Genres are constituted by many things besides form. Miller suggests that readers' and writers' expectations are much more important to constituting a genre than is form--expectations about language, rhetorical stance, what problems are legitimate to talk about, what claims may be advanced to solve them and what proofs may be used to support them. But form is nonetheless an extremely important component of genre. As Coe argues in "An Apology for Form," form in this sense is not just a matter of learning how to "get on" in a discourse community. Form as genre is heuristic: by providing the "empty spaces" that writers must fill in with certain kinds of knowledge and not others, it models an epistemology.
Form is both generative and constraining--or, better said, generative because constraining. Form is empty, an absence. But this emptiness has shape (i.e., form). In human beings, at least, this emptiness creates a desire to find what might fill it.
("An Apology for Form," 17)
Form is thus a socially significant way of thinking:
Formal structures are in this sense the social memory of standard responses to particular types of rhetorical situations and subject matter. . . . One function of discourse communities is to provide, prescribe, and prefer forms. Learning conventional forms . . . is a way of learning a community's discourse, gaining access, communicating with that community.
("An Apology for Form," 19)
This puts a significantly different spin on the idea of genre, locating the social practices of a discourse community more securely in the formal shape of its documents than other genre theories might wish to.
In one sense, hypertexts have no formal shape. The reader finds her own way through the text, creating a non-linear garden of forking paths as she reads. Form is written on the fly by the reader rather than coming the writer who in turn received it from a discourse community. This "do-it- yourself" aspect of hypertext form could mean that, if argumentative, scholarly, philospohical (call it what you will) discourse migrates to hypertext (not just linear text posted on the web), we will lose an important part of our social memory.
But is this really a good analysis of hypertext? Certainly hypertext has no form in the sense of what ought to be read first, second, third. But perhaps this is too restrictive an analysis of form.
The rhetoric of hypertext is still evolving (as are all rhetorics). But there are already some fairly well defined conventions about how information in a hypertext may be structured. And the reader's choices, while highly varied, are far from infinite. They are constrained by the shape of the nodes and the links that the author has chosen. Barnes, for instance, argues that
Replacing ambiguous choices with predictable ones, improves reading in general. According to Smith (1985), prediction is an essential reading skill. We predict because the world is full of ambiguity and we can become overwhelmed with possibilities. A second "reason for prediction is that there would otherwise be far too many alternatives to choose among" (p. 77). Hypertext designs eliminate alternatives by linking together texts based on a theme or topic. By enabling students to select texts based on a related topic, they will not become overwhelmed with information. In contrast, students can gain a greater sense of control that will maintain their interest in the texts.
("Hypertext Literacy" 27)
A reader can always find a way to defeat authorial intention and read in an unpredictable way. However, by making it easy to follow preordained links and harder to chase after other material, authors can indeed impose a degree of order. Ideally, this order would be a set of significant thematic associations (at least, associations which the author thinks are significant) and not a linear path through the material that is so tight that it defeats the purpose of hypertext.
The writer, too, is constrained in some of the ways that Coe attributes to form. This web, for instance, is constrained by the discourse conventions and composing practices that I imagine to be appropriate to native hypertext (even though I am making some of them up as I go along). I have found this myself in writing this text.
David Kolb lists a number of different possible forms that hypertext might adopt in response to different purposes. As these forms settle into conventions, readers will come to recognize them and and writers can use these expectations to convey meaning.
And of course, non-linearity is itself a form. By internalizing these forms, the students absorbs ways of thinking no less surely than by absorbing the forms of print culture. They're just different, more exploratory and associative, more in keeping with the ways of thinking of a postmodern culture, and not necessarily disharmonious with the academic discourse that we presume to teach. Perhaps these forms are in fact more appropriate because they emphsise the collaboration and intertwining of voices that we are only now coming to value and to encourage in our students' discourse.
So what am I afraid of? Why am I worrying that we may be losing something important by losing more linear rhetorical forms? Because thematic connections and non-linear form may be fine for information retrieval but may still not provide sufficient movement from claim to claim to allow a reader to retrieve a coherent argument as opposed to a set of thematic connections.