Date of Degree
Matthew K. Gold
Digital Humanities | Intellectual History | Literature in English, British Isles | Literature in English, North America
algorithms, poetics, history of computation, mathematics and literature, philosophy and literature, history of disciplines
In recent decades, scholars in both Digital Humanities and Critical Media Studies have encountered a disconnect between algorithms and what are typically thought of as “cultural” concerns. In Digital Humanities, researchers employing algorithmic methods in the study of literature have faced what Alan Liu has called a “meaning problem”—a difficulty in reconciling computational results with traditional forms of interpretation. Conversely, in Critical Media Studies, some thinkers have questioned the adequacy of interpretive methods as means of understanding computational systems. This dissertation offers a historical account of how this disconnect came into being by examining the attitudes toward algorithms that existed in the three centuries prior to the development of the modern computer. Bringing together the histories of semiotics, poetics, and mathematics, I show that the present divide between algorithmic and interpretive methods results from a cluster of assumptions about historical change that developed in the eighteenth and nineteenth centuries and that implicates attempts to give meaning to algorithms in the modern narrative of technological progress.
My account organizes the early-modern discourse on algorithms into three distinct intellectual traditions that arose in subsequent periods. The first tradition, which reached its peak in the mid-seventeenth century, held that the correspondence between algorithm and meaning was guaranteed by divine providence, making algorithms a potential basis for a non- arbitrary mode of representation that can apply to any field of knowledge, including poetics as well as mathematics. A second tradition, most influential from the last decades of the seventeenth century to around 1800, denied that the correspondence between algorithm and meaning was pre-ordained and sought, instead, to create this correspondence by altering the ways people think. Finally, starting in the Romantic period, algorithms and culture came to be viewed as operating autonomously from one another, an intellectual turn that, I argue, continues to inform the way people view algorithms in the present day.
By uncovering this history, this dissertation reveals some of the tacit assumptions that underlie present debates about the interface between computation and culture. The reason algorithms present humanists with a meaning problem, I argue, is that cultural and technical considerations now stand in different relations to history: culture is seen as arising from collective practices that lie beyond the control of any individual, whereas the technical details of algorithms are treated as changeable at will. It is because of this compartmentalization, I maintain, that the idea of progress plays such a persistent role in discussions of digital technologies; similarly to the Modernist avant garde, computing machines have license to break with established semantic conventions and thus to lead culture in new directions. As an alternative to this technocratic arrangement, I call for two complementary practices: a philology of algorithms that resituates them in history, and a poetic approach to computation that embraces misalignments between algorithm and meaning.
Binder, Jeffrey M., "Symbols Purely Mechanical: Language, Modernity, and the Rise of the Algorithm, 1605–1862" (2018). CUNY Academic Works.