珞襁褞瑣 蒻褞璋韜 チナムマヒタメヘタ゚ チネチヒネホメナハタ ミホムムネネ

ホヘヤナミナヘヨネネ, ハヘネテネ, マホムホチネ゚, ヘタモラヘロナ ネヌトタヘネ゚

<< テヒタツヘタ゚
テミホネヘニナヘナミネ゚
ムメミホヘホフネ゚
ナヌホマタムヘホムメワ
ネホヒホテネ゚
ナフピ
ヘヤホミフタメネハタ
ムハモムムメツホツナトナヘネナ
ムメホミネ゚
モヒワメモミホヒホテネ゚
タリネヘホムメミホナヘネナ
ナトネヨネヘタ
ナメタヒヒモミテネ゚
ナユタヘネハタ
ナトタテホテネハタ
ホヒネメネハタ
ミネチホミホムメミホナヘネナ
ミホトホツホヒワムメツネナ
ムネユホヒホテネ゚
タトネホメナユヘネハタ
ナヒワムハホナ ユホヌ゚ノムメツホ
ホヨネホヒホテネ゚
メミホネメナヒワムメツホ
ナユヘネラナムハネナ ヘタモハネ
ミタヘムマホミメ
タミフタヨナツメネハタ
ネヌネハタ
ネヌネホヒホテネ゚
ネヒホヒホテネ゚
ネヒホムホヤネ゚
ネフネ゚
ハホヘホフネハタ
ヒナハメミホメナユヘネハタ
ヘナミテナメネハタ
ミネムマミモトナヘヨネ゚
ヌロハホヌヘタヘネナ
タヌヘホナ
ハホヘメタハメロ


Pages:     | 1 |   ...   | 30 | 31 || 33 |

ォツナムメヘネハ マナミフムハホテホ モヘネツナミムネメナメタ 末末末末末末末末末末末末末末末末末末末末末末末末末末末末末末末末末末 ミホムムネノムハタ゚ ネ ヌタミモチナニヘタ゚ ヤネヒホヒホテネ゚ ...サ

-- [ ム瑙頽 32 ] --

. .)? マ褞糺 珮蒟 琅 鈔瓊裨 褶 珸籵, 蒡粽 瑰 胛粽韜 襃韈粽 韈頸 () 鈿, 磊 ォ粡 瑪銛サ, 韃頏籵, 胛粽頸 萵. ゚, 胛 鈕褥 鈞粨頸 裲, 糂褶瑯 萵 .

ナ蒻粢, 粢褊 跫 珸瑣: () 鈿 粱褪 蓖韲 韈 褌褊, 瑩褞 瑙 褶, 琺 瑰瑙褊 璧裝頷 褞裼胛 頌瑙 瑪 頸褞瑣.

フ瑣褞鞨 蓁 瑰裙 頌裝籵 跖 瑙 肛 韈 矜瑙頏籵 瑰 ヌ糒粽胛 胛 鍄 (ヌハミ゚)4. タ琿韈 葢褞肭頌 瑰頡籵 裲 瑙 褶 竟瑙-頌 (肛 褞褥珸 趺胛 褥趺胛 裝裲, 瑕趺 粽碚蓖 瑰珸)5 竟瑙-蒟 (肛-頌瑙 趺胛 褥趺胛 韈碣琥褊韜)6. ツ 碼裨 跫 磊 瑙琿韈頏籵 176 瑙 胛, 褊 褌 珸 韭瑣鞣 褊瑩. タ琿韈 瑣褞鞨 糺粨 珸 韋 韋 () 鈿. ミ瑰韲 頷 裝籵褄.

1. ハ褊 粽褪瑙 鈞粨韲 琅褂 ( 韲褞瑾 蔔褞):

(1) 裨鈞 / 珸籵 鈿 / 粽 / 蔘磊 珸籵褪 (竟.-趺., 頌瑙韃 褥. 韈碣琥褊);

(2) -- 韈碣琥褊 蒟裘 / 褥 胛粽 褶 胚 鈿 頷 珸籵 / 萵 / 蔘磊 (竟.-., 頌瑙韃 褥. 韈碣琥褊);

(3) -- 瑰褊 / 珸籵 瑰褊韜 鈿 / 珸瑣 胚 / 瑕- 珞 (竟-.., 頌瑙韃 褥. 韈碣琥褊).

