Page 40 - Fall 2011

Fig. 5. Oral-glottal alignment configurations from EGG signal. VCLS represents voicelessness during closure. B represents the voiceless burst duration following closure release. Figures from DiCanio (in press). Reprinted with permission. like the one in Itunyoso Trique as strength contrasts, i.e. “fortis” and “lenis” stops. DiCanio (2012) examined the tim- ing of vocal fold vibration in Trique consonants using electroglottography (EGG) in order to determine exactly what accounted for this variability in voicing in short stops and to investigate if another explanation could account for the pattern. With EGG, sensors are placed on opposite sides of the speak- er’s neck, just over the thyroid cartilage (below the Adam’s apple) through which a weak electrical current is passed. When the vocal folds are closed, more of the current can pass from one side to the other. When the vocal folds are open, less current passes through. EGG maxima correspond to the moment of maximum contact between the vocal folds while minima correspond to the moment of mini- mum contact between the vocal folds (Childers and Krishnamurthy, 1985; Childers and Lee, 1991; Heinrich et al., 2004). The presence of EGG maxima and minima indicates that there is vocal fold vibration. EGG data are typ- ically collected along with acoustic recordings for the identification of acoustic-phonetic boundaries. The advantage to using an elec- troglottograph is that the EGG signal is unaffected by acoustic disturbances in field recordings. In rural villages, recording is often done in quiet spaces in private homes. Typically, houses are constructed by community members without the help of an electrician. Thus, in addition to the external noise found in these communities, there are often ground loops due to the use of low current wiring. Ground loops in AC power lines can produce an unwanted signal in acoustic recordings, with a fundamental frequency of 60 Hz and its associated harmonics. To adjust for this effect, a low stop filter can be applied to the recordings; yet, this filter also eliminates low frequency voicing from the signal. Devoicing typically involves low frequency and low ampli- tude glottal pulses. EGG is ideal for examining devoicing in these less- than-ideal recording conditions because it accurately captures low amplitude voicing. There are three ways in which vocal fold vibration may be aligned to the clo- sure of a stop in running speech. If the sound preceding the stop is voiced, voicing may cease coincidental with the closure of the following stop. This is simultaneous oral-glottal alignment. Voicing may also cease during the vowel duration prior to stop closure. The time between devoicing and clo- sure is called voice offset time (Vo). This is the gestural configuration used for Phonetics of Endangered Languages 39