Consonant recognition with two levels of decision complexity

The perceptual tuning hypothesis can be expressed as a prediction that discrimination of a particular stimulus dimension will be enhanced if it is the only dimension requiring discrimination and not one of a set of dimensions in a multidimensional discrimination task. In previous tests of the tuning hypothesis, using speech material, the size of a message set has been manipulated as an indirect means of controlling the number of dimensions an observer is required to discriminate. In the present experiments the particular items appearing in a message set have been selected in order to control precisely the number of dimensions relevant to the discrimination task, and performance is estimated in terms of the discrimination of each relevant dimension. No evidence is found for a perceptual tuning effect even though it is argued that the conditions of the present experiment represent those most favourable for its elicitation.