Page 1

The Structure of The Korean Sijo

David R. McCann

Harvard Journal of Asiatic Studies, Vol. 36. (1976), pp. 114-134.

Stable URL:

http://links.jstor.org/sici?sici=0073-0548%281976%2936%3C114%3ATSOTKS%3E2.0.CO%3B2-P

Harvard Journal of Asiatic Studies is currently published by Harvard-Yenching Institute.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained

prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in

the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/journals/hyi.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic

journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,

and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take

advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org.

http://www.jstor.org

Sat Jan 26 14:01:34 2008

THE STRUCTURE OF

T H E KOREAN SIJO

DAVID

R. McCANN

CORNELLUNIVERSITY

OREAN verse is most often described as syllabic, largely,

IK

it seems, because of the influence of Chinese and Japanese

verse analysis. But although it has long lived in the shadows

of its two better-known neighbors, Korean verse is quite unlike that

of either Japan or China.

One of the most characteristic Korean verse forms, the sijo, is

still practiced today, to the accompaniment of a musical setting

which actually constitutes a detailed analysis of the form, the terms

of which are readily translatable to those of literary prosody. Using

an analysis of this musical setting as a keystone, this study of the

sijo has five main divisions:

(I) a critical examination of the traditional description of the sijo

form ;

(2)

an analysis of the four-part syntactic structure immediately

apparent in any sijo text;

(3)

analysis of the musical setting of the sijo;

(4)

translation of the terms of the musical setting to those of

prosodic analysis as presented in the articles "Prosodic Notation"

and "Prosody" in the Princeton Engclopedia of Poetry and Poetics

(Princeton

:

Princeton University Press,

I

965) ;

(5) in the light of the preceeding analysis, an examination of the

pattern of syllabic distribution within the four-part sijo line.

The conclusion of this study is that the sijo must be seen as an

accentual verse form, not syllabic.

This analysis of the sijo form is based on the twenty-nine texts by

kisaeng

#&,

professional women entertainers, that are collected in

the Sijo munhak sajdn (hereafter SMS)

@32@$&

(Dictionary of

Sijo Literature), compiled and edited by Chijng Pyijng-uk

m@iE

114

KOREAN S I J O

115

(Seoul,

1966).

There are several reasons for choosing this particular

group of texts. The selection arises naturally from within the

sijo

tradition as it operates, rather than by the imposition of some

external criteria of selection. The number of texts is of a manageable

size; at the same time, these twenty-nine texts represent fifteen dif-

ferent authors, a sampling diverse enough to remove the individualis-

tic aspects of

sijo

style that appear in other samples of comparable

size, such as the forty verses of the "Fisherman's Calendar"

("busashisa"

&A@zfl)

by Yun S6n-do

93%.

Finally,

I

assume that the

kisaeng,

as professional entertainers, would have

composed

sijo

as they would have performed them, in accordance

with the expectations of their patrons. It seems reasonable to expect

that their approach, with respect to the literary form, would have

been conservative rather than innovative, and that their

sijo

works,

therefore, will adequately represent the structural principles of the

sijo

as a received form.

Although the

sijo

was traditonally written for sung performance

to a musical accompaniment, descriptions of its structure have in

general treated it either as a musical or as a literary piece. We shall

follow this example for a time, distinguishing the "work of literary

art.

. .

from its production or reciting" in reviewing the major

features of typical descriptions of the

sijo

as a literary f0rm.l We

shall then turn to its performance as a song, however, to consider

aspects of its dynamic structure which may be used for a more

phonologically oriented description of the

sijo

as a verse form.

I .

Traditional Descriptions of the Sijo

There are three verse forms called

sijo:

the standard, three-line

p)dng sijo

;

the slightly expanded variant

dssijo

W

@a;

and

the greatly expanded narrative

sasdl sijo

%%@a.

All of the

sijo

by

kisaeng

are standard; for contrast we shall consider one

dssijo

by

a "palace woman."

The standard form has generally been described as a three-line

verse form, each line being composed of four "groups" of from three

to five syllables each. These "groups"-the

nature of which we shall

consider later-are in pairs, two in each hemistich, with intervening

A. W. De Groot, "Phonetics in

its

Relation to Aesthetics," B. Malmberg, ed., Manual

of Phonetics (Amsterdam: North Holland Publishing Go.,

1968),

p. 533.

116

DAVID

R.

