The overall aim of the present article is to evaluate the potential of conceptual metaphors as reliable research tools to be applied in glottodidactics. The main premise adopted is the cognitive principle of inseparability between conceptual mechanisms on the one hand, and the richness of human experience on the other. Consequently, man and his language are viewed through the prism of multifarious and multi-layered encounters involving individual as well social aspects (see T. Siek-Piskozub i A. Strugielska 2007, 2008, 2008a and A. Strugielska i T. Siek-Piskozub 2008). This, in turn, is closely intertwined with the main tenets of ecophilosophy, which highlights interdependencies between biological, social and mental occurrences and language studies. The interdisciplinary character of linguistic research has been evidenced, among others, in publications by Ronald Langacker and George Lakoff. Undoubtedly, the theory of conceptual metaphor has offered a particularly promising testing ground for applied linguistics. The question, however, remains which of the many versions of Conceptual Metaphor Theory should, if at all, be transferred to usage-oriented studies. Therefore, the more specifi c aim of the current paper is to postulate that the classic version of metaphor theories, most frequently exploited by applied linguists (see, for instance, L. Cameron 2003), is not the most valid model. Instead, we propose that metaphor should reflect the dynamics and multicontextuality of human experience. Consequently, there is a need for a more syncretic and systematic bottom-up approach, which does not necessarily validate the results of the dominant deductive theories. Instead, our approach offers a qualitative presentation of data situated within a multitude of relevant contexts.