34 A Grammar of Spoken English Discourse
they could have produced (55) instead of (54). Furthermore, the speaker
could have unambiguously signalled the key and termination selections by
producing utterances with a single extended tonic segment:
(56) // i COULDn’t go ↑COULD i //
(57) // i ↑COULDn’t go COULD i //However, Tench (1996: 38) reminds us that these examples are unlikely.
Checking tags, if made prominent, have a tendency to form their own tone
units. Therefore it appears that if speakers wish to unambiguously label
their utterances as having separate key and termination values they must
produce utterances such as (53) and (54). This in turn suggests that key
and termination as well as operating in tone units and pitch sequences
also have the potential to operate in increments. Examples (53) and (54),
presented below as increments in (58) and (59) respectively, suggest that
the initial key serves as the key for the entire increment as likewise does the
fi nal termination e.g.^14
(58) // P/R i COULDn’t go // P/R ↑COULD i //
N V V' V N #
(59) // P/R i ↑COULDn’t go // P/R COULD i //
N V V' V N #The fi nal high-termination choice in (58) invites adjudication of the entire
increment while the initial mid key projects the increment as neither con-
trary to expectations nor equative. The initial high key in (59) presents the
increment as contrary to expectations and the fi nal mid termination expects
concurrence.
Moving away from tag questions we fi nd example (60) from Brazil,
Coulthard and Johns (1980: 168) which is simultaneously an increment
and a pitch sequence.
(60) //..... ↓FRICtion // r and when we ↑RUBbed our PEN //
# c w+ N V d N
o on OUR //
p d
// r JERsey // p we were CAUSing ↓FRICtion //
N+ N V V' N #