McCANN

caesura. The usual pattern of syllabic distribution among the groups

is generally presented as follows

:

"Group"

I

I1

I11

IV

No. of syllables

:

line

I

3 4

3 o r 4

4

line

2

3

4

3 o r 4

4

line 3

3

5

4

3

This kind of description suggests that each line is composed of

fourteen or fifteen syllables; much analysis of the sijo form proceeds

as if this were indeed the case-as if syllable counting were the

primary structural feature of the sijo. But variation in syllable count

is the rule, not the exception, in sijo verse. Various accounts give

somewhat different ranges to the variations within the groups; for

the twenty-nine texts which form the basis of this analysis, the fol-

lowing variations are found

:

"Group"

I

I1

I11

IV

line

I

2-4

4-5

2-4

4-5

line

2

2-5

3-4

3-6

3-4

line 3

2-3

4-7

3-4

3-5

To assert that so diverse a series of numbers is the basis for con-

structing a verse system has led, in general, either to hedging about

the description, or to variously disguised apologies for the Korean

versifier's inability to keep to the pattern. We shall find, however,

that the syllabic configuration can be analyzed only within another,

overriding structural design.

The rhetorical form of the sijo is said to follow that of the Chinese

quatrain: statement and elaboration in the first two lines, a "twist"

in the first part of line three, and resolution in the remainder of the

third line. For an illustration of this rhetorical structure, and as

the basis from which to begin our analysis of the sijo form, we shall

consider a well-known text by the kisaeng Hwang Chin-i

AEfj+

(SMS, p. 482)

:

Summarized in Richard Rutt, "An Introduction to the Sijo," Transactions of the

Korea Branch of the Royal Asiatic Society, 34 (Seoul. 1958)7-1

1

;

Yi T'ae-giik

$%&,

Sijo

kaeron

@a@&%

(Seoul. 19 74),

pp.

79-1

10.

KOREAN SIJO

"Group"

I

I1

I11

IV

ch'bngsan.

pybkkyesu.

su.i kamal

charang

ri (within

ya (jade blue (easy going) mara

green hills) stream

0)

(don't boast)

ilto

ch'anghae

tora.ogi

bry6 wbra

(once reach) hamybn

(to return) (is hard)

(wide sea, when

you do)

mybngwb1.i man-kongsan suyb kandal bttbri

(bright

hani (fills

(pausing,

(how

moon)

empty hills,

then going about ?)

since it does) on)

The first line makes a statement, and the second line elaborates

on the circumstances of that statement. The first part of the third

line introduces a new element, the moon, for the "twist," while the

remainder of the third line resolves the poem. These sections are

more conveniently discernible in translation

:

Deep blue stream, don't boast so loud

of your passing through these green hills.

Though your way runs swiftly down to the sea

there is no such easy return.

While the bright moon floods these lonely hills

why not pause? Then go on, if you will.

The syllabic distribution among the "groups" of this sijo is in the

following pattern :

Traditional analysis will note that the only variation from the

"basic" pattern of syllabic distribution is in the first "group" of

line two, where there are two syllables rather than three. But what

does this reveal about this particular text? The question needs to

be raised: just what are these "groups"? They have been described

in quasi-prosodic terms as "breath groups" and "feet" (ambo

g$)

in several accounts; another refers to them in more grammatical

118

DAVID R. McCANN

terms variously as "phrases," "subsections," "sections," and

6 6

part^."^

The terms for these units aside, what is their relation to

units of the natural language? Are they metrical feet, having

no direct correspondence to syntactic or phonological units, or are

they something else entirely?

2.

Syntactic Structure of the Sijo

For those who read the texts in the original, the problem of

defining their constituent "groups" has most likely seemed in-

consequential: each line is simply divided into four syntactic units.

The syntactic divisions in the text by Hwang Chin-i are as follows:

line

I

ch'bngsan.ri

prep. phrase

pybkkyesu.ya

noun+

interject.

su.i kamiil

vbl. noun

charang mara

verb

line

2

ilto

verb

ch'anghae

hamybn verb

tora.ogi

vbl. noun

brybwbra

verb

line 3 mybngwb1.i

noun

man-kongsan

hani verb

suyb kandal

vbl. noun

bttbri

verb

This pattern of syntactic division might be described as a series

of four "slots" in each line. The specific type of syntactic form which

fills a slot within a given text is not predetermined; it is to be de-

fined, rather, by internal contrast. That is, in each line there will

be four syntactic units distinguishable by mutual contrast. In one

line, for example, the four slots might be composed of two modifier

plus noun pairs, as in the first line of a sijo by the kisaeng Maehwa

@E

(SMS, p.

