The Shadow of an Informand: A Rhetorical Experiment in Hypertext

Stuart Moulthrop

Textuality Intertextuality Narrative and Consciousness

[1] technical introduction

The literature of hypertext research contains many papers that describe experimental work; this paper is itself an experiment in thinking and writing. I attempt to construct a discourse about conceptual changes necessary to expand our idea of hypertext rhetoric. I also try to apply those changes by modifying the conventions of traditional print discourse. I have worked from the assumption that some discourse about hypertext needs to originate in hypertextual form. Although that form cannot be duplicated in print, it may be approximated. The structure of this paper reflects as fully as possible that of the electronic hypertext from which it was derived [45].

The linking scheme employed here was invented by the novelist Julio Cortzar for his fiction, Hopscotch. That book consists, Cortzar writes, "of many books," that is, many possible sequences in which the chapters may be read. Two of these sequences are outlined in an introductory note: one that proceeds straight from chapter 1 to chapter 56 (at which point the reader is supposed to stop midway through the book), and a second, non-serial reading delineated by markers in the text (1966).

I have implemented both of Cortzar's methods here. You may read this paper straight through to lexia 46 ("Perfunctory Closure") and stop there, or you may follow the alternative transitions indicated in brackets at the end of each section. There are, of course, other ways to read as well. [>>2]

[2] dangerous questions

At the 1990 European Conference on Hypertext, Mark Bernstein posed a now notorious question: where are the hypertexts? This was a bit like asking about the Emperor's fashion sense a dangerous move, because such questions always lead to other questions. In the case of hypertext, the inquiry might continue as follows: Where are the original writings made possible by this medium? What new genres, what new modes of thought does this kind of writing open up? In science, in technology, in culture, what difference does hypertext make?

These are all excellent questions, but now I want to ask something else, a related but more immediate question: Why aren't you reading this document in a hypertext system? How is it that those of us who analyze hypertext, even those of us who promote and proselytize for it, carry on our communications primarily in print? What does this preference imply, both about the organizations interested in hypertext and about the systems they develop and study? [>>8]

[3] hypertexts on hypertext

Of course a large body of technical literature about hypertext is already available in hypertextual form. To name but four very interesting examples: the Guide hypertext version of Nelson's Literary Machines (1987), the Hypertext on Hypertext published by the Association for Computing Machinery (Yankelovich 1988), Jay David Bolter's Storyspace version of his book, Writing Space (1991b), and the ACM Hypertext Compendium (Akscyn 1991a). Each of these projects makes a significant contribution, both in form and content, to our understanding of hypertext and its rhetoric (Akscyn 1991; Bolter 1991a; Rous et al. 1989); but each of these texts is limited in one crucial respect.

All of the works mentioned above reprocess material that originally existed in print. Decisions about the organization of the text, the number and kinds of links, and the design of user interfaces were made after initial, linear composition. Despite the addition of hypertext apparatus, their logical and discursive structures remain fundamentally those of print. They are not true hypertexts but cradle-works or incunabula (McDaid 1991), retrofit projects much like early scribal redactions of oral poems. [>>4]

[4] what kind of hypertext?

So where are the hypertexts? It seems that we first need to ask another question: what sort of hypertexts are we looking for? Certainly the existing hypertexts-on-hypertext satisfy at least part of the definition of hypertextual writing (Berk and Devlin 1991; Nelson 1987). They can be read in multiple sequences and they allow readers to articulate transitions through a scheme of specified links. In what sense then do these works seem inadequate?

It's because these texts were printed documents before they were hypertexts, for which reason they retain traces of singular and hierarchical composition. To be sure, these writings represent hypertext of a kind hypertext as repurposed print. This is the kind of hypertext Nelson has in mind in his early descriptions of a "docuverse" (1990) of interlinked linear discourses. It is also the conception Bolter uses when he describes the ultimate hypertext as a vast linked encyclopedia or "great book" (1991). [>>7]

[5] traces of singular and hierarchical...

So what work doesn't retain such traces? Is this document that you're reading now "true" hypertext simply because it was written as a network of nodes and links? Doesn't it have its own linear and hierarchical strucures? The difference between documents composed in print and documents authored in hypertext may be one of degree rather than kind. The "traces" of linearity will always be present: it's a question of how important they are to the structure of the work and to the way it achieves its purposes (its rhetoric). [>>6]

[6] this text

This document was conceived and written as a hypertextual network in Storyspace, after which I converted it to a word processor format and implemented the Hopscotch links [>>1].

[7] going native

I'm interested in another kind of hypertext, one that denies the priority of print structures. This would be an original or native hypertext composed in and for the electronic medium, and therefore situated outside the literary conventions of print conventions like linear coherence and hierarchical argumentation.

Are you reading such a text now? Is this kind of writing a real possibility or just a theoretical fiction? Assuming for a moment that it's the former, could such writing be used to develop a discourse about hypertext and its rhetoric? Could it be the hypertext Bernstein is looking for or perhaps a kind of hypertext none of us expected? [>>12]

[8] statement of intent

What I'm mainly trying to do here is mix it up i.e., use the hypertextual medium to assemble a wide range of ideas and discourses which I wish to hold, if not in suspension, then at least in a kind of dynamic balance or chaotic (dis)equilibrium. I am trying to deliver and process information, at times rationally and perhaps even with intent to persuade. But I am also trying to provide conditions under which this project of persuasion will break down into something else not a textual object but an object-event or an informand. [>>26]

[9] disclaimer

"If our argumentation for designing an authoring system this way did not convince you to the extent we hoped it would, this might be due to the circumstance that, unfortunately, the system we propose is still under development and therefore has not been at our disposal for writing this paper" (Streitz et al. 1989, 361).

As inventors and experimentalists, hypertext people often get a bit ahead of themselves. Where are the hypertexts? no glot, clom Fliday. Check with us again Real Soon.

