“-nun” and “-un”, two of the relative clause markers used in Korean, exhibit some interesting peculiarities yet to be investigated. They cannot combine with the past tense marker “-ess”. This is why many researchers consider that they are relative clause markers which express tense; “-nun” is a relative clause marker expressing [-Past] tense, and “-un” is a relative clause marker expressing [-Past] tense or [+Past] tense according to the semantic type of the predicate with which it combines: “-un” expresses [+Past] tense when it combines with an event predicate while it expresses [-Past] tense when it combines with a state predicate. However, the view that they are relative clause markers which express tense cannot account for the fact that the morpheme “-un” can express not only [-Past] tense but also [+Past] tense. This paper proposes a new approach which presupposes the existence of a unique pure relative clause marker in Korean. We argue that this is “-un”, and that tense interpretation in relative clauses containing “-nun” or “-un” is due to the [-Past] zero morpheme or the elided [+Past] morpheme “-ess”. More concretely, “-nun” is considered through analysis to be a complex composed of the [-Past] zero morpheme + the progressive morpheme “-nun” + the pure relative clause marker “-un”, where “-nun” and “-un” are contracted to “-nun”. “-un” when combined with state predicates is considered through analysis to be a complex composed of the [-Past] zero morpheme + the pure relative clause marker “-un”. As for “-un” when combined with event predicates, it is considered through analysis to be a complex composed of the elided [+Past] morpheme “-ess” + the pure relative clause marker “-un”. Note that one of the important assumptions in this paper is that the [+Past] morpheme “-ess” is elided before the pure relative clause marker “-un” for a morpho-phonological reason. Given this assumption, the question arises as to why the “-ess” ellipsis does not occur when “-ess” comes before the “-un” which combines with a state predicate. If the “-ess” ellipsis did occur, this would allow for a [+Past] interpretation, contrary to fact. To resolve this problem, we propose an ambiguity elimination strategy: when a form allows for more than two interpretations, only one should be taken and the others are eliminated. The situation where “-un” combines with state predicates should be applicable to this ambiguity elimination strategy since the same form of “-un” can be derived from the complex [[-Past] zero morpheme + “-un”], which would produce a [-Past] tense interpretation, or from the complex [elided “-ess” + “-un”], which would produce a [+Past] tense interpretation. Which one should be settled upon and which should be eliminated? Note that in relative clauses containing a state predicate, “-un” is the only means to express a [-Past] tense interpretation while there are other means for expressing a [+Past] tense, such as “-ten”, which is the complex [retrospective morpheme “-te” + pure relative clause marker “-un”]. Therefore, a [-Past] tense interpretation should be taken to the detriment of a [+Past] tense interpretation. That is why, when “-un” combines with a state predicate, there is only a [-Past] tense interpretation. On the other hand, when “-un” combines with an event predicate, for which there is only a [+Past] tense interpretation, a different question arises: why can it not have a [-Past] tense interpretation even though it can be derived from the complex [[-Past] zero morpheme + “-un”]? Note that in relative clauses containing event predicates, there is another means for a [-Past] tense interpretation, which is “-nun”. This is why, when “-un” combines with an event predicate, a [+Past] tense interpretation should be settled upon rather than a [-Past] tense interpretation.