Whether pragmatic or Romantic, the projected benefits of hypertext follow from certain assumptions about how people read or should read. The belief that readers can select for themselves which links in a network to follow rests on the assumption that readers know best what information they need and in what order they should read it. The goal of creating paths for different readers assumes that hypertext designer/writers can predict readersŐ needs well enough to create the "right" set of paths and direct each reader onto the appropriate one. The very notion that hypertext designer/writers can create meaningful, useful networks in the first place depends on a whole range of assumptions about how to divide up and relate parts of texts, including what segments of text constitute meaningful nodes, what types of links are meaningful and important, and what types of texts can or ought to be read non-linearly. In fact, many of these assumptions contradict current thinking in rhetorical theory, cognitive psychology, and document design, where the evidence suggests that, as currently conceived, hypertext may in fact dramatically increase the burdens on both readers and writers. My purpose in this essay is to review relevant educational and psychological research on reading that bears on the problems hypertexts may pose for readers and writers. The purpose of this evaluation is not to accept or dismiss hypertext in principle, but rather to point to specific aspects of reading and writing processes that hypertext designers must consider if they are to serve readers and writers effectively.
See also the talk that this article was based on:
"Comprehending non-linear text: the role of discourse cues and reading strategies," Conference on Hypertext and Hypermedia. Proceeding of the ACM conference on Hypertext 1987 , Chapel Hill, North Carolina.