But with so many systems now available it gets increasingly difficult to use the Streitzian disclaimer to explain why we aren't pursuring or at least presenting our work in hypertext. Indeed, some researchers do use the medium at least on an experimental basis (see Marshall et al., 1991). But why do so many continue to use hypertext (as in systems like NoteCards, Writing Environment, SEPIA, Aquanet) as a facility for print production? [>>12]

[10] containing multitudes

Why is it important discriminate electronic incunabula from native hypertext? Being by definition large, the discourse of hypertext can surely contain multitudes and so we have "Romantic" versus "pragmatic," "exploratory" versus "constructive," "smooth" versus "chunky" hypertexts (Charney 1992; Joyce 1988; Nelson 1987) why add one more taxonomic pair?

Because there's no way around it. The distinction is important. As I have argued elsewhere, hypertext rhetoric needs to expand beyond literary categories imported from print (1991). Trigg and Irish were right when they noticed "an inherent contradiction in the notion of using hypertext for writing papers" (1987, 89). Hypertext isn't just for papers anymore. If we continue to regard it as such we may find ourselves in an intractable position. [>>14]

[11] rhetorical question

Why do I keep asking so many questions? [>>53]

[12] looking at and looking through

Let's be skeptical for a moment. In a sense, the idea of "native" hypertext springs from a gratuitous distinction. Why can't we accept the ACM Compendium as or some other print-derived hyperdocument as an adequate model? After all, we know that all hypertexts are linear at some level of granularity. The Compendium retains linearity at the level of fairly long, extended discourse (individual papers); the "native" text you are now reading retains its linearity over smaller fragments, what we might call micro-lexia [>>36]. Isn't this just a change of scale, and do scalar differences of this sort mean anything?

Possibly they do. Lanham has sugested that electronic technologies like hypertext re-awaken an understanding of media (1989). When a medium is predominant (as print has been the predominant medium for writing), its structures and conventions go unexamined. The medium becomes a transparent force, or in McLuhan's terms an "environment" (McLuhan and Fiore 1967). But with the advent of electronic technologies, Lanham argues, we begin to look "AT" rather than "THROUGH" writing structures (1989, 286). [>>13]

[13] looking the other way

Hyperdocuments composed of long linear nodes leave the transparency of print conventions largely unaffected. Even with an overlay of link anchors, texts such as these can still be read with all the local coherence and discursive stability of a conventional paper. Such structures let readers default to a reassuring but inappropriate form of reception. In effect, documents of this sort let us act as if we were still reading a conventional print text. Since there is no pressing local need to deal with questions of linking or navigation, we can defer these problems or give them short shrift.

So the distinction between hypertext incunabula and native hypertext remains valid. All hypertexts manifest disjunctive breaks at some level of their structure, but native hypertexts manifest these flaws and gaps on a fine-grained scale where they cannot be ignored. The finer the grain of hypertextual language (or we might say, the closer its contours), the greater the difficulty of interpretation. But if we wish to engage the most significant issues in hypertext rhetoric we will not be able to evade difficulty. [>>11]

[14] charney's dilemma

We're told that hypertext can't do the work of print very well and it doesn't take a cognitive scientist to figure this out. Hypertext gives readers more choices at the same time that it takes away the linear structure and simple closure of the codex. A number of researchers have suggested that hypertext complicates the reading process in information-retrieval tasks associated with education and training (Carlson 1990; Charney 1992; Landow 1987; Wright 1991). Hypertexts can be disorienting and confusing; this now passes for documented fact. [>>15]

So, as theorists like Davida Charney see it, hypertext developers face a dilemma. Our documents must function in the real world, where people need systems that can deliver information quickly and efficiently. Our documents must deliver the goods, they must adapt to constraints. We can create rhetorical conventions to restrict the complexity of hypertext documents, but as we do so we must give up many of the features that make hypertext so interesting, such as implicit and associational links. The choice as Charney outlines it lies between a "Romantic" view of hypertext as totally unconstrained, free-associative discourse and an instrumental approach in which hypertext looks very much like print (1992). [>>16]

[15] ...documented fact

Compare the results of Bellcore's Chemistry Online Retrieval Experiment (CORE): "The SuperBook hypertext system studied here proved superior to the printed indexing system and the document retrieval system for finding research articles related to known topics for browsing, search, and essay writing" (Egan et al. 1991, 310). [>>16]

[16] cognitive flexibility

The force of Charney's dilemma depends on our expectation that hypertext perform the same functions as print. Cognitive complication is only a problem if we expect our media to make the world simple. For some applications, for instance those involving knowledge transfer rather than data acquisition, hypertext may be a better choice than print precisely because it models "cognitive flexibility" (Spiro and Jehng 1990). But cognitive flexibility hypertexts are not likely to resemble books either in their structure or their operation. Spiro and Jehng's experiments in this genre have indeed produced something that more rsembles a film director's storyboard or shooting notes than it does any formal print genre.

The solution to Charney's dilemma is easy enough bid adieu to print and go hypertextually native. We need to develop a strain of hypertext rhetoric that explores post-print concepts in electronic writing, one that is more at home with databases and multivalent networks than with books and papers. [>>21]

[17] fragmentary

Fragmentariness may be (if I'm lucky) the most serious charge that can be brought against this text: it doesn't deliver on its points, it doesn't follow out its argumentative leads.

Fragmentation, ellipsis in the words of the poet, why are these words so nasty? Aren't research writings always, in their nature, fragmentary? Science after all (so the Textbooks say) is a constant cyclical reiteration of hypothesis and test. Normal science punctuated by paradigm shifts. The subject of rational inquiry is never closed.

