Abstract: The purpose of this study, an expansion of Nobuyoshi and Ellis's (1993)experiment, was to examine the output produced by 46 university students of Spanish at the intermediate level during focused or unfocused communication tasks that required the use of the command forms in Spanish. The study also examined whether the learners who experienced the focused communication tasks would maintain linguistic accuracy during an unfocused communication task after five weeks had passed. Learners were randomly assigned to one of three groups. Two of the groups participated in focused communication tasks and received feedback on form by means of clarification requests or corrective confirmation checks. The other group participated in unfocused communication tasks and did not receive feedback on form. A one-way ANOVA revealed that differences in mean scores on a delayed posttest were not statistically significant. However students who received confirmation checks scored higher than did students in the other groups. Nevertheless, communicative tasks designed to elicit a more accurate output from learners may be better suited for reinforcing linguistic features that have already been introduced to and internalized by learners.