Franklin Cooper invented the Pattern Playback, a machine that converts pictures of the acoustic patterns of speech back into sound. With this device Alvin Liberman, Cooper, and Pierre Delattre (later joined by Katherine Safford Harris, Leigh Lisker, and others) discovered acoustic cues for perception of phonetic segments (consonants and vowels). They found that segments are not usually isolated bits in the speech stream, and that cues vary widely with context due to coarticulation, that is, to the overlapping actions of larynx, soft palate, tongue, jaw, and lips within and across syllables. Liberman, Cooper, and Delattre concluded that the perception of phonetic segments is more simply related to articulation than to acoustic signals. They proposed a “motor theory” of speech perception to resolve the acoustic complexity: we perceive speech, they hypothesized, by learned associations between speech sounds and sensory feedback from their articulation.

Liberman, Harris, and colleagues, working with synthetic speech, discovered that listeners discriminate a given acoustic difference between consonants that belong in different categories more easily than they discriminate the same difference between consonants in the same category. Dubbed “categorical perception” and initially believed peculiar to speech, the phenomenon inspired years of research by Peter Eimas, Michael Studdert-Kennedy, David Pisoni, and others at Haskins and elsewhere. Today, though categorical perception is no longer seen as peculiar to speech, its experimental paradigm retains utility as a measure of phonological skills in young children.

Liberman, aided by Frances Ingemann and others, organized the results of the work on cues into a groundbreaking set of rules for speech synthesis by the Pattern Playback.