Right. Except that, formally speaking, the rhetorical structure of every research paper is very carefully closed, summed up, neatly concluded. Call it complementarity, ying for yang. Call it irony. There's something about that yawning gulf of uncertainty we call the Real World that makes us rage for order in the scientific court. [>>18]

[18] nonfeasance

This experiment violates the closure convention repeatedly. It therefore opens itself to the charge of nonfeasance failure to deliver the Object, i.e. a duly supported singular Claim.

The charge may be a valid one; but please do consider that it may not be in the nature of hypertextual discourse to deliver such objects. [>>21]

[19] why print?

Instead of asking why you aren't reading this paper electronically, perhaps it is more useful to reverse the logic: why are you reading a print document? What does print have to offer that proves so important to hypertext researchers, or indeed to any intellectual community? [>>20]

[20] pragmatics

One answer might be based simply on professional pragmatics. This is a printed paper because only printed papers constitute creditable work in the discourse community of rhetoric and technology, which is where I earn my keep. Where I work, writing is a major element in our (written or unwritten) job descriptions. For many of us, especially theorists, printed work is the most important criterion of professional judgement. By publishing in journals, books, and conference proceedings we earn recognition from our colleagues and superiors. We work in print because our professional communities (the university, the research institute, the corporation) authorize no other options. [>>17]

[21] the literature

But why don't we have other options besides print? What does print offer that other modes of intellectual discourse (lecture or oral debate, for instance) do not? The answers to these questions have less to do with pragmatics and more to do with the social organization of discourse. As Latour points out, publication in print allows a research community to construct its specific literature, a system of references that validates or invalidates a particular claim: "the status of a statement depends on other statements" (1987, 27). Any given article in the literature is linked both retrospectively and prospectively to other articles, either by references to work previously published or by implicit links to future developments. Our print-based academic culture has developed numerous ways to promote this process, including genres of writing the research report, the theoretical paper, the summary of research, the textbook. [>>22]

[22] intertextuality

Latour's understanding of technical literatures draws on what literary theorists call intertextuality, the notion that writings derive their meaning not by reference to external reality but from their relationships to other writings. This idea maps neatly onto the basic structural principle of hypertext, the potential relationship of any element in a document to an undefined number of other elements. Landow and Bolter both describe hypertext as the technological apotheosis of intertextuality (Bolter 1991a, Landow 1992).

But if there is such a strong link between intertextuality and hypertextual discourse, then why has hypertext writing not been embraced by the people who study and promote it? The lack of such engagement seems most unaccountable. As researchers we surely depend on just the sort of intertextual citation practices Latour discusses; and what community is better prepared or equipped to make use of hypertext systems? [>>23]

[23] summary findings

As Latour describes it, the formation of technical knowledge depends on our ability to represent complex claims as discrete propositions. In her paper on fuel cells, Researcher X may discuss a number of conflicting observations and findings, but she will ultimately advance a single proposition (or a generalizable range of propositions) as a summary of her work. Though later writers may return to the anomalous details, they are more likely to regard X's general summary as her contribution to the literature.

The printed research report, with its formal progression from abstract to conclusion, reinforces and authorizes this social convention. In the paper information comes neatly packaged in a single, linear sequence. Any ambiguities we encounter along the way must be resolved in the summary, or at least accounted for as areas for further inquiry. [>>24]

[24] incompatible

Perhaps the reason hypertext has not caught on with hypertext researchers lies in the fact that this medium supports intertextuality all too well so well in fact that it breaks down the particularity of discourse, interfering with the mechanism of summary representation.

Compare the print paper to a truly multi-linear hypertext where the writer explores with equal discursive emphasis the ambiguities and alternative hypotheses that derive from her research. Depending on their interactions with the text, subsequent readers might form very different conclusions about the researcher's findings, making it impossible to represent her article in terms of a simple, particular proposition. If the article cannot be thus represented, it cannot effectively contribute to the stream of citations from which technical consensus emerges. Its intertextuality is simply too pronounced, or its focus too weak, to serve the dialectical process of science-in-action. This, then, may be a major reason why hypertext researchers do not work primarily in hypertext: because the work we do and the institutions in which we work are both hierarchical, and because a fully realized or "native" hypertext is incompatible with hierarchical discourse. [>>40] [>>48]

[25] prospecting

At least as far as rhetoric is concerned, most attention to hypertext has taken the form of integration, an attempt to understand this technology in terms of existing models, to make the theory and practice of hypertext co-extensive with existing notions of reading, writing, and cultural practice. Thus Bolter's conclusion that "[t]he computer is simply the technology by which literacy will be carried into a new age" (1991, 237). While there is much to be said in favor of this strategy, its value is ultimately limited. Integration strategies tell us where we are in terms of where we have been. They cannot say much about what we can expect in the "new age" that lies ahead. We cannot assume that reading and writing will go on without major transformations. Bolter also tells us that hypertext means the end of literacy as it has been defined by the medium of print (1991, 2). To understand what this means, we have to consider the future. [>>26]

[26] in the workplace

What is the future of hypertext and hypermedia systems? Set aside the 2020 visions of Xanadu or the even wilder prospects of cyberspace or next-stage consciousness. Let's think not about science fiction but about engineering realities. Consider the future of hypertext as it looks to information system designers at Boeing Computer Services: "Certainly hypermedia will continue to be an effective way of presenting static reference information.... But a larger role for hypermedia requires eliminating the distinction between authors and readers. We assume that all members of engineering teams will be able to create and access information in a shared, distributed environment" (Malcolm et al. 1991, 15; see also Parunak 1991a).

Where are the hypertexts? Try over in the engineering department. [>>27]

[27] informating

The collaborative, distributed, multimodal hypermedia application envisioned by the Boeing communication designers represents a use of technology that Zuboff has called informating (1988). Informating strategies emerge as industries move beyond an emphasis on productive efficiency, where automating strategies predominate, to an emphasis on innovation and competitive advantage in which manipulation of information (texts) can be as important as the production of material goods.

