The present study extends previous research on motoric activity and imagery production to the text processing of Native American learning-disabled students and third-grade regular-education students. Two experiments were developed to test predictions derived from Glenberg's (1997) "indexical hypothesis". Experiment 1 was performed with learning-disabled Native American students listening to narrative passages under one of three randomly assigned listening strategies: free-study, visual, and manipulate. Experiment 2 was performed with regular-education Native American third graders reading similar passages under one of three randomly assigned reading strategies: reread, observed manipulation, and manipulation. With the learning-disabled students, statistically significant improvements in memory for story events, locations, objects, and actions were observed on cued- and free-recall outcomes when toys representing story characters and settings were present during encoding. Facilitative strategy transfer was not apparent when the toys were removed. With the third-grade students, similar benefits were found when the toys were present. In addition, students who had access to the toys during a training period performed significantly better on cued- and free-recall measures relative to reread students when the toys were no longer present.