VOWELS IN STANDARD AUSTRIAN GERMAN - Acoustics ...
VOWELS IN STANDARD AUSTRIAN GERMAN - Acoustics ...
VOWELS IN STANDARD AUSTRIAN GERMAN - Acoustics ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
177<br />
Vowels in Standard Austrian German<br />
A subsequent study (Brunner et al. 2005) showed that subjects with a flat palate are<br />
allowed less articulatory variability because slight displacements would result in great<br />
changes in the acoustic output, whereas subjects with dome shaped palates are allowed<br />
more articulatory variability, which they can employ or not.<br />
These results strongly suggest that “the objective of speech articulation is to<br />
produce an acoustic signal with properties that will enable the listener to understand<br />
what is said” (Perkell 1997: 363). Within the Acoustic Invariance Theory, some of these<br />
properties have to be invariant. Blumstein & Stevens (1979), who showed that the stop<br />
place of articulation can be arranged according to spectral properties of the burst<br />
(diffuse-raising, diffuse-falling and compact), conclude:<br />
“In particular, it has been shown that the speaker provides the listener with invariant acoustic<br />
cues, cues which can be directly derived from the speech signal itself. Thus, the interface<br />
between the perceptual and production systems resides in the acoustic signal where the<br />
properties of speech can be uniquely and invariantly specified.” (Blumstein & Stevens 1979:<br />
1015)<br />
Despite these initially promising results with respect to the stop place of articulation<br />
(see also Blumstein & Stevens 1980, Stevens & Blumstein 1978, Lahiri et al. 1984,<br />
Blumstein 1986), the search for an invariant acoustic property was unsuccessful.<br />
Löfqvist (1986) argued for a separation of phonetics and phonology in order to get away<br />
from a concept that sees invariance in terms of static entities. In a similar way, Fant<br />
(1986) doubts “the absolute invariance of feature correlates irrespective of context”<br />
(1986: 486) and suggests a context-dependent analysis instead:<br />
“In my view, human speech perception relies on gestalt decoding rather than on isolated<br />
short-time spectral patterns or templates. […] The auditory system probably makes efficient<br />
use of the entire evidence available. Why should we limit our descriptive work to less<br />
precise specifications or to a diluted specification which can operate in all contexts?” (Fant<br />
1986: 487f)<br />
Yet, results on the role of context in vowel perception are not unanimous. The theory of<br />
the dynamic specification of vowels (Strange 1998, Strange & Bohn 1998, Jenkins et al.<br />
1999) states that vowels in continuous speech are better identified when context is<br />
available, whereas the results of Nearey & Assmann (1986) and Andruski & Nearey