Automating technologies seek to extract productive value (skill, technique, strength) from human bodies, investing them in machines in order to make human labor superfluous. Informating technologies for example electronic mail and bulletin boards, telecommunications networks, distributed hypertexts also extract or externalize workers' activities (planning and discussion) but they do not seek to replace workers. Rather, as Vannevar Bush and Douglas Engelbart have envisioned, they are intended to expand or augment people's performance (Bush 1945; Engelbart 1991). [>>28]

[28] the informand

Automating technologies externalize the work of the body, producing as their ultimate object the robot (a name which we ought always to remember originally meant worker). Informating technologies externalize the work of the mind or of language, but in an important sense they do not produce an object. What they produce is rather an intersection of objects and events “ a text “ or as Barthes calls it, a "social space" of communication (1986, 64). Following an idea given me by my collaborator Nancy Kaplan, I will call the object-event an informand. The informand is the virtual space or dynamic network of linguistic possibilities created when workers collaborate by means of an informating technology. Examples of informands include a telephone conversation, an email discussion thread, a jointly authored memo or report, and a collaborative hypertext. [>>51]

[29] the title of this paper

The Shadow of an Informand: A Rhetorical Experiment in Hypertext

An informand is the object-event produced by an interactive communications technology in its social application. A hypertext at work in its discourse community (i.e. one that is being read, discussed, criticized, and rewritten) is an informand. The contours of an individual hypertext (the momentary structures of coherence and possibility apparent to the reader as she interacts with the structure) are also informands.

However, the document you are now reading is an object, not an object-event. It is therefore like the shadow of a three-dimensional object, simply a planar approximation, flat and obscure. If you are reading this discourse in print you are dealing with the shadow of a shade, as Borges or Nabokov or someone like that may have said. [>>32]

[30] hypertext and informand

The informand is not itself an object but rather a socially distributed, virtual text comprising an unspecifiably dense network of propositions and links. The informand is not an autonomous system but a medium, both in the conventional sense of a communications tool and in McLuhan's specialized sense of prosthesis or extension of human function (1964). Thus a true artificial intelligence would not qualify as an informand, but some kinds of expert systems might: what distinguishes the informand is its lack of objective closure. The informand is not intended for independent operation. It does not exist outside the context of its use. It is not so much a tool as an occasion.

Informands are relevant to hypertext research because certain kinds of hypertext are also species of informand. These would be hypertexts of the type referred to by the Boeing group: distributed, collaborative, and multimodal. Such hypertexts also fall within the purview of Joyce's "constructive" hypertext, which are "versions of what they are becoming, a structure for what does not yet exist" (1988, 11). The idea of the informand provides a powerful metaphor for understanding such texts a metaphor, most important, which does not reduce the text to an object or an essence. [>>36]

[31] an experiment

A Rhetorical Experiment in Hypertext. The idea behind this writing is to explore some way in which the rhetoric of hypertext might escape its shadow status and take on the full three-plus dimensions of hypertextual discourse. Even in print? Who knows... [>>32]

[32] non sequitur

I have never really understood what Nelson meant when he called hypertext "non-sequential writing" (1987). In what sense can writing ever be non-sequential? If we dispense with normal word order? If we stop reading from right to left? Consider the following example:

Even when scramble we of syntax English) sentence (especially re-imposes reader in re- linearity interpret in order to the the the.

Despite its thorough jumbling, the sentence above can still be understood with a certain amount of effort. Using familiar associational cues like subject-verb argreement ("we scramble"; "reader re-imposes") and likely phrase combinations ("even when"; "in order to"), a proficient reader should be able to piece out a translation. Even with rules of normal sequence suspended at one level, the sentence is still intelligible because it remains sequential on a smaller scale that of individual words or tokens. [>>33]

[33] polylogue

I've always thought polysequential is a better way to describe hypertext. [>>36]

[34] ein norqusut

However, if we were to abolish all sequentiality from the sentence, turning it into a mere random string of characters, then interpretation really would be impossible:


Intelligibility in language demands some kind of sequence; or in the case of hypertext, some level of absolute sequentiality. No one would seriously propose a hypertext system where variability or randomness obtain at the character level. It may be possible to conceive of languages without syntax or spelling, but they don't seem very useful unless you happen to be a cryptographer. [>>35]

[35] hyperbaton

Jay Bolter has an interesting approach to the dialectic of singular and multiple sequences in hypertext. He invokes the rhetorical figure of hyperbaton or scrambled syntax (think of the way Yoda speaks in the Star Wars movies). According to this view, hypertexts (e.g., interactive fictions where linear order alternates with disjunctive jumps) represent deliberate deformations of a primary discursive or narrative sequence (Bolter 1992). Readers traverse the network in different ways, but they understand these traversals in terms of an intuited ur-sequence, in much the same way that we understand what Yoda means when he says, "Sorry I be but go you must." We register the disorder in the text but make sense of it by extrapolating a more conventional pattern. [>>36]

[36] lexia and dialectics

All hypertexts retain some degree of absolute sequence, some minimal dimension or granularity word, phrase, paragraph, chapter at which the reader is momentarily disengaged from the variability of the text. Following Roland Barthes, Landow calls these units lexia (1992). Lexia are units of local stability in the general flux of the hypertext, invariant moments in the larger pattern of textual displacement. The relationship between the lexia and links that connect them is implicitly dialectical, a dynamic opposition of forces. The local stability of the lexia arouse expectations of coherence and internal consistency, familiar hallmarks of print; but the operation of the link overturns these expectations, constantly throwing the reader into unfamiliar discursive territory, invalidating apparent structures of causality and necessity. No wonder hypertext seems so problematic to researchers interested in coherence and unity: the experience of hypertextual reading is fundamentally dissonant. [>>37]

