Jim Rosenberg
R.D. #1 Box 236
Grindstone, PA 15442
E-mail: jr@amanue.pgh.net

Navigating Nowhere / Hypertext Infrawhere


Various non-linear methods of structuring the lexia are discussed, including simultaneities and polylinearity. The simultaneity is similar to Aquanet relations. A distinction is drawn between the typical disjunctivity of the hypertext link and conjunctivity of simultaneities and relations. We begin the process of exploring the rhetoric of the conjunctive hypertext relation. Finally, the structuring of the lexia is intensified and extended into the fine structure of language itself: hypertext infrawhere.

1. Must the lexia be linear?

It is a truism of hypertext rhetoric that navigation is the user's choice. If we truly believe this, then among these choices we must place the null choice: the choice to go nowhere. This presentation begins with the rhetorical question: "When you go nowhere, where are you?" The conventional term in literary hypertext for "where you are" in a hypertext when you haven't gone anywhere is lexia ([Lan92]). Thus we begin with a discussion of the structure of the lexia. It is customary to avoid this question: the lexia is typically considered ordinary linear writing, with the "real" issues of hypertext relating to how lexia are organized; the lexia itself is not really considered as a hypertext. (E.g.: "[The within-component] layer is purposefully not elaborated within the Dexter model" [Hal90]; "The local stability of the lexia arouse expectations of coherence and internal consistency" [Mou92a].) However, non-linear writing, particularly in poetry, has a rich and varied history; poets who are comfortable with extending that history have reason not to accept an inevitable linearity of the lexia.

Intergrams [Ros93] illustrates an example of a non-linear, non-link structuring method, the simultaneity: the literal layering on top of one another of language elements. (See Figure 1.) The hypertext link does not really express this structure, particularly as it is actually realized by existing user interfaces. A much closer match is provided by the concept of an Aquanet relation [Mar91][note 1]. A simultaneity may be thought of as a relation in which all slots are unnamed, equivalent, accessible, and located in approximately the same graphical position on the computer's display. In Intergrams, simultaneities are implemented through the use of "tactile", no-click buttons in which there is screen behavior from simple mouse-cursor movement through hot-spots on the screen. Such devices may have uses beyond the artistic. The concept of simultaneity is the direct analog for writing of juxtaposition concepts which have a rich and varied history throughout the twentieth century in visual arts and music, and is clearly related to simultaneous-voice work (both live and on magnetic tape) in poetry. (See e.g. [Ros75a].)

Figure 1: Screen dump from Intergram 9. The top rectangle shows a simultaneity which is "closed" -- all phrases are visible simultaneously -- while the bottom rectangle shows a simultaneity which has been opened to show only one phrase. The rectangle on the right is a button serving as a link to a relation among three simultaneities; the button at the lower right is linked to the parent screen. This screen shows a ternary syntax relation with the "verb slot" (to use Aquanet terminology) toward the right.

Another example of how the lexia may be structured nonlinearly is polylinearity: the stringing of word skeins in a graphical space where normal print conventions establish no clear ordering among the skeins. The simultaneities in [Ros94] include both polylinear text and "single-card" relational syntax diagrams. (See Figure 2).

Figure 2: Polylinear text from Diffractions through: Thirst weep ransack (frailty) veer tide elegy. This is only one plane in a simultaneity consisting of 2 other polylinear screens and 3 relational syntax diagrams.

Is "the lexia" a single plane in a simultaneity, or the totality of all planes stacked together? In a work like Intergrams, one could easily argue there are "conventional" lexia, that a single plane in a simultaneity acts as the lexia. The same could be said of Diffractions through, though in that case the lexia is fractured by polylinearity and sub-diagrams. Is "the lexia" in a polylinear case a single word skein, or the whole net? This question is a bit more troublesome. One could be comfortable arguing that "the lexia" is whatever you see on the screen without moving the mouse or touching the keyboard. On the other hand, given a representation of a hypertext as a network of nodes, the traditional concept of lexia is: one of the nodes. Should we insist that lexia be node-focused in the case of a typical hypertext but insist on accepting all skeins simultaneously in the polylinear case? This seems arbitrary.