256)

:

saldiirhbn

nae ma.iim kwa

aldiirhtin nim.iii ch6ngiil

(affectionate) (my heart and)

(fond)

(love's feelings)

Another text by Maehwa has a line composed of four verb forms

(SMS7 P. 444)

:

chug6

nijbya harya

sara

kury6ya harya

(dying)

(must

I

forget)

(living)

(must

I

yearn)

Ch6ng Py6ng-uk

@pfig,

Kungmunhak san'go

@*@gg

(Seoul: Shin'gu Munhwa Sa,

1g5g),

p. 163; Kim S6g-y6n

&$$R

"Sijo unyul fii kwahak-ch6k y6n'gu" &?&@LL)

119

KOREAN SIJO

To describe the four-part syntactic structure of the sijo line, and

to prepare for relating that structure to the prosodic, rhythmic

features of the sijo, we shall examine the "deep" syntactic structure

of two texts. Each will be taken as a single, whole "message" and

then analyzed into its grammatical immediate constituents (IC),

from the sentence level (labelled

"4"

in the diagram analysis), to

the phrase

("3"),

word

("2"),

and morpheme

("I

"), commonly

defined as the "smallest unit of grammatical analysis."*

I

shall

consider first the text by Hwang Chin-i, and then a variant dssijo

by a "palace woman.'' The latter, despite great divergence from the

usual syllable count, shows a pattern of syntactic structure clearly

uniform with that of the standard type represented by Hwang

Chin-i's piece.

The three lines of the text by Hwang Chin-i correspond to three

sentences, with sentence-final verb forms signalling the line ends.

The first sentencellhe may be divided into two immediate con-

stituents

:

ch'6ngsan.ri-pybkkyesu.ya

su.i kamid-charang mara

(green hills in-jade blue

(easy going-boast don't)

stream

0)

The interjection ya signals the end of the first constituent, a noun

phrase and interjection; the sentence-final verb form mara signals

the end of the second.

These two constituent elements may be further analyzed into two

units each:

ch'bngsan.ri pybkkyesu.ya

and

su.i kamal charang mara

The syntactic divisions are signalled by the preposition

ri

("within") and the direct object particle zil, respectively.

We shall analyze these four units in turn.

$$@&I@%,

Aseay8n'gu

$j@?$@%

(Korea University),

I

1.4 (1968), 7-9; Peter H. Lee,

Korean Literature

:

ToPics and Themes (Tucson: Univ. of Arizona Press, 1966), p. 38;

Richard Rutt, The Bamboo Grove: An Introduction to

Szj'o

(Berkeley: Univ, of Calif. Press,

'971)P.

'0.

David Crystal, Linguistics (London: Penguin Books, 1971)

p.

187.

120

DAVID

R.

McCANN

ch'dngsan.ri: a phrase composed of a bipartite noun phrase, ch'dng

("blue") +sun ("mountain") followed by the preposition ri, graphi-

cally represented as

ch'6ng san ri

u

I

pydkkyesulya: a phrase composed of a three-part "proper noun,"

pydkkyesu ("jade blue stream") followed by the interjectionya:

py6k kye su ya

w

I

su.i kamal: a phrase composed of the adverbial su.i modifying the

verbal noun kam with the object particle

iil

appended:

su i ka m 81

(m

:

nominalizing suffix)

charang mara: a verbal phrase composed of the substantive charang

("pride") +the auxiliary verb form mara ("do not")

:

charang ma ra

(ma: verb base;

I

v

ra

:

imperative ending)

This analysis for the complete line is diagrammed as follows

Following the same procedure for the second and third lines:

121

KOREAN SIJO

While the diagrams are schematic representations of a grammati-

cal analysis of these lines, the dotted line cutting across at the

"3"

level intersects the four syntactic slots which form the basic pattern

of the sijo line. This four-part division of the line is a basic pattern,

immediately noticeable in any text, but we must also be aware of

the great variety of composite syntactic structures which underlie

the basic pattern. We find that syntactic units of such disparate size

and complexity as ch'bngsan.ri and man-kongsan hani are identical,

nevertheless, with respect to their relevance as constituent units in

the structure of the line.

