This paper investigates whether the same assumptions are at work when we use irony (a) implicitly, e.g., by saying "Good job" to comment ironically on an action we do not approve of, or (b) explicitly, e.g., by using phrases like "Isn't it ironic that..." to comment on a situation perceived as ironic. When irony is noted explicitly, we usually find one participant (P1) involved in two actions, (A1a) and (A1b), which an observer considers conflicting. But there exists another type of explicit irony, namely a semantic expansion which includes the feature of coincidence. In this case, we find two participants, (P1) and (P2), whose respective actions, (A1) and (A2), are seen in conflict. We have recognized here an addition to the irony inventory. The analysis of explicit irony markers thus may reveal also semantic change. (Katharina Barbe)