Our models of the world develop through a recursive trial-and-error process, so naturally no step in the process starts “from scratch,” although the whole process must have had a beginning. Each step in an evolutionary, developmental or growth process is represented in our modeling by an experiment on a diagram. Any such process requires continuity with variation: variation without continuity is inconceivable, and continuity without variation is inertia.
Continuity is also … the basis for Peirce’s ‘medieval’ realism with regard to the existence of real universals which refer to natural habits and the continuity of their possible instantiations. But diagrams are intimately connected to symbols, as we have seen, in the diagrammatic reasoning process. Concepts are ‘the living influence upon us of a diagram’ – this should be compared with Peirce’s basic pragmatist meaning maxim, according to which the meaning of a concept is equal to its behavioral consequences in conceivable settings. This implies that signification of a symbol is defined conditionally: ‘Something is x, if that thing behaves in such and such a way under such and such conditions’ – ‘Something is hard, if it is not scratched by a diamond.’ But this maxim, developed on the basis of a conception of scientific experimenting, is formally equal to the idea of diagrammatic experiments: the signification of the concept is the diagram of the experiment. The aim of science is to try to make such conditional definitions as diagrammatic as possible. This is the diagrammatic component in Peirce’s laconic enlightenment maxim, ‘symbols grow’: new symbols arise through diagrammatic experimentation.
— Stjernfelt 2007, 115