Character Networks and Culprit Concealment in Classical Detective Fiction, 1868–1930
Detective fiction invites readers to use narrative detail as evidence, but the most visible character is not necessarily the culprit. This article examines twenty English-language detective novels published between 1868 and 1930 through character co-occurrence networks, applying digital humanities methodologies to classical narratology. The first 80% of each novel forms a pre-denouement reader network. Mention frequency, however, places a culprit among the top three candidates in 13 of 20 novels (65%; random expectation 31.4%, p = .002), with ten culprits ranked second or third (p = .003). Culprits often appear early, but so do other major characters. Within individual novels, early appearance does not distinguish culprits from other candidates (p = .918). In 13 of the 20 novels, the culprit ranks among the top three candidates in the change of narrative visibility between the investigation and the ending (p = .002). This result also holds under alternative measures. Furthermore, the novels reveal two distinct strategies for concealing a culprit: keeping the character relatively inconspicuous (under-exposure) or making the character familiar enough to escape suspicion (over-exposure). These empirical findings help explain how detective novels strategically guide readers' attention and manipulate narrative focus. Ultimately, the study demonstrates that while character networks illuminate structural patterns of concealment, computational metrics alone cannot definitively identify the culprit without qualitative contextualization.
Authors
- Xinyi Fan
Institutions
- Beijing Normal University (CN)
Publication Details
- Journal
- International Journal of Law Ethics and Social Sciences
- Published
- 2026-11-16
- DOI
- https://doi.org/10.70088/67vejz93
- Primary Topic
- Crime and Detective Fiction Studies
- Type
- article
- Field-Weighted Citation Impact
- 0.00