Vocabulary-Based Classification and Contact-Induced Formation of Neologisms in Two Standard Varieties of Karelian

Abstract

This paper investigates the lexical similarities and formation of neologisms of two written standard varieties of Karelian, North and Livvi Karelian, spoken in the Republic of Karelia, Russia. Firstly, a naïve Bayes statistical model was generated to classify North and Livvi Karelian newspaper texts automatically. Secondly, the word formation strategies of neologisms from the classified newspaper texts were studied. The strategies between the two varieties were compared in terms of the code-copying framework. The results from the automatic classification and the investigation of neologisms show that the standards differ in lexicon and phonology, but the strategies of forming neologisms are similar: the most common strategy is to form words by language-internal means, and the other strategies are selective and global copying from Finnish, Russian, and English. The similar strategies in both standards suggest similar language planning.