Abstract
The article surveys the principal computational models applied to literary texts and weighs their analytical value. Stylometry, distant reading, macroanalysis, topic modelling, sentiment based plot analysis, character networks and machine learning are examined through their founding studies and measurable results. Each method is described by its unit of analysis, its representative tool and the empirical scale at which it has been tested, from the 150 word feature set of Burrows’s Delta to corpora of several thousand novels. The critique advanced by N.Z.Da is set against the constructive replies of T.Underwood and A.Piper. The survey shows where quantitative evidence genuinely extends literary history and where its claims remain fragile.
References
This work is licensed under a Creative Commons Attribution 4.0 International License.