This paper compares the Korean past tense marker -ess with another past form -essess (double past), the distinction between which has been controversial among Korean linguists, and provides a discourse-based semantic analysis of them. It is argued that -ess and -essess are logically distinguished in terms of dynamic versus stative information in dynamic semantics, which is more or less in line with the distinction between the English past and the past perfect. The simple past in English and the Korean -ess at least have the potential to give dynamic information, updating the given context with a new event and shifting the current temporal perspective. The English past perfect and the Korean -essess, on the other hand, provide stative, background information, preserving the given context. However, while the English past perfect is ambiguous between preterit and aspectual interpretations, Korean -essess has only the preterit interpretation, triggering a flashback effect in narratives. Their semantic properties and differences are represented in Discourse Representation Theory.