Research on the interaction of keyword treatment with materials or student characteristics in learning foreign languages was reviewed. Most of the studies implied that foreign language vocabulary learning was related to verbal, not visual-spatial ability or elaboration, conditions, and that student-generated keywords were more effective than instructor-generated keywords. There were four suggestions for future research: (1) develop an instrument for predicting performance in visual elaboration conditions not involving picture production; (2) use pretests to explore possible aptitude treatment interaction and to reduce within-cell variability; (3) determine if cognitive ability can be modified to improve performance; and (4) determine if keyword instruction benefits or hinders specific foreign languages. (CP)