2. Conjunctive vs. disjunctive structure

Notwithstanding the celebrated remark of J. Yellowlees Douglas [Dou91]:

They are all laid out before us: the genuine post-modern text rejecting the objective paradigm of reality as the great "either/or" and embracing, instead, the "and/and/and."

in fact, the typical hypertext link may be described as a disjunctive link: if lexia X has links A, B, C, D, the user may choose A or B or C or D (or to go nowhere, of course!). Almost the entirety of hypertext rhetoric surrounds what may be called "the confrontation with or" -- how to assist the reader in coping with the volume and structure of choice. A simultaneity may be disjunctive or conjunctive: the whole of a simultaneity with planes A, B, C, D may be A and B and C and D. One might envision hypertext links as being conjunctive also. Surely in many cases Aquanet relations must be described as conjunctive, not disjunctive. Consider, for instance, an Aquanet schema for diagramming sentences in which a Sentence object has slots for Noun Phrase and Verb Phrase. May we describe the verb phrase slot as optional?

While the literature concerning the rhetoric of the hypertext link has become amazingly voluminous considering how few literary hypertexts there still are, the rhetoric of the conjunctive hypertext relation is nearly non-existent. Discussions of hypertext consistently use travel vocabulary: links are "followed"; [Lan89] speaks about "arrivals" and "departures" -- all terms that are quite cogent when applied to disjunctive links, but which may have limited relevance when applied to conjunctive relations. Perhaps we should speak of gathering a conjunctive relation, rather than following, as in the case of a disjunctive link. This has major implications for discussion of the lexia. "Following" a link means leaving it for another lexia; "gathering" a relation means bringing things to a central place: whereas the disjunctive link is associated with travel, the conjunctive relation is associated with locus, with an inherently structured lexia. Stuart Moulthrop, in [Mou92b], concurring with [Bol91], seems to apprehend the problem, but has proposed the wrong solution: a kind of Wittgenstein duck-rabbit flip in functionality between node and link. While this is a fascinating and compelling metaphor, the structural inadequacy of having nothing but nodes and disjunctive links is not to be solved by having nodes and links philosophically trade places but by the much more obvious expedient of providing more explicit structure. (But Moulthrop and I come out the same place in the end, see below.)

That the concept of relation is closely tied to spatialized text has been studied in detail by Marshall and her colleagues ([Mar92], [Mar93]). She relates that Aquanet users frequently constructed piles where it had been anticipated they would construct relations.[note 2] Perhaps systems such as Aquanet should formalize piles as legitimate objects. Certainly the pile is a worthy artistic device in its own right. Piles are in widespread use in the visual arts, and e.g. [Ros73] and [Ros75b] used word piles as a formal device, in the latter case including word piles into higher-level relational structures.

The proper user interface behavior in the face of conjunctive structure is an interesting challenge. It is likely that writers will want different word object behavior from conjunctive structures than from disjunctive structures. Note this has obvious ramifications for formal theories of hypertext. The assumption is explicit in [Grø94] that the Dexter concept of composite adequately formalizes Aquanet relations, and there is an implicit inference in [Mar91] that hypertext links are special cases of Aquanet relations. Composites are arguably conjunctive, but it is not clear that the conjunctive relation has the same object behavior requirements as the composite (which may have no behavior requirements at all), and certainly disjunctive links and conjunctive relations are vastly different kinds of objects, and deserve to be recognized as such both in software object models and formal theories. Indeed, as a writer, I would make an urgent plea to software developers to include as many object types as possible in your models, and absolutely to provide some form of extensibility so that writers can program their own way to workarounds when off-the-shelf object behaviors are not adequate.[note 3]

The concept of contour, [Ber92], which I would describe as a supra-lexial projection by the reader of a geography underneath pathways through the or-cloud, is perhaps intermediate between the purely disjunctive and purely conjunctive. Can we describe the contour as the attempt to resolve disjunctive experience into conjunctive resonance?

3. From intensity of structure to hypertext infrawhere

The author's current work in progress (as yet untitled) carries the concept of simultaneity still further in the idea of a nested simultaneity. In some cases this work carries the simultaneity inside the sentence. Hypertext is carried into the fine structure of language. Where is "the lexia" now? Is there really a concept of lexia when we are inside the sentence?

A hypertext may be thought of as a kind of virtual diagram, with software for navigating the diagram. If the diagram is small enough it may be presented in a single graphical space, without the aid of software. The author's Diagram Poems, e.g. [Ros79], [Ros84], are examples of such works. These present an explicitly relational syntax notation, still used in Intergrams and Diffractions through. The structural atoms in the Diagram Poems are small clusters of words; the relational (i.e. hypertext-on-paper) structure is the sentence structure.[note 4] What shall we say is "the lexia" here? In the Diagram Poems, the diagram notation carries syntax itself. Executed on a larger scale, this concept leads to the use of hypertext to carry the very infrastructure of language. Such works would have hypertext infrawhere: a structural underneath so fine and so pervasive, a lexia so completely fragmented, that the concept of lexia ceases to have any meaning: a completely dematerialized lexia, as in [Mou92b] after all.

