The tone system of Ikema Ryukyuan Yuka Hayashi (Kyoto University / JSPS) Thomas Pellard (EHESS-CRLAO) Yukinori Takubo (Kyoto University) Tomoyuki Kubo (Kyushu University) Yosuke Igarashi (Hiroshima University) International Workshop on The History & Reconstruction of Japanese Accent 3-4 September 2009

1

Introduction • Previous studies on the tone system of Ikema, a dialect of Miyako Ryukyuan, have been based on incomplete and problematic data • We will propose a better analysis of Ikema tone based on newly collected data, and show its consequences for the reconstruction of the proto-Japonic accent/tone system

2

3

4

Amami

Okinawa

Miyako Yaeyama

5

Ikema Ryukyuan • Spoken in Ikema island, Sarahama in Irabu island and Nishihara/Nishibe in Miyako island

• Our data is mainly from Nishihara

• Ikema

• Sarahama

• Nishihara

6

Genetic relationship Southern Ryukyuan Yaeyama

Miyako Common Miyako Irabu-Ikema Ikema

Tarama

Ikema Sarahama

Central Miyako

Nishihara 7

Data collection • Our description is firmly based on instrumental analysis • Our data comes not only from the elicitation but also from the natural discourse of several speakers

8

Claims • Hirayama et al. (1967, 1983), the most influential studies on the tone system of Ikema, should be corrected from both synchronic and historical points of view Our analysis on Ikema tone (contra Hirayama) • The tone system of Ikema can only be described by two factors: • a two-way lexical tonal distinction (α vs. β) • a basic H(igh)-L(ow) alternating melody • The distribution of words between tone classes in Ikema supports Hattori’s theory about proto-Ryukyuan 9

Hirayama et al. (1967, 1983) • Ikema has a two-pattern accent • The distinction is that of presence vs. absence of an “accent” realized as a pitch fall • Accented words have the pitch fall between the second and the third mora

10

Hirayama et al. (1983) Mora Phonological analysis

Phonetic realization

2

/oo˥/

○●, ●○

/oo/

○○

/oo˥o/

○●○, ●●○

/ooo/

○○○

/oo˥oo/

○●○○, (●●○○)*

/oooo/

○○○○

3 4

o = mora

* Only for nouns

○=L, ●=H

11

Problems • Hirayama’s analysis cannot fully explain the tone system of Ikema • There are discrepancies between their and our data regarding the phonetic pitch shapes • Their analysis is only valid for two- to three-mora words with one bimoraic particle • Their analysis cannot account for words with more than three morae

12

Two- to three mora words ● α: mayu ‘cat’ x β: nabi ‘pan’

170

140 F0 (Hz)

Hiyayama Phonological Phonetic analysis realization

110

80

ma na

α

/oo˥/

○●, ●○

β

/oo/

○○

yu bi

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Normalized time (sec)

13

Two- to three mora words Citation forms + particles

● α: mayu=mai ‘cat=also’ x β: nabi=mai ‘pan=also’

170

140 F0 (Hz)

Hiyayama

110

80

ma na

yu bi

m m

a a

Phonological analysis

Phonetic realization

α

/oo˥/

○●, ●○

β

/oo/

○○

i i

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Normalized time (sec)

14

Two- to three mora words ● α: yabizi ‘Yabiji coral reef’ x β: yumunu ‘mouse’

170

Hiyayama

F0(Hz)

140

110 ya yu

bi mu

Phonological analysis

Phonetic realization

α

/oo˥o/

○●○, ●●○

β

/ooo/

○○○

zi nu

80 0.00.10.20.30.40.50.60.70.80.91.0 Normalized time (sec)

15

Two- to three mora words Citation forms + particles 170

● α : yabizi=mai ‘Yabiji, too’ x β: yumunu=mai ‘mouse, too’ Hirayama

F0(Hz)

140

110 ya yu

bi mu

zi nu

m m

a a

Phonological analysis

Phonetic realization

α

/oo˥o/

○●○, ●●○

β

/ooo/

○○○

i i

80 0.00.10.20.30.40.50.60.70.80.91.0 Normalized time (sec)

16

Four-mora words ● α: bakamunu ‘youngster’ x β: sarahama ‘Sarahama (place name)’

170

Hirayama

F0(Hz)

140

110 ba sa

ka ra

mu ha

Phonological analysis

Phonetic realization

α

/oo˥oo/

○●○○, ●●○○

β

/oooo/

○○○○

nu ma

80 0.00.10.20.30.40.50.60.70.80.91.0 Normalized time (sec)

