This study aimed to isolate the effects of the two input activities in Processing Instruction: referential activities, which force learners to focus on a form and its meaning, and affective activities, which contain exemplars of the target form and require learners to process sentence meaning. One hundred and twenty 12-year-old Taiwanese learners of English as a foreign language were assigned to one of four groups: Referential + Affective, Referential only, Affective only, or Control. The treatments were computer-based. Pretests, posttests, and delayed posttests, including a timed grammaticality judgment, a written gap-fill, an oral picture narration, and a short semistructured conversation, measured learning of the "-ed" past tense verb inflection. Findings suggested that referential activities were responsible for the learning gains observed, that affective activities did not provide additional benefits in terms of learning "-ed," and that the gains observed displayed some broadly defined characteristics of explicit knowledge.