Abstract:
The main objective is to study and compare the behavior of fundamental frequency on vowels influenced by voiceless and voiced initial sonorants in three tonal languages: Hmong, Mien and Mal, in order to prove the hypotheses that (1) the fundamental frequencies of vowels following the voiceless initial sonorant are higher than those following voiced initial sonorants, (2) similar behavior will be found in all of the three languages investigated, and (3) the F0 difference is statistically significant. The data was recorded from three female native speakers per language using the Cool Edit Pro v. 2.5, acoustically analyzed using Praat v.4.4.04. and statistically analyzed using SPSS for Windows v. 13.0 with the two-tailed (p<0.05). The investigation was carried out in two main stages. In the first step, the acoustic characteristics of the tones were analyzed from 330 test tokens in order to give an overview of tones in the three languages. An analysis of the fundamental frequency at every 25% (0%-100%) of normalized times was done, while duration was taken into account to help interpret the fundamental frequency. In the second step which is the main objective of this study, 2,250 test tokens were used to investigate the fundamental frequency of vowels followed by voiceless and voiced initial sonorants. Measurements were done every 25 milliseconds. This research reveals that Hmong, Mien and Mal have the same behavior of fundamental frequencies, i.e. those of vowels following voiceless initial sonorants are higher than those following voiced initial sonorants thus supporting the hypothesis. As expected, the F0 difference is statistically significant. Generally, it can be concluded that initial sonorants have an influence on the fundamental frequencies of the following vowels: voiceless initial sonorants cause higher frequencies, effectively increasing the pitch level, whereas lower frequencies are generated from voiced initial sonorants resulting in a lower pitch level. This finding suggests that in the future some tones in Hmong, Mien and Mal will probably split into two groups: a higher set and a lower set.