Jean-Pierre Marquis has made a detailed and convincing case for conceiving the development of category theory as an extension and generalisation of Felix Klein's Erlangen Program in geometry. The Erlangen Program aimed to resituate the entire set of linked problematics concerning general geometrical objecthood and proof around a more abstract and algebraic conception of groups of transformations. In the face of a proliferation of different and at times competing geometries in the mid-nineteenth century, Klein proposed an intrinsically variable and pluralistic reconceptualisation of geometric systems and relations. For Klein's approach, geometries are understood as systems of relations of figures that remain invariant under possible transformations. Or to put it otherwise, geometries are coherent selections of transformations of figures that are allowed to collapse into indifference in order that the remaining transformational differences might be effectively theorised. Geometries thus stipulate what does and does not matter from the standpoint of what turns, slides, shrinks and reflects in a generic space of any and all such figural transformations. A geometry (Euclidean, Lobachevskian, Riemannian, projective) becomes an abstract object that is variable and relatively underdetermined in itself, remaining structurally invariant only at the level of certain group-theoretical axioms, properties and operations defining ‘indifferent’ transformations (effectively, isomorphisms). Then at a second level, the variable differentiations among such systems of variable differentiation schematise a space of possible geometries as such, including utterly non-intuitive, unimaginable and interestingly pathological cases.