ツ 糂襄 韲褞瑾 鈞粨韲 褞襄蓖胛 肭璢 頽瑙韃 珸琿 褥粨褄 蒻褄 琅褂7.

2. マ裝韭瑣鞣 籵 肭珞胛 裝趺 跫蔔竟褊胛 鞴瑣 韈頸褄. ヘ琲碚裹 瑰瑙褊 裝粽 鉅 萵 裝趺 磊 鈿 粽 (裝籵 鉅 韲褞瑾 蔔褞):

(4) 鈕瑯 褊韃 / 萵趺 瑩竟 胙瑶 / 籵頸 / / 鈿 襌 珸瑣 (竟.-., 頌瑙韃 褥. 韈碣琥褊);

(5) 硴瑕 糅蒟 粨蓖 // 褥 褥 鈿 蒡矜粨 萵 / 瑕韃- 鞣 頷 蒟裘 (竟.-., 頌瑙韃 褥. 韈碣琥褊);

(6) 粨蓖 頌 / / 韈 琅頌 リ顏竟 / 瑩頌籵 褊 跖粽/ 鈿 襌 珸瑣 (竟.-., 頌瑙韃 褥. 韈碣琥褊);

(7) 褥 褊 磊 鈿 瑟 蒟琿 (竟.-., 粽碚蓖 瑰珸).

マ韲 鈿胛 籵 , 韲褞瑾 萵 糂褪齏頌 蓿肛 籵鞨 裝 鉅 . (8)-(15):

(8) 韈碣琥褊 / 琿鞣瑯 葢 褄粢 裔 / 粽 / 鈿 籵頸 糂襄 / 粽 (竟.-., 頌瑙韃 . 韈碣琥褊);

(9) 粽 褊 / -- 鈿 / 裼瑙 粽褌 / 肭 / -- 裙琥萵褪 韲 蒡胚 (竟.-., 頌瑙韃 . 韈碣琥褊);

(10) 聿 鈔褞 瑾蒻 / 裙 蓍籵 / 鈿 瑕韲 籵 / 聰褄 齏 糂-瑕 碚裹 褊裹 琿 (竟.-., 頌瑙韃 . 韈碣琥褊);

(11) 頸 瑕-鞦蕈 矜 瑕 粽 / 鈿 鉅 齏 瑙珸 瑕 粽 頸琿 (竟.-., 褞褥珸 . 裲);

(12) -- 肄顆褥韲 粨蒡 / 胛 褊 / 趺 趺糒韲 / 萵趺 鈿 / 瑕 裔頸 褞竟 瑕 / 粽 (竟.-., 頌瑙韃 .

韈碣琥褊);

(13) 褄 鈔頸 / 萵 鈔頸 鈿 / 珞褞 糂褌 蓴粢韭瑟 (竟.-趺., 頌瑙韃 . 韈碣琥褊);

(14) 琿韭 齏 / 鈿 / 瑕- 鞣瑯 粢 (竟.-趺., 頌瑙韃 . 韈碣琥褊);

(15) 糂 裹 // - 粽 萵趺 鈿 珸瑣 / 粽 襌 赳 瑕- (竟.-., 頌瑙韃 褥. 韈碣琥褊).

ツ 齏璋 (12)-(15) 褪鞣 粱 砒 瑙胛 褶襃趾褊. メ瑕, 韲褞 (12) 蓖 肭珞 裝趺韋 頌銛 珸 葢 鈿 裝籵 瑕 瑕, 胛粽韜 裲頏褪 粽 褶 褥 胛粽褊:

鈿 / 瑕 (珸瑣) 鈿 / 瑕 裔頸 褞竟.

ツ 韲褞瑾 (13)-(14) 跫 珮萵 竟粢韶 (褞裘褞, 碣瑣) 蒡 瑰裨 跫蔔竟褊胛 裝趺: 鞴瑣 裝褥糒褪 肭珞, 糀 頸 瑙 瑩瑕褞 褊 胛 聰. ツ 韲褞 (15) 褊 裙 粽瑙珞鞣瑯 裲. マ蒡硼 裼瑕褊 竟瑕顆褥韃 瑕趺 糀 粽鴦粢 瑙 褶. ヌ蒟 碣籵褪 葢琥蕘: 鈿 裝聰 .

