Sibilants and Affricates and Stops, Oh My!

I was able to run the experiment in the new Computational and Experimental Linguistics Lab (CELL) facility on campus. There is a sound booth where I give the participants instructions and a consent form, then leave them to run through the experiment. I originally allotted a half an hour of time for each participants, but they went through the experiment surprisingly fast, and rarely stayed more than 10 or 15 minutes. The sound booth was a great place to run the experiment, as it prevented external noises from interfering with the participants results. When the booth was not available, as another student often needed it, I found another quiet room close by. I used the same pair of headphones, in an attempt to keep each run through as similar as possible.

I found 6 different native Korean speakers on campus who were willing to participate. After these six participants went through the experiment, I brought the data to Professor Lunden so that we could analyze it. Quite quickly we saw that one individual did not identify the unspliced, normal words correctly, so their results had to be thrown out.

First we analyzed the three different [t]s. For this group, the data did not match up as close to Kim et al (2002). data as we had hoped. It was however, similar in some aspects. This could have been due to our very small sample size. Also, when we compared the characteristics of our stimuli and Kim et al (2002), we noticed some differences, namely the length of aspiration of lenis consonants. This may prove to be important later. When we analyzed the [tʃ] series, we found that the data lined up quite well with Kim et al (2002). This suggests that the hierarchy of cues found in the stop series by Kim et al (2002) also extends to these affricates. Using this information, we then looked at the sibilants. Unlike the other series in Korean, the sibilant series contains only two consonants. However, the status of one of the consonants is debated. Assuming the sibilants follow the same pattern that the other consonants have been shown to follow, we can analyze the data and determine which consonant type (lenis, fortis, or aspirated) it acts like.  We found that the debated sibilant [s], which is traditionally described as lenis actually follows the same pattern as other aspirated consonants. This gives evidence that the name of this sibilant should be revised.