Paper Short Abstract:

Digital tools have a great potential in helping us to better understand our archival collections and the historical cultures—both elite and vernacular—that produced these. This paper concentrates on how a set of similarity recognition methods can be used both in analysing historical oral poetry and in estimating the characteristics of archival collections.

Paper Abstract:

Variation is a core characteristic of folklore: in oral tradition it is typical that the information is changing, due to various reasons such as creativity, adaptation, misunderstanding, and forgetting. This results in various degrees of similarity between the performances and records of folklore. In order to explore the patterns of variation in the vast material of Finnic oral poetry, Maciej Janicki along with the rest of the FILTER group configured a set of methods enabling us to recognize similar verses, passages and poems and measure their similarity.

Similarity assessment methods have enabled us to get better overview of the patterns of variation in oral tradition, for example, 1) recurrent units (or basic building blocks) at the line and passage level and the use of these in oral composition; 2) degrees of similarity of song texts and various reasons affecting these 3) oral-literary relationships and processes of modernisation; 4) variation of versions of poem types in multilingual and multidialectal continua; and 5) regional verse and song repertoires and their interconnections.

Besides explorations of the tradition itself, similarity calculations are useful in analysing overall characteristics and biases of archival collections. Very similar texts can often be explained by different versions by one singer, family or village tradition, or very popular and stable short poems, but many cases also derive from singers adapting songs or motifs from literary sources, basic education or school song books, collectors copying texts from each other or literary sources, and curating and editing processes causing duplicates.