Ep 18: Digital Humanities Project
Hello and welcome to my podcast! Today I have an exiting announcement to make: I am going to pursue a masters program at Cambridge University in Digital Humanities! It is quite a new program that just opened last year, so I will be joining their second cohort.
According to the department: “The new MPhil in Digital Humanities at the University of Cambridge explores ways in which the humanities can engage with digitally enabled research approaches, considers the impact of digital innovations on cultural forms and practices, and explores digital futures. This one year taught masters programme is designed to be inter-disciplinary and caters for different skill levels in digital humanities methods and approaches.”
In my application essay, I propose using sentiment analysis to determine the shapes of the stories of four of the pre-Modern Chinese texts Water Margin (水滸傳), Romance of the Three Kingdoms (三 國演義), The Plum in the Golden Vase (金瓶梅), and Dream of the Red Chamber (紅樓夢). As of 2023, there has been no published critical study where computational sentiment analysis has been applied to these four texts, which have been considered the core of the Chinese literary canon for centuries. The extensive literary discourse around these texts allows for the opportunity to compare the results of sentiment analysis to long-recognized critical cruxes of each work and the most frequent reasons the works are read or compared together in the history of their literary study. In addition, sentiment analysis models now offer us a new approach for comparative literature in translation, where language models can be used to compare the shapes of the stories in their English translations versus in Chinese. Sentiment analysis is a syntagmatic approach to understand why these four texts are considered paradigmatic examples of Chinese literature. By graphing the underlying shapes of these texts allows us to chart - line by line - arcs of lexical urgency and affect.
Sentiment analysis can yield new approaches in literary scholarship by using language models to map the affective arc, the rise and fall of positive and negative emotions of a literary work. These underlying shapes help identify new or re-evaluate familiar key passages of a text as the subjects of literary criticism and scholarship. Once the specific “shape” of an individual literary work is produced, it can be used to assist in furthering narrative analysis of the text and comparative literary criticism. I intend to build upon the most recent work by Katherine Elkins and Jon Chun published in The Shapes of Stories (2022) who are pioneers in the field of sentiment analysis in Digital Humanities and whose research I assisted with as an undergraduate. In my graduate studies in Pre-Modern and Modern Chinese literature at Columbia University, I became familiar with the critical landscape around these four literary works.
I hope to develop and advance the approaches by Elkins, Chun, et al, both are my college professors, with the opportunity to introduce a digital humanities approach to these classics of Chinese literature by mapping their story shapes for the first time using sentiment analysis. I intend to use a collective approach of sentiment analysis machine-learning models such as RoBERTa and FLAIR and ensemble tools that combines multiple models like SentimentArc first to chart each work’s shape in their English translations and then to use separate Mandarin sentiment analysis tools to chart their shapes in Chinese. The visualizations produced should allow for the first time a reading approach that combines both the practical application of computational methods and close reading so that scholars can not just view the results of sentiment analysis as raw text processed algorithmically, but to read specific passages and re-evaluate the significance of character, plot, story, and discourse in each of these great works anew.