The purpose of this paper is to clarify the role of curriculum evaluation in Japan, by examining Michael Scriven's framework of evaluation, especially on formative/ summative distinction. Today we are confronted with the necessity of school-based curriculum development, for example, cross-curricula, subject-choice system in secondary, which are almost novel and most schools must be involved. In curriculum development, curriculum evaluation plays an important role as well as design or making. However, in Japan, curriculum evaluation is relatively a new study area, and most theories or methods have been imported. For curriculum evaluation in Japan, I think that firstly we have to connect our traditional evaluation study with foreign-made study. Secondly, it is important to balance the role of curriculum evaluation with the methodology, since their unbalance may make curriculum evaluation invalid. Based on these backgrounds, I focus on a researcher of evaluation, Michael Scriven and his framework of evaluation, for he has suggested a famous distinction about the role of evaluation, not limited to curriculum nor education, formative/ summative. This paper has two significance points. The first is to clarify the role of curriculum evaluation in Japan, by examining directly Scriven's original sense without Benjamin Bloom's indirect interpretation. And the second is to relate formative/ summative distinction with Scriven's another well-known theory, goal-free evaluation. According to Scriven, including curriculum evaluation, the role of evaluation can be divided into two. One is formative, to improve during development, the other summative, to help decision-making based on evaluation. And there is no difference between the two in terms of merit, both of them are important. Because one evaluation activity can be interpreted as formative/ summative, that depends on the context in which the activity is placed. Further, Scriven divides evaluation into two, the role and the methodology, for he thinks that their contamination can cause problems, especially the methodology may be weakened in value judgment. As a result, I make a sketch of Scriven's framework of evaluation. Based on this sketch, I point out the relationship between formative/ summative distinction and goal-based/ goal-free evaluation, and a tendency of relativism in Scriven's framework. In conclusion, I suggest following points for the role of curriculum evaluation in Japan. 1 . Formative role is important as well as summative. This view is not new itself, but through this study, newly we can add two points: indexicality (or context-dependency) in evaluation, and relative importance in formative/ summative distinction. 2. We should continue to refine our methodology of evaluation, balancing with the role of evaluation. Especially, in terms of indexicality in evaluation, we can find "evaluation" in daily school activities and teachers' decision-making processes which have been not named "evaluation".