The interaction between spatial cases and relational nouns in the Uralic languages of the Volga–Kama area

In this paper we take a look at the variation between plain spatial cases and relational nouns in expressions of two spatial relations, namely CONTAINMENT and SUPPORT, in the Uralic languages of the Volga–Kama area (Erzya, Moksha, Meadow Mari, Hill Mari, Udmurt, Komi-Zyrian, and Komi-Permyak). By studying the variation in the coding of these relations, we tested the proposition put forward in literature that structurally more simple forms are usually used to code semantically more simple relations. In the most prototypical case of CONTAINMENT there is a three-dimensional landmark in which the trajector is situated, and in the most prototypical case of SUPPORT there is a two-dimensional landmark which shields the trajector from the effects of gravity. To test the proposition, we gave the landmarks in our data  prototypicality scores based on how well they conform to the spatial prototypes of CONTAINMENT and SUPPORT. Next, we looked at what means (plain case or relational noun) was used to code spatial relations with diff erent prototypicality values. The results show that in the studied languages the proposed generalization of simpler meaning coded by simpler form does not explain the variation of means of coding very well. Especially, the CONTAINMENT relation has almost equal numbers of prototypical and non-prototypical relations coded by plain case and relational nouns. Finally, we propose an alternative explanation based on the focus of attention as the motivation of the variation.