The Perception of Speech

None of the acoustic constituents of speech is unique to speech, although some features of speech are characteristic: a cyclical rise and fall of energy associated with a train of syllables, amplitude peaks, and valleys in the short-term spectrum, and variation over time in the frequency at which the peaks and valleys occur. Despite all, a perceiver often tracks the speech of a specific talker sampling by ear and eye, two kinds of perceptual organization that also combine multimodally and resolves the linguistic properties in the sensory effects. That is to say, perceptual analysis of the symbolic properties of speech succeeds. To gauge the means of resolving the sound produced by a single individual, the contrast between visual and auditory attention is instructive. In attending to a visible object or event, a perceiver typically turns to face it bringing the light reflected by the object of interest to the fovea of the retina. A listener's attention to the audible world achieves spatial and spectral focus psychologically, without the selective benefit of a heading at which auditory pattern acuity peaks. Despite all, perception often reciprocates the patterned variation of a speech stream with its discontinuities—that is, dissimilarities among components, and similarities among its components and those of unattended utterances and other events. This perceptual function is fast, unlearned, keyed to complex patterns of sensory variation, tolerant of anomalous sensory quality, nonsymbolic, and dependent on attention whether elicited or exerted.