The differences of connotations between two flowers,plum and cherry, in classical Japanese poetry, 10th century.Hilofumi Yamamoto Tokyo Institute of Technology

Introduction• This project addresses an analysis of connotations of flowers in classical poetry: i. e., ‘ume’ (plum) and ‘sakura’ (cherry) .

• We will identify the characteristics of two flowers by computer modeling.

• Using parallel texts of original texts and contemporary translations of classical Japanese poetry, the Kokinshu, we will clarify

the details of connotations in an objective procedural manner that is not influenced by human observations.

• The aim is to examine whether or not the residual of CT −OP gives information on the non-literal elements of OP .

Problem1. What is the difference between ume (plum) and sakura (cherry)?

2. What kind of connotations does each flower contain?

3. Which picture is that of cherry flowers?

a. b. c. d. e.

MethodsMaterial: Kokinshu a.k.a. Kokinwakashu is:

the first anthology compiled by the order of Emperor Daigo

(ca. 905), which contains about 1,111 poems. And 10 sets of

their Contemporary Japanese Translations (CT)

Kaneko(1927)◦

1930

Kaneko(1933)◦⊳

Kubota(1935)◦

1940 1950

Saeki (1958)•

1960

Kubota(1960)◦⊳

Matsuda

(1968)◦⊳

1970

Ozawa(1971)◦⊳

Takeoka

(1976)◦⊳

Okumura(1978)◦⊳

Kyusojin

(1979)◦⊳

1980

Kom

achiya

(1982)◦⊳

Kojima&

Arai (1989)◦⊳

1990

Katagiri (1998)◦⊳

2000

Yusa(2000)◦

Fig. 1: Dates of publication of annotations of the Kokinshu: ◦indicates that it has CT; • indicates that it does notinclude CT; ▷ indicates that it is used in this project.

poet write OP read expert reader

write

CT

read

novice reader

compare

10th century

Field of experience

20th century

Field of experience (expert)

20th centuryField of experience

(novice)

R = CT - OPwhere, CT , Contemporary translation Texts; OP , Origi-nal classical Poetry texts; R indicates a residual obtainedby subtracting elements of OP from those of CT .

Fig. 2: Schema of relationship between theoriginal poem texts (OP ) and thecontemporary translation texts (CT )based on Schramm (1954).

Result

plum (20/20,3.91): OP cw.>5.6;non-dist=off; idf=on(2)

that

see.1

2

plum3

old age

woven hat

1

wear in (my) hair1

hide.vi.1

1

sew.111

1

1

1

1

2

dark

1

flower

2

1

lodge.n

fragrance.12

3 2

10

sorrow.1

2

who.13

2

colour

4

3

snow

4

break off5

spring

4

be.3

4

scatter.1

2

see.3

3

smell.vt.2

2

2

2

4

sleeve.1

24

hisakatano.pw

cloud.vi

1

1

in whole area

12

2

spring season

1

evidence

1

715

4

1Kurabu.PN

1

2

2

warbler

4

2

go over 1

fall

3

3

separate.3

2

2

2

1

1

sing.vi

3

3

do.1

3

bloom

2Fig. 3:Plumby OP

plum (20/198,3.71): CT cw.>15.5;non-dist=off; idf=on(2)

which

plum16

woven hat10

distinction

17garden

17

dark 13

fragrance.1

65

fragrance.240

flower157

look away

8

smell.vt.1

14

Kurabu.PN8

old age

6

warbler

10

degree

stop by5

cling

6

8

snow

17

break off10

33

sleeve.142

31

lingering fragrance

18

5 9

blame

7

7

daybreak

7

get dark

7

4

7

around5

moonless night

5

mountain

8

go over

8

clearly

7

wear in (my) hair

sew.24

6

hide.vi.2

6

7

4

6

a bit5

7

7

break off.1

13

discern

11

Fig. 4: Plumby CT

plum (20/198,3.71): OP-5.6-1; CT-15.5-2; (op(IS)ct)

sleeve.1 plum42

dark

13

fragrance.1

65 flower157

smell.vt.1

14

warbler

31break off

33

Fig. 5: Plumby intersectionof OP and CT

plum (20/198,3.71): OP-5.6-1; CT-15.5-2; (op(is)CT)

which

plum16distinction 17

woven hat

10

garden17

fragrance.240

look away

8

Kurabu.PN

8

degreestop by5

cling6

8

snow17

break off

10

warbler

10lingering fragrance

18

59blame

7

7

daybreak7

get dark

7

around

5

moonless night

5

mountain

8

go over

8

dark

7

clearly7

hide.vi.2

7

4

6

a bit57

7

break off.1

13

discern

11

sew.2

6

smell.vt.1

4

Fig. 6: Plumby subtracting OP from CT

plum-CT-20-198-3.71-15.5 cherry-CT-43-298-3.30-15

whichplum 16

woven hat10

Kurabu.PN

8

distinction

17

garden

17

dark

13 fragrance.165

fragrance.2

40flower

157

look away

8

smell.vt.1

14

way home

lodging4

scatter.2

5

village.1

6

old age6

warbler

10

this year

dyed inky black

6

edge

Yoshino.PN

5

mountain

19

cherry

13

go over7

only13

hearsay.2

4

number7

9

7

scatter.19

compare

9

without cease

6

around

5

moonless night

5

love.2

98

clearly

7

degree

stop by5

cling6

8

snow

17

break off 10

33

sleeve.1

42

31

lingering fragrance

18

59

blame

7

7

spring hazemountain cherry7

hide.vt

19

trail

19

mist7

6

how much

7

19

field

Fukakusa

7

10

6

daybreak

7 get dark

7

4

22

6

25

27

77

200

142

break off.1

15

13

cherry blossom

7

31

48

bloom25

60

wear in (my) hair

sew.2 4

6

hide.vi.2

6

7

4

6

a bit57

7

5

17

road

12

increase

7

8

7

76

6

discern

11

keep listening.2

5

faint

4

Fig. 7: Combined Network Modelof Plum and CherryConclusion

• It will be necessary to examine not only common nouns but also the distinctive characters ofproper nouns in order to further examine the connotative associations of poetic vocabulary.

• We observed proper nouns such as place names, Kurabu, Tatsuta, Otowa, Yoshinoin the network models of common nouns, and concluded that they seem to strongly influence the associations of poetic vocabulary.

• The relative salience clearly indicates that both ume (plum) and sakura (cherry) share Kurabu yama (Mt.Kurabu), which comprisesa cluster of nodes in the sub-network.

Reference• Schramm, W. L. 1954. How communication works. The process and effects of mass communication. 3–26. University of Illinois Press.

• Yamamoto, H. 2006. Extraction and Visualisation of the Connotation of Classical Japanese Poetic Vocabulary. Symposium forComputer and Humanities, 2006. The information processing society of Japan. Vo. 2006, No. 17, 21–8.

1