Abstract
Theories in physics require reference to manifolds of locations and events. Abstract versions of these manifolds, ‘space’, ‘time’ and ‘space‐time’ are frequently used as reference systems. Should they be included in the ontology of physics as well as the material manifolds from which they are abstracted? This problem can be approached through a study of the identity conditions of events. The argument is offered that neither an abstract ‘time’ of moments is viable, nor is the assumption that events are individuated by their locations necessary. Taken together these conclusions would require reinterpretation of relativity as a grammar of histories and of the Einstein—Podolski‐Rosen experiment as displaying the possibility of spatially distributed events.