Advances in Sociophonetics

112 Rosalind A. M. Temple

have occurred in delivered by in (24) resulting in the percept of deletion despite a

full or lenited alveolar gesture for the final singleton /d/. We return to the matter

of such residual gestures in the discussion of assimilation in §2.5.

Relative timing of gestures may also account for deleted tokens with preceding

sonorants. In (38) there is coronal closure for the preceding lateral consonant; it is

possible that the sides of the tongue were raised before the release of this closure,

essentially forming a [d] or [d̥]:

(38) but there was all old carpets [ɔ̞lkʰa̱pʰɪʔ͡ts] and pictures

In (39) the timing overlap is between the transition from alveolar to bilabial clo-

sure on the one hand and the raising of the velum for the cessation of nasality on

the other:

(39) something like eight thousand people [θaʊz̥n̩pʰiːpəlˠ]

Example (39) contrasts with (40), where nasality ceases before the bilabial closure:

(40) they were rather like unmanned bombs [ʊnmand ̚bɒmz]

Examples (39) and (40), which are illustrated in Figures 4a and 4b, are directly

analogous to Nolan’s hypothetical continuum for hundred (see above, p. 106),

suggesting again that the most straightforward account of deletion or not here

is a CSP one.

In similar vein, CSPs towards the natural end of the scale provide a straight-

forward account of deletion between nasals. The velum is known to move more

slowly than other articulators. It would therefore require extra articulatory effort

to produce (41) with an oral [d] closure (released or unreleased) between the

alveolar and nasal preceding and following consonants:

(41) then it’ll have locked behind me [bɪɦɐɪnmi]

2.2.4 Co-occurring patterns of lenition

Finally, if these cases of lenition are a function of general CSPs on a continuum of

decreased phonetic explicitness, one would expect that they would co-occur with

characteristics of lenition in other segments, and this is indeed what we do find.

In (20) and (22), above, lenition of final /t/ is accompanied by lenition of the first

consonant of the cluster, /k/, which in (20) is fully elided and in (22) is realised

with glottalisation, in the form of creaky voice, but no acoustic evidence of a velar

gesture. l-vocalisation is not a common feature of the York dialect, but there are

ten tokens (out of 130) where /l/ in (t,d) clusters is vocalised, and some where it is

elided altogether. When there is a following word-initial consonant, these always