posted on 2025-05-10, 11:28authored byJack Elliott
Stylometry has mostly confined itself to seeing a text as a summary of word-counts: this thesis takes these ideas one step further and develops practical algorithms for analysing the flow of literary language. These models seek to retain all words in the corpus, rather than focusing primarily on common words (as stylometric models have traditionally done) or on infrequent words (as information extraction has usually done). These algorithms also consciously exploit the scale-dependent nature of literary language, building models on small (from word-to-word), large (throughout a text) and massive (across an entire genre) scales. Although new results are presented for 19th century fiction and Arden of Faversham, these techniques are mainly applied to the study of Harlequin Presents, the most popular and prolific form of romance literature in the world.
History
Year awarded
2015.0
Thesis category
Doctoral Degree
Degree
Doctor of Philosophy (PhD)
Supervisors
Craig, Hugh (University of Newcastle); Pender, Patricia (University of Newcastle)