The production of coherent and causally-motivated narrative plots invoking internal states of the characters, in particular epistemic ones, to account for the characters' behavior, appears around 6-7 years in typically developing children, but it is not until 9 years that most children produce evaluative, mind-oriented narratives. Previous studies have shown that, children aged 6-7 years produce more coherently structured narratives after intervention procedures consisting in a conversation on the causes of the story events or in the narration of the story by the experimenter. The present study aims at applying these same two intervention procedures to 20 SLI children aged between 9 and 10; 6 years. All children were first requested to tell the experimenter the story they understood after seeing the set of five pictures presented sequentially ("the stone story" based on a misunderstanding), and again after one of two intervention procedures. One group participated in a conversation soliciting the reasons of the events of the story; a second group heard the experimenter narrate the story and a third group played a "Memory" game with the pictures of the story and some similar ones (a control group). Results show improvement in coherence and in evaluative components after the two intervention procedures. Contrary to studies of typical children, the improvement is more marked after the narration of the model story. In both cases, the immediate effects are maintained one week later. These results show the importance of intervention procedures and the necessity of using various ways to evaluate SLI children's competences.