3. マ裝韭瑣鞣 籵 鈞粨韲 鈿 粽, 砒 鞴瑣胛:

(16) 趺 / 瑟 鞣 鈞珸籵 / 粢粢 瑩 碣琿 / 瑩 趺 褌 褊 珞齏 / 鈿 褌 / 粽 (竟.-., 粽碚蓖 瑰珸).

テ瑟瑣韭 胛 鍄 蒡 瑕 粽銕跫 (蓁 肭璢 () 鈿 褌 蒡硼) 瑕 蓁 頸褞瑣, 瑕 蓁 珸胛粽 褶, .:

ツ萵 粨蒟 粽錮 瑙, ツ 裙瑾, 胛頷 瑕 琿珸, ム裝, 裼硴褌 ハ珞珸;

ネ 磊 褞蔕 褌 ヒ裙, 鈿 褌 (ヒ褞.);

メ 碚顏, 瑟 鈿瑯 裙 (マ頌褌.);

ツ 齏 竟 褊 琅 褌 糂裙萵 瑙粨 胙, 瑟 鈿琿 裙 (ト.);

ホ褊 褪 頌瑣, 鈿 (ヒ. メ.);

[ヒ韈瑙:] ム萵 籵 矜 鈞褄, 碆褞;

ツ褞褄瑰 褞裝 韲, 糅琿 (テ鞦.);

[ト テ瑙:] ト タ ト ム籵! 瑕? 聰 瑙蒡 モ礪胛... 褌? (マ.);

ネ 籵 鞴褪 粢 珸胛粽 裴鈔褥 褌 (ホ蓙.)8.

ツ 瑣褞鞨 頌裝籵 瑕 韲褞 糂褪齏 糂裙 蒻, ハ 褥 蓿肛, .:

鈿 - 韵 琿璧 齏 蒡韭 瑕胛- (竟.-趺., 頌瑙韃 褥. 韈碣琥褊);

褥 / 琅頸 / 碯蒟 赳 // 鈿 / // 糂ク 珞褞 (竟.-趺., 褞褥珸 褥. 裲);

磊 / 裨瑰 鈿 (璢褊 ホミト);

跛 / 鈿 / 粽 跛 (璢褊 ホミト);

/ 裹 瑟 粹 蓴鈿瑙 珮瑣籵褪 / 褪 / / // 鈿 褌 / 蔘 (璢褊 ホミト);

聿 褞 磊 (...) 裨瑰 鈿 瑕 / 聿 磊 褞 (璢褊 ホミト);

褄裘韈 / 瑕 萵粹 / 糯褞 / 齏 鈞糯褞 / 粽 瑕- 瑟 蓆 褄 / 鈿 / 聿- / 胛粽 萵裲 / 鈿 聿 (璢褊 ホミト).

4. ヌ瑜頸褄 瑪鍄 裼頸璋韋. ン 琲碚裹 竟褞褥 頌裝褌 韋, 糂褶瑯 頌頸褄 瑙 褶 琥瑯 蓖糅褌褊 褥 裹 砌籵 趾褊. ム 銜褊 胙瑟瑣韭, () 鈿 頌銛褪 鈕褥 瑕 瑟褄 裝趺韃, 硴琅裹 粹 糺琥褊 裼頸璋韶 瑩瑕褞. ン 瑩瑕褞 瑰 蔔褞鞣瑯 蒡頸褄 裼頸璋韶 瑩褞, 瑕頷 瑕 ;

瑟;

瑕 磊, - ., 蓁竟褊 瑪鈞, 糘瑟 齏 碣籵 ( 韲褞瑾 糂 蔔褞), 裝糒 頌裝褌 韃 齏 萵趺 糂琲籵 裹, .:

(17) 瑕- 碎 瑕 硴竟 蔘璋 / 鈿 襃 (竟.-趺., 頌瑙韃 . 韈碣琥褊);

(18) 粽 瑟 (...) 鈿 / 鈞胛 / // 蒡 (竟.-趺., 粽碚蓖 瑰珸);

