Abstract The selective adaptation paradigm was used extensively for about 5 years following its introduction to speech research in 1973. During the next few years, its use dropped dramatically; it is now little used. Several reasons for the abandonment of the paradigm are discussed, and theoretical and empirical justification is provided for rejecting these reasons. Experiment 1 demonstrates that “acoustic similarity” of an adapting sound and test items cannot account for the observed results. Experiments 2–4 demonstrate that adaptation effects are not equivalent to simple contrast effects. These experiments indicate that selective adaptation produces robust reaction time effects—items in the adapted category are identified more slowly than unadapted items. The effects found in a simple paired-contrast procedure differ from those found with selective adaptation. Most strikingly, contrast effects are extremely ear dependent—much larger effects occur if testing is conducted in the right ear than in the left; adaptation effects are relatively symmetrical with respect to ear. The empirical and theoretical analyses suggest that the selective adaptation paradigm can be a powerful tool for investigating the perception of complex acoustic stimuli like speech.