This paper describes one way of representation of expert’s knowledge aimed at revealing implicit semantics associated with the chosen domain. For this, two important techniques are combined. First, a space of metaphors is selected to complement a given set of the domain stimuli. An expert then performs a subjective scaling quiz on both domain stimuli and metaphors. Second, a geometrical representation of expert’s answers is built and analyzed. The result is interpreted in terms of theories of psychological meaning. The paper discusses the approach and proposes a template for arbitrary experiments of such ways of knowledge elicitation.