Philosophical discussions of depiction sometimes suffer from a lack of differentiation between several questions concerning the ‘nature’ of pictorial representation. To provide a suitable framework I distinguish six such questions and several levels on which one might want to proceed in order to answer some of them. With this background, I reconstruct Goodman's and Elgin's answer to the specific question: ‘What distinguishes the pictorial from the verbal or linguistic?’ I try to reveal some major motivations behind their system-oriented approach and to indicate some reasons why a strategy of this kind is to a certain extent mandatory to grasp the ‘nature of the pictorial’. The system-relative and functional character of depiction has to be captured by every adequate theory.