Representing Change

Abstract

This paper evaluates knowledge representations for time-dependent information. It compares recent work by Moore, McDermott, and Allen with an ealier proposal by McCarthy and Hayes. Moore's formalism is faulted for its needless and unmotivated complexity and a simpler alternative is outlined. McDermott's formalism is proved inconsistent and unintuitive. Allen achieves the most by attempting the least. He proposes a simple plausible formalism, which makes few ontological or computational commitments. The paper concludes with a high-level discussion of the merits formal logic as a representation for empirical knowledge.