TOWARD A THEORY OF HYPERTEXTUAL DESIGN
Communication, Information & Library Studies
_Postmodern Culture_ v.3 n.2 (January, 1993)
Copyright (c) 1993 by Kathleen Burnett, all rights reserved. This text may be freely shared among individuals, but it may not be republished in any medium without express written consent from the author and advance notification of the editors. While the study of the temporal and spatial distanciation of communication is important to the concept of the mode of information the heart of the matter lies elsewhere. For the issue of communicational efficiency . . . does not raise the basic question of the configuration of information exchange, or what I call the wrapping of language. --Poster, 8HYPERTEXT/HYPERMEDIA  What distinguishes hypermedia from other modes of information is not that it is computer-driven--after all, the browsing and retrieval mechanisms of Vannevar Bush's memex were non-electronic--nor that it is interactive, since the entire history of oral communication, whether electronically mediated or not, might be characterized as interactive; nor even that it includes navigational apparatus such as links and nodes, which might better be thought of as symptoms than causes, or buttresses rather than groundwork. What distinguishes hypermedia is that it posits an information structure so dissimilar to any other in human experience that it is difficult to describe as a structure at all. It is nonlinear, and therefore may seem an alien wrapping of language when compared to the historical path written communication has traversed; it is explicitly non-sequential, neither hierarchical nor "rooted" in its organizational structure, and therefore may appear chaotic and entropic. Yet clearly, human thought processes include nonlinear, nonsequential, and interactive characteristics which, when acknowledged by traditional information structures, are not supported. In fact, one might characterize the history of information transfer as a tyranny against such characteristics, that is, a tyranny against the rhizome.  Hypermedia might be understood as one manifestation of the struggle against this tyranny. In current parlance, hypermedia is used to describe both applications which make use of navigational tools such as links and nodes to form "texts" or databases, and the organizational principles of such "texts" and databases. Hypertext is also used to denote these same meanings. When a distinction is drawn between the two, it normally focuses on content--"hypertext" is used to refer to hyper-structures consisting exclusively of written texts, while "hypermedia" denotes similar structures built around multiple media. Others have noted the artificiality of such a delineation. "Text" is also used as a synonym for a "written work" or "book" which may or may not be limited to alphanumeric characters. A "text" may included charts, graphs, illustrations, photographs, and other visual media in its expression of meaning. Why then should a "hypertext"--which has the potential for incorporating an even wider range of expressive media (sound, animation, etc.)--be limited to alphanumeric characters in its expression?  A more useful differentiation might be drawn along structural rather than contextual lines. Hypertext demonstrates "traits that are usually obscured by the enforced linearity of paper printing"; it is text--only more so--because it participates in a structure that resonates asynchronous and nonlinear relationships. Hypertext is a kind of weaving--"text" derives ultimately from the Latin %texere%, and thus shares a common root with "textile"--a structuring with texture--web, warp, and weave, allowing for infinite variation in color, pattern and material; it is the loom that structures the "text-ile." Hypertext is the organizational principle of hypermedia. Hypermedia is the medium of expression of a given hypertext structure. When that medium mirrors the singularity of the print medium of alphanumeric text, it may be properly called either "hypertext" or "hypermedia"; when the medium reflects an "intertwingling" (Nelson 31) of what we understand as separate "media" in the analog sense of the term, it should perhaps be referred to as "hypermedia," but might equally be acknowledged as "hypertext." Neither hypertext nor hypermedia is an object, rather the former is a structure, and the latter a medium, of information transfer. HISTORICAL CONTEXT  All electronically mediated exchange participates in hypertext, though the degree of participation varies enormously. Some electronically mediated exchange is "hypertextual" only to the degree that it is virtual--that it consists of a series of switches or codes (binary or otherwise) which are, in and of themselves, unreadable (and, therefore, nontextual), and which contain "pointers" to their reconstruction as meaningful exchanges. The switches or codes are "nodes" which are "linked" to a "textual" form which, at any given moment may exist only "hypertextually." Electronically mediated exchange is therefore paradigmatically different from other modes of information precisely because it participates in the organizing principle of hypertext.  In _The Mode of Information_, Poster proposes a concept which plays on Marx's theory of the mode of production:
By mode of information I similarly suggest that history may be periodized by variations in the structure in this case of symbolic exchange, but also that the current culture gives a certain fetishistic importance to 'information.' Every age employs forms of symbolic exchange which contain internal and external structures, means and relations of signification. Stages in the mode of information may be tentatively designated as follows: face-to-face, orally mediated exchange; written exchanges mediated by print; and electronically mediated exchange. (Poster 6)
