Although the details vary from study to study, the basic procedure is to present listeners with phonetically ambiguous stimuli, with some source of contextual information that disambiguates the stimuli. Perceptual learning is subsequently indexed by a shift in phonetic categorization toward the contextually defined speech environment.