Patterns of Hypertext and their Impact on Reading Activities

Karin Wenz
English Department
University of Kassel
Georg-Forster-Str. 3
Kassel, 34119

Following Mark Bernstein's taxonomy of patterns of hypertext I will investigate their specific influence on the user's actions and reader's experiences reading a printed version and a multimedia HTML-hypertext version of the same text. The aim of my investigation is to highlight the reading of hypertext by (1) recording and following the ways of navigation in hypertext by using log files as a primary source to communicative processes and (2) coordinating these results with narrative reports governed by a questionnaire as a secondary reconstructive action.

The text I've chosen is Carolyn Forché's epic poem "The Angel of History". One group of students has read the printed poem first and the hypertext version second; this setting was repeated with a second group in reversed sequence (reading first the hypertext and second the text). The different strategies can be followed by log files, which record the number of screens visited and the amount of time spent with each screen. These data will provide means for capturing processes of document search, reading, contemplation, and exploring action. However, measuring only the reading time and path does not suffice to interpret the data unambiguously. This is the reason why a second step, the analysis of "reader" reports related to their navigational experience, motivation and intention, has to be added. There will be an additional setting with a group of reader's reading the hypertext version without any multimedia effects and one reading the poem as a digital text in linear version in front of the screen. This extension is necessary to judge, in how far not just the medium used but the mixture of sign systems lead to different results. The storing of navigation paths as a primary resource for the user's action and the narrative reports about the experience with the printed text and the hypertext version leads to protocols about a natural setting of media reception .

Mark Bernstein developed a taxonomy of patterns of hypertext, which have a specific influence on the user's actions. These patterns are elements hypertext consists of. Their presence or absence influence the reader's choices and offer specific interactive possibilities. This interaction is governed by the rules of the software which the specific hypertext requires. The following patterns are components my hypertext consists of. These are seen in relation to the specific activity they seem to support: (1) The tangle offers a variety of links without specifying information where the link is pointing to. Therefore tangles lead to a disorientation of the reader. The tangle supports the search strategy of browsing through the text. (2) A counterpoint is given by the two voices that alternate in the poem. A third voice is added by using elements (as background color of one voice and text color of the second) to highlight the convergence of both voices, as they seem to merge in the second part of the poem. Another possibility to differentiate between both voices is layers that I used in one and the same lexia to maintain the closeness spatially by dividing the voices at the same time. (3) Neighborhood, and therefore coherence, is created by the use of a specific design (color, visual motifs) for each voice in the poem. This supports the reader's understanding of who is speaking. (4) Mirrorworlds are parallel narratives or intertextual lines, which echoes one theme more in depth. The mirrorworld I added to the poem consists of passages from Walter Benjamin's work and documentary material about the holocaust, which both add new voices to the poem. The mirrorworld offers a possibility of entering and leaving a new space, which leads to an exploring behavior more than it supports close reading. (4) The montage, which opens up several writing spaces at once, is a possibility to highlight a motif or a metaphor used in the text. Montage is used here to focus the user's attention and to develop a special motif in depth. This complexity supports a closer reading of the lexia instead of a browsing activity. (5) The cycle is used to mirror the dialogic cycle and repetition of elements in the printed poem. The cycle leads the reader to a central page, which then will offer new reading possibilities. The cycle fosters browsing strategies as soon as readers recognize that they already have seen these lexias before. The log files will show if the reader's behavior is changing from their first visit of the lexia, which usually means reading and spending more time with each lexia, and the second visit, which lead to a browsing strategy.

Bernstein's categories concern different aspects of hypertext. Cycle and counterpoint have clearly to do with the shape of the underlying map of nodes and links and can be described as structural or syntactic patterns. But the difference between counterpoint and mirrorworld is more of a thematic or semantic nature: counterpoint means two alternating voices talking about different topics, while mirrorworld means different perspectives on the same world. Montage is mostly a matter of visual presentation and seems compatible with various types of map structure. Therefore I suggest to relate Bernstein's patterns to different levels of hypertext.

Structural PatternsThematic PatternsVisual PatternsExpectational Patterns
cyclemontagemontagemissing link

Montage and feint are clearly patterns related to two different levels of hypertext. Both mark a transition: montage from the thematic to the visual level and feint from the thematic to the expectational level.

As the patterns of hypertext the multimedia elements have a significant influence on the reader's activity, if the time spent with a single lexia is regarded. The amount of text offered in a single lexia (from 5 words to 250) has no influence on the time spent with "reading". Pattern recognition seems to be a strategy chosen by most readers. As hypertexts seem to foster a type of reading between improvisation and planning, we need more open models of reading, which a comparison of reading a text, playing a game, and browsing a screen may offer. Both, the hypertextual structure, the sign systems used and the technology chosen, have an important impact on the reading process.


Aarseth, Espen. 1997. Cybertext: Perspectives on Ergodic Literature. Baltimore: John Hopkins.

Andersen, Peter Bogh 1995. The Force Dynamics of Interactive Systems: Toward a Computer Semiotics. Semiotica 103, 1-2: 5-45.

Bernstein, Mark. 1998. "Patterns of Hypertext." In Proceedings of the Ninth ACM Conference on Hypertext and Hypermedia, (Pittsburgh, June 20-24), 21-29. Pittsburgh: ACM Press.

Burbules, Nicholas Constantine. 1998. "Rhetorics of the Web: Hyperreading and Critical Literacy." In Page to Screen, edited by Ilana Snyder, 102-122. London: Routledge.

Hess-Lüttich, Ernest W.B. 1997. "Text, Intertext, Hypertext: Zur Texttheorie der Hypertextualität." In Androiden: Zur Poetologie der Automaten, edited by Jürgen Söring, 53-83. Frankfurt/Main: Lang.

Landow, George P. 1997. Hypertext 2.0 . John Hopkins University Press, Baltimore.

Larsen, Deena. 1998. Workshop: Toward a Hyperliterate World: Expectations of Readers, Writers, and Programmers. Siglink Newsletter 7, 1-2: 21-24.

Santaella, Lucia. 1998. "Der Computer als semiotisches Medium." In Medientheorie und die digitalen Medien
Stiftung Lesen, ed. 1998. Lesen im Umbruch: Forschungsperspektiven im Zeitalter von Multimedia. Baden-Baden: Nomos.