Poster's periodization suffers from the coarseness of any totalizing metaphor. While he stresses the trans-historical nature of his classification of symbolic exchange, the metaphor is only as effective as it is historically informed. As outlined, the third stage--written exchanges mediated by print--is not only Western in its bias, but fails even within this bias to recognize a rather large chunk of history--the manuscript period (circa 4th century AD through the mid-fifteenth). An examination of the influence of the mode of information on social structure can only be enriched by the recognition of the impact of mass-production, in the form of the mechanized reproduction of written language, on that structure. It is impossible, however, to understand the full significance of this impact, either historically or theoretically, unless its contextualization is carefully discerned. For example, contrast these two very different experiences of the introduction of the hand-press and its effects on social stratification.  The pre-Reformation Church was able to maintain a restrictive social stratification largely because of its ability to control the production and comprehension of written communication--those who could read and write belonged to a privileged elite, while those who could not had to be satisfied with acquiring their information from those who did. Through most of the Medieval period and well into the Renaissance, the Church was able to control the size and membership of the elite through two mechanisms: Latin education and limited distribution of written communication. The latter was facilitated by production limits imposed by the rigorous and time-consuming process of hand-copying, which in turn limited the supply of reading material. Without supply, the demand for education was kept to levels that the Church could manipulate and control. The introduction of the hand-press in the mid-fifteenth century was accompanied by a precipitous erosion of that control which led decisively to the Reformation. Once reading material could be produced in large quantities in a relatively short period of time--500 to 1000 copies of an average-length manuscript could be produced by a printer owning two hand-presses within the space of less than a month, as compared to the production of a single copy of a manuscript, which could take up to a year--in other words, once the non-elite were able to acquire material to read, they began to do so. Printers, recognizing the commercial potential of this new market, began to produce material in the vulgate, which in turn expedited exponential growth in the educated population, since it facilitated the process of self-education. As this population grew, demands for equity in education across social classes escalated. The earliest signs of this movement are evident in the growth of the popular and self-help literature markets, and the introduction of mass communication, across time and distance, over which the Church could ultimately exercise little effective control (Eisenstein).  Contrast this experience with that of the introduction of a hand-press in colonial Massachusetts in 1660 for the express purpose of propagating the gospel among the Indians, who had no written language. The social stratification which existed within the tribe prior to the introduction of the press was anchored in the individual's ability to communicate with the spiritual realm and was maintained through oral mediation of the ritual culture. After the introduction of the press, the very foundations of that stratification were undermined. A schism developed between those who subscribed to the gospel, and thus to the notion of a single god, and those who continued in the old beliefs. Since the introduction of the very act of written communication was inextricably tied to the new religion, many who did not endorse the Christian faith simply refused to acknowledge the new mode of information.  Clearly the introduction of the hand-press in this context did not have the effect of popularizing written communication that it had in western Europe on the eve of the Reformation. While differences in the social structures of the two cultures might be cited as the major contributing factors in this differentiation, the privileged status of chirography in pre-Reformation Europe clearly at least served to buttress the social structure of that culture, while the absence of any form of written culture in the case of the Native American tribe equally served to buttress a quite distinct social structure. Both structures were undermined by the introduction of a new mode of information, but in very different ways. While a totalizing metaphor may be put to effective use in an account of this differentiation, Poster's four-stage delineation is simply too coarse to serve. Clearly, a distinction must be drawn between a culture which partakes only of oral exchanges and one in which oral exchange is coupled with some form of written exchange. Equally clearly, a similar distinction needs to be drawn between written exchanges mediated by chirographic writing and written exchanges mediated by typographic writing. The latter of these could be further subdivided into two stages: the first mediated by hand-press reproduction, and the second by machine-press reproduction. The importance of this latter distinction is borne out by the study of the growth of literacy in nineteenth-century Europe following the introduction of the mechanized press (cf. Altick and Eisenstein).  Between Poster's third stage--written exchanges mediated by print--and his fourth--electronically mediated exchange--lies much of the nineteenth and twentieth centuries, for although he does at one point acknowledge the nineteenth century origins of electronically mediated information systems in the telegraph and photography (19), his analysis of such systems is limited to the telephone, television advertising, databases, computer writing and computer science. The inclusion of the machine-press production stage suggested above accounts for a large share of the information technology of the nineteenth century, but the end of that century and the first half of the next, it seems to me, several quite distinct modes of information transfer have emerged which may help to provide a bridge from written exchanges to electronically mediated exchange and, particularly, to multimedia exchange mediated electronically.  We might group the various non-computer modes of information available in the twentieth century in a variety of ways; I would like to propose one such classificatory scheme based, as is Poster's, on the wrapping of exchange:
the stages are not 'real,' not 'found' in the documents of each epoch, but imposed by the theory as a necessary step in the process of attaining knowledge. In this sense the stages are not sequential but coterminous in the present. They are not consecutive also since elements of each are at least implicit in the others. The logical status of the concept of the mode of information is both historical and transcendental. In that sense the latest stage is not the privileged, dialectical resolution of previous developments. In one sense, however, a sense that Marx anticipated, the current configuration constitutes a necessary totalization of earlier developments: that is, one cannot but see earlier developments from the situation of the present. The anatomy of the mode of electronic information . . . necessarily sheds new light on the anatomy of oral and print modes of information . . . . I prefer to consider the present age as simply an unavoidable context of discursive totalization, not as an ontological realization of a process of development. (6-7) THEORIZING  From within this context of discursive totalization, other possibilities suggest themselves. In _A Thousand Plateaus_ (1970), Deleuze and Guattari propose a different history of written exchange. "Writing," they claim, "has nothing to do with signifying. It has to do with surveying, mapping, even realms that are yet to come" (4-5). Their history is delineated in terms of types of books. There are three types of books, the first being historically the earliest and the third the most recent, but all three are coterminous in the present. The first type they describe as the root-book. The root-book "imitates the world, as art imitates nature: by procedures specific to it that accomplish what nature cannot or can no longer do" (5). The second type is the radicle-system, or fascicular root book. "This time, the principal root has aborted, or its tip has been destroyed; an immediate, indefinite multiplicity of secondary roots grafts onto it and undergoes a flourishing development" (5). The approximate characteristics of Deleuze and Guattari's third book type--the rhizome--clearly indicate a departure from the book as printed codex to electronically mediated exchange: 1. and 2. principles of connection and heterogeneity; 3. principle of multiplicity; 4. principle of asignifying rupture; and 5. principles of cartography and decalcomania. (7-9) The significance of this taxonomy for this discussion is that its classification, unlike Poster's, is entirely media-independent, gaining its meaning, so to speak, from a delineation of structure or design.  The root-book roughly corresponds to written communication prior to the development of the paste-up technique (which Deleuze and Guattari refer to as assemblage; 4) in the early part of the twentieth century. Its history is one of linear production. In its earliest form, the writing of the root-book was synonymous with its publication. Today, the production of the root-book is still characterized as a linear process consisting of five steps: 1. writing of a manuscript; 2. submission/editing of the manuscript; 3. the composition of the manuscript in type; 4. the proofing of the type sheets; and 5. the dissemination of the publication. The production process for the radicle-system book is much lengthier, requiring the addition of at least two additional steps, the first, the mock-up or layout stage normally falling between the second and third root-book steps; and the second, the paste-up stage falling between the third and fourth steps in the production of the root-book. In its earliest manifestations (and still today in the certain fine-printing and vanity publishing circles), the production of the root-book is characterized by oneness and stability. Even in its more recent manifestations, the root-book strives to be an exact replica of the author's words, a representation or signification of an individual's thoughts. Even as the production process has fragmented (through the intervention of editors, publishers, printers who are not the author), it has maintained its linearity. Likewise, the publication has retained its insularity and rootedness.  In contrast, the design of the radicle-system book is fragmented and multifarious, and while representation is still employed as an element, it is only one of many couched in layers that problematize its signification. Interestingly, the technology which initially enabled this kind of production was photography. The production process is less emphatically sequential, the organizing principle being collage or assemblage which allows for alteration and reorganization at almost every stage of the production process. In some cases this process has extended even to the composition of the manuscript itself, as in the case of William S. Burroughs's cut-up texts, or, in a less mechanical implementation, in the poetry and critical writings of Rachel Blau DuPlessis.