17

Hirayama’s distinction in four-mora words class membership Squares of solid line: Hirayama’s distinction Squares of dotted line: our distinction

kannai ‘thunder’, sakazïki ‘sake cup’, bikidun ‘guy’, kanamai ‘head’

α

kyaudai ‘brother/sister’, katauta ‘one’s side’ sinsii ‘teacher’, nukuzïï ‘saw’ sinbun ‘news paper’, gakkoo ‘school’

β 18

Our analysis of the lexical tone of Ikema • Two-tone system (word tone) • Tonal distinction appears at the rightmost position of a word • Citation forms do not always exhibit the tonal distinction • The tone classes can be identified easily when followed by another element (e.g. particle)

19

Tonal distinction • Type α words have an underlying low tone at the right edge of the lexical word • Type β words have an underlying high tone at the right edge of the lexical word -> The edge tone will be linked to the rightadjacent element if there is any

20

HL alternation • Trochaic rhythm pattern realized as iteration of HL tone • The TBU of each tone is a foot consisting of two to three morae: tonal foot cf) The HL alternation in Irabu No lexical tone in Irabu but the HL alternation can be observed ex: (kjav)(dai)=(kara)=(mai) ‘from siblings, too’ (μμ)(μμ)=(μμ)=(μμ) H L H L (Shimoji and Hayashi 2008: 1(1a)) 21

Lexical tones with HL alternation • The lexical tonal distinction is not the only factor determining the pitch realization of words in Ikema • A HL alternating melody interacts with the tonal distinctioin • The tone system can be properly described only if we take these two factors into consideration 22

Four-mora words with a particle ● α: bakamunu=mai ‘youngster=too’ x β: sarahama=mai ‘Sarahama=too’

170

F0(Hz)

140

110 ba sa

ka ra

mu ha

nu ma

m m

a a

i i

80 0.00.10.20.30.40.50.60.70.80.91.0 Normalized time (sec)

23 23

Verbs Two-mora words nii vs. mii

α: nii ‘cook’

200

β: mii ‘look’

F0(Hz)

160

120

80 0.0

SENTENCE

ni i mi i 0.2

0.4 0.6 0.8 Time (msec)

1.0

A. nii B. mii

24

Verbs Two-mora words + particles nii hazï vs. mii hazï nii hazI vs. mii hazI

α: nii=hazї ‘cook=must’

200

β: mii=hazї ‘look’ ‘look=must’

F0(Hz)

160

120

80 0.0

ni i mi i 0.2

ha ha

SENTENCE

zї zї

0.4 0.6 0.8 Time (msec)

1.0

A. nii B. mii 25

Verbs Three-mora words budui vs. tabai 200

α: budui ‘dance’ β: tabai ‘cling’

F0(Hz)

160

120 SENTENCE

bu du i ta ba i 80 0.0

0.2

0.4 0.6 0.8 Time (msec)

1.0

A. budui B. tabai 26

Verbs Three-mora words budui vs. tabai hazI budui hazï vs.hazI tabai hazï

α: budui=hazї ‘dance=must’ β: tabai=hazї ‘cling=must’

200

F0(Hz)

160

120

bu du i ta ba i

80 0.0

0.2

SENTENCE

ha zї ha zї 0.4 0.6 0.8 Time (msec)

1.0

A. budui B. tabai 27

Verbs Four-mora words cImmui vs. kammui cïmmui vs. kammui

α: cïmmui ‘pinch’

200

β: kammui ‘bite’

F0(Hz)

160

120 SENTENCE

cï m mu i 80 ka m mu i 0.0

0.2

0.4 0.6 0.8 Time (msec)

1.0

A. cImmui B. kammui 28

Verbs Four-mora words + particles cïmmui hazï vs.hazI kammuivs. hazï kammui hazI cImmui

α: cïmmui=hazї ‘pinch=must’ β: kammui=hazї ‘bite=must’

200

F0(Hz)

160

120

cï m mu i ka m mu i

80 0.0

0.2

SENTENCE

ha zї ha zї

0.4 0.6 0.8 Time (msec)

1.0

A. cImmui B. kammui 29

Lexical tones with HL alternation • Some derivational and inflectional suffixes of verbs and adjectives are part of the leftadjacent word. PW (prosodic word): word that has the lexical tone on its right edge [{Root}+{derivational suffixes}+{inflectional suffixes}]