(19) 蕘 / 蓴粢韭瑟 裙 赳 / 赳 襄 胚 籵 珸瑣 / 褊 赳 / 裙 蓴粢韭 褊 趺 // 粽 (...) 鈿 (...) 跖粢 襄 粨蒻 蒡褌 鈿瑯 (竟.-趺., 粽碚蓖 瑰珸);

(20) 跖 - - 鈞褄 瑟 // 褶瑣瑣 -鞦蕈 / 粡 瑜頌琿 // 襄 - 鈿 瑕 - 褪 裙 瑕 糂褌 頸 - 褊 ... 頌硴 ... 珸 頸璋 磊 裙 瑩蔘 鞣 (竟.-趺., 粽碚蓖 瑰珸);

(21) 璞竟瑯 鈿珞瑣 粽 / 韈 裙 瑕 // 鈿 // 瑕 /// 褞蒟 (竟.-趺., 頌瑙韃 . 韈碣琥褊);

(22) 粽 竟韲琿 褶 籵 珞褞 韲 瑕 磊 - 裼 鈿 粽 褊 瑕- (竟.-., 粽碚蓖 瑰珸);

(23) 鈿璞頸 蒟 糺蓖 / 跫 珞褥頸 / 齏 瑕- 粽 (...) 蓴粢韭 瑟 鈿 / 粽蔘 磊韜 趺 磊 粢ク (竟. 趺., 粽碚蓖 瑰珸);

(24) 蕘瑣 硴 / 鈞褶瑣褄 / / 瑕 磊 鈿 / 鞣蒻 裔 瑟 裔 瑟 蒡 (竟.-趺., 粽碚蓖 瑰珸);

(25) 頷 鈑 / 跫 鞴褪 褥 / 鈿 / 褌 蔘瑣 (竟.-., 粽碚蓖 瑰珸);

(26) 褂瑣 / 鈿 瑟 / 糺頸 鈞.. 鈞頸 / 萵 (竟.-., 頌瑙韃 褥. 韈碣琥褊);

(27) 胛 褊 蒟裘 萵 瑕 褌 頸璢鞣琿 / - 鈿 萵趺 / 磊 褥 胛粽 磊 珸 瑕 - 琅 瑙 -- 蒟赳 磊 胛 褊 磅 (竟.-., 頌瑙韃 . 韈碣琥褊);

(28) 粽銕跫 碣瑣 /-- 粽頷 蓿鈑 / -- 瑕-鞦蕈 瑟 聰韈籵 / 蒻韜 韭韭 / -- 鈿 / 瑙瑟 / 鍄 (竟.-趺., 頌瑙韃 褥.

韈碣琥褊);

(29) 趺 瑩竟 リ顏竟 褌- 珸 糂齏 ォモ 粽 褥サ 瑟 / 鈿 瑟 韲頸鞣 瑙 褊 胛粢 瑩竟 韈 蒟籵 / 粽 (竟.-趺., 頌瑙韃 褥. 韈碣琥褊);

(30) 韃琿 赳 // 瑟 蒟 珞褞 [釿.] 糺鈔瑣 瑕-鞦蕈 / 鈿 韈 頷碚頽 胛-鞦蕈 (竟.-趺., 頌瑙韃 .

韈碣琥褊).

ツ鞴, 粽 糂襄 韲褞瑾 頌鉋籵 韋 () 鈿 璞褥粢 鈞頸褄 瑪鍄 裼頸璋韋 頌糒褪 襌 褌褊 褪瑕韭璋韋: 胛粽韜 粢矜韈褪 粽褓碣珸, 竟韲琿 磅褌, ォ裲 裲蟒 粽 褫粢褊, 褥硼 蒡赳 褶 齏 珞齏 糺頸 韭瑣鞣 鈞萵韃. ン 糂褶瑯 褌 瓊, 褌 蓖裹 瑟 鈞萵韃 瑜韲褞, 瑯 褞褥珸 褥趺胛 裲 齏 頌瑙 褥趺胛 韈碣琥褊.

