Media theorists as far back as Marshall McLuhan and as recent as Friedrich Kittler and Jonathan Crary have asserted that since the photograph's and gramophone's invention in the nineteenth century, technical apparatuses have increasingly come to extend and simulate the humans senses. My dissertation argues that the convergence of these technical media with automated technical apparatuses begins a recursive cycle. The technical apparatus encodes itself through various technical media, and as a result, produces new technical media. This cycle of encoding and re-encoding is cognitive, and constitutes what I call a nonhuman cognitive system. I trace this evolution of this nonhuman cognition from early cybernetic devices like Vannevar Bush's Differential Analyzer to the contemporary digital computer.

This evolution of technical bodies is still closely intertwined with human ones, because the technical media enabling the nonhuman cognition's evolution also encode images, sounds, and narratives for human cognitions. The so-called realism that technical media enable (i.e., a greater degree of verisimilitude to what humans actually see, hear, feel, etc.) is actually what I call fractal realism. It is fractal because humans and nonhumans use different cognitive schemes for mediating or encoding their reality, and so fracture their world as both cognitions work on it. It is fractal as well, because the nonhuman cognitions rely on statistical measures rather than narrative chains of cause-and-effect to order their encodings, just as fractal sets are defined by their statistical self-similarity. Under fractal realism, mediation is not a means of conveying information between parties but rather the conversion of one medium into the pattern of another medium. I argue that we understand mediation as, what Gilbert Simondon calls, a transductive process.

Given the tight link of cognition and mediation, I parallel the account of the nonhuman cognition's evolution with an examination of how its statistical mode infiltrates and affects media as humans understand them. We can then see fractal realism as a new genre that cuts across realism and postmodernist literature and cinema. Some of the works considered include Joseph McElroy's Plus (1977), Thomas Pynchon's Gravity's Rainbow (1973), Michelangelo Antonioni's Blow-Up (1966), and Mark Z. Danielewski's House of Leaves (2000).

Despite the recent rise of interest in digital humanities, the long history and influence of computerized humanities scholarship remains unrecognized. In response, this dissertation reads novels written over the past five decades as reactions to humanistic computing. Though recognized as important examples of postmodern literature, these novels also belong a category of "parsing fiction": novels written partly in response to projects that analyze language and literature with computers. This dissertation constructs a tradition of novels that consciously develop aesthetic techniques that subvert the theoretical rubrics such projects embrace.

The existence of parsing fiction demonstrates that humanistic computing has influenced the development of postmodern fiction, having inspired authors to explore the implications of computerizing linguistic and literary study. A primary reason for this engagement between novelists and scholars has been that most humanistic computing projects follow hegemonic scholarly methodologies rather than challenging them. Parsing fiction reacts against this high-tech conservatism by questioning these common theoretical underpinnings in light of the computer.

Each of the four chapters pairs one novel with a group of projects to which it responds. This dissertation begins with John Barth's Giles Goat-Boy, which critiques Literary Data Processing in the 1960s. In the 1970s, Italo Calvino's if on a winter's night a traveler responds to Mario Alinei's linguistic research. A decade later, Michael Joyce's afternoon, a story advocates the use of hypertext over humanistic Artificial Intelligence. Finally, at the end of the millennium, Mark Z. Danielewski's House of Leaves argues for the specialness of print against digital hypertext. The conclusion offers some final reflections on the digital humanities' influence on contemporary literature.

Reading novels as parsing fiction demonstrates that understanding the computer's role in literature requires attention to specific technologies, such as Artificial Intelligence and hypertext. Ultimately, this dissertation argues that humanistic computing has had a more significant impact on postmodern fiction than scholars recognize. While scholars who design projects and write programs remained comfortable with the computer as a tool for reinforcing accepted theories, novelists who sometimes only glimpse computers from a distance have imagined them as agents of radical transformation for literature.