Haifeng Hui: Topic modeling and BookNLP: Seeking the emotional turn in the history of eighteenth-century English fiction

Join us on April 2nd at between 12 and 1:30 pm for a paper presentation titled "Topic modeling and BookNLP: Seeking the emotional turn in the history of eighteenth-century English fiction" on topics modelling, emotions, and data visualizations by Professor Haifeng Hui from School of Foreign Languages, Huazhong University of Science and Technology, China. Professor Hui is also currently a Visiting Scholar at Center for East Asian Studies at Stanford, and an affiliated researcher with CESTA. Lunch will be served for in person participants. If you cannot make it in person, we will have a Zoom option available as well. RSVP via this link to receive the Zoom link or to make sure that we have enough food.

Dr. Haifeng Hui (惠海峰) is Professor of English at the School of Foreign Languages, Huazhong University of Science and Technology, China. He researches children’s literature from diverse theoretical perspectives, including narratology, stylistics, adaptation studies, and digital humanities. He serves as an Advisor Board member of International Research in Children's Literature, and an editor of Jeunesse: Young People, Texts, Cultures. He received his B.A. (2003), M.A. (2006) and Ph.D. (2012) from Peking University. He is also a visiting scholar at University of California at Los Angeles (2014-2015). Haifeng’s recent publications include Adaptation of British Literary Classics for Children (Peking University Press, 2019), “Canon Studies in China: Traditions, Modernization and Revisions in the Global Context,” Poetics Today (2021), “Embedded Mental States in Percy Jackson and the Lightning Thief and Uneven Distribution of Narratorial Attention,” Orbis Litterarrum (2023), “What Can Digital Humanities Do for Literary Adaptation Studies: Distant Reading of Children's Editions of Robinson Crusoe,” Digital Scholarship in the Humanities (2023).

Abstract of the Paper

While topic modeling has been widely used in NLP tasks, its application to literary texts has encountered challenges and dilemmas. In this seminar, I will begin by conducting topic modeling on a single novel, Robinson Crusoe, as a case study to demonstrate how we can use topic modeling for fiction analysis with the help of word2vec to differentiate different topic words in space, which makes it easier to interpret their significances. The result is further validated by readings of different editions of the novel through topic modeling. In the next step, I will apply topic modeling to the history of eighteenth-century English fiction, where I have found an interesting phenomenon of the topic of 'cry'. Following this clue, I have studied the evolution of emotional expression in English novels in the 18th century. By using BookNLP, I extract the usage of nouns and verbs related to emotions and feelings over the course of time, and explore how this trend is influenced by gender. The digital evidence amassed in this research contributes to understanding issues concerning the emergence of sentimentalism and emotions in literary works.