As for the rhythmical implications of this pattern, the first IC cut

in a line (level "4") corresponds to the line-medial caesura mentioned

in most descriptions of the sijo form. For the most part this corre-

spondence between syntactic and rhythmic units prevails in each

hemistich as well, so that syntactic and rhythmic juncture are

superimposed.

This is not invariably the case, however. In the first hemistich of

line two, for example, both syntactic and semantic segmentation

would divide the hemistich at a point between the Chinese phrase

"once reach (the) wide sea" --$IJ$$& and the auxiliary verb

hamybn, "when (you) do." In performance of this line, however, two

factors lead to placement of the juncture in the middle of the phrase,

after "once reach." The first factor is a dominant tendency in the

Korean language to form verbs by combining a bisyllabic Sino-

Korean term with the auxiliary verb hada. Thus kaga

q%

"praise-

worthy" with hada becomes kaga hada, "to be praiseworthy." In

122

DAVID

R.

McCANN

general practice, then, two syllables, rather than four, would be

anticipated before hamydn, producing ch'anghae hamydn as a rhythmic

unit, even though it makes no sense out of context. The second

factor is a general tendency in sijo verse to have, within a hemistich,

a longer group follow a shorter one. General rhythmic patterns,

then, of both the language and the verse form, take precedence in

a case like ilto ch'anghae hamydn over syntactic and semantic segmen-

tation in determining the rhythmic grouping of the line.

The text by the "palace woman" provides confirmation of the

primary importance of the four-part syntactic and rhythmic struc-

ture of the line. The syllable count of this text is quite different from

the one by Hwang Chin-i, although it does return to the "normal"

pattern in the final line (SMS, p. 320)

:

ap m0s.e tiin kogidii1.a ne wa tiinda

nui nbriil morbdaga ybgi nal tiinda

pukhae ch'bngsoriil bdi tug0 i m0s.e wa tiin

tiilgodo mot-naniin chbngiin ne.0 nae.0

tariirya

g

8

da

5

3

llables

4

4

5

7

3

4

5

3

3

0 fish placed in the pond before me, why have you come?

Did someone gather and put you here?

What clear pool have you left somewhere on the Northern Sea

to enter this pond?

Once in, is the feeling of no escape

any different for you or me ?

The first line is composed of two sentences:

nui nb ral morb daga ybgi nal tiin da

(1)

(2)

(3)

123

KOREAN SIJO

The four units which compose this line are clearly of

a

different

order from those in the text by Hwang Chin-i. Two potentially

distinct units, up mos.e tiin ("placed into the pond in front") and

kogidii1.a ("0 fish"), which are syntactically analogous to the two

units of Hwang Chin-i's text, ch'bngsan.ri and pybkkyesu.ya, function in

the palace woman's text as a single, composite unit, bound together

by intonation, at the sentence ("4") level, rather than word accent

at the

"3"

level. The primary importance of the four-part structure

of the line is made all the more apparent by the differences in

syntactic and syllabic form.

The second and third lines of the palace woman's text are more

regular in syntactic pattern, being single sentences, and in syllabic

distribution, particularly in the final line

:

nae o tarti rya

I

The notion of four syntactic slots in each line is helpful in analyz-

ing the syntactic structure of the sijo form, and particularly in com-

paring the structure of various texts. But it should not be viewed

alone as the primary structural principle of sijo verse. Syntax is to

be seen more generally as the framework for any given verse system;

whatever schematization it appears subject to usually reflects the

organization of prosodic

feature^.^

There may be a high degree of

Cf. William Skillend, "Words from the Heart," Tonga Munhwa

xi$*ik

No. 5 (June

1966), pp. 157-87; John Lotz, "Elements of Versification," W.

K.

Wimsatt, ed., Ver-

sijcation: Major Language Types (New York: New York Univ. Press, 1g72), p.

10.

124

DAVID R.

McCANN

correlation between syntactic and rhythmic units in a text (or none,

as in English metrical verse), but where this correlation is relaxed,

as in ilto ch'anghae hamy2in, or where greatly disparate syntactic units

function as corresponding composite units in the line, as in ap mos.e

tiin kogida1.a and ne wa tanda, the predominance of rhythmic grouping

as the relevant structural criterion is clear.

3.

