This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Lecture notes: 5.2 July 7, 2009 1 Phonology problems Let’s look at a problem we saw (if we did the reading for today) – Mokilese. Mokilese is an Austronesian language of the Malayo-Polynesian family, spoken in Micronesia. We want to explain the distribution of the voiced and voiceless vowel pairs [i, i ˚ ] and [u, u ˚ ]. (Reminder: voiceless vowels have a circle under the phonetic vowel symbol.) (1) Mokilese voiceless vowels: a. [pi ˚ san] full of leaves g. [uduk] flesh b. [dupuku ˚ da] bought h. [kaskas] to throw c. [pu ˚ kol] basket i. [poki] to strike something d. [ki ˚ sa] we two j. [pill] water e. [su ˚ pwo] firewood k. [apid] outrigger support f. [kamw O ki ˚ ti] to move l. [lu > d Z uk] to tackle There are no minimal pairs in the data where [i] vs. [i ˚ ] are the only different sounds between the pair, and none where [u] vs. [u ˚ ] are the only different sounds. Even a near-minimal pair would cast grave doubt on the idea that [u], [u ˚ ], e.g., are allophones of one phoneme. Imagine if we found the pair [dupu ˚ kda] / [bupukda] (hypothetical). This would be hard to reconcile with the idea that u and u ˚ belong to the same phoneme, because it is extremely unlikely that the decision whether or not we get [u] or [u ˚ ] in position #4 depends on whether or not we find [b] or [d] in position #1. Why is this extremely unlikely? Because the conditioning environments that determine the distribution of allophones 1 of a phoneme are overwhelmingly local – very close to the allophone in question, not far away. We had concluded that [i ˚ , i] are not in contrastive distribution: no minimal pairs. And the same for [u ˚ , u]. Our hypothesis: [i, i ˚ ] are allophones of one phoneme, and [u, u ˚ ] are allophones of one phoneme. To see whether that’s true, we need to try the local environment method. So we can next test the hypothesis that the pairs [u], [u ˚ ] and [i], [i ˚ ] are each allo- phones of one phoneme, /i/ and /u/. We proceed to look for complementary distribution, keeping in mind what we learned about natural classes. Our hypothesis: [i] and [i ˚ ] are allophones of one phoneme, and [u] and [u ˚ ] are allo- phones of one phoneme. Are the sounds in complementary distribution? What is the rule? Lets call these phonemes /i/ and /u/, respectively, assuming that [i] and [u] are the default allophones and that [i ˚ ] and [u ˚ ] are derived. We could also have tried /i ˚ / and /u ˚ /, but /i/ and /u/ seem more plausible. And we can always reconsider our choice. We collect [i ˚ ] and [i], then the same thing for [u ˚ ] and [u ˚ ]. Second step: evaluation of local environment tables. Evaluation of preceding context for [i ˚ , i]....
View Full Document
This note was uploaded on 04/17/2010 for the course LINGUISTIC 117 taught by Professor Farkas during the Spring '09 term at University of California, Santa Cruz.
- Spring '09