“…The evaluation of speech errors from audio recordings, however, may be influenced by listeners' perceptual biases, as well as by the conventions used for phonetic transcription (Cutler, 1981;Frisch & Wright, 2002;Pouplier & Goldstein, 2005). Previous work using electromagnetic articulography (EMA), with normal auditory feedback, has suggested that many errors perceived to be categorical may involve subphonemic articulatory errors, including the coproduction of multiple segments (Goldstein, Pouplier, Chen, Saltzman, & Byrd, 2007;Pouplier, 2007;Pouplier & Hardcastle, 2005). Thus, it is important to determine if errors that are induced by DAF and judged by listeners to be categorical represent pure sequencing errors or involve subphonemic alterations to the articulatory output because the two classes of errors may arise from different mechanisms, as noted previously.…”