The

Musical Setting of Sijo

In preparation for a more detailed examination of the dynamic,

rhythmic structure of the sijo, let us recall that all of these texts were

written with known melodies in mind that would have imposed a

rhythmic pattern of their own on the linguistic material. From syn-

tactic analysis, then, we turn now to a consideration of the musical

performance in order to develop a model of the rhythmic structure

of sijo song which may be used, with some terminological transla-

tion, for prosodic description.

However many sijo melodies there may have been in the past, at

present there is but one. This standard sijo setting treats the text in

three major sections, corresponding to the lines of the literary text,

with instrumental transition between sections. Each section is in

turn treated in subsections, lines one and two with five, and line

three, with three. These subsections correspond to the four line-

constituent units we have found in the literary text, with the fol-

lowing exceptions: the final (fifth) subsection in lines one and two

is used for the last syllable of the line, which is detached and sung

as a separate, somewhat shortened musical unit; in line three, the

final (group IV) unit is omitted in performance. Numbering the

musical subsections sequentially and arranging them in the four

line-constituent slots, we have the following pattern:

I

I1 I11

IV

line^:

1 2 3

4

5

linez:

6

7

8

910

line

3: 11 12 13

(omitted)

The musical subsection is a dynamic unit sung as a single phrase.

It is characterized by a singer's strong initial attack, reinforced by

combined hand and stick beat on the accompanying drum, and

extended to the close of the singer's breath contour. The focus of the

125

K O R E A N

S I J O

performance is on this unit

as

". .

.

a single continuum of sound-

as a chain or a stream" rather than as a "succession of separate

sounds"

:

. . .

the singer uses two or three syllables to set up a pitch upon which he elaborates

with the singularly Korean vocal techniques. This is by no means a recitation

tone, since it is sung on a single syllable. The listener's attention and the singer's

concentration are centered upon the treatment of that single tone, the technical

term being 'YE-am' (sic)

(&%)

or 'Aftertone.' Especial interest centers upon the

wide vibrato and the microtonal inflections obtained with the 'last gasp' of the

singer's breath for each sung component.'

Just as traditional description of the literary form of the sijo

employs a somewhat misleading syllable count, the terminology of a

beat or metrical measure has been transferred to analysis of the

dynamic units of the musical performance, commonly in the form

below

No.ofbeats:

I

5

5

5

I1

5

8

8

I11 IV

8 5+3

8 5+3

8

(omitted)

But in actual performance it is the singer's sense of the breath

<<

contour," a combination of physical capacity and aesthetic taste,

rather than a set of equidistant, regular beats, that determines the

lengths of the dynamic units. The accompanying drummer depends

entirely upon his rapport with the singer for the placement of his

"beats" and the success of the performance. The drum beats merely

divide a given unit into segments; they do not determine the unit's

duration,

The dynamic structure of the sijo performance is fully analyzed in

the traditional transcription system, an example of which is shown

below. Breath pauses are marked in this transcription with the

inverted caret. The temporal segments of the performance-

<<

beats9'-are marked off as a sequence of rectangular sections, akin

Charles Seeger, "Prescriptive and Descriptive Music Wtiting," David McAllester,

ed., Readings in Ethnomusicology (New York: Johnson Reprint Corporation, 1g71),p.

24.

'

Robert Provine, "The Si-jo as a Musical Form," unpublished study paper, Harvard

University, 1971,p. 13.

Cf. Richard Rutt, "An Introduction to the Sijo," p. 88; Yi Hye-gu

SB*,

Han'guk

Emak yTn'gu

@Zfi@s@f%

(Seoul: Kungmin nmak Yon'gu Hoe, 1957), p. 153.

126

DAVID R. McCANN

to, and sharing the term for, the sections of a bamboo stalk: chdl

@.

In thirteen sections

(5+8)

located at the top of the transcription

(labelled "Drum" to the left of the transcription) of Hwang Chin-i's

Ch'dngsan.ri, the drum accompaniment is notated with the following

set of symbols for performance on the double-ended Korean changgo

%&k

:

0

: a beat on the low end of the drum, called kung.

I

: a beat on the high end of the drum, called ttak.

:

the two ends struck together.

'..,

":

i

:

:

stacatto, c%lled ttadak.

a muted or silenced beat, called tta.ak.

Beneath the transcription of the drum accompaniment, which

repeats in units of five or eight "beats" during the performance,

the three lines of the text are transcribed in a corresponding series

of temporal sections. The melody is represented graphically as a

line, the relative height of which indicates pitch, and the wavy por-

tions, vibrat .