30

Verbs with multi-foot PW ‘cook-cause-passive-past=must’

200 F0 (Hz)

α: nii-ssa-hai-tai=hazï

L nii

ssa

hai

tai

hazI

75 0.5 Time (sec)

1

β: mii-ssa-hai-tai=hazï F0 (Hz)

‘see-cause-passive-past=must’

200

H mii

ssa

hai

tai

hazI

75 0.5 Time (sec)

1

31

Adjectives • Adjective as one PW • Adjective with a particle The lexical tonal distinction clearly exists in adjectives contra Hirayama et al. (1983), who say that the two tonal distinctions are on the way to decay

Note: Stems of Ikema adjectives usually cannot stand alone and must appear with suffixes

32

Adjectives Two-mora stem + suffix garukai vs. imikai

α: garu-kai ‘light-SUF’ β: imi-kai ‘small-SUF’

200

F0(Hz)

160

120 SENTENCE

ga ru 80 im i 0.0

ka ka

0.2

i i 0.4 0.6 0.8 Time (msec)

1.0

A. garu B. imi 33

Adjectives Two-mora stem + suffix garukai hazï vs. hazI imikai hazïvs. imikai hazI garukai

α: garu-kai=hazї ‘light-SUF=must’ β: imi-kai=hazї ‘small-SUF=must’

200

F0(Hz)

160

120 SENTENCE

ga ru 80 im i 0.0

ka ka

0.2

i ha zï i ha zï 0.4 0.6 0.8 Time (msec)

1.0

A. garu B. imi 34

Hirayama et al. (1983) Phonological analysis and phonetic realization of words with four morae for adjectives

Mora Phonological analysis

Phonetic realization

4

/oo˥oo/

○●○○

/oooo/

○●○●, ●●○●

-> Again, like the case of four-mora nouns, Hirayama can not specify how the lexical distinction appears 35

Nouns followed by two particles HL alternation in the sequence of particles ● α: mayu=kara=mai ‘cat=from=also’ HLH

x β: nabi=kara=mai ‘pan=from=also’ HHL

ma yu na bi

ka ka

ra ra

ma ma

i i 36

● α: mayu=kara=mai ‘cat=from=also’ x β: nabi=kara=mai ‘pan=from=also’

HLH HHL

Lexically specified tone

H

L

H

H

H

L

ma yu na bi

ka ka

ra ra

ma ma

Alternating tone (different tones from the leftadjacent foot)

i i

The place of the Ikema tone system in the history of Japonic • Hirayama’s theory: Ryukyuan tone/accent system is derived from a Kyushu-like dialect

• Hattori’s theory: Ryukyuan tonal classes differ from the Japanese ones (Hattori 1932, Hattori 1958) cf) Matsumori (1998, 2000b)

-> Many scholars have dismissed Hattori’s argument, but Ikema data supports it 38

Terminology ProtoJaponic ProtoJapanese 1 2 3 4 5

ProtoRyukyuan A B C Ikema α

β 39

Hirayama’s historical analysis (for disyllabic words) Ryukyuan varieties of three tonal distinctions:

Ruiju Myōgishō:

Ikema:

A

B

C

1 2 3 4 5 α

β

40

Hattori’s historical analysis (for disyllabic words)

Ruiju Myōgishō:

1 2 3 4 5

Proto-Ryukyuan Ryukyuan varieties of three tonal distinctions:

Ikema:

A

B

α

C

β

41

The wordlists for our historical research • Uwano wordlist (based on Japanese dialects) cf) Uwano (1985)

• Matsumori wordlist (based on Northern Ryukyuan dialects) cf) Matsumori (2000a)

42

Results -Uwano list • Monosyllabic words

1.1 1.2 1.3

α 3 4 9

β 0 0 2

α: 16 (89%) β: 2 (11%) 木 ‘tree’, 田 ‘rice field’ 43

Results -Uwano list • Disyllabic words

α

β

2.1

31 (82%)

7 (18%)

2.2

9 (82%)

2 (18%)

2.3

28 (85%)

5 (15%)

2.4

12 (60%)

8 (40%)

2.5

5 (45%)

6 (55%) 44

Results -Uwano list • Trisyllabic words α

β

3.1

20 (95%)

1 (5%)

3.4

7 (41%)

10 (59%)

3.5

2 (33%)

4 (67%)

3.7

1(50%)

