A common sense picture of space portrays it as a collection of entities (places) which stand in unchanging spatial relations to one another and which may or may not have objects located at them. To impose manageability on the project to be pursued here, I will assume additionally that on the common sense or prescientific view, which is as much as I will be concerned with, each place is at a definite distance from every other, distinct places are distinguished by their distance relations to other places, and places do not stand in causal relations either to themselves or to material objects.1 This common sense picture is congenial to two kinds of realist about places, the absolutist (or substantivalist) and the reductionist. Both kinds of realist are happy to say that are places, and mean nothing unusual by there are, but the reductive realist offers an identification of places with other, purportedly less problematic, entities; or, to look from the bottom up, he or she offers a logical construction of places from the less problematic entities. Reductive realism is midway between absolutism and eliminativism, the latter position offering a translation of talk about places into a language free of any ontological commitment to them. This paper investigates the prospects for a traditional version of reductive realism, known as relationism.2 Relationists hold that places are possibilities of location. The fundamental idea is Leibniz's; he held that we arrive at our conception of the spatial manifold by some kind of abstraction from the conception of objects standing in a system of spatial relation-