Semantic categorizations and encoding strategies

Languages differ with respect to the lexicalization of the conceptual distinction between the region SUPERIOR & CONTACT (cf. English on), i.e. in a place which is higher than the place occupied by the landmark and in contact to it, and the region SUPERIOR & NON-CONTACT (cf. English above), i.e. in a place which is higher than the place occupied by the landmark and without contact to it. Both concepts are instances of the superordinate concept of SUPERIOR, i.e. the space that is in the positive domain of a vertical coordinate which originates at the landmark. The typological variation concerning the encoding of these concepts contains: (a) languages that lexicalize the subordinate concepts, e.g. German and Russian; (b) languages that only lexicalize the superordinate concept, e.g. Korean and Yucatec Maya; (c) languages that lexicalize the superordinate concept and one of the subordinate concepts, e.g. Nanafwe; and (d) languages that lexicalize the superordinate concept in one paradigm of local relators and the subordinate concepts in another, e.g. Modern Greek. Speakers of these representative languages have participated in a number of interactive games, in which either the superordinate or the subordinate concept of the taxonomic relation at issue was required in order to fulfill the game tasks. The collected results show that the diversity in semantic categorizations partially determines the encoding strategy used. In particular, languages differ in (a) the lexicalization pattern they choose in order to encode identical concepts, (b) the concept they choose in order to conceptualize identical situations, and (c) the semantic vs. pragmatic conveyance of the same concept.