 Deleuze and Guattari describe a third type of book: A system of this kind could be called a rhizome. A rhizome as a subterranean stem is absolutely different from roots and radicles. Bulbs and tubers are rhizomes. Plants with roots or radicles may be rhizomorphic in other respects altogether . . . . Burrows are too, in all their functions of shelter, supply, movement, evasion, and breakout. The rhizome itself assumes very diverse forms, from ramified surface extension in all directions to concretion into bulbs and tubers . . . . The rhizome includes the best and the worst: potato and couchgrass, or the weed. (6-7)
Telecommunications systems are rhizomorphic, as are computer networks. Think of maps you have seen and descriptions you have heard of the internet--a rhizome. If we accept the rhizome as a metaphor for electronically mediated exchange, then hypertext is its apparent fulfillment, and Deleuze and Guattari's "approximate characteristics of the rhizome"-- principles of connection, heterogeneity, multiplicity, asignifying rupture, and cartography and decalcomania--may be seen as the principles of hypertextual design. PRINCIPLES OF CONNECTION AND HETEROGENEITY  The principles of connection and heterogeneity state that "any point of a rhizome can be connected to any other, and must be" (Deleuze & Guattari 7). In this sense a rhizome is very different from a tree structure, where the order is fixed by a hierarchy of relationships. Cognitive jumps, which must be mechanically forced in an hierarchy, are intuitively sustained in a rhizome. A rhizome is the only structure which can effectively sustain connections between different media without giving hegemony to language. Many current relational and flatfile multimedia database applications support the storage of multiple forms of media, and some will even display different types contiguously, but keyword searching is the only mechanism provided for cross-type searching. Like film and video, they support synchronous display (but then, so can the book, albeit with limitations), but they do not support nonverbal access. Traditional hierarchical database structures are even more problematic in their support of nonverbal expression. Meaningful formation of hierarchies across media boundaries can be accomplished only through the use of language, since hierarchy is itself a creation of language, and therefore, language is the only universal tool available within an hierarchical structure. A rhizomorphic structure, on the other hand, does not rely on language for its ordering, although many of the linkages in a given structure may be linguistic.  A rhizome ceaselessly establishes connections between semiotic chains, organizations of power, and circumstances relative to the arts, sciences, and social struggles. A semiotic chain is like a tuber agglomerating very diverse acts, not only linguistic, but also perceptive, mimetic, gestural, and cognitive; there is no language in itself, nor are there any linguistic universals, only a throng of dialects, patois, slangs, and specialized languages (7).  Hypermedia design is rhizomorphic in its sustenance of heterogeneous connection, because there is no systemic hierarchy of connection. The perception of connectivity is entirely left to the user, though the pre-existence of particular connections may foster varying user perceptions of overall structure. At its most political, connectivity is a democratizing principle. It functions as a structure of individuation since at any given moment the "center" of any rhizomorphic structure is the individual's position in relation to that structure. Distinctions between author and reader, constituent and politician, even intermediary and end-user disintegrate as the reader participates in authorship, constituent in %polis%, and end-user in the search itself. At its worse, connectivity inspires anarchy. Witness (as we all did) the impact of limited connectivity (exclusive of the important element of interactivity) via the broadcast of a videotape of the arrest in the case of the aftermath of the Rodney King verdict.  As the distinctions between participant/viewer, author/reader blur, the concept of authorship itself will be problematized. All paths through hyperspace are equally valid to the individual traveller. As the "reader" negotiates hyperspace he/she becomes a navigator--traversing established links to pre-existent nodes; but also an explorer--creating new links to previously known, but unrelated territories; a pioneer--venturing forth into uncharted realms; and a visionary--imagining and giving shape to the as-yet unknown. PRINCIPLE OF MULTIPLICITY Act so that there is no use in a centre . . . . --Stein, 63 A multiplicity has neither subject nor object, only determinations, magnitudes, and dimensions that cannot increase in number without the multiplicity changing in nature . . . . An assemblage is precisely this increase in the dimensions of a