, what must they do to orient readers and help them read efficiently and with pleasure? Second, how can they inform those reading a document where the links in that document lead? Third, how can they assist readers who have just entered a new document to feel at home there? (Landow, 1991, p. 82)

Link distractability: i.e., the surfing syndrome -- the grass is alway greener on the other side of the link. As I have put it elsewhere, hypertext invites a mode of reading that is unsettled, restless, meandering, insatiable. Internet readers are surfers. Is this how reading is, or how it must become in the electronic future?

Words, once they are printed on a page, cannot but remain fixed and static. Not so with electronic text. Anyone who uses a word processor knows how malleable text can be: deleting, cutting and pasting, relocating whole paragraphs. Similarly, documents in a hypertext such as the internet can be changed as easily by their authors. Another kind of fluidity is given by the multiplicity of links available at any given point: readers can take different possible pathways through the nodes. This fluidity of the word changes the nature of writing:

An electronic book is a fragmentary and potential text, a series of self-contained units rather than an organic, developing whole. But fragmentation does not imply mere disintegration. Elements in the electronic writing space are not simply chaotic; they are instead in a perpetual state of reorganization. They form patterns, constellations, which are in constant danger of breaking down and combining into new patterns. . . . The unity or coherence of an electronic text derives from the perpetually shifting relationship among all its verbal elements. (Bolter, 1991, p. 9)

Is a literary experience still possible, if the nodes of a text can be read in any different order and if the text itself is subject to change? Does this call for a kind of attention which is different from that we give to the literary? As Hassan notes, the literary is destabilized, unmargined:

Without formal or generic constraints, our ability to make specific sense of literature diminishes even as our freedom, as readers, swells. (Hassan, p. 201)

Ihab Hassan, The Postmodern Turn (Ohio State UP, 1987)

A spatial universe seems inescapable (e.g., Bolter's "writing space"): geometric metaphors abound in accounts of hypertext, as though what cannot be realized in spatial terms doesn't count. Moreover, there is a danger that the transfer of the non-spatial to hypertext tends to spatialize and hence diminish it. Coleridge:

Under that despotism of the eye . . . under this strong sensuous influence, we are restless because invisible things are not the objects of vision; and metaphysical systems, for the most part, become popular, not for their truth, but in proportion as they attribute to causes a susceptibility of being seen, if only our visual organs were sufficiently powerful. (Biographia Literaria (1817), Ch. 6)

While it doesn't follow that any text on the internet necessarily degrades the non-spatial, the graphical interface, the presence of links (quick: click 'em), and the narrowness of the screen, each tend to encourage a short attention span and a form of writing that calls for the tangible, immediate, and vivid. Can you get subtle on the net? Yes, of course: but, notice that those who do tend to write long, linear essays. Nice to have them available, but is this really where you want to do your serious reading?

Is interaction limited merely to choosing links? Is this empowering? Only if you become a hypertext author yourself. And for the reader on the web, current tools don't allow for genuine interaction, not even highlighting, notetaking, annotation, or marginalia (what would Coleridge have done?). Compare this to reading a book (that you own, preferably). In which kind of reading are you more active? Recent hypertext rhetoric is dismissive of the book.

A specific case of interactivity. Does it favour the transmission model of learning or a creative model? If hypertext is used merely as a medium to present information, then the authority of the medium must outweigh that of the student, just as textbooks, teachers, school timetables, and education boards have always done. Extending the hypertext available to include the internet is a difference in degree, not in kind: more of the same will not end in giving control over knowledge to the student. Let the student be an author in her own learning, then the hypertext tool offers a different potential. Students can either author their own hypertexts, or a teacher can use hypertext as a medium to report work that students have done.

student authors: see projects done for a Romanticism course