The dynamic elements of this transcription are abstracted below.

The inverted caret is retained to mark the position of breath pauses,

but

I

have squared off the melodic lines, eliminating the vibrato,

in order to show clearly the location of pitch change with respect

to the dynamic units.

-

mybng wbl

i

man kongsan

,

kamil

The transcription shows an analysis of the text into what have been

termed dynamic units. The two important characteristics of these

units are, first, that they include at least one shift of pitch, up, down,

or both; second, they are bounded by breath pause.

In the sijo repertoire today, the same musical setting is used for

any text, including the much longer sasdl sijo, so we cannot say that

9

Transcription from Ch6ng Ky6ng-t'ae

@?flijl%,

ed., Sijo

po

w&g

(Ch6nju: Ch6nju

Kugak HagwGn, 1g55),p. 7.

128

DAVID R. McCANN

the connection between a particular text and the dynamics of its

musical phrasing represents an analysis of distinctive aspects of that

particular text. Accordingly, this analysis of the musical setting

indicates that in principle the basic constituent elements of the

rhythmic structure of the sijo are dynamic units of one or more syl-

lables, segmented by breath pause, with a pitch change analogous

to that which marks accent in the Korean language.1° Groups of

syllables, in other words, rather than syllables as such, comprise the

building blocks in the rhythmic structure of the sijo.

4.

The Prosodic Structure of the Sijo

For purposes of prosodic analysis of the sijo, the proper term for

the basic rhythmic entity equivalent to the musical dynamic unit

is the "rhythmic group," which the article on "Prosody" in the

Encyclopedia of Poety and Poetics (hereafter

EPP)

defines as

". . .

a

sound cluster spoken with a single phonologically conventionalized

pattern or 'contour' of accentuation and intonation, the nuclear

center or peak of which in the free or full form

is

a phonologically

maximal or 'primary' accent" (EPP,p. 674).

As suggested in the discussion of the syntactic structure of the

sijo, there tends to be a general correspondence between syntactic

units and rhythmic grouping, but not an absolute one. The rhythmic

groups are not simply the phonetic "realization" of syntactic

units, the "phonetic [or phonemic] word," but are "essentially

phonetic entities whose correspondence with syntactic units is not

fixed"

(EPP,

p. 676). The boundaries of the rhythmic group,

unlike the divisions between syntactic elements, occur as relative

diminutions, or even silences, in the continuum of the sound flow.

( 4

In the group unit it is the nuclear center that is most distinctive;

the limits of the group may be marked by any degree of openness in

juncture, and they are often, as in the syllable, obscure" (EPP, p.

674).

In linguistic studies of prosodic features the practice is generally

to present a phonetic transcription of the text, accompanied by

symbols representing the prosodic elements and values. For the

presentation of prosodic features in a literary text, however, the

lo

Cf. Samuel Robert Ramsey, "Accent and Morphology in Korean Dialects: A De-

scriptive and Historical Study" (unpubl. Ph.D. diss. Yale Univ. 1g75),p p 4-34; 280-81.

KOREAN SIJO

129

qualitative contrasts represented by phonetic transcription are

superfluous, since qualitative contrasts are not rhythmically relevant.

For literary purposes, again, it is relative prominence, rather than

the type of quantitative relation which may signal it-pitch, inten-

sity, or duration-that is relevant.

We shall use the system of prosodic analysis and notation outlined

in the

EPP

to make prosodic abstracts of the two

sijo

we have

examined. In this notational system, a single o represents the syllable

or "minimal aggregatory unit." Four levels of prominence are

notated: o for weak or unemphatic stress;

6

for primary;

6

for

secondary; 6 for tertiary. Where the stress may be either primary or

secondary, the symbol

6

is used. Additional marks are the straight

comma for "open juncture or break between stress groups," the

caret

,,

for "longer pause at more open juncture," a caret with a dot

beneath it

,

to indicate protraction of pause, and the dot

.

for the

location of potential juncture (EPP, p.

668).

Before proceeding to the analysis of the texts, let us first illustrate

the use of this system of notation by contrasting it with a phonetic

transcription. A recent article on "Suprasegmental Morphemes in

Korean"ll contained the following example:

2 +3>

+4>3 +4> +3 +2

+2

1

/ulika hayaya hallilil ma ni chaca poaya haciaqkhessirnnikkaI

/

("Shall we not have to seek out the many things we must do?")

