Previous studies have shown that motor encoding of verbal phrases facilitates recall or recognition performance comparing with verbal or imagery encoding. This phenomenon is called subject-performed task (SPT) effect. Although many researches confirm the effect under various conditions, most of the research employs first language (L1) of the participants and concrete words. An experimental study was conducted to examine whether motor encoding is superior to verbal or imagery encoding even in second language (L2) and with abstract words, adverb. The participants were twelve college students learning Japanese as a second language in Japan. They were all in the first level of The Japanese-Language Proficiency Test. The participants were asked to encode three lists of verbal phrases presented visually on a monitor. Three conditions were used as encoding tasks. Under the SPT condition, the participants enacted the denoted action. Under the imagery task (IT) condition, the participants drew visual images of someone, as though they were really watching him or her performing the action in front of them. Under the verbal task (VT) condition, the participants wrote down the verbal phrases shown on a monitor. They were required to recall freely the verbal phrases by writing after learning each list. As a result, there was significant SPT effect in free recall test. It was suggested that the SPT effect were observed even in L2. That is, recall performance of SPT condition was superior to that of IT and VT conditions. The results were discussed from the viewpoint of the function of motor encoding.