This thesis combines methods from sociology of literature and natural language processing to answer the questions: what is the relationship between author gender and the perceived literary quality of her work? And to what extent can textual qualities be ascribed to author gender? I first assess Dutch female authors’ chances of gaining literary prestige. Even though female authors publish many literary works, they still have a harder time climbing the literary ladder. Results of the 2013 National Reader Survey mirror this skewness. Respondents were asked to supply ratings of literary quality, on a list of 401 recent, bestselling Dutch-language novels in several genres (the Riddle of Literary Quality corpus). Even within genre works by female authors’ are judged to be of lesser quality, and ‘feminine’ novels are seen as the worst; formulaic detectives score better. Female author gender is not a conscious motivation, but analysis of respondents’ motivations shows that instead, the text is associated with ‘femaleness’ – through genre, topics or style. Such associations lead to perception of lower literary quality. I then analyze the text of the novels themselves to examine to which extent such femaleness of text exists. First, computational analysis of the Riddle corpus indicates that author gender is too easily assumed to be the cause of text differences. Moreover, additional visualizations show that gender group differences are often caused by outliers. In the final chapters, I focus on a ‘feminine’ topic, attention to characters’ physical appearance. I show that it is not exclusive to the genre of chick-lit, nor to female authors. In fact, male literary authors write most about physical appearance, in describing female love interests. This novel approach shows that female author gender is connected to the text differently than expected. By reading differently, literary quality can be judged separate from female author gender.
|Award date||18 May 2018|
|Publication status||Published - 18 May 2018|