Hearing ’’words’’ without words: Prosodic cues for word perception

We found that the stress pattern and rhythm of speech were prosodic cues for word perception in trisyllabic adjective–noun phrases. When listeners heard a ’’ma mama’’ nonsense phrase that mimicked an actual English adjective–noun sequence such as ’’new result,’’ they correctly parsed (divided) the phrase into the words ’’ma’’ and ’’mama’’ corresponding to ’’new’’ and ’’result.’’ Listeners must have used prosodic features to parse the phrases because the speech had the stress pattern, rhythm, and pitch prosody of normal speech, but had none of the usual meaning and sound cues for words. The stress pattern was a prosodic cue for word perception because the phrases were easy or hard to parse depending on their stress pattern. And rhythm was also a prosodic cue for word perception because the parsing of a phrase was changed when its rhythm was changed via synthetic speech. But the pitch and amplitude contours were not prosodic cues for word perception because the parsing of a phrase remained unchanged when its pitch and amplitude contours were changed via synthetic speech. The results suggest that, in the rule synthesis of speech, the rules for stress and rhythm must be carefully formulated because they affect not only the naturalness of the speech, but also the ease with which it can be understood.