A quantitative approach to doublets in Russian denominal adjective construction

This article is devoted to the rivalry between suffixes in the derivation of denominal adjectives in Russian. In particular, it proposes a large-scale quantitative analysis based on the Russian National Corpus. Its main goal is to contribute to identifying the properties that determine the choice of the suffix in these derivatives. Denominal adjectival derivation in Russian makes use of a wide variety of exponents. Most of them are phonological variants (extensions) of three main suffixes, - n-, - sk- and - Ov-. These three, which can be considered as basic, constitute the focus of our analysis. Two datasets were built for this research, a general one containing one of the suffixes above, and a more specific one containing doublets, which is to say, adjectives constructed on the same base with different suffixes. Data from the two sets are analyzed by means of various statistical models. Our results globally provide a quantitatively robust confirmation of observations previously made in the literature. In particular, we show that -n- occupies a specific position in the derivational system of Russian, as it is less productive and its derivatives are tendentially less transparent and more prone to display lexicalized meanings that point towards the qualitative pole of the qualitative-relational semantic spectrum. Moreover, - sk- and - Ov- are more likely to form doublets (be attached to the same bases), a further argument in favor of a greater homogeneity between them as opposed to - n-.