In 3D virtual environments (3DVE), we need to know what an object looks like (i.e. geometric information) and what the object is, what are its properties and characteristics and how it relates to other objects (i.e. non-geometric information). Several interactive presentation techniques have been devised to incorporate non-geometric information into 3DVEs. The relevance of a technique depends on the context. Therefore, the choice of an appropriate representation technique cannot be done once for all and must be adapted to the context. In this paper, we first present a preliminary classification of representation techniques for non-geometric information in 3DVE. Then we propose a formalism, based on description logics, to describe the usability of a technique in a given context. We show how these descriptions can be processed to select appropriate techniques when automatically or semi-automatically generating a 3DVE.