multiplicity that necessarily changes in nature as it expands its connections. There are no points or positions in a rhizome, much as those found in a structure, tree or root. There are only lines. --Deleuze & Guattari, 8  Hypertextual design is able to support non-hierarchical thinking and cognitive jumping because it recognizes the diversity of multifarious modes of information. Information may be structured hierarchically within a hypermedia system, but only to the extent that such a structure exists in a coterminous relationship with other structures. In other words, hypertextual design presupposes not only that multiple points of access are preferable to a single point, but by extension, that multiple structures are preferable to a single structure. Information retrieval studies have shown that a single user's selection of access points for a given topic may vary over time and space, making it difficult for an indexer to predict potential user vocabulary. The principle of multiplicity is reflected in hypertextual design by the coterminous presence of varying modes of access to a single structure on the one hand, and of varying structures on the other.  Landow and others have noted the hypertextual nature of pre-hypertext literary projects from Sterne's _Tristram Shandy_ to Robert Coover's _Pricksongs and Descants_. Yet the lists I have seen are conspicuous in their omission of female writers and feminist critics, not to mention writers of color. I have already mentioned Rachel Blau DuPlessis, but there are others who might be mentioned as well--Emily Dickinson, Gertrude Stein, Zora Neale Hurston--all of whom practice a writing of inclusion and fragmentation, of absent centers and centered absence. Multiplicity, as a hypertextual principle, recognizes a multiplicity of relationships beyond the canonical (hierarchical). Thus, the traditional concept of literary authorship comes under attack from two quarters--as connectivity blurs the boundary between author and reader, multiplicity problematizes the hierarchy that is canonicity. PRINCIPLE OF ASIGNIFYING RUPTURE  Hypertextual design intuitively supports two forms of access which must be forced in hierarchical structures: user-generated access and mapping. The principle of asignifying rupture supports the former, and those of cartography and decalcomania, the latter. In an hierarchical structure, a user-generated access point may cause a rupture in the system. For example, in a database search, a user may, through the process of serendipity, arrive at a particular point in a hierarchy, even though her departure-point has no apparent hierarchical relationship to that arrival point. If she is allowed to introduce a link from her departure term to her arrival point into the hierarchy without further evaluation, the very structure of that hierarchy might well be undermined. One might view the project of feminist criticism in this light. The introduction of non-canonical texts and authors into the canon disrupts the foundations of the canon altogether. In contrast, hypertextual design encourages such disruptive activity while rendering it insignificant. Since the structure does not rely on any given theory of relationship, it cannot be affected by the characterization of a new relationship previously alien to it. The potential for any relationship exists within the hypertextual structure; some simply await unmasking. PRINCIPLES OF CARTOGRAPHY AND DECALCOMANIA  The second form of access not easily supported within an hierarchy is mapping. Tracings or logs of an individual's progress through an hierarchical database are of course possible and may help a user to retrace a given path, or provide useful data for research in human-computer interaction. Current maps of search paths exist in the form of recordings of transactions, though the best systems record only the user query and the system response, without making a record of the context of either query or response. The records thus constructed are divorced from context, non-relational, and perhaps most importantly, non-spatial. They are grammatic, rather than diagrammatic. They perpetuate the hegemony of language and de-emphasize the sense of a journey through space and time. Deleuze and Guattari's notion of mapping is, however, quite different, and presupposes the operation of the principles discussed previously.  