Choosing the size of the chunks
Each intonation phrase presents one piece of information. The speaker has to break the message up into chunks of information - into IPs - and has considerable freedom of choice in how to do so. Typically, an IP lasts for between one and two seconds (Tench, 1996: 31). An utterance, or a speaker's turn in a conversation, may consist of one or many such IPs.
The chunk also reflect the speaker's decision about focus. Each IP covers a single focus domain (culminating in the nucleus, as seen in
In an alternative spoken version of the same words, the speaker might decide to place an additional focus on this year, perhaps wishing to imply that there is a difference between this year and last year. Dividing it into two chunks, with an intonation break after progress, gives:
IPs tend to be longer, and have more accents, in scripted material, and in material read aloud. In spontaneous conversation they tend to be shorter, with fewer accents. This maybe because the IP is not only the basic chunk for intonation purposes but also the basic chunk for mental planning. In our minds as we speak we plan one IP at a time. Reading from a script, we can look ahead and plan ahead; in ordinary conversation we may not have decided what we are going to say until just before we say it.
Tonicity thus varies considerable according to the style of speech. In some styles IPs may be short, with nearly every accent being nuclear. Here is a genuine, if extreme, example. At the church concert the choral conductor, a tall man, apologized for fussing about a footstool to stand on. He said: