Sizeable expertise? Measuring the focus of professional critics and 'fans' in the online talk of literature

In our contribution, we reflect on what the usage of artificial intelligence for the large-scale study of reader responses may bring to the table of the empirical study of readers in the near future. Over the years, we collected reader/viewer reactions to (live) jury discussions in the context of multiple literary prizes, mostly with the aim to engage in aspect-based sentiment analysis. Through the use of this fine-grained ABSA, we were able to gain insights into readers’ literary criteria and the evolution of their literary criticism over time. It might be argued that "second screen" participation, at the heyday of the broadcasting and Twitter alignment during the noughties, did not constitute "reading" at all. However, the debate of what constitutes minimal reading seems to have recently shifted in new directions. Furthermore, the study of devoted followers of a high-brow literary competition (extant since 1977) over a longer period of time did indeed bring into focus past and ongoing transformations of the media ecosystem, opening up pathways for the study of what may count as "readerly or reader-like engagement". Those practices might, in the end, comprise and engender not just "criticism", but also more generative types engagement, like "fan fiction" or "reading for specific purposes", in spite of the ritual denouncement of these readers as an "uncritical fandom" by stakeholders entrenched in the more traditional and established newspaper and television ecosystem. The marked change in the landscape of literary awards (and literary criticism in general) and the rise of social media, with a digitalisation-led democratisation and globalisation of the public sphere, enables reader-reviewers can now share their own literary judgement via digital channels and to employ prize- or book-related hashtags to signal their participation in the literary debate. At the end of the presentation, we will also briefly discuss the case of the digital social reading communities that have propelled Sarah J. Maas and other romantasy authors to global readership. Building upon the results of the tortuous "match made in heaven" between blurb-like, Twitter-based endorsements/takedowns and BERT-transformers' architecture's windows of 512 tokens viable in 2022, we sketch some avenues for future research that seek to do justice to the multimodal nature of "live" global readerships in the years to come.