1(50%) 45

Results -Matsumori list

A B C

α β 18 (78%) 5 (22 %) 17 (70 %) 7 (30 %) 7 (14%) 43 (86%)

• Clear trend for AB-α C-β relationship 46

Exceptional cases for both Hirayama and Hattori’s theory (disyllabic words) • β in words of class 1 and class 2 2.1 -β: 姉 ‘sister’, 箱 ‘box’, 底 ‘bottom’, 横 ‘side’, 効 ‘reward’, 釜 ‘pot’, 道 ‘road’ 2.2 -β: 橋 ‘bridge’, 為 ‘sake’ 47

Discussion • Basically the data from Ikema supports the hypothesis of Hattori • Exceptional words might be loan words

48

Reconstruction Reconstruction of proto-Miyako “accent”: Shimabukuro (2007) and Bentley (2008) • All the data comes from Hirayama et al. (1967, 1983) • They do not incorporate Matsumori (2001)’s important finding: there are more tonal distinctions in the Tarama dialect than previously thought 49

Reconstruction • Shimabukuro (2007) and Bentley (2008) reconstruct a fixed-position “accent” (=pitch fall) and distinctive initial tone *HH-L > Ik. α (L) *LH-L > Ik. α (L) *LLL-L > Ik. β (H)

*HHH-L *LHH-L *LLL-L

> Ik. α (L) > Ik. α (L) > Ik. β (H)

50

Reconstruction • Shimabukuro (2007) and Bentley (2008) ignore the HL alternation • They miss some distinctions like “sea” vs. “dog” →Their reconstruction cannot be accepted

51

Conclusions • Hirayama’s analysis is unable to account for the pitch shape of words and contains factual errors • The data of Ikema Ryukyuan basically confirm Hattori’s and Matsumori’s hypotheses on protoRyukyuan • A historical reconstruction cannot be done on the basis of wrong and misanalyzed data • We need first to do more (field)work in order to collect accurate data and produce reliable analyses on the synchronic system 52

References BENTLEY, John R. 2008. A Linguistic History of the Forgotten Islands: A Reconstruction of the Proto-Language of the Southern Ryukyus. Folkestone: Global Oriental. HATTORI, Shirō. 1932. Ryūkyū to kokugo no on’in hōsoku. Hōgen, 2–7, 8, 9, 10, 12. HATTORI, Shirō. 1958. Amami guntō no shohōgen ni tsuite – Okinawa/Sakishima shohōgen to no hikaku. Jinrui kagaku, IX. HIRAYAMA, Teruo, ŌSHIMA, Ichirō, & NAKAMOTO, Masachie. 1967. Ryūkyū Sakishima Hōgen no Sōgōteki Kenkyū. Tokyo: Meiji shoin. HIRAYAMA, Teruo (ed.) 1983. Ryūkyū Miyako Shotō Hōgen Kiso Goi no Sōgōteki Kenkyū. Tokyo: Ōfūsha. MATSUMORI, Akiko. 1998. Ryūkyū akusento no rekishiteki keisei katei – ruibetsu goi 2-hakugo no tokui na gōryū no shikata o tegakari ni. Gengo kenkyū, 114 : 85–114. MATSUMORI, Akiko. 2000a. Ryūkyū akusento chōsa no tame no ruibetsu goi no kaihatsu – Okinoerabu-jima no chōsa kara. Onsei kenkyū, 4(1) : 61–71. MATSUMORI, Akiko. 2000b. Ryūkyū no takei akusento taikei ni tsuite no ichi kōsatsu – ryūkyū sogo ni okeru ruibetsu goi 3-paku go no gōryū no shikata. Kokugogaku, 51(1) , 93–108. MATSUMORI, Akiko. 2001. Historical tonology of Japanese dialects. In KAJI, Shigeki (ed.) Cross-linguistic Studies of Tonal Phenomena, Tokyo: ILCAA, 93–122. SHIMOJI, Michinori, & HAYASHI, Yuka. 2008. Tonal Alternation and Rhythmic Structure in Irabu Ryukyuan. Paper read at Phonology Festa, 3rd Joint Meeting of PAIK and TCP (Atami, Japan). SHIMABUKURO, Moriyo. 2007. A Reconstruction of the Accentual History of the Japanese and Ryukyuan Languages. Folkstone: Global Oriental. UWANO, Zendō. 1985. Murakami hogen no akusento shiryo – one- to three-mora words. Tokyo daigaku gengo ronshu ’85, 25-60. 53