Seoul Korean is generally thought to have lost its historical vowel length distinction. This has been confirmed by a recent phonetic experiment by Blumstein and Magen (1994). On the other hand, it is generally agreed that the vowel length distinction is still maintained in Chonnam dialect, although there has been few published literature on the prosody of Chonnam except for Jun (1994).
In Jun (1994), Chonnam Korean is described to show some tonal patterns that can be characterized by a specific set of pitch-accent assignment rules. She argues that the position of the H tone is fixed on the second mora of the word, except when a word begins with a laryngeal consonant (i.e. aspirated or tense), in which case the H tone of the second mora spreads to the word initial one, resulting in consecutive word initial HH tones.
(1) Examples (Jun 1994: 53): Italicized
segments (or moras) bear a H tone
(Geminated vowels indicate long vowels)
a. sicang-hata 'be hungry'
(H fixed on the second mora)
siicang-kata 'go to market' (initial HH due to the [+asp]
of /s/)
b. caangsa 'funeral'
(H on the second mora)
cangsak'un 'business man' (H on the
second mora)
Compare the minimal pair 'sicang-(hata)' and 'siicang-(kata)' in (1a) in the picture provided below.
It looks like her theory predicts the tonal pattern properly.
Go
back to the top of this page
Her theory would predict the following
tonal pattern for the names 'Ye:ng-Mi' and 'In-A'.
L H
L
L H
| |
|
| |
m m
m
m m
| |
|
| |
y e e N m i
i n a
The difference between the above
two examples is that 'Ye:ng-Mi' is composed of a long followed by a short
syllable, while 'In-A' consists of two short syllables. She indeed represents
the tonal pattern of 'Ye:ng-Mi' as LHL, although she does not specifically
mention the example 'In-a'. But this is fully predictable from her analyses.
Now, compare the prediction her theory produces with the spectrogram below.
Many other examples with a nasal coda illustrate that in such syllables, the pitch track shows a smooth rising until it reaches the target of the high pich at the end of the syllable. In other words, the locus of the pitch peak is aligned with the nasal coda instead of the second half of the initial vowel.
To represent such alignment pattern, her theory should provide what exactly counts as mora in Korean tonal phonology (i.e. whether nasal counts as mora, or it's just vowels that are moraic) as far as she assume the mora to be the TBU. Also, in such a theory, it needs to be clarified what kind of prosodic hierarchy is assumed with regard to the association of segments to the syllable structure (i.e. whether direct linking of segment to higher prosodic structure such as the syllable is allowed or every segment should be dominated by a mora before being connected to a hier prosodic unit as in Zec (1988)).
Although Jun does not discuss the tonal patterns that Chonnam exhibit in calling contours, her analysis would predict the following tonal patterns for the names 'Sun-Ja (/suunja/)' and 'Su-Jin (/sujin)' in declaratives:
HH L
HH
suunja
sujin
Interestingly, only 'sujin' loses its initial H tone when used in vocative chant. The following illustrate how these names are realized in declaratives and vocative chants.
 |
 |
|
|
|
 |
 |
|
|
|
Go
back to the top of this page
In an analysis where the TBU in Chonnam is syllable, instead of mora,'Young-Mi' would be categorized as simply H.L ('.' indicates syllable boundary) in declaratives instead of LH.L, while 'In-A' is L.H. Likewise, 'Sun-Ja' would be H.L both in declarative and vocative, while 'Su-Jin' is L.H in both declarative calling. In other words, there is no tonal difference between the declaratives and the calling-contour in Chonnam.
One might ask why 'Su-Jin' in declarative is LH instead of HH. This is because I interepret the laryngeal effect as phonetic, whereby the following vowel is realized with a higher pitch. Although Jun treats this laryngeal effect as phonologization, thus interpreting the consequent tonal pattern as H, I argue that this is phonetic as can be evidenced by its relatively weaker influence in the calling contour (i.e. although the vowel following a laryngeal is realized higher than that does not, it is still much lower than a phonological High tone as can be seen in 'Sujin-a' (vocative).
More importantly, it will be shown that a high pitch caused as a result of the laryngeal effect does not have any correlation with the duration of the vowel, whereas a phonological H tone does have one. This strongly supports my argument that the laryngeal effect is phonetic.
To reflect the relatively strong laryngeal effect in
the declaratives compared with the vocatives, later I assume a phonetic
rule that changes the L to H for the initial laryngeal syllable, thus eventually
making Su-Jin (declarative) as H.H at the phonetic level. However, this
is just a convenient label for mnemonic purposes. Again,
however, it should not be confused with its phonological identity as
L.H.