This study shows the multimodal analysis of the process of instructing and learning shamisen skills. It extends to the specific instructive actions during the simultaneous playing between the participants. Accordingly, it examines the interactions for synchronization to start the simultaneous playing at the beginning of the practice session and restart it after the instruction during the session. The examination reveals the difference between the above interactions, especially for devices that match the first sound. Finally, through a discussion focusing on the abrupt resumption of co-playing in practice sessions, it is confirmed that this practice needs to assume co-presence, i.e., the complementary and intersubjective framework for the active interaction between the participants unique to the shamisen lesson.