When does rhythm emerge in speech?From the little we have seen here, it looks like a safe bet that rhythm, or a regular series of beats, appears in the speech signal under just those speaking conditions where subjects have to be precise about the temporal alignment of events across diferent syllables or groups of syllables.In the speech cycling experiment, we explicitly asked subjects to place one event (the beat of duck) at a particular phase of the overall repetition cycle. This meant that they had to coordinate the beats associated with different syllables (big and duck).
| |
Why does rhythm emerge in speech?There are many answers to this question. From the point of view of speech production there is a clear advantage to providing a framework or reference (the regular beat period) with respect to which events can be timed. Other beats (typically, syllable onsets) can be placed at phases of 0, 1/3, 1/2 or 2/3 of the regular period. The rhythm emerges under just those conditions which require long-distance coordination, to wit speech cycling and the production of rapid fluent speech.Speech cycling is an artificial task. However, we all speak rapidly sometimes, and we move into and out of fluency as we speak continuously. Under these conditions, rhythm may serve the listener as well as the speaker. Some theories suggest that we need a temporal pointer when events are happening rapidly, so that we can focus our attention on just the right moments. Fluent and rapid speech is a problem for the perceiver , and rhythm can help to structure the listening process as much as the speaking process. Furthermore, once a rhythm has been established, the listener immediately begins to expect the next beat at a particular point in time. Confounding this expectation by the jusicious use of pauses and lengthenings can be used communicatively, signalling to the listener that this is an important part of the speech message. Other theories have claimed that a rhythm is established across speakers to facilitate the taking of turns in a conversation. So if I set up a series of beats and then stop, you will time your responce to come in on the beat. Perhaps rhythm also helps us to break up the speech signal into just the right units necessary for decoding. There are many ideas here. Few of them have been tested using the beat-based approach we advocate here, so there is a lot of work to be done. And at the end of it we will only have begun to understand speech rhythm. The expressive exploitation of rhythm, the asthetics of poetry, these will not yet have been addressed. | |
| This short foray into the experimental investigation of speech rhythm is now over. If you need more details, feel free to consult the thesis, or our personal web pages, from where other publications can be accessed. |