A Concept Formation Experiment on the Nasality of Vowels in Taiwan Min

In Wang 1996 we performed an experiment to test if the subjects treated the oral voiced stops and their corresponding nasal stops as the same phonemes. We found that the [b, m] pair and the [g, N] pair were unambiguously accepted, while the [l, n] pair was equivocal, and tended to be regarded as different phonemes. We concluded from the results that the theoretical predictions were generally correct, but the /l/ phoneme was on the way to be split in two. In this present study we pursued the corollary question: whether the oral and nasal vowels were considered different. Four groups of subjects were recruited, with each group involving in one of four test conditions. In each test condition, the subjects were required to respond positively to the oral vowel [i], [e], [a] or [ç] respectively, and negatively to other vowels. In the test session, the subjects were tested on how they respond to the nasal counterpart of the target vowel in the learning session, i.e., [i)], [˜e], [˜a] or [ç)] respectively. The results showed that the test vowels received much more ‘yes’ answers than ‘no’ answers, meaning that they considered the nasal vowels to belong to the same category as the oral vowels. In other words, nasality is not important in vowels either. Therefore, we conclude that the segmental theory of the nasal phonemes in Taiwan Min is not supported by the results of this experiment. We propose an autosegmental theory in its stead, as outlined in Wang 1995. The validity of such a claim, however, is a subject for further experimental research.