The Conventionality of Temporal Relations in Relativity Theory

Abstract

1. The Conceptual Innovation in Special Relativity. One of the most interesting claims that has been put forward for Einstein's Special Theory of Relativity is that it involves a drastic revision of the classical concept of time and it is this conceptual advance that seems to distinguish Einstein's Relativity Theory from the theories of other Relativists, e.g. Lorentz, Poincare. However, it is a matter of some dispute just what this conceptual advance was. While it is generally agreed that the classical concept of time was, in some sense, that of an absolute time, it is not clear as to which of the absolute/nonabsolute distinctions reflects the difference between classical time and relativistic time. One distinction is that between absolute time and relational time as represented by the differing theories of Newton and Leibniz. While Einstein's theory is often taken to support a relational theory, it is by no means clear that it does; and, in any case, if it were relational and nonabsolute in this sense it would involve no great conceptual innovation. The second sort of relevant distinction is between absolute simultaneity and relative simultaneity (or as we might say, between universal time and relative time). Einstein's theory seems to entail that spatially separated events will not be simultaneous relative to all frames of reference, but will bear different temporal relations in different frames of reference. This aspect of relativistic time is commonly taken to represent the conceptual innovation of Einstein-i.e. the claim that there is no universal time t, but a set of different times t, t', t etc.-but this view has been challenged by Griinbaum [1] who regards the central innovation wrought by Einstein to be the recognition of the conventionality of time (in particular, the conventionality of the simultaneity of spatially separated events). It is because light is the fastest causal process that can link spatially separated objects that, with respect to any given event in the history of one object, there is a set of events in the history of the other object which bears no temporal relations at all to the given event E in the history of the object. That is, the given event E is neither earlier than, nor later than, nor simultaneous with, any of the events in the other set; however, one can by convention choose any one event from the set and assign it the same time coordinate as that of the event E; and thus it will be by convention that these two events are simultaneous and not because they bear any physical relations to each other.