[37] synthesis and contour

To say that hypertexts are dialectical implies that their nature may admit of a synthesis, some plane or aspect in which the play of opposing forces is harmonized or balanced. Is there some element of hypertexts, or of reading and writing hypertexts, which reconciles the stable sequentiality of the lexia and the unstable randomness of the links? Joyce proposes to understand hypertext in terms of the contour, a virtual representation of the reader's experience of the hypertext as it unfolds in time. The contour is not "non-sequential" at all, since it articulates the connectedness or "coextensivity" of the text; in fact, like syntax, the contour is time-dependent. It is perceived in time. But neither is the contour sequential in the exclusionary sense of older coherence structures (cadence, plot, argument). The contour does not exclude possibilities; it remains contingent as the reader moves back and forth through the text, revealing or realizing new connections. The contour unfolds each time through a process of replacement (Joyce 1992). [>>38]

[38] virtuality

So much for theory, but what does the contour mean in terms of hypertextual practice? What exactly is a contour? Can it be represented in concrete terms or shown by example? What are the contours of this text? Is a contour some specific set of relations between lexia, either realized or potential? Is it therefore a computable object?

These questions are not easy to answer. The contour is real but virtual, like the "virtual machine" created by a time sharing system: it exists only in actu as an articulation of the reader's interaction with the text. It corresponds in a rough way with what theorists of conventional reading call the "virtual work," a discursive construct arising out of the intersection of the reader's experience (or "repertoire") and the semantic content of the text (Iser 1978). [>>39]

[39] contour rhetoric

Does the contour have any value as a rhetorical concept? A rhetoric of contours might do considerably more than print-based conceptions to explain how hypertexts work and how we can make them useful. In particular, the idea of the object-event or contour might help us derive a better understanding of the semiotics of links, a subject on which more conventional approaches shed little light. [>>43]

[40] print analogues

Previous rhetorics of hypertext have proceeded on the assumption that links are directly analogous to prose transitions, page sequence or other connective structures in print (Slatin 1990). Thus Landow in his groundbreaking discussion of hypertext rhetoric proposes a system of "arrivals and departures" (1987) and Carlson insists that hypertext must evolve structures that duplicate the functionality of print (1990). These approaches conceptualize the link as a direct connection, a linear conduit from one lexial stability to another not surprising given that early hypertext models tended to implement links as fairly simple, hard-wired transitional functions. [>>41]

[41] beyond node/link

Later thinking has moved beyond the simple dualism of stable nodes and dynamic, overdetermined links. Halasz as early as 1987 calls for node/link aggregates or "composites" (1987) to expand the functionality of hypertext. DeRose outlines an extensive taxonomy of link types as a basis for creating implict (dynamically computed) links (1989). Most recently researchers have proposed discarding the node/link scheme altogether in favor of a set-based model of direct node intersections (Parunak 1991b) or "complex relations" in which the hypertext takes on many properties of a sophisticated database (Marshall et al. 1991).

These developments invalidate any rhetoric based on a simple congruence between links and transitional devices inherited from print. In what sense is a dynamically computed, implicit link analogous to turning a page? How can one specify a "rhetoric of arrivals and departures" for a multi-user system like Aquanet in which changes can "ripple through" multiple levels of hierarchy (Marshall et al. 1991)? [>>42]

[42] links and gaps

If rhetoric is to understand these more sophisticated structures emerging in hypertext theory, it must move beyond exclusively linear concepts inherited from conventional writing. As Landow points out, poststructuralist literary theory, itself an attempt to critique conventional writing structures, provides a rich conceptual base for hypertext rhetoric (1992). Terry Harpold's deconstructive approach to the linking problem suggests an important point of departure: "Wandering in a hypertext discovers traits of the text as a fabric of representations that are not evident if you don't dwell on the gaps in the fabric. Hypertext navigation means not only traversing a space between two points... it means as well electing to diverge from a predetermined course" (1991, 129).

Hypertext rhetoric must examine the gaps as well as the integument of links, the unrealized or virtual possibilities as well as those represented in present structures, the event of unfolding as well as the object that is disclosed. [>>3]

[43] object-event

The contour is not entirely an object and therefore it is not strictly speaking computable, though there may be ways to approximate or represent contours in terms of computable objects (Bernstein 1991). The contour is more like an event, an interactive transaction between the reader and the hypertext. But unlike the "virtual text" of print reading (which is all event), the contour does have a materialized component: the backtrail of lexia visited by the reader, a history which virtually all hypertext systems maintain as an articulable record. The contour is thus both object and event, a concept that defies strict categorization. It is in this regard that the contour becomes useful as a synthesis, a structure for apprehending both the dynamism and the stability of hypertext discourse. [>>44]

[44] contour and informand

A rhetorical theory of the contour augmented, perhaps, by a practical technique of contour representation and navigation could yield an important shift in our understanding of hypertext. It could allow us to move beyond the concept of the text as a fixed hierarchy (a transformation which collaborative, multi-user hypertexts will demand) while at the same time retaining a sense of the text as an articulated process or object-event. It could give us an important new way of understanding the text as what Roland Barthes called "the social space of writing" (1986, 64). Recognizing that the text has both a social and a spatial element is crucial to our understanding of hypertext as a kind of informand. [>>19]

[45] electronic version

Readers who would like to obtain the original hypertext version of this paper may download it by anonymous FTP from the directory pub/lcc on If you can't do this, send an 800k Macintosh disk and a stamped, self-addressed disk mailer to Stuart Moulthrop, School of Literature, Communication, and Culture, Georgia Institute of Technology, Atlanta, Georgia 30332-0165. [>>46]

[46] perfunctory closure

[47] contradiction

"Do I contradict myself? Very well, then, I contradict myself. I am large, I contain multitudes."

