Emotions are communicated through the prosody of speech. Are such expressions perceived categorically or rather in a multidimensional space without clear foci? The former alternative is in line with longstanding views about the existence of some basic emotions. Recent reports present evidence that facial expressions are perceived categorically [de Gelder et al., Cognit. Emotion (in press)]. To explore this issue for auditory speech, a continuum between happiness and fear was created by varying pitch excursion, pitch height, and duration of an utterance via PSOLA. Categoricity of perception for auditory expressions was examined in (a) an auditory-only presentation and (b) an auditory--visual condition of simultaneous presentation of the auditory stimuli and facial expressions corresponding to the either of the two expressions. In the auditory-only conditions a clear CP phenomenon is observed. In the bimodal conditions the visual information modulates the perception of the speech emotion. The latter phenomenon is not observed in a patient suffering from prosopagnosia. The findings raise the issue of a common biological basis for facial and vocal expression perception.