The Mysterious Letter J

Abstract

Ekavian and Ijekavian are two different variants of the contemporary standard Serbian language. The difference between them is related to the reflex of the old Slavic vowel jat and it influences both the speaking and writing language norms. The sensibility of existing language identification tools for both variants is of great importance for building representative corpora and development of relevant linguistics resources and tools underlying an automatic text processing. In this paper we present the results obtained after testing the three popular tools for language identification on corpora containing documents from each of the two variants. As it will be reported, the identification of Ijekavian variant is a much more difficult task since the observed tools are not adopted to it at all.