| |
Abstract:
Alveolar to velar place assimilation commonly occurs in
English, particularly at faster rates of speech. Empirical
studies have shown that assimilation is gradual on the basis of
evidence of partial alveolar assimilations. This suggests that
assimilatory variability is phonetic and not phonological in
origin. Partial assimilations have been mainly identified from
tongue-palate contact data (EPG), however, there has been
speculation (and only limited evidence) that more subtle reduced
coronal gestures, which do not result in contact with the
hard-palate, can be identified from EMA data. It has been
hypothesised that assimilation never reaches completeness. First,
a number of speakers' multiple productions of /n#k/ within
sentences were recorded using EPG. In fast speech, partial
alveolars were indeed identified. However, the distribution of
these was restricted to only some speakers who produced them as
part of a continuum of alvoelar reduction. Surprisingly, for
other speakers, when assimilation occurred it was always complete
suggesting that a high-order phoneme substitution rule is at work
for these individuals. A follow-up study using EPG in combination
with EMA was carried out to look again for reduced coronals in
the latter group. No such EMA-defined partial gestures were found
confirming the preliminary findings. This indicates that
contact-only data is sufficient to reveal assimilatory variants
and to distinguish important inter-speaker strategy differences.
Also discussed are the psycholinguistic implications of these
different assimilation strategies.
|
