2024/03/01 by Avery Blankenship, Ryan Cordell · 1 voice
Arts and Humanities · Computer Science · #Art #Artificial intelligence #Authorship Attribution and Profiling #Computer science #Digital Humanities and Scholarship #Embedding #History #Hybridity #Linguistics #Literature #Media studies #Natural Language Processing Techniques #Newspaper #Philosophy #Sociology #Word (group theory)
paper · doi:10.1093/ahr/rhad493
openalex publication_date 2024/03/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/06/16
Genre is contextual. One text can move through genre categories over time or even between readers, and different media can facilitate these metamorphoses. Genre theorist John Rieder argues that the concept of genre is best thought of as a historical process, one that defies fixedness or definition.1 Nineteenth-century newspapers, in particular, challenge twenty-first-century understandings of genre. Nineteenth-century newspapers are densely packed with information, often with little delineation between texts, topics, or genres. Methods such as computational text analysis can help researchers restore order to the radical hybridity of the newspaper page through categorization, classification, and sorting. Yet the very hybridity of the nineteenth-century newspaper page that makes computational text analysis compelling also complicates its use. Tanya Clement writes that “text mining presupposes a binary logic; there is meaning in the results or there is not.” Text mining thus assumes the existence of a “ground truth” regarding the text’s significance—a consistent and verifiable interpretation of the “facts” at hand.2 Because each researcher determines the contents of their corpus and the logic of their program, their interpretation of genre risks missing the reality of newspaper genres as intertwined and interrelated.