Locke's account of how we construct the idea of space from the perception of distance between objects is circular and therefore fails.

Daniel Warren clarifies this argument in an especially helpful way (Warren 1998; cf. Allison 2004, 100–104). Certainly, it is not true in general that in any order to represent any two entities, A and B, as related in some way, I must represent them as falling into a larger “space” of some relevant character. Warren gives a useful example: in order to represent A as “brighter than” B, I need not represent A and B as being part of a larger “brightness” space. I could do so: I could represent A an