Each user's path of connection through a database is as valid as any other. New paths can be grafted onto the old, providing fresh alternatives. The map orients the user within the context of the database as a whole, but always from the perspective of the user. In hierarchical systems, the user map generally shows the user's progress, but it does so out of context. A typical search history displays only the user's queries and the system's responses. It does not show the system's path through the database. It does not display rejected terms, only matches. It does not record the user's psychological responses to what the system presents. On additional command, it may supply a list of synonyms or related terms, but this is as far as it can go in displaying the territory surrounding the request. It can only understand hierarchy, so it can only display hierarchical relationships. What distinguishes the map from the tracing is that it is entirely oriented toward an experimentation in contact with the real. The map does not reproduce an unconscious closed in upon itself; it constructs the unconscious. It fosters connections between fields, the removal of blockages on bodies without organs, the maximum opening of bodies without organs onto a plane of consistency. It is itself a part of the rhizome. The map is open and connectable in all of its dimensions; it is detachable, reversible, susceptible to constant modification (12).  A hypertextual map is more closely related to geographic maps than to search histories. It shows the path of the user through the surrounding territory, but always from the point-of-view of the user. It is as though the map were perpetually shifting as the traveller moved from one quadrant to the next. Some of that territory is charted--it is well mapped out in terms that the user understands, and connected to familiar territory or nodes--and some is uncharted, either because it consists of unlinked nodes that exist in the database much as an undiscovered island might exist in the sea, disconnected from the lines of transfer and communication linking other land areas, or as an unidentified planet in space, with the potential for discovery and even exploration, but as yet just a glimmer in the sky--or because it is linked in ways that are meaningless to the user in his present context. The user can zoom in on zones of interest, jump to new territories using previously established links or by establishing new links of his own, retrace an earlier path, or create new islands or nodes and transportation routes or links to connect them to his previous path or the islands or nodes charted by others.  The rhizome operates by variation, expansion, conquest, capture, offshoots. Unlike the graphic arts, drawing, or photography, unlike tracings, the rhizome pertains to a map that must be produced, constructed, a map that is always detachable, connectable, reversible, modifiable, and has multiple entryways and exits and its own lines of flight. It is tracings that must be put on the map, not the opposite. In contrast to centered (even polycentric) systems with hierarchical modes of communication and preestablished paths, the rhizome is an acentered, nonhierarchical, nonsignifying system without a General and without an organizing memory of central automaton, defined solely by a circulation of states (21).  Hypertext is rhizomorphic in all its characteristics. Its power derives from its flexibility and variability; from its ability to incorporate, transmute and transcend any traditional tool or structure. Like the rhizome, it is frightening because it is amorphous. The hierarchical systems we are accustomed to are definitional--they are centers of power. Knowledge of the hierarchy engenders authority; corrupted authority breeds despotism. Knowledge of the rhizome as a totality is impossible, precisely because "totality" and other absolutes have no meaning in a rhizome. The rhizome is as individual as the individual in contact with it. It is that individual's perception, that individual's map, that individual's understanding. It is also, and at the same time, a completely different something--another individual's perception, another individual's map, another individual's understanding. It provides no structure for common understanding. It is a state of being, reflective always of the present, a plateau in a region made up entirely of plateaus--"a continuous, self-vibrating region of intensities whose development avoids any orientation toward a culmination point or external end" (Deleuze & Guattari 22). -----------------------------------------------------------
WORKS CITEDAltick, R. _The English Common Reader_. Chicago: U of Chicago P, 1957.
Bush, V. "As We May Think." _Atlantic Monthly_ 176 (July 1945): 101-8.
DuPlessis, R. _Tabula Rosa_. Elmwood, CT: Potes & Poets Press, 1987.
---. _The Pink Guitar: Writing as Feminist Practice_. New York: Routledge, 1990.
Deleuze, G. & Guattari, F. _A Thousand Plateaus_. Minneapolis: U of Minnesota P, 1987.
Eisenstein, E. _The Printing Press as an Agent of Change_. Cambridge: Cambridge U P, 1980.
Landow, G. _Hypertext: The Convergence of Contemporary Critical Theory and Technology_. Baltimore: Johns Hopkins UP, 1992.
Nelson, T. _Dream Machines_. Redmond, WA: Tempus, 1987.
Poster, M. _The Mode of Information_. Chicago: U of Chicago P, 1990.
Stein, G. _Tender Buttons_. Los Angeles: Sun & Moon Press, c1914.
Vandergrift, K. "Hypermedia: Breaking the Tyranny of the Text." _School Library Journal_ 35:3 (Nov. 1988): 30-35.