Argument: Korean foot structure is unbounded rather than iambic

1. Overview: Chonnam Korean is a stress language
2. Is Korean Iambic?
3. Problem with iambic analysis: cannot explain calling contour patterns
4. A way out

1. Overview
In conjunction with the second question on the TBU, I want to investigate whether Chonnam Korean is a stress language or not. There can be both phonetic and phonological factors to help determine this property. If Chonnam were a pitch-accent language as Jun argues, what is perceived as prominence should be mainly a matter of melodic variance rather than involving other acoustic factors such as duration or amplitude.

Many of the data show that the syllables that are perceived as stressed involve higher pitch as well as greater amplitude, which suggests that Chonnam Korean is a stress-accent language according to the definition of Beckman (1987). Test this hypothesis in the above spectrogram; you'll notice the duration and amplitude difference associated with the pitch in the first and the second syllable of each word. This suggests that in 'sicang' (hungry), the stress is on the second syllable, while in 'siicang', the stress is on the first.

Moreover, the H pitch in (sino)Korean words are limited either to the first or second syllable, which is a characteristic of a stress language rather than a pitch accent one. Later I will come up with an unbound foot analysis of Korean, where the locus of the stress is decided by the lexical tone. (footnote)

The above facts suggests that Korean is a streee language.

2. Iambic?

Then, what kind of foot structure does Korean have? Let's assume for now that it is iambic, and the coda contributes to the weight of the syllable, following Kim (1987). Then, we would expect the following stress pattern for 'Sun-Ja', 'Su-Jin' and 'Sang-Won', respectively.

su'nja, suji'n

This is insufficient to predict the locus of the high pitch. So let us apply the 'laryngeal effect' discussed in Jun (1994).

H H
sunja, sujin

Applying the 'laryngeal effect' will redundantly predict a H on the initial syllable of 'sunja', but will predict a more or less correct pitch shape for the above data.

H H
su'nja, suji'n

Pitch track of 'Sunja' and `Sujin' in Declarative


Sunja (Declarative)	Sujin-i (Sujin-citational suffix) (Declarative)

Recall that according to Jun, 'Sun-Ja', 'Sujin' and 'Sangwon' should be 'HH.L', 'H.H', and 'H.H'. Let's assume that this is the correct annotation of the pattern. (footnote)

In stress approach, the second syllables will get a H as one of the acoustic factors in realization of the stress, and then the first syllable of 'sujin' and 'sangwon' will also get a H due to the laryngeal effect suggested by Jun (1994).

Thus even though the sources are different, both syllables will get a H tone in 'sujin' and 'sangwon' in either analysis.

3. Problem

However, let us see whether this stress approach can provide an explanation for the calling contours.

Pitch track of 'Sunja' and 'Sujin' in Vocative


Sunja-ya (Sunja-vocative suffix)	Sujin-a (Sujin-vocative suffix)

For the above examples 'sunja-ya' and 'sujin-a', we could simply say that the laryngeal effect is not at work in calling-contour. (In OT terms, the ranking between the Align (H, Laryngeal) >> Align (H, FtHead) is reversed in calling contour, i.e. Align (H, FtHead) >> Align (H, Laryngeal)). This predicts the following calling contour:

H H
'su:nja-ya', sujin-a.

So far, it seems to work fine.

However, the following examples challenge this analysis:

Examples that show LH in calling contour


Jung-Min-a Nol-ca (jungmin-voc, play-let's)	Dong-Chul-a nol-ca (Dongchul-voc, play-let's)

Jung-Min, Dong-Chul, Myung-Jun, etc.

The above examples are all Heavy syllable followed by a Light one, just as in the case of 'Su:nja'. According to the proposed analysis that H is attracted to the Heavy syllable, without the laryngeal effect, we expect HL both in the declarative as well as calling. However, they show LH shape in calling.

So, we face a dilemma in stress appraoch:

We cannot say that the H tone is assigned to the syllable with a laryngeal onset in Chonnam calling contour since there are problematic data such as 'sujin' whose initial syllable is not realized as 'H' in vocative.

Meanwhile, if we try to explain the H tone assignment purely in terms of the syllable weight, we are again faced with the data that do not conform the generalization such as 'Jung-Min, Dong-Chul, Myung-Jun', etc., whose initial syllable is not realized as 'H' even though they are Heavy. One could alternatively suggest that it is only vowel length that counts toward the syllable weight: but this is not plausible given the relatively free performance of the vowel length.

4. A way out

As such, neither pitch-accentual nor stress approach alone can explain the pitch variation found in Chonnam declarative and vocative chant.

However there is a way out:

Pitch accentual consideration is out given the obvious correlation between the High pitch and the duration as will be statistically proved in the next section.

Stress approach is still more plausible: then the only way out is to discard the iambic system where stress is assigned by the weight of the syllable.

In the next section, I propose a new analysis where the stress assignment is decided by the lexical tone of the syllable.

Footnotes
1. In literatures of Korean phonology, it has been argued that Korean has a vowel length shortening, which can be explained by rhythmic rules such as shortening and lengthening. Namely, Korean is argued to have an iambic foot system, and vowel length is often adjusted to meet this foot type. For example, `so: (cow)' is shortened when compounded with 'kacuk (leather)' as in 'sokacuk (leather of cow)', which, as a result, contains two iambic feet without any unparsed syllable. (NB. I do not agree with this iambic analysis of Korean foot structure. As will be discussed later, I assume an unbounded foot for Korean, and the length of the syllable is closely associated with its lexical tone. I will develop a full picture of the Korean foot system in a subsequent work.

2. I don't think this is the right annotation of the pitch contour. I will argue that the high pitch of 'su' in 'sujin' is a phonetic effect, thus cannot be given the phonologcal tonal description as H. After all, I will argue that the stress is on 'jin' which is the source of the H tone in the second syllable. More importantly, pay attention to the 'length' associated with this H tone, especially in the calling contour.