74 A Grammar of Spoken English Discourse
that he found ‘an overwhelming number of level tones (17 out of 19)’ in
Dylan Thomas’s authorial reading of his Prologue; a fi nding which he claims
‘clearly indicate[s] Thomas’s perception of the level tone’s dramatic
effect’.
Tench (1990: 502) found that speakers reciting the Lord’s Prayer in
unison at a non-conformist service broadcast by BBC Radio Wales selected
level tone on all tone units except the fi nal one. Crystal (1975: 102) in his
discussion of prayer in unison likewise claims that:
The introduction of variation in nuclear tone-type (e.g. rising, falling-
rising tunes) or in pitch-range (e.g. high falling or low falling) is optional,
and usually not introduced.He, also, notes that the fi nal word of the prayer (amen) ‘is given a marked
drop in pitch’. It appears that falling tone may signal the completion of a
piece of language which the speaker is unable or unwilling to assess as
either new or part of the common ground. Crystal further maintains that in
individual liturgical prayer level tones are more frequent than in other
modalities of speech. He claims, however, that speakers engaged in bible
readings or in making sermons tend to select tones analogous to those
found in conversation. It is tempting to explain the preponderance of level
tones found in prayer in unison and in recited individual liturgical prayer
as instances of scripted/learned stereotypical language. Yet Tench’s (1990:
505–6) fi nding that public unscripted prayer contains a preponderance of
level tones indicates that a different explanation is required. He argues
that: ‘Linguistic communication with God does not anticipate a linguistic
response’ (ibid. 513).^25
Ladd (1980: chapter 8) and Gussenhoven (1983: 221) discuss the intona-
tion of calling contours^26 which are realized phonetically by a step down
from one fairly level pitch to another (ibid. 169). Ladd argues that the
communicative signifi cance of the calling contour is that speakers label
their speech as containing a predictable or stereotypical element (1980:
173). Similarly, Gibbon (1976: 279–80) describes calling contours as low
in information value. Ladd (1980: 185) also describes stylized rises^27 as
signalling less information and more predictability. Gussenhoven (1983: 222)
argues that the modifi cation stylization labels the content of an utterance as
a matter of everyday occurrence or routineness.
Gussenhoven (ibid. chapter 7) conducted a small experiment which
attempted to explore the semantic relationships between tones. He hypoth-
esized that his subjects would fi nd the semantic distance between the rise