In [Mou92b] Stuart Moulthrop asks: "Why does the hypertext research community publish its work in print?" At the risk of seeming glib, the answer is obvious: because hypertext is not our native tongue. Many will surely balk at the idea that this needn't be so, that there can exist a natural language in which hypertext carries the very structure of syntax itself: hypertext not as a medium of organizing thoughts, but as a medium of thought. Perhaps in the end this will turn out to be unachievable, but as a focus for poetic experimentation it provides this author with a sustaining vision.


[Ber92] Mark Bernstein, Michael Joyce, and David Levine, "Contours of Constructive Hypertexts", ECHT '92 Proceeding of the ACM Conference on Hypertext, ACM, New York, 1992.

[Bol91] Jay David Bolter, Writing Space, The Computer, Hypertext, and the History of Writing, Lawrence Erlbaum Associates, Hillsdale, New Jersey, 1991.

[Dou91] J. Yellowlees Douglas, "Understanding the Act of Reading: the WOE Beginner's Guide to Dissection", Writing on the Edge, Volume 2, Number 2, 1991.

[Grø94] Kaj Grønbaek and Randall H. Trigg, "Design Issues for a Dexter-Based Hypermedia System", Communications of the ACM, February, 1994.

[Hal90] Frank Halasz and Mayer Schwartz, "The Dexter Hypertext Reference Model", Hypertext Standardization Workshop, NIST, 1990.

[Lan87] G. P. Landow, "Relationally encoded links and the rhetoric of hypertext", Hypertext '87 Proceedings, Chapel Hill, NC, 1987.

[Lan92] G. P. Landow, Hypertext: The Convergence of Contemporary Critical Theory and Technology, Johns Hopkins University Press, 1992.

[Mar91] Catherine C. Marshall, Frank G. Halasz, Russell A. Rogers, and William C. Janssen Jr., "Aquanet: a hypertext tool to hold your knowledge in place", Proceedings of Hypertext '91, ACM, New York, 1991.

[Mar92] Catherine C. Marshall and Russell A. Rogers, "Two Years before the Mist: Experiences with Aquanet", ECHT '92 Proceeding of the ACM Conference on Hypertext, ACM, New York, 1992.

[Mar93] Catherine C. Marshall and Frank M. Shipman III, "Searching for the Missing Link: Discovering Implicit Structure in Spatial Hypertext", Hypertext '93 Proceedings, ACM, New York, 1993.

[Mou92a] Stuart Moulthrop, "Shadow of the Informand: A Rhetorical Experiment in Hypertext", Perforations 3, Public Domain, Atlanta, GA, 1992.

[Mou92b] Stuart Moulthrop, "Toward a Rhetoric of Informating Texts", ECHT '92 Proceeding of the ACM Conference on Hypertext, ACM, New York, 1992.

[Ros73] Jim Rosenberg, Temporary Poetry 10/73, word environment, Les Salons Vides, San Francisco, 1973.

[Ros75a] Jim Rosenberg, "Intermittence", poem for four simultaneous readers and conductor, in Roger Johnson, ed., Scores: An Anthology of New Music, Schirmer Books, New York, 1981.

[Ros75b] Jim Rosenberg, Permanent and Temporary Poetry 5/75, word environment and performance, The Kitchen, New York, 1975.

[Ros79] Jim Rosenberg, Diagrams Series 3, published on-demand by the author, Grindstone, PA, 1979. Exerpts appeared in Interstate 14, Austin Texas, 1981.

[Ros84] Jim Rosenberg, Diagrams Series 4, published on-demand by the author, Grindstone, PA, 1984. Exerpts appeared in Tyuonyi 1, Santa Fe, NM, 1985. Available on-line in ACEN Conference, The WELL, Sausalito, CA.

[Ros93] Jim Rosenberg, Intergrams, Eastgate Systems, Watertown MA, 1993.

[Ros94] Jim Rosenberg, Diffractions through: Thirst weep ransack (frailty) veer tide elegy, to appear.

Jim Rosenberg's home page o Poetics