The parallels (and oppositions) between hypertext and AI are brought out in section five.
In the range of perspectives on new media, that of the hypertext tradition is distinguished by its focus on authored rather than emergent experiences. To put it another way: the hypertext project's focus is antithetical to that of artificial intelligence. While Nicholas Negroponte, prior to founding the MIT Media Lab, dreamed of an intelligent machine that could learn to be an ideal -- hyperpersonalized -- architect's assistant, hypertext figures such as Ted Nelson and Doug Engelbart dreamed of systems no more "intelligent" than power tools or movie cameras -- and just as effective when placed in trained, talented hands.
More recently, some thinkers have attempted to bridge these perspectives. (First Person contributors Phoebe Sengers, Michael Mateas, and Warren Sack combine approaches from both groups.) Yet in general AI and hypertext theorists do not simply diverge, but in fact begin their practices from radically different points.
For the authors in this section, Nelson and Engelbart's hypertext concepts are not simply an unspoken background; these essayists are well-known for their engagement with hypertext, and have directly addressed Nelson and Engelbart's work, as well as that of hypertexts both preceding and outside of the Web.
Thus we have our hypertexts. We find "interactives" in this section's third essay, by J. Yellowlees Douglas and Andrew Hargadon. Interactive is a candidate term for what we might otherwise call interactive narratives, hypertext fictions, or videogames. Like most such terms ("cyberdrama" comes to mind), its use implies a focus on certain types of examples and a general approach to their interpretation.
One might ask, do we require a term like "interactive"? This is a more interesting question in this case than it is for most neologisms because, as Mark Bernstein and Diane Greco demonstrate in their essay, systems that might at first seem beyond the bounds of hypertext, which might seem to cry out for the label "interactives," may prove on closer examination to be hypertexts. Thespis, one of the systems Bernstein and Greco discuss, is an example of this phenomenon which also upsets a number of common assumptions about writing and play. While many forms of language play (from Scrabble to the "exquisite corpse") are rule-bounded, multiplayer games, we do not generally think of preauthored texts as being such. Yet this is what Thespis creates -- a multiplayer performance in which an absent author's text becomes crafted, through gameplay, into a narrative.
Arguments such as this operate against the unfortunately narrow popular conception of hypertext. Still, it may yet prove to be the case that this constricted idea of hypertext is now too well-cemented -- and that a neologism such as Douglas and Hargadon's "interactive" is required to explain to audiences that the systems under discussion are not limited to web-based hypertexts.
Textually and theoretically, hypertext poet Stephanie Strickland is situated somewhere between our two other essays. She turns our attention to the experience of writing and reading hypertexts, particularly hypertext poetry. Strickland's examples further extend what types of systems and artifacts might be included in our categories of hypertexts or interactives: from systems that emphasize the performance of direct configurative practice by the reader, to texts one explores and reconfigures with the eye and mind (and not with a mouse, keyboard, or console controller). Understanding the experience and pleasure of interacting with these new forms -- seen through schemas related to authorship rather than emergence -- is also the explicit topic of Douglas and Hargadon's essay. And the opening of Bernstein and Greco's essay is designed to correct what they perceive as the major critical misunderstandings of the reader's experience of hypertext literature.
It could be argued that this focus on understanding the reading experience is a natural outgrowth of creating new media from a hypertext perspective. Freed from worrying whether system intelligence has been achieved, those who work in hypertext can instead focus on developing new formal crafts, on new methods of creating pleasure and meaning.
Adrian Miles responds to Hypertexts and Interactives