Beschreibung:
The goal of the paper is to provide a guide for dealing with challenges which occur when dealing with multilingual data at the semantic level. Three dilemmas are pointed out which relate a) to decisions on the size of coding units, b) to the integration of parts of meaning deduced from inference and c) to the diversity of theoretical categories used for describing semantic concepts. Problems and possible solutions are illustrated by looking at two semantic domains: space and time. Recent cross-linguistic studies on spatial and temporal categories are taken as examples for suggesting a four step model for multilingual coding.