40 A Grammar of Spoken English Discourse
accent in the utterance. While neither the system of key nor termination
is found in Pierrehumbert and Hirschberg (1990), Wennerstrom (2001a:
278 fn6) notes that Pierrehumbert (1980) includes a phrase-initial high
or low boundary tone which, she states, is similar to high and low key
respectively.
It is widely reported in the literature that speakers raise and lower the
level of their pitch range to express emotion. For example, Brazil et al.
(1980: 23) state that speakers may expand their pitch range to express
excitement, surprise and anger and that they may narrow their pitch range
to express boredom and misery. However, they go on to state that regardless
of whether their pitch range is narrow or wide, speakers use the same
number of pitch contrasts to express linguistic meaning (ibid. 24). They
choose high, mid and low key within the expanded or narrowed pitch
range to convey their meaning.
The work cited above has discussed the intonation of pre-planned discourse:
e.g. Brazil, Coulthard and Johns (1980) a teacher’s lesson; Thompson
(2003) and Pickering (2004) academic lectures; Wichmann (2000) news-
reading. It is possible that in naturally occurring spontaneous multi-party
interacts that there may not always be clear evidence of phonological
paragraphing because speakers do not have suffi cient time to prepare pre-
planned global planning units for a number of reasons such as processing
hitches, competition for the fl oor or interruptions.
To summarize, Section 2.3.2 has shown that there is extensive support in
the literature for the existence of phonological planning units such as pitch
sequences and paratones. Pitch sequences, while similar to paratones, are
not always identical to them. It discussed the communicative signifi cance of
high pitch on the initial onset of a paratone as signalling the introduction
of a new topic into the discourse. While the pitch level of onsets tends
to decline across a paratone, the decline in the pitch level of onsets is
not uniform. To emphasize or indicate the rhetorical relations within
paratones, speakers have the freedom to raise or lower their onset pitch
levels. The end of the paratone is signalled by a fi nal low pitch which is
the lowest in absolute terms. Evidence for the existence of phonological
paragraphing is stronger in pre-planned discourse than in spontaneous
discourse. It is not possible to draw more explicit links between Brazil’s
theory and the wider literature for a number of reasons: the domains in
which the initial pitch level serves a communicative purpose differ: pitch
sequences, paratones and spoken sentences are not necessarily identical in
extent; some scholars propose only two values for key high and low though
others have found evidence of three values for key.