In this article, the function of utterance in explaining action-sequences was studied from two different points of view: 1) mutual knowledge; and 2) evaluation of the explanation. In the first experiment, two groups of subjects (they were mutually known or unknown) were asked to explain how to fold paper into ‘Turu’, and their performances were recorded on video tapes. Then, their utterance was put into five classes and the frequencies of utterance in each class were compared between the two groups. In the second experiment, subjects evaluated the explanations acquired in the first experiment, and the relation between the evaluation and the utterance in the explanations were examined. Results were as follows: 1) Subjects, mutually known, produced more abstract utterance; 2) Subjects, mutually unknown, gave more additional utterance; and 3) The explanations containing additional utterance were set at a high value while the explanations containing unrelated utterance for folding paper were set at a low value.