While there has been descriptive research done on the phonetic and phonological properties of Shoshoni, little to no acoustic analysis has been done thus far. This paper analyzes the production of initial stops by a native speaker of the Northern dialect of Shoshoni. Previous phonetic and phonological work defines Shoshoni as a single stop series contrast (Gould & Loether, 2002). However, given relatively limited acoustic work on single stop series languages, and Shoshoni’s high level of contact with English, which has a two-way stop contrast, it was unclear what this would mean on the acoustic level. Furthermore, in Shoshoni language classes, English stops are described as voiced and voiceless aspirated with Shoshoni stops differing from English by being voiceless unaspirated. However, acoustically, it is widely acknowledged that English has a contrast, not between voiced and voiceless aspirated, but rather between voiceless aspirated and voiceless unaspirated (Lisker & Abramson, 1964). Therefore, it was unclear how this contrast would manifest in Shoshoni acoustics. Because of the influence of English on Shoshoni, this study also analyzes the speaker’s production of initial stops in English to compare and contrast the production.
The study found that initial stops in Shoshoni do indeed appear to be voiceless unaspirated with an average VOT of approximately 19 ms. However, the average voicing onset times were more complicated for the velar stops as velar stops appear to be produced with glottalization. Interestingly, this glottalization is also evident in the speaker’s production of initial stops in English for both /k/ and /g/. Initially, it was hypothesized that the speaker’s production of English stops /b/, /d/, /g/ would be voiced to explain the distinction between English and Shoshoni that is sometimes described by language teachers. However, while it did seem that initial stops in English were more likely to be voiced by a Shoshoni speaker than by the average native English speaker, this was not always the case.
Theories on single stop contrast languages often conclude that the range of values for voiceless unaspirated stops in a single stop contrast language will be broader than values for language with a two way contrast (Brookes & Kempe, 2014). These data may support this hypothesis as the average values for voiceless unaspirated stops in Shoshoni is higher than the average for voiceless unaspirated stops in English. However, the range of values seen in these data is not especially broad when the velars, which should be categorized differently due to the glottal features, are excluded. Furthermore, the average VOT for voiceless unaspirated stops in Shoshoni is still within the range of acceptability for voiceless unaspirated stops in English. Therefore, this paper proposes a variety of options to further investigate this theory, including a broader sampling of speakers and tokens as well as perception tests in which speakers could rate tokens with varying VOT.
Please join us for refreshments before the talk at 3:30 pm in 571 EPB.