The letters of writers naturally tend to be considered to a lesser degree than the actual literary work. Yet some collections of letters can be regarded as "works alongside works" [1], both quantitatively and in terms of their aesthetic scope. The letter can appear equally as a medium of pure information exchange as well as an expression of a literary-aesthetic aspiration - and quite often both types combine to form a double function. The high variance in content, coupled with the occurrence in large, formally constant amounts of data, predestines the epistolary work as an object of investigation for quantitative computational approaches, which will form the methodological framework of this project. In the project, which will initially be based on a corpus of about 14.000 letters by Charles Dickens, quantitative methods will be used for the first time to trace possible references or - if these are further consolidated - reference systems between epistolary and literary works which can reveal themselves, for example, in common motifs, figures and central concepts as well as indicate a similar aesthetic category formation. In a second step, the possibility of automatically detecting textual manifestations of presumed references and applying them to further epistolary and literary corpora by means of machine learning methods will be examined in order to reveal generic (i.e., cross-corpus) reference points within a framework of aesthetic communication.
[1] Irmgard Wirtz and Alexander Honold, for example, situate Rilke's extensive epistolary work in this way, cf. Honold, A. & Wirtz, I. M. (2019). Rilkes Korrespondenzen: Das Briefwerk als Medium kommunikativer Selbstentwürfe und literarischer Interaktion. In A. Honold & I. M. Wirtz (Eds.), Beide Seiten - Autoren und Wissenschaftler im Gespräch: Bd. 6. Rilkes Korrespondenzen (pp. 7-32). Wallstein; Chronos, p. 7.
