Metre and Semantics in the Poetry of Czech Post-Symbolists Accessed via LDA Topic Modelling

Abstract

The article deals with the relationship between semantics and poetic meter in the works of Czech post-symbolist poets and their predecessors. We access the phenomena by means of a machine-driven meter recognition on one hand and LDA topic modelling on the other. We first show how the poetic groups differ in their general preferences for particular topics. Next we analyze the topic distributions in two dominant metres (i.e. iamb and trochee) across the poetic groups.

The fall of genres that did not happen: formalising history of the “universal” semantics of Russian iambic tetrameter

“…We rely on topic modeling to find more abstracted groups of words that share similar behavior, appear in similar contexts with each other, which often is a signal of an underlying shared theme, or semantic domain. Topic models are widely used in information extraction tasks and were multiple times proven to be a robust method of inferring themes in poetry (Navarro-Colorado 2018; Plecháč, Haider 2020), including its application to the problem of semantic halo (Šeļa et al 2022;Plecháč, Kolár 2022).…”

“…Although this article takes the development of Russian meters as the main object of investigation, we claim that the theoretical and methodological framework we use is applicable for modelling the history of universal meters in other poetic traditions (see an example from Czech poetry in (Plecháč, Kolár 2022)).…”

The Metrics of Four Czech Poets in Russian Translations

“…14 According to the data retrieved from the database of Czech verse (Plecháč -Kolár 2014) for the 19th to early 20th centuries, the ratios of masculine, feminine, and dactylic clausulae in dactyl verse are 18 : 55 : 27, in dactyl-trochee 12 : 69 : 19, while the same ratios for Dyk are 20 : 56 : 24 (dactyl) and 13 : 40 : 48 (dactyl-trochee). 15 Kruzhkov's translations include five texts (out of 24) that differ from their respective originals in length.…”