The symbols used in this transcription are

+

for juncture,

>

for

stress,

I,

2,

3, 4

for the four pitch levels, and

for sentence-final

intonation contour. The information presented in this phonetic

transcription would be abstracted in EPP notation as:

The

EPP

system notes the occurrence of syllables, the location of

prominence, and the boundaries of the group of syllables that are

aggregated around a prominent, nuclear syllable. Omitted as ir-

relevant to rhythm are the phonetic quality of the syllable, and the

identity of the quantitative feature-stress or pitch-associated with

syllabic prominence. Using this system to make an abstract of the

l1

Pak Ch'ang-hae

BIZJgfl,

"KugG Gi Gnjch'im Gmun e kwanhan yGn'gu"

q q

-1

?&!3

43$011

&a*?,

Tongbang ha&

Rk@&

6 (1963), 29.

130

DAVID R. McCANN

text by Hwang Chin-i:12

6 o blo 6 o 6

o b old o b o

a

(ch'6ngsan.ri pybkkyesu.ya

su.i kamnl charang mara)

o 6/6 o b oAo 6 o b16 o b oa

(ilto ch'anghae hamybn

tora.ogi brybwbra)

0 6 0 1 6 0 0 b o A 6 0 b o 1 6 0 0 , (my6ngwbl.iman-kongsan

hani suyb kandtil bttbri)

With the exception of the first hemistich of line two, as we have

noted, juncture and pause coincide with the syntactic divisions.

Although these points of division may for the most part be inferred

from the syntactic analysis, such analysis does not reveal the shifting

rhythmic pattern with its changes from groups with accents on the

final syllables (606 and 060b), to groups with unaccented final

syllables (6060). These modulations, in conjunction with variation

in the span of the groups-their lengths in terms of syllable count-

contribute to the slightly syncopated rhythmic effect of this piece.

The text by the "palace woman" has similar rhythmic modula-

tions within the basic four-part line pattern

:I3

The four rhythmic groups of line one in this case are marked as

being divided by pause

A ,

rather than juncture

.

The line is ac-

tually composed of six potentially independent rhythmic groups,

and in another phonological environment the first and third groups

might have functioned rhythmically as hemistichs themselves-

that is, as two independent groups each-but in this text they are

bound under intonation curves into the "composite groups" (EPP,

p.

675)

ap m0s.e tan kogida1.a 6 o o

b.

6

o o b and nui ndrnl moradaga

b 6 0 . 6 o o b .

The evidence of syntactic analysis, of investigation of the dynamic

structure of the musical performance of the sijo, and prosodic analysis

of the sijo as a literary text, all indicate the importance of the four-

l2

Rhythmic abstracts prepared in consultation with Professor Sin Tong-uk

@!xfi,

Department of Korean Literature, Korea University.

l3

Ibid.

131

KOREAN

SIJO

part division of the line as the primary structural principle of the

sijo.

What of syllables, though? The identification of the four-part

line structure of the

sijo

as its fundamental structural feature per-

mits us to ask, not whether there is variation in syllable count, for

all accounts do agree that there is, but whether this variation is in

any way patterned within the sequence of four line-constituent

rhythmic units.

5.

Syllabic Distribution in the Sijo

We have noted that most descriptions of the

sijo

give a range of

variation in the number of syllables occurring in the line-constituent

groups. For the twenty-nine

p'ydng sijo

texts by

kisaeng

the ranges

are :

1

I1

I11

IV

Line

I

2-4

4-5

2-4

4-5

Line

2

2-5

3-4

3-6

3-4

Line3 2-3

4-7

3-4

(0);3-5

The most frequently occurring syllable count (modal) for each

of the groups is:

3 4 4 4

3 4 4 4

3 6 4 3

The modal figures correspond fairly closely to those in the basic

pattern given in most accounts, but none of these descriptions

shows which of the groups might have more and which less varia-

tion in syllable number. To represent the pattern of variation, first

we show the range of variation and number of occurrences by posi-

tion

(I,

11, 111, IV) in the three lines.

Syllabic Distribution and Frequency in Kisaeng Sijo

I

I I

I11

IV

Line

I

No. of syll.:

2 3 4

4 5

2 3 4

4 5

Frequency:

12 16 1

20

9

2 12 15

27 2

Line2 No.ofsyl1.:

2 3 4 5

3 4

3 4 5 6

3 4

Frequency:

8 1 6 4 1

11 18 7 2 0 1 I

1 2 8

Line 3 No. of syll.

