15th International Congress of Phonetic Sciences (ICPhS-15)
Most previous research on the effect of alcohol on speech has focused
on read speech. On that basis, prosodic changes have been demonstrated.
Some findings, however, were not unanimous, e.g. those on speaking
fundamental frequency (F0) and F0 variability. Another
of most previous studies is that they use a very small number of subjects
(Chin & Pisoni, 1997).
The present study is based on an experiment involving 33 male subjects who described a picture first in sober condition and subsequently at different levels of intoxication. Their individual BRAC (breath alcohol concentration) was measured at peak alcohol levels as well as once every hour during the decay.
Results on semi-spontaneous speech (i.e. a picture description) will be presented. The variables studied include verbosity, voice fundamental frequency (average and variability), and parameters of speaking tempo (syllable rate, articulation rate, and pausing). Results indicate that all speakers produce more text when they are intoxicated than when they are sober. The effect on average F0 is quite clear-cut: whereas subjects with less than 0.08% maximum intoxication level show no significant change, F0 will increase significantly for higher levels of inebriation. The same is true for F0 variability. The temporal organization of the speaking process is also severely affected by BRAC levels above 0.08%: Speech tempo in terms of syllable rate is generally slowed down. For lower levels of intoxication, however, the opposite effect occurs for some of the subjects.
Bibliographic reference. Braun, Angelika / Künzel, Hermann J. (2003): "The effect of alcohol on speech prosody", In ICPhS-15, 2645-2648.