ム頸 褪頸 襌 蓖 竟褞褥 砒 裔褊 萵 韋. ネ聿 糂褶瑯 糺珸籵, 糺 褌 瑟, 韲 裼頸璋韶 褪瑕韭瑣鞣, 襌 蒻鞣 , 褥 瑩頏褪 胙瑙頽 璢褊 齏 糂裙 聰, . (蒡頸褄 裼頸璋韶 瑩褞 韲褞瑾 褂褌 蔔褞):

(31) 趺 韭 琅褪 褥 / 鈞褌 裙 聿 粽碼 瑰琿 / 粽 / 粽碼 鈿 (竟.-趺., 頌瑙韃 . 韈碣琥褊);

(32) 粽 裼赳 褄 粽碼 / 蓁 褊 ミ 磊 (...) 鈿 // 瑟 瑕 粽 (...) 鈿 (...) 褊韃 褄粢 瑟 糂褌 蓿胛 (竟.-趺., 粽碚蓖 瑰珸);

(33) 裔 磊 頌 / 砒鉋硴璞 / 粢齏 () 鈿 / 琅 頌籵 瑕- (竟.-趺., 褞褥珸 褥. 裲);

(34) - 蒟瑯 蒡 / - 聰褸 蒻褄 / 聰褸 裔 / - 鈿 褶 ホラナヘワ 趺 糺鶯 (竟.-趺., 粽碚蓖 瑰珸);

(35) 褞裝 粹 瑕- 瑩竟 // 粨蒻 瑕胛- / -- /// 趺 磊 琿韈粢胛 蒡跫韭 鈿 // - // 瑩頌籵 -褌 萵趺 瑰 瑰瑟 (竟.-., 頌瑙韃 褥. 韈碣琥褊).

ツ琥 褪頸, 瑕 胛韶琿 瑣 韋 () 鈿 萵褪 籵 瑰瑣鞣瑣 裹 蓖 蔘 蓿肛 矜 粢矜韈籵胛 鈞褊 瑪 裼頸璋韋 瑙 肄9.

ツ 褄 瑙琿韈 瑣褞鞨, 粲裙 () 鈿, 鈔齏 蒟瑣 裝韃 糺粽蕘. ツ 176 裲瑾 磊 糺粱褊 瑯 裔褊 韋. ミ瑰裝褄褊韃 頷 韵瑟 裲 裝珞褊 珮. 1.

メ珮頽 ム褊韃 韶琿 珸粨蓖裨 韋 () 鈿 珸 韵瑾 瑙 裲 メ韵 裲 メ韵 タ碵. ホ.

マ褞褥珸 マ褞褥珸 ホ頌瑙韃 ホ頌瑙韃 ム粽. -粽 -粽 韋 . 褥. . 褥. 瑰珸 (1) ハ褊 粽褪瑙 - - - 3 - 3 7, 鈞粨韲 琅褂 (2) マ裝韭瑣鞣 1 2 6 7 - 16 41, 籵 鞴瑣 韈頸褄 (3) マ裝韭瑣鞣 - - - - 1 1 2, 籵 鈞粨韲 鈿 粽 (4) ヌ瑜頸褄 - 1/5,3 5/26,3 4/21,1 9/47,4 19/100,0 48, 瑪鍄 裼頸璋韋 (珮./.) ツ裙 韵瑟 1 3 11 14 10 100, 裲 ネ 珮頽 1 粨蓖, 琲碚裹 瑰瑙褊 瑙 褶 珸琿瑰 () 鈿 粽鴦粢 蒻頽頏籵 鍄 韋 鈞頸褄 瑪鍄 裼頸璋韋 (48,7 %). ツ 糂襄 磊 裝竟粢 糺琥褊韃 裔瑙韜 胛粽裙: 裝粢 韃 磊 褶褊 蓿肛 裼頸璋韶 褌褊 (.

韲褞 糺).

ネ 珮頽 1 粨蓖 瑕趺, 琲碚裹 裔頸褄 頌裝褌 萵 韋 珸琿瑰 粽碚蓖 聰-瑰珸瑾 (47,4 %). ン 跫 磅頸 褌 碵褄粽, 粽碚蓖 琲褊裨 襃褊 鞣頏籵 瑕韲- 頌蓖 韲 琲碚裹 瑙褊, 裙 珞韃 琲碚裨 襃褊 琥 瑙 瑩瑕褞 褶.

フ跫 瑰褪 瑰裝褄褊韃 韋 () 鈿 珸 韵瑟 裲 褪 珸顆 竟瑙 (. 珮. 2).

