Evolution of Russian Novel Titles (1763–1917): Computational Analysis of Continuity and Innovation

Value of the Information & Strength of the Argument

The presentation offers valuable insights into the evolution of Russian novel titles, combining quantitative analysis with literary interpretation. The argumentation is solid, grounded in a large corpus and clear methodological choices. The use of POS tagging to identify structural patterns and BERT embeddings to measure semantic novelty is appropriate and well-explained. The comparison with British titles helps distinguish universal trends from national specifics. The interpretation of the 1860s as a period of radical novelty, correlated with socio-political changes, is plausible but carefully framed as correlational. The discussion of canonical titles as traditional rather than innovative is a thought-provoking contribution. The speaker effectively argues that the shortening of titles was driven by thick journals and critics, providing evidence from publication patterns and contemporary criticism.

Scientific Rigor, Source Quality, Title Accuracy

The presentation demonstrates scientific rigor through a systematic corpus, transparent methodology, and careful interpretation. The speaker cites relevant prior work, including Moretti’s corpus study of British titles, and builds on her own previous research. The quality of sources is high, with references to established scholars and the provision of a GitHub repository for data and code. The title accurately reflects the content, focusing on computational analysis of continuity and innovation. The discussion by Daniil Skorinkin adds critical perspective, addressing potential limitations and alternative interpretations. The presentation avoids overclaiming and acknowledges the complexity of causal relationships.

The title accurately reflects the content: the talk presents a computational analysis of the evolution of Russian novel titles, focusing on continuity and innovation.

8/10

The presentation is based on a systematic corpus of 2000 Russian novel titles, uses established computational methods (POS tagging, BERT embeddings), and includes comparison with a British dataset. The methodology is clearly explained and results are presented with quantitative evidence. The speaker is a PhD student at HSE, and the discussant is a recognized digital humanities researcher. Minor limitations include the lack of peer-reviewed publication details and potential biases in corpus selection.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction to the seminar and speaker
• 05:00Overview of research questions and corpus
• 10:00Early period: long titles and European influence
• 20:00Shift in 1830s-1850s: shortening of titles and role of thick journals
• 30:00Use of POS tagging to identify structural patterns
• 40:00Introduction of BERT embeddings for semantic analysis
• 50:00Analysis of novelty and the 1860s peak
• 1:00:00Discussion of canonical titles and their traditional nature
• 1:10:00Comparison with British titles and universal trends
• 1:20:00Discussion and Q&A session

• GitHub repository of Daria Chelnokova — Speaker's GitHub with data and code for the study
• System Block (Системный Блокъ) — Online publication on digital humanities, edited by discussant Daniil Skorinkin
• Telegram channel 'Humanities in Digital' (Гуманитарии в цифре) — Official channel of the Digital Humanities Research Institute (DHRI) at SFU
• Telegram channel 'Digital Philologist' (Цифровой филолог) — Telegram channel of discussant Daniil Skorinkin

• Moretti, F. (2009). Graphs, Maps, Trees: Abstract Models for a Literary History — Moretti's work on quantitative literary analysis, which the speaker cites as a key influence.