Walt Whitman, Song of Myself [>>1]

[48] science in action

But wait - can we really conclude that hypertext is not a useful medium for research communities simply because our own applications have been so limited and conservative?

We need a subtler analysis than the one just offered, one that regards research discourse as something other than a collection of inert report-objects. Latour develops such an analysis: though the literatures of research communities tend to be constructed out of generalized claims, Latour also maintains that science is ambiguous or Janus-faced. That is, science recognizes both the object and the process nature of knowledge. One the one hand scientific discourse avers that "When things are true they hold"; but at the same time it also affirms that "When things hold they start becoming true" (1987, 12). [>>49]

[49] i could be wrong

Drexler provides a vision of scientific hypertext completely at odds with my conclusions here he foresees a linked worldwide knowledge bank on nanotechnology research that would serve as a non-hierarchical medium for the ajudication of scientific claims. Drexler is so convinced of the usefulness of such a system that he describes it as the only way to control an apocalyptic knowledge explosion in nanotechnology (1987).

Perhaps Drexler is right and I am wrong. However, if we are to consider hypertext as a useful medium for research discourse, we will certainly need a better understanding of hypertext rhetoric. [>>42]

[50] adequate reading

Can we really do serious intellectual work in a polysemous, de-centered medium? If we wish to do so, we must develop conventions for review and assessment of hyperdocuments. This task in turn poses a more basic problem: what constitutes an adequate reading of a polysequential text? As Lai and Manber observe (1991), existing hypertext systems provide no analogue for the kind of quick survey that can be performed by flipping or scanning the pages of a print text. How can a reader derive a sense of the text as a whole if all its discursive pathways are not immediately apparent? Even if the options are comprehensively mapped, is a critical reader responsible for exploring every possibility? If we commit ourselves to this standard, might do we risk being overwhelmed by the convolutions of our own discourses?

The notions of the contour and the informand may be of some use here. The hyperdocument is an informand, an externalization of an evolving, indefinite "structure for what does not yet exist." It is a social space, a ground of possibility. The reader's encounter with the text takes the form of a contour, or more likely of several contours, which are specific expressions of the informand within the language of the text. The work of the critic is contiguous with both the contours of the text and with the informand those represent but the critic does not exhaust the text. We need not become paralyzed by totality. To the extent that it deals in evaluation, the rhetoric of hypertext is a rhetoric of approximation and representation, not one of totalization or summary. [>>40]

[51] robot, automaton

Zuboff's distinction between automating and informating parallels Baudrillard's discussion of "robots and automata" as representatives of two different uses of technology (1983). Though the overlap of terms is a little confusing, the parallel between Baurdillard's and Zuboff's schemes is exact: robotic technologies replace human workers while automata augment their abilities.

I applied this distinction in my critique of Joyce's hypertext fiction, afternoon, arguing that hypertext systems which do not allow what Nelson was later to call "free and knowing user movement" might have limited value for the future (1989). [>>52]

[52] i, robot

Then two years later I wrote my own hypertext fiction,Victory Garden, using a system of computed conditional links very much like the one in afternoon. In that particular rhetorical context (a long work of fiction) sustaining reader interest was very important and keeping the text "robotic" (i.e. mysterious in its operations) seemed a good strategy.

I admit to all kinds of hypocrisy. For a view of hypertext fiction that is closer to Zuboff's notion of informating, see Guyer and Petry 1991. [>>29]

[53] a serious question

Is this all a joke? Seriously, folks: this paper violates academic decorum in many, many ways. It's informal. It's elliptical. It's most egregiously un-empirical. It's self-reflexive and self-contradictory to boot. Even fragmentary [>>17]

Can you take this writing seriously? Should you even try? [>>54]

[54] q &a

I do find myself asking a lot of questions here. What does that mean?

Possible answers: (a) I don't really know what I'm doing; (b) I'm more concerned with the gaps in my knowledge than with its apparent continuities; (c) questions are a way of holding the discourse open: they may in fact be literal invitations to a subsequent reader/writer to intervene in my discourse (specifically in my electronic text); (d) rhetorical questions are a sneaky way to imply that the answers the writer provides are logically necessary without offering any proof but you knew that, didn't you? [>>55]

[55] mixed up

Though the "pragmatists" and the "artists" of the hypertext community have lived in amity up to now, they've largely kept on their own sides of the cultural fence. There's been reason enough for this diffidence. What computer scientist or knowledge engineer wants to dwell on the fact that the most popular multimedia hypertexts are things like Cosmic Osmo and Spaceship Warlock (Nielsen et al. 1991)? By the same token, what avant-thinking artist or cultural critic wants to embrace a technology whose strongest proponents are Boeing and Bellcore (Malcolm et al. 1991; Egan et al. 1991)? We've lived up to now in a kind of polite detente: one side has its visions and fictions, the other side its systems and science.

Now here I come trying to mix it up. Is this a good idea? [>>56]

[56] compare...

To compare small things with great, remember Ted Nelson's Computer Lib/Dream Machines, the flip book that (with due deference to Vannevar Bush, Douglas Engelbart, and Andries van Dam) helped start all this. Those Nelsonian manifestoes aren't very decorous though that was the point, wasn't it? You were supposed to see Nelson's self-published bricolages as great seminal hypertexts stuck in the glacier of the Batch Processing Age, waiting for the first spring thaw.

If it's animated by any reasonable impulse, this paper is inspired by the example of Nelson's crazy mixed-up books. [>>57]

[57] a theoretical joke

So what if this writing is indeed a kind of joke?

A practical joke: Ulmer proposes (and he's serious about this!) a rhetoric of multimedia which he calls "euretics" after the Greek heuresis, the faculty of invention or discovery. According to Ulmer, euretics is the result of a cultural revolution. Euretics embraces absurdist practices like puns and jokes because these uses of language have a positive potential to "blow up theory" and leave us sitting (eyebrows singed) in a world beyond rationalist Method. "Euretics," Ulmer says, "does not interpret art; it uses art for the making of theory" (1990, 20).

