dc.description.abstract | While Flanders and The Netherlands share the same language and a large part of their history, the independence of Flanders/Belgium in 1830 has led both countries to develop largely independently – politically as well as culturally. Many scholars have theorized about the question whether modern-day Dutch and Flemish literature could nonetheless be considered one of the same. However, little is known about the actual differences and similarities between both literary fields, since structural comparisons are lacking.
I employ computational methods to compare post-war Dutch and Flemish literary criticsm on a large corpus containing 22.361 reviews. I use and compare both state-of-the-art and well-established Natural Language Processing techniques (most notably: Word2vec, Bi-GRU and BERT) to automatically classify and categorize evaluative statements in literary reviews based on a classification scheme developed by literary scholars and communication researchers. This classification scheme categorizes both the aspects critics use when evaluating literary prose, such as the style, the author or the characters, as well as the characteristics that are assigned to those aspects, e.g. humoristic, emotional or realistic. The employed machine learning models achieved a relatively high accuracy, allowing us to classify and compare which aspects and characteristics Flemish and Dutch critics take into account.
The comparision reveals that Flemish reviewers are significantly more concerned with the ideological (e.g. the religious, political, moral and didactic) characteristics of literary works than Dutch reviewers. When discussing these political characteristics, Dutch critics tend to highlight the relationship of the work to current events, whereas Flemish critics speak more about (political) engagement. Furthermore, the diachronic analysis shows that extra-literary evaluations have dominated the nature of Flemish literary reviews longer. In Dutch reviews, extra-literary evaluations have steadily decreased since 1946, subsequently allowing for the rise of more autonomous evaluations (e.g. structure, style). In Flemish reviews, this is only the case from the second half of the sixties. The results show how the 'variable of nation' plays a role in literary criticsm. They give insight into the extent of which reception and evaluation is embedded within a cultural-societal framework. | |