Stylometry in TV shows - my work so far

My interest in the topic was first picked up by study of Agata Hołobut, Monika Woźniak and Jan Rybicki, presented during DH 2016 conference Hołobut et al. Stylometry on the Silver Screen: Authorial and Translatorial Signals in Film Dialogue.
Encouraged by Jan, then supervisor of my Master’s thesis, and interested as a fan, I built a corpus of fan-transcribed dialogue lines of the shows, which I subjected to various types of analysis, mainly related to the switch from character- to showrunner-oriented production mode and discourse analysis. Parts of this study are currently under review for publication in scholarly journals.

Main research questions

Are there stylistic similarities on textual, visual and audio levels?
What’s the level of diversity on each of them?
How does it affect the concept of authorship as we know it?


As part of my thesis, I compared cross-author relations and showrunners’ influence in Doctor Who, discovering that some showrunners (i.e. Davies) would excercise more control over subordinate writers, whereas others (Moffat) would have very distinct style of their own, but not impose it on their writing team.

Character idiolects - discourse analysis

I use contrastive analysis with Craig’s Zeta and classic discourse analysis based on (in)frequent phrases to examine the speech of particular characters. These classic methods I combine with classification and network analysis as used in stylometry to detect less visible differences and similarities.

Creative relations

Preliminary study presented during EADH 2018 conference Who is the author? Modeling creative relationships in TV studies


  • fan-created transcriptions, revised and normalized
  • why not subtitles - Bednarek 2010, 2018
  • atm ~90 series, ** balanced for genre, ** not for length (varying from ~20 to ~60 min.), ** more US than UK
  • metadata drawn from IMDB API

Future plans

more soon…