:

2

3

4

5

6 7

3 4

0 3 4 5

Frequency

:

1 2 8

2 1 0 1 1 6

1 2 8

4 2 0 4 1

132

DAVID R. McCANN

From the table it appears that the relative amount of variation

alternates through the lines, with increased variation in the odd-

numbered groups of lines one and two, and the reverse in line three.

Counting the number of variants for each group, we have the fol-

lowing pattern:

Total Number of Variants

I

I1 I11 IV

3

2

3 2

4 2

4 2

2

4

2 4

This measure is somewhat crude, however. The value of

"2"

for

the number of variants in the second group of line one, for example,

represents twenty occurrences of a four-syllable group in that posi-

tion, and nine of a five-syllable group, whereas the

"2"

in the fourth

group of line two stands for twenty-eight four-syllable groups and

only one other variant. This incompatibility can be corrected by

weighting, by excluding the modal figure for each group position-

"16" in line one, group

I,

for example-and recording the remaining

number of variants-thirteen,

for the same example. In tabular

form

:

Total Number of Variants from Modal

I I1

I11 IV

Line1 13

9

14

2

Line2 13 11

9 1

Line3

1 18

1

9

The most striking features of this table are the low numbers for

the final groups of lines one and two, and groups

I

and

I11

of line

three (2,

I, I, I,

respectively), and the high number of variants in

group

11,

line three (I 8). The low number of variants-or, conver-

sely, the greater regularity-at the ends of lines one and two con-

forms to the general principle that, in a verse line, the "terminal

sections are their most determinate parts; freedom of variation is

more often permitted in the earlier part than at the end" (EPP,

p. 677). Metaphorically, the greater regularity at the line ends cor-

133

KOREAN SIJO

responds to the tightening of the reins as the farmerlpoet prepares

the oxlreader for the re-verse at the end of the furrow.14

The syllabic regularity in groups

I

and I11 of line three is directly

related to the great degree of variation found in group

I1

of that

line. Regularity in group I prepares for the variation to follow; it

establishes the principle of regularity so the rhythm does not be-

come lost in group

11.

Regularity in group I11 once again establishes

the principle. In more general terms, "it almost goes without saying

that expressive metrical variations are possible only in verse con-

ceived in a tradition of more or less regular base rhythm: we can

have no variations in entirely 'free' or 'cadenced' verse."15

The fairly large amount of variation (9) in the final group of line

three where we should, on the contrary, expect greater regularity,

suggests that in some way the third group of line three is to be con-

sidered "final" for the literary form as well as in the musical per-

formance. The fourth group of line three is, in fact, rhetorically

conventionalized: fifteen of the twenty-five texts that have a fourth

group in line three end with some form of the verb hada, "so be it"

or "that's how it is." Of the other ten, all are verbs either in the

imperative or the interrogative. This type of conventionalized

ending makes the actual meaning of the final group of little impor-

tance in the development of the sijo's point. Such rhetorical detach-

ment of the final group accounts for its syllabic irregularity.

6.

Conclusions

The standard

p'ydng

sijo may be summarized as a verse form of

three lines, having general correspondence between line boundaries

and syntactic units, and consequent tendency to end-stopped lines.

Each line is composed of four rhythmic groups which correspond,

in the main, to the four grammatical immediate constituents of the

line as a syntactic unit. Exceptions to this correspondence occur in

association with certain types of diction, notably the use of a four-

syllable phrase in Chinese. The syllable count of the line-constituent

groups is more regular at the ends of lines one and two, and both

l4

Cf. John Thompson, The Founding of English Metre (London: Routledge and Kegan

Paul, 1966),

p.

4.

l6

Paul Fussell,

Jr.,

Poetic Meter and Poetic Form (New York: Random House, 1965), p. 39.

134

DAVID

R.

McCANN

before and after the syllabically irregular second group of line three.

That the degree of variation in syllable count depends upon group

position (I, 11,

111,

IV) indicates the primary importance of the

rhythmic group as the fundamental constituent unit in the rhythmic

structure of the

sijo.

Syllable count may be optionally more or less

varied as a secondary rhythmic element, but as a verse system,

Korean is accentual, not syllabic.