Is that what this is about? [>>58]

[58] but marshall mcluhan said...

"Humor as a system of communications and as a probe of our environment - of what's really going on - affords us our most appealing anti-environmental tool. It does not deal in theory, but in immediate experience, and is often the best guide to changing perceptions. Older societies thrived on purely literary plots. They demanded story lines. Today's humor, on the contrary, has no story line “ no sequence. It is usually a compressed overlay of stories" (McLuhan and Fiore 1967, 92). [>>59]

[59] so relax!

This project is more like McLuhan's "probe" than Ulmer's absurdist bomb. It isn't trying to blow anything up. It's playful (to a point) but also serious (to a point): if I've managed the trick, it should read more like satire than slapstick.

I ask you to consider only one outlandish thing before breakfast: that the critique of western rationalism that started in the sixties with people like Bateson, McLuhan, and Marcuse and became hermetic mystery during the seventies in the hands of Barthes, Derrida, Foucault, et al. “ this thing is not yet over. It has not gone away. Inchoate and self-defeating at first, the critique may (as Landow and Mark Poster both tell us) just recently have found its Text, which is to say, its technology (Landow 1992; Poster 1990).

If so, we have some thinking to do about our horizons. [>>60]

[60] whose revolution continues?

You say you want a revolution? Well, you know...

I am not saying that we should all, Tim Learylike, go tripping off to cyberspace just because somebody once called it a "hallucination" (Gibson 1984, 51). I do not suggest that hypermedia is the psychedelia or higher consciousness of the nineties, nor am I advocating some kind of hypercounterculture. Such ideas may be fun at parties but they're notoriously hard to live with, especially when you're an info-working man.

I've staked out a "radical" position here. How radical, and in what sense?

The anarchocritic Hakim Bey has proposed a "festal" withdrawal from the 21st century. He speculates that "the web," his fantasy of a global electronic underground, may be used to create "temporary autonomous zones," islands of "freedom" in an otherwise monolithic information culture. Could hypertext be part of this movement? Might what I am calling the informand be cognate with what Bey calls the Temporary Autonomous Zone (1991)? [>>61]

[61] negative

The Temporary Autonomous Zone, writes Bey, "is like an uprising which does not engage directly with the State, a guerilla operation which dissolves itself to re-form elsewhere/elsewhen, before the State can crush it. Because the State is concerned primarily with Simulation rather than substance, the TAZ can 'occupy' these areas clandestinely and carry on its festal purposes for quite a while in relative peace" (1991, 101).

No, no, and again no.

Elliptical and irreverent though it might be, my discourse of informands, object-events, and fragments is not dedicated to partying in the interstices or any other kind of dropping out. Though I believe that hypertext and other new media require us to re-think the boundaries of reason, science, and intellectual decorum, I would never assert (as does Bey) that "fight for the right to party" is a radical statement. [>>63]

[62] radical?

2-58-8-24-50-43-28-37-29-45-40-19-16-44-26-27-51-52-30-3-12-13-53-55-32-35-17-56-33-36-18-5-47-57-11-20-54-31-7-1. [>>2]

[63] substance abuse

"Because the State is concerned primarily with Simulation rather than substance..."

We are all concerned with Simulation, or we ought to be. Though the approach to hypertext I suggest here is divergent and perhaps (within the terms of our discourse community) radical, it makes no egregious "revolutionary" claims on "substance."

This is rhetoric, not revolution.

Substance was the domain of Automation. You and I inhabit the domain of Information and the informand, or as someone has memorably called it, Writing Space. And writing, as someone else says, "is pre-eminently the technology of cyborgs, etched surfaces of the late 20th century. Cyborg politics is the struggle for language and the struggle against perfect communication, the one code that translates all meaning perfectly..." (Haraway 1990, 165).

The rhetoric of hypertext. [62]


Akscyn, Robert (1991b). "The ACM Hypertext Compendium." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 403-405.

Akscyn, Robert, ed. (1991a). Hypertext Compendium [electronic document]. Association for Computing Machinery.

Barthes, Roland (1986). The Rustle of Language. Hill and Wang.

Berk, Emily and Joseph Devlin. "What Is Hypertext?" In E. Berk and J. Devlin, eds. The Hypertext/Hypermedia Handbook. McGraw-Hill. 3- 9.

Bernstein, Mark (1991). "Architectures for Volatile Hypertext." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 243-61.

Bey, Hakim (1991). T.A.Z.: Temporary Autonomous Zone, Ontological Anarchy, Poetic Terrorism. Autonomedia.

Bolter, Jay David (1991a). Writing Space [electronic document]. Lawrence Erlbaum and Associates.

Bolter, Jay David (1991b). Writing Space: The Computer, Hypertext, and the History of Writing. Lawrence Erlbaum and Associates.

Bolter, Jay David (1992). Private communication.

Bush, Vannevar (1945). "As We May Think." Atlantic Monthly (July, 1945): 101-08.

Carlson, Patricia (1990). "The Rhetoric of Hypertext." Hypermedia 2: 109-31.

Charney, Davida (1992). "The Impact of Hypertext on Processes of Reading and Writing." In S. Hilligoss and C. Selfe, eds. Literacy and Computers. Modern Language Association. [in press]

CortÁzar, Julio (1966). Hopscotch. trans. Gregory Rabassa. Avon.

DeRose, Steven J. (1989). "Expanding the Notion of Links." In N. Meyrowitz, ed. Hypertext '89 Proceedings. Association for Computing Machinery. 249-59.

Drexler, K. Eric (1987). Engines of Creation: The Coming Era of Nanotechnology. Anchor Books.

