This presentation offers preliminary findings and future plans for a text analysis project that examines Virginia Woolf’s irregular uses of quotation marks throughout her corpus as the moments in which her interests in heard sound, print text, and the politics of recording intersect. Most digital text analysis protocols treat punctuation marks as extraneous information, focusing on the vocabulary itself as the tokens of interest. My project explores just these often-overlooked markings as the most representative examples of the human voice in a text and, accordingly, deeply embedded in questions of power, politics, and what it means to be human. The preliminary findings show a sharp decrease in the proportions of text that Woolf devotes to quoted speech over her career, and I argue for an analogous stylistic shift in her work from more Victorianist prose to more modernist aesthetics. This general trajectory is accompanied by a shift away from interests in society and towards the individual, and I argue that this decline corresponds to a considerable increase in the amount of speech that Woolf does not flag with quotation marks, moments that closely mirror modernist experiments in narrated thought. By adapting machine learning and natural language processing techniques, the project trains the computer to identify speech-like moments where no quotation marks appear. Bringing such unpunctuated speech to the surface stands to reshape our understanding of Woolf’s writing practice as one as much centered on the radical inscription of sound as on a new aesthetics of the text.
