Creating the cognitive form of phonological units: The speech sound correspondence problem in infancy could be solved by mirrored vocal interactions rather than by imitation

Theories about the cognitive nature of phonological units have been constrained by the assumption that young children solve the correspondence problem for speech sounds by imitation, whether by an auditory- or gesture-based matching to target process. Imitation on the part of the child implies that he makes a comparison within one of these domains, which is presumed to be the modality of the underlying representation of speech sounds. However, there is no evidence that the correspondence problem is solved in this way. Instead we argue that the child can solve it through the mirroring behaviour of his caregivers within imitative interactions and that this mechanism is more consistent with the developmental data. The underlying representation formed by mirroring is intrinsically perceptuo-motor. It is created by the association of a vocal action performed by the child and the reformulation of this into an L1 speech token that he hears in return. Our account of how production and perception develop incorporating this mechanism explains some longstanding problems in speech and reconciles data from psychology and neuroscience.