Spontaneous and explicit speech imitation

The single word speech shadowing task typically produces spontaneous imitation of the speech of the shadowed talker (the model). This task has been used as a tool for examining speech convergence in a non-social setting and has provided data for claims that the mental lexicon is constituted from instance-based exemplars. We examined whether the speech of participants who shadowed or explicitly imitated a model talker would converge on similar properties of the model’s speech. The model talker produced speech in two styles (normal and clear speech) that differed in word duration, intensity and F0. Participants produced immediate and delayed naming responses. The results suggested that spontaneous and explicit imitation tap different processes. For spontaneous imitation only word duration showed convergence with model speech and this effect was reduced with delayed naming. Explicit imitation showed an association with model speech both for duration and intensity and this effect was unaffected by the delayed naming. The pattern of partial correlations between the imitation conditions and the model speech provided further evidence that the spontaneous imitation was based upon different processes than those used in explicit imitation.