Relative effects on adult learners who received experimenter-provided elaborations or who generated personal elaborations, using strategies taught by means of detached versus embedded training, were examined in 80 employees in a professional-development course at a large corporation. Subjects were assigned to a control group, a group receiving experimenter-provided elaborations, and groups receiving training to produce learner-generated detached or embedded elaborations. Recall, recognition, and application of the lesson content (professional-development topics) were measured in a posttest. Subjects generating elaborations surpassed control-group subjects on three or four comparisons. However, there were no significant advantages for either learner-generated-elaborations group versus the experimenter-generated-elaborations group. It may be that the level of generative processing achieved in the experimenter-generated category was sufficient to produce elaborations that function similarly to learner-generated elaborations. Contrary to the hypothesis, the embedded-training group did not demonstrate higher achievement than did the detached-training group. One table summarizes means and standard deviations. (Contains 29 references.) (SLD)