メ珮頽 ム褊韃 裔頸褄 韋 () 鈿 珸 韵瑾 裲 肄蒟 竟琅褂 竟瑙 メ韵 裲 ツ裙 ネ瑙 マ褞褥珸 マ褞褥珸 ホ頌瑙韃 ホ頌瑙韃 ム粽碚蓖 (珮./.) . 褥. . 褥. 瑰珸 18/46, フ踟竟 1 - 5 8 21/53, ニ褊竟 - 3 6 5 ハ瑕 粨蒻, 褥粢 珸頽 顆褥粢 裔褊韜 頌裝褌 韋 踟竟瑟 趺竟瑟 珮萵褪. ユ 褥 鈞褶瑙韜 褊韋 肄蒟 胛粽頷 糂 趺 跫 蒟瑣.

ヘ瑜韲褞, 瑕 蓁 踟竟, 瑕 蓁 趺竟 褞褥珸 趺胛 裲 珸琿 琲碚裹 鈞萵韃, 顆褥粽 胛, 裙 竟瑙 ォ 鈿サ, 鈕褥 竟韲琿 珞褊 頌瑙韃 褥趺胛 裲 (. 珮. 2). ツ 褶 踟竟 頌裝褌 瓊 裔瑰 頌瑙韋 褥趺胛 韈碣琥褊 (44,4 %). ツ 褶 趺竟 褪 褪頷 褞襃琅 蓖胛 韵 裲 蓿胛, 瑕韲琿 顆褥粽 頌裝褌 韜 頸肄 珸琿 粽碚蓖 瑰珸瑾 (33,3 %).

メ瑕韲 碣珸, 葢 褞糺 頸肛 頌裝籵, 跫 蒟瑣 糺粽, , 粽-褞糺, 蓖 竟瑕顆褥 硼 糺瑣 瑙 褶 珸顆 . ツ-糘, 瑰 裹 頌鉋籵 珸籵褪 粱韃 韵 韭瑣鞣胛 褊瑩. ネ, 瑕褻, -褪頷, () 鈿 瓊 裔褪 頌褌 襄 , 粽鴦粢 頸褞瑣 鍄, 鈞頸褄 瑪鍄 裔瑙, 裙 頸褞瑣 褶 裝璢瑯. ヘ琿顆韃 蒡硼 磅褌, 粢矜韈籵 瑪鍄 裼頸璋韋 顏韜 珸 蔔褞鞣瑯 瑙 瑩瑕褞 瑙琿韈頏褌 褶.

マ韲褶瑙 ネ裝籵韃 糺褊 蒿褞跛 胙瑙 ミヤヤネ ォネ銛褊韃 鈞粨韲 褶裘 瑩瑕褞頌韭 粨 韭璋韋 ( 頌裝籵韃 瑣褞鞨 糂裝裘 褶)サ (裲 10-06 00300).

ム. 褌 蓿硼裹: チ聿瑙籵 ヘ. ツ., タ竟糂韜 タ. ム., ミ瑕籵 フ. ツ., ム襃瑙籵 ム. チ., リ褞竟籵 メ. ゙. ヌ糒粽 胛 鍄 糂裝裘胛 碼褊 ォホ蒻 褶裘 蒟サ:

襃 韃 頏籵 // ハ褞 竟聹頌韭 竟褄裲琿 襄肛.

ツ 7 (14). マ 瑣褞鞨瑟 褂裙蓖 褂蔘瑩蓖 褞褊韋 ォト鞨羹 (2008) / テ. 裝.

タ. ナ. ハ鞦韭. フ., 2008. ム. 488-494.

マ 粨 瑙 裲 粽碼 萵 琺 胙瑙頽 琅頽韶 裝趺韜 . 蓿硼裹: チ聿瑙籵 ヘ. ツ. ホ 裝竟頽 頌瑙 竟瑕顆褥 胛 瑙胛 聰: 硴褌, 褪蒻, 肛裼 // ムヒホツホ ホメヌホツナメム゚. マ瑟 タ ム粹 リ褞 ヒ褓鞴 ツ粨 ム瑾瑩胛. マ褞, 2006. ム. 288-293.

