
7.2.3. Spontaneous talk
The presence of phonation periods uttered without an inner pause was also found in spontaneous talk, mass media interviews, TV news reading and weather forecast. However their lengths may vary considerably.
As we listen to a text read aloud or to someone talking we expect the reader or speaker to make pauses. Pauses are necessary for the speaker to take a breath and to collect one's thoughts. For the listener they give a chance to hear and process the speech signal, in other words, digest what he hears. Phonation and pausation time periods are balanced but they are not equal at all. In reading a fable, for instance, most English readers kept the ratio at 2:1 which means that speech takes twice as much time as pause. That was probably the easiest thing to listen to, which even a child could understand. The following data indicates how information is squeezed into phonation periods of longer duration, and what a challenge it is to humans: public speech - 3.3:1, spontaneous talk - 3.6:1, mass media interview — 4:1, TV news — 13:1, weather forecast —15:1. No matter how difficult it is for us to follow news reading presented that way we can still cope, for some time. We have to get used to accelerating rhythms of our time or be selective and protect ourselves by ignoring part of the information poured on us.
For the present part of our discussion the relevant conclusion is that time intervals also serve as elements of speech rhythm but not in absolute duration. It is the relative lengths of uninterrupted talk and pause that matters.
There are sociocultural norms which prompt you that in a conversation people ought to take turns, and holding the floor for a long time is not polite unless we deal with people of unequal social statuses. In a conversation between two people a theme is developed by two people at least, and both speakers are supposed to contribute (see Part V for an example of a polilogue). When the theme is dropped off, there may be a long lapse of pause but not obligatorily. People normally find a way of opening a new theme and signal that by increased loudness and higher pitch. Thus even in spontaneous talk when there are common themes for people to discuss there are prosodic signals to separate one part of the talk (Theme 1, for example) from the other (Theme 2). The difference between monologue and dialogue (or polilogue) is that each participant takes a share of time. Time balance between the contributors depends on the social conventions, including the situation, the status, the abilities and the pragmatic aim of each individual.
Now we can sum up the features of rhythm which have been found in conversation (the basic ones are given in bold):
rhythmic group,
Intonation group,
phonation period (phrase, utterance),
theme talk.
Although the rhythm of spontaneous talk is not so obvious, it certainly is more varied and more flexible in many ways, its basic units are similar to that of reading, for example, or verse. It was proved experimentally that the average length of a rhythmic group does not exceed one second, the average length of an intonation group is around 2-3 seconds. There are physical and cognitive conditions of speech production and speech perception which determine the quantification of speech into similar commensurable units. It was also hypothesized that stress time is equal to heart beat, while intonation group (or phonation period) is equal to the time of breath group. We talk as fast as our mind works, as it monitors the motor activity of articulators synchronizing it with breathing activity.
Psycholinguistic and phonetic experiments suggest that there is a multi-level time grid with syllable as the minimal unit of motor activity (the average length of a syllable is around 200 ms), then a higher level stress group (foot) unit synchronized with pulse (the average length is in the range of 400-600 ms) and above it is the intonation group (phrase, tone group) synchronized with a breathing period (around 2-3 sec). The length of a theme (supra-phrasal unity) is around 30 sec {Antipoval984, Lehiste 1973, Uldull 1971). и*