The Cartography of Narrative

Regardless of whether one has read a single word of critical theory about hypertext, one cannot long write it and think about it at all seriously without beginning to need to visualize its structure. This is true even for the conventional hierarchical structures that predominate on so many Web indices and other reference sources. But once you get into the complex relationships within large narratives, the need for the writer to somehow "see" his structure becomes urgent. And, in many cases, it it urgent that the writer consider this for the reader's sake, too.

Mapping a Narrative on the World Wide Web

For the writer used to working solely with text, mapping a narrative currently poses a number of (heretofore) non-writerly problems in mounting a piece of fiction on the World Wide Web. Consider, for example, the main map for navigating "Holier Than Thou," shown below in Figure 1.

Figure 1

This map demonstrates one possible spatial structure of "Holier Than Thou." Spatial visualization of hypertexts is not impossible on the Web, but current browsers certainly obscure these structures, both to readers and writers. Since there is currently no direct way to create such an image with a Web browser, other software must be employed both for creating the image and, if necessary, converting it to a format that Web browsers can interpret. In this case the image is a screen capture of a document created in Storyspace, a hypertext writing system developed by Eastgate Systems. I then converted into the .gif image format common on the Web. Now I had an image showing the structure I wanted to reveal and was ready to make the image click sensitive. This is done by defining certain areas (using pixel addresses on the horizontal and vertical axes) to be linked to certain nodes.

In a system such as Storyspace, this feature is built in. Its usefulness is that it lets you easily see the spatial nature of a hypertext, which can be advantageous for a number of reasons. It is first of all a visual reminder that, unlike in print, thinking a single sequence of pages (or files or nodes) is misleading. The diagram in Figure 1 is, to my mind, a truer depiction of the structure of "Holier Than Thou," though I stress that this is only one possible arrangement of the nodes. I've chosen to put the node "Nelson" in the middle since the character Nelson Tucker is the ostensible reason for the existence of all these narratives. The blue nodes in the inner circle are various recurrent images, actions, characteristic traits, and so on that the various narrators note about Nelson or connect to him in some way. For purposes of visual clarity, I have chosen not to show the actual links between (or within) nodes. Note that these nodes have others (in green) inside them--these are individual instances of the motif named in the parent node, and are the timed links I discuss elsewhere. Currently, there are some eighty nodes and perhaps 350 links in "Holier Than Thou." This is not an especially large or complex web compared to many published by Eastgate (or mounted on the Web), but an image of the same map with the links showing gives one a sense of the complex relationships that can be visualized in even a relatively brief and sparse hypertext. The spatial rendering also gives one a sense of the number of arrival links to a node. On the Web, readers normally only see departure links, and this can be very misleading in trying to discern the crucial episodes in a hypertext fiction. Which is more important, the section of narration that refers to many others, or to which many others keep leading the readers back? Allowing the readers to see the density of links gives them the knowledge to consider such questions.

This is all rather mundane stuff for people who spend a lot of time enlivening their web sites with graphics, and my description of the process is not meant to be a technical one. That's not my point here. The point, rather, is that while spending all this time locating and learning (or, more accurately, familiarizing myself with) three different software packages (Storyspace to create the image, Graphics Converter to get it into the proper format for the Web, XV to define the pixel coordinates to attach to links) I expended a lot of time and energy and wandered far away from what, in print, I would have considered "writing." Such methods demand a considerably larger investment of time than if spatial writing and reading were incorporated into HTML and the various browsers. And the final product is dynamic only for the reader, not the writer. Once I have painstakingly created this image map I cannot change links by simply dragging a box to a different place with my mouse--so God forbid I should change my mind. Obviously, this cannot be a writing process except on special occasions, and I only offer it here for readers as an example of the spatial or topographic nature of hypertext.

Mapping the Links

In the map in Figure 2 the nodes are shown in the same spatial relationships as in the main map (Figure 1), but reduced in size so that there is more space to show the arrows or links. (Note: this image has not been converted into an image map and thus cannot be used for navigation). Like the image map, this one was created in Storyspace. Many of the relationships between nodes would be much easier to discern if one were actually viewing the map in the Storyspace software, as it has built-in features that allow a reader to view lists of the arrival and departure links for an individual nodes. Even without that software, however, one can gain a general sense of the structure of the story by noting where the arrows tend to cluster.

Figure 2

Note a couple of other things about this map. One is that not all the links are shown here. Within the blue nodes in the inner circle are many other nodes (in green) that also link to each other. If I brought those nodes out of their current hierarchical arrangement in this diagram and made all of them visible at the same size, the picture would be much denser with links and nodes. It would also give the reader a sense of how, as I discuss in talking about the changing nature of the writer's workspace, hypertext is having a Big Bang effect on writing. Contrary to the goal in writing for print of herding all ideas into one document with one recommended reading sequence, writing in hypertext tends to explode the text. Related themes and motifs and mini-narratives go flying out from the center--or, perhaps, they simply stay out in their natural orbit, where they first formed in the writer's conception of the textual web, related to other nodes but not held rigidly in place.

A second point of interest in this representation is that no arrows point to the "node" labelled "Nelson." Representing the character of Nelson as a node is actually misleading on my part--there is not a node named Nelson--but I have done it for a particular purpose. His character is at the center of the story. Link arrows simply pass over this center rather than connecting directly to it, suggesting that the character Nelson acts as an intersection in which different intepretations of him--and the world--cross each other's path.

Navigation

Abstract | Bibliography | Glossary | "Holier Than Thou" | Project Entrance

Version Notes

Initial release: May 27, 1996
Last updated: July 7, 1996