Recent data (T. J. Perfect, C. J. A. Moulin, M. A. Conway, & E. Perry, 2002) have suggested that retrieval-induced forgetting (RIF) depends on conceptual memory because the effect is not found in perceptually driven tasks. In 3 experiments, the authors aimed to show that the presence of RIF depends on whether the procedure induces appropriate transfer between representations and competition rather than on the nature of the final test. The authors adapted the standard paradigm to introduce lexical categories (words that shared the first 2 letters) at study and practice. Direct and indirect fragment completion tests were used at retrieval. The results showed significant RIF effects in perceptually driven tasks. Furthermore, they indicated that the presence of RIF effects depended on using adequate cuing to induce competition during the retrieval practice and on the final memory test tapping the inhibited representation.