The Relative Frequency of Inner-Life Verbs as Signifier of Change in 19th- and 20th-Century Fiction? Distant Reading of a Corpus of Hungarian Novels
The Relative Frequency of Inner-Life Verbs as Signifier of Change in 19th- and 20th-Century Fiction? Distant Reading of a Corpus of Hungarian Novels

Summary/Abstract: This study reexamines a key hypothesis of computational literary studies (CLS): that long-term historical shifts in narrative style – particularly the so-called “inward turn” associated with modernist fiction – can be detected through the relative frequency of verbs related to mental and emotional states (“inner-life verbs”). Building on the methodology proposed by Radak et al. within the Distant Reading for European Literary History project, this research significantly expands both the corpus and the linguistic precision of the analysis. While the original study was based on 100 Hungarian novels from the ELTeC corpus, we extend the analysis to over 1,200 Hungarian novels spanning the 19th to the 21st century, annotated with enhanced natural language processing tools and corrected for systematic tagging errors – particularly those affecting past tense verb forms in historical prose. Our findings suggest that the overall frequency of inner-life verbs remains strikingly stable over the two centuries studied, undermining claims that this metric can effectively signal major literary-historical transitions. However, by examining semantic subgroups of inner-life verbs, we identify statistically significant trends that align with established periodizations in Hungarian literary history. Specifically, verbs of perception and affect peak around 1900, while volition-related verbs rise sharply between the 1920s and 1930s before declining from the 1960s onward – mirroring the “second modernism” and the later prose turn of the 1970s. At the same time, the growing prevalence of cognition-related verbs suggests a long-term, less period-specific trend. Although our findings confirm that inner-life verb frequency alone is insufficient to redefine literary periodization, they do reveal intriguing patterns at the level of individual works and authors, calling for further investigation. The study thus demonstrates both the limitations and the untapped potential of CLS methodologies. While statistical verb analysis may not yet replace traditional literary-historical reasoning, it can productively complement and challenge it. For CLS to fulfil its promise of offering new historical insights, more sophisticated semantic models – such as diachronic vector space embeddings – must be integrated into future research.

• Details
• Contents