In reviewing the literature on the vocal expression of emotion, a discrepancy between reported high accuracy in vocal-auditory recognition and the lack of clear evidence for the acoustic differentiation of vocal expression is noted. The latter is explained by (a) a paucity of research on voice quality, (b) neglect of the social signaling functions of affect vocalization, and (c) insufficiently precise conceptualization of the underlying emotional states. A "component patterning" model of vocal affect expression is proposed that attempts to rink the outcomes of antecedent event evaluation to biologically based response patterns. On the basis of a literature survey of acoustic-phonetic evidence, the likely phonatory and articulatory correlates of the physiological responses characterizing different emotional states are described in the form of three major voice types (narrow-wide, lax-tense, full-thin). Specific predictions are made as to the changes in acoustic parameters resulting from changing voice types. These predictions are compared with the pattern of empirical findings yielded by a comprehensive survey of the literature on vocal cues in emotional expression. Although the comparison is largely limited to the tense-lax voice type (because acoustic parameters relevant to the other voice types have not yet been systematically studied), a high degree of convergence is revealed. It is suggested that the model may help to stimulate hypothesis-guided research as well as provide a framework for the development of appropriate research paradigms.