Egan, Dennis, Michael Lesk, Daniel Ketchum, Carol Lochbaum, Joel Remde, Michael Littman, and Thomas Landauer (1991). "Hypertext for the Electronic Library? CORE Sample Results." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 299-313.

Engelbart, Douglas (1991). "Knowledge-Domain Interoperability and an Open Hyperdocument System." In E. Berk and J. Devlin, eds. The Hypertext/Hypermedia Handbook. McGraw-Hill. 397-415.

Gibson, William (1984). Neuromancer. Ace Books.

Halasz, Frank G. (1987). "Reflections on NoteCards: Seven Issues for the Next Generation of Hypermedia Systems." In N. Yankelovich, ed. Hypertext '87 Proceedings. Association for Computing Machinery. 345- 67.

Haraway, Donna (1990). Simians, Cyborgs, and Women. Routledge.

Harpold, Terence (1991). "The Contingencies of the Hypertext Link." Writing on the Edge 2(2):126-39.

Iser, Wolfgang (1978). The Act of Reading: A Theory of Aesthetic Response. Johns Hopkins UP, 1978.

Joyce, Michael (1988). "Siren Shapes: Exploratory and Constructive Hypertext." Academic Computing, November, 1988: 11+.

Joyce, Michael (1992). "A Feel for Prose: Interstitial Links and the Contours of Hypertext." Address given at SUNY Buffalo, February 10, 1992.

Lai, Patrick and Udi Manber. "Flying through Hypertext." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 123-33.

Landow, George (1992). Hypertext: The Convergence of Contemporary Critical Theory and Technology. Johns Hopkins UP.

Landow, George P. (1987). "Relationally Encoded Links and the Rhetoric of Hypertext." In N. Yankelovich, ed. Hypertext '87 Proceedings. Association for Computing Machinery. 331-45.

Lanham, Richard (1989). "The Electronic Word: Literary Study and the Digital Revolution." New Literary History 20: 268-89.

Latour, Bruno (1987). Science in Action: How to Follow Scientists and Engineers through Society. Harvard UP.

Malcolm, Kathryn, Steven Poltrock, and Douglas Schuler (1991). "Industrial- Strength Hypermedia: Requirements for a Large Engineering Enterprise." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 13-25.

Marshall, Catherine, Frank Halasz, Russell Rogers, and William Janssen (1991). "Aquanet: A Hypertext Tool to Hold Your Ideas in Place." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 261-77.

McDaid, John G. (1991). "Breaking Frames: Hyper-Mass Media." In E. Berk and J. Devlin, eds. The Hypertext/Hypermedia Handbook. McGraw- Hill. 445-59.

McLuhan, H. Marshall (1964). Understanding Media: The Extensions of Man. McGraw-Hill.

McLuhan, H. Marshall and Quentin Fiore (1967). The Medium is the Massage. McGraw-Hill.

Moulthrop, Stuart (1991). "Beyond the Electronic Book: A Critique of Hypertext Rhetoric." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 291-99.

Nelson, Thedor H. (1987a). Literary Machines [electronic document]. Owl International.

Nelson, Theodor H. (1987a). Computer Lib/Dream Machines. Tempus Books.

Nelson, Theodor H. (1990). Literary Machines. Mindful Press.

Nielsen, Jakob, Lynda Hardman, Anne Nicol, and Nicole Yankelovich (1991). "The Nielsen Ratings: Hypertext Reviews." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 359-61.

Parunak, H. Van Dyke (1991b). "Don't Link Me In: Set Based Hypermedia for Taxonomic Reasoning." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 233-43.

Parunak, H. Van Dyke (1991b). "Toward Industrial Strength Hypermedia." In E. Berk and J. Devlin, eds. The Hypertext/Hypermedia Handbook. McGraw-Hill. 381-97.

Poster, Mark (1990). The Mode of Information: Poststructuralism and Social Context. U. Chicago Press.

Rous, Bernard, Ben Schneiderman, Nicole Yankelovich, and Elise Yoder. "Lessons Learned from the ACM Hypertext on Hypertext Project." In N. Meyrowitz, ed. Hypertext '89 Proceedings. Association for Computing Machinery. 385-87.

Slatin, John M. (1990). "Reading Hypertext: Order and Coherence in a New Medium." College English 52: 870-83.

Spiro, Rand J. and Jihn-Chang Jehng (1990). "Cognitive Flexibility and Hypertext: Theory and Technology for the Nonlinear and Multidimensional Traversal of Complex Subject Matter." In D. Nix and R. Spiro, eds. Cognition, Education, and Multimedia: Exploring Ideas in High Technology. Hillsdale NJ: Lawrence Erlbaum Associates. 163-205.

Streitz, Norbert, J‹rg Hannemann, and Manfred ThĖring (1989). "From Ideas and Arguments to Hyperdocuments: Travelling through Information Spaces." In N. Meyrowitz, ed. Hypertext '89 Proceedings. Association for Computing Machinery. 343-65.

Trigg, Randall and Peggy Irish (1987). "Hypertext Habitats: Experiences of Writers in NoteCards." In N. Yankelovich, ed. Hypertext '87 Proceedings. Association for Computing Machinery. 89-109.

Ulmer, Gregory (1990). Teletheory: Grammatology in the Age of Video. Routledge.

Wright, Patricia (1991). "Cognitive Overheads and Prostheses: Some Issues in Evaluating Hypertexts." In P.D. Stotts and R.K. Furuta, eds., Hypertext '91 Proceedings. Association for Computing Machinery. 1-13.

Yankelovich, Nicole, ed. (1988). Hypertext on Hypertext [electronic document]. Association for Computing Machinery.

Zuboff, Shoshana (1988). In the Age of the Smart Machine: The Future of Work and Power. Basic Books.