Are your minimal pairs too neat? The dangers of phonemicisation in phonology therapy.

Using Conversation Analysis (CA) to look at the interactional dynamics of therapy repair sequences, this paper shows exactly what happens for children when idealised misrepresentations of their phonology are used in therapy. The primary video extract involves a 4-year-old phonologically disordered boy and his therapist. It is not new to warn of the dangers of overlooking any subtle phonetic distinctions that the phonologically disordered child may have but this paper adds something new to the debate. The child's interpretation of the adult's prior turn in the sequence is made explicit and the likelihood of the child producing phonetic revision following an error is seen to be affected by the way the therapist chooses to initiate repair. This is especially true where a 'redoing' of the error is incorporated into such an initiation, when the child's phonetic output is 'tidied up' so that it fits in with the neat minimal pair which forms part of the therapy programme. This is one factor that can be seen to militate against appropriate phonetic repair by the child, especially when used in certain turn structures that are routinely associated with lexical rather than phonetic matters. When the adult 'redoing' more accurately reflects the child's output phonetic revision is more likely to occur. Implications for assessment and therapy are drawn from this evidence, with accurate phonological assessment and continuing interactional error analysis being recommended.