Intonational Phonology (Cambridge Studies in Linguistics) by D. Robert Ladd

Author:D. Robert Ladd [Ladd, D. Robert]
Language: eng
Format: mobi
Publisher: Cambridge University Press
Published: 2008-12-04T05:00:00+00:00

5.1.2 Segmental anchoring

While the facts discussed in the preceding subsection may be relatively theory-neutral, better evidence for a specifically AM interpretation of F0 segment alignment comes from a large number of controlled ‘production’ studies of the acoustic details of intonation contours. Like Bruce’s early work, many of these studies have been based on the alignment of F0 maxima and minima – ‘turning points’ – which, as we saw in section 4.1.2, are readily interpretable as the phonetic reflexes of phonological tones. These studies have repeatedly found that the alignment of F0 turning points is extremely consistent and predictable.

Such consistencies were apparently first noted for English by Ashby (1978). Ashby was investigating the acoustic correlates of the ‘high-fall’ and ‘low-rise’ nuclear accent configurations of the traditional British analysis of English intonation. He had three phonetically trained speakers read a long randomised list of sentences, with no instructions on how to produce them except to specify one or the other nuclear accent type. Ashby found to his surprise that, while there was a good deal of variation in the prenuclear stretches of the sentences, the nuclear accent realisations were extremely consistent. For the high-fall contours, for each speaker considered separately, both the scaling and the alignment of the peak showed very little variation: for example, one male speaker had peaks of about 200 Hz aligned about 35 ms after the onset of the vowel. The final low of the high-fall contours on sentence-final nuclear monosyllables was also extremely constant: the same male speaker reached a low of about 100 Hz 150 ms after the vowel onset. The low-rise contours were characterised by a low, very slightly falling ‘plateau whose length is a linear function of the total voicing time after vowel onset’ (p. 334), followed abruptly by a rise of fixed slope (6.7 octaves/sec.) that continued to the end of voicing.

A decade after Ashby, the first AM-inspired instrumental work demonstrated that the alignment of turning points can be predictably variable as well as predictably stable. Two studies around this time (Silverman and Pierrehumbert 1990; Prieto, van Santen, and Hirschberg 1995) investigated the systematic influences of prosodic context – specifically, the proximity to other pitch accents (‘stress clash’) and prosodic boundaries – on the alignment of accentual pitch peaks. Silverman and Pierrehumbert manipulated the distance between a prenuclear high accent and the beginning of the following (nuclear-accented) word, using made-up names like Mom Lemm, Mama Lemm, Mamalie LeMann. They showed that the peak of the prenuclear accent is aligned proportionally earlier in the syllable the closer it is to the following word – that is, the upcoming word boundary and/or upcoming nuclear accent seem to exert some backward time pressure on the alignment of prenuclear peak. (They also found effects of speech rate, with the prenuclear peaks being aligned later in the syllable when the speech rate was increased.) In much the same vein, but with a rather more complex set of variables, Prieto, van Santen and Hirschberg (1995) studied the alignment of accentual high peaks in Mexican Spanish.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.