The study of textual style by computational means has a long tradition, but only quite recently have such approaches been able to leave the sphere of forensics to become an instrument of legitimate literary analysis of style.
This new line of research has already yielded a considerable range of digital style studies of epochs, genres, and authors (e.g., Egbert, 2012, Hoover, 2007, Houston, 2013, Mahlberg, 2012). At the same time, questions of conceptualization (e.g., Herrmann et al., 2015) and method (e.g., Jockers, 2013, Moretti, 2013 and 2005) have quickly arisen, and it has become clear that the use of software and computational means does not necessarily imply the adherence to a strictly experimental methodological approach (cf. Ramsay, 2011).
In this workshop, we shall explore novel empirical findings as well as methodological and epistemological questions connected with the computational study of style. In particular, we intend to address topics that fall within the following domains:
(A.) The technical domain. Tools/methods of analysis (computational stylistics, corpus stylistics, literary authorship attribution, NLP, digital hermeneutics); research infrastructures and corpus building;
(B.) The empirical domain. Style variation across factors such as genre, author, gender, epoch and period; narrative perspective and characterization; non-literary registers;
(C.) The conceptual domain. Notion(s) of style; style and aesthetics; typology of style indicators; style production and reception; research design (“close”, “scalable”, and “distant reading”); epistemological status of data; issues of representativeness.
First, the workshop will feature a set of papers focused on the technical side of digital literary stylistics. Jean-Gabriel Ganascia will present a tool for establishing stylistic distinctiveness on the basis of syntax (“Towards a computational and syntax-based stylistics”) and Tomoji Tabata will evaluate a series of Eder's (2015) “Rolling Stylometry” techniques (“Experimental Stylistics: A Meta-analysis to Evaluate Rolling Stylometry”).
Next, we turn our attention to empirical findings about style. Papers in this category include Christof Schöch’s computationally-informed revision of Spitzer’s take on Racine (“Spitzer on Racine, digitally revisited”); Natalie Houston’s discussion of period-specific styles in nineteenth-century British poetry (“Towards a Computational Poetics: Some Features of Nineteenth-Century Poetic Style”); Anne Bandry-Scubbi’s presentation of stylistic differences between male and female authors in a corpus of novels between 1750-1830 (“Women’s Novels 1750s-1830s and the Company They Keep: A Computational Stylistic Approach”); Jan Rybicki’s analysis of how most frequent word usage changes with time in the oeuvre of authors of various languages and literary periods (“Authorial chronology by most frequent words: do writers’ stylometric thumbprints evolve with age?”), and Christine Knoop’s analysis of the distribution of rhyme and cadence schemata in German lyrical poetry between 1700-1930 (“Rhyme and Cadence Distribution in Poetry”).
Finally, the workshop will address ‘style’ conceptually within the framework of computational analysis, with papers including a discussion by Mike Kestemont of different ways of understanding artistic authenticity in various humanistic disciplines (“The Matter of Art: Authenticity Criticism in the Humanities”); a paper by Sarah Allison on what the humanistic adoption of the statistics concept of a ‘proxy’ means for stylistic research (“A Proxy for Style”); and Fotis Jannidis on whether the concept of “historical period” is a meaningful notion in stylometry (“Period Styles”). Expanding this conceptual discussion, Hugh Craig will present some possible remedies to critiques of descriptive quantitative stylistics (“Beyond Authorship”), and Mark Algee-Hewitt will examine how two common assumptions in digital literary stylistics (that authors are individuals with identifiable mentalities, and, in the pragmatics of authorship attribution, that these individuals can be reduced to probabilities of word frequencies) affect our notion of the author (“The Author: Between Style and Substance”).
Insights from the three domains will be drawn together in a discussion panel. Here, we will identify current topics and trends, define the horizons of “Digital Literary Stylistics” and map out the avenues of collaboration between its different branches. Our aim is to define a shared roadmap for a Special Interest Group (SIG) within the ADHO for the field of Digital Literary Stylistics. A special issue of a DH journal will comprise the papers as well as a report of this discussion.