チ裹 蓿硼 硴 ヌハミ゚ ., 瑜韲褞: チ聿瑙籵 ヘ. ツ., チ蔗 ネ. ム., ハ瑙籵 ツ. ツ., マ珞籵 ホ. ツ., ム瑜籵 ナ. フ., ヤ齏韵籵 ヘ. ム. ホ ォ蟒 裲 跖粽 褶: 竟韵 頏籵 粽銕跫 頌瑙 // ハ褞 竟聹頌韭 竟褄裲琿 襄肛. ツ 7 (14). マ 瑣褞鞨瑟 褂裙蓖 褂蔘瑩蓖 褞褊韋 ォト鞨羹 (2008) / テ. 裝. タ. ナ. ハ鞦韭. フ., 2008.

ム. 57-61;

チ聿瑙籵 ヘ. ツ., ハ瑙籵 ツ. ツ., ム瑜籵 ナ. フ., ヤ齏韵籵 ヘ. ム. ヌ糒粽 瑕 頌瑙 鍄 糅褌褊胛 胛萵 // ミ褶裘 韭璋 糅褌褊 ミ韋: フ瑣褞鞨 I 褂蔘瑩蓖 瑪 褞褊韋 (ホ, 27-29 瑜褄 2009 .) / マ 裝. ホ. ム. ネ褞, ヘ. タ. ハ錮竟.

ホ, 2009. ム. 43-48;

チ聿瑙籵 ヘ. ツ. ホ 裲 跖粽 褶: 糺 褊 褞糺 裼瑣 頌裝籵 // ハ褞 竟聹頌韭 竟褄裲琿 襄肛. ツ 9 (16). マ 瑣褞鞨瑟 褂蔘瑩蓖 褞褊韋 ォト鞨羹 (2010) / テ. 裝. タ. ナ. ハ鞦韭. フ., 2010. ム. 35-40;

ホ 趺. マ竟韵 裘萵 竟韵 糯裙 頌裝籵 褶 // ム粽 褥 蒟.

゙礪裨 碚韭 瑪 蒡 褥 褥 ネ竟 マ珞粹 ヒ瑕粽. メ 褞糺. ムマ., 2010. ム. 82-91.

ム.: ミ 瑙 褶. ム粽碚蓖 肛-瑰珸 鈞萵 褌. メ裲. ヒ裲顆褥韃 瑣褞鞨 / ム. ツ. ツ. ハ瑙籵 / ホ. 裝. 珞 裝頌粨 ヘ. ツ. チ聿瑙籵. ムマ., 2008;

ミ 瑙 褶. フ肛-襃蔘鞣. メ裲. ヒ裲顆褥韃 瑣褞鞨 / ム. ツ. ツ. ハ瑙籵 / ホ.

裝. 珞 裝頌粨 ヘ. ツ. チ聿瑙籵. ムマ., 2010.

ム. 齏趺韃 蒻褞璋韋: ヤ齏韵籵 ヘ. ム. マ竟韵 褊 胛 頌瑣褄胛 蒻 ( 瑣褞鞨 瑙 褶). ト頌. 瑙. 齏. 瑪. ムマ., 2010 (璧竟頌).

ム.: ォマ 褞襄蓖 肭璢瑾 頽瑙韃 齏珞褌 韲 鈿璞. 磅裲 趺 韲褪 . 齏 粨. サ (www.rusgram.narod.ru, メ. 2, ァ 2667).

ム.: www.rusgram.narod.ru, メ. 2, ァ 2873.

ム., 瑜韲褞, 跖 珮蒟 琅 裔褊韃 瑙 褶 韋 () 蔘 ():

チ聿瑙籵 ヘ. ツ. ハ () 蔘 () 瑙 褶: 褊韃 珸 韶琿 韵 // フ瑣褞鞨 瑪 褞褊韋 ォマ褞 韶頷竟聹頌顆褥 :

ネ蒟 襄 褊韜サ. ハ 70-褪 蓖 趾褊 タ ム粹 リ褞. マ褞, 1 裘琿 2011 .

マ褞, 2011 ( 褶瑣);

ホ 趺. ホ 蓖 韵 裼頸璋韶 糂珞 瑙 褶 // フ瑣褞鞨 IX フ褂蔘瑩蓖 