Latviešu 20. un 21. gadsimta dzejas vārdu krājumi statistiskā griezumā: pētījuma metodoloģija un provizoriskie rezultāti

Abstract

The paper discusses an ongoing study of Latvian poetry of the 20th and 21st century, explains its methodology and interprets some preliminary results. The aim of the study is to explore the dynamics of usage of common poetic concepts (such as natural phenomena, concepts of emotions etc.) and look for other patterns of lexical and semantic change that have occurred in the course of time. The study is conducted by statistically analysing a digital corpus of Latvian poetry. The paper shortly discusses current trends in digital humanities and argues that, although computational methods can prove to be forbiddingly difficult to use, there are tools that are comparatively accessible even without prior training in computer programming. In addition to that, even simple inquiries in the statistics of word usage can provide interesting results and allow building a case for further research. Three computational tools have been used for this paper—Stylo package for stylometric analysis, web-based text reading and analysis environment Voyant, and the corpus analysis toolkit AntConc. The results of statistical analysis suggest that certain trends in the dynamics of usage of concepts in the 20th century can be identified. Among the concepts of meteorological phenomena, ‘wind’ has been by far the most popular concept; furthermore, the usage of meteorological and emotion concepts most prominently fluctuates in the early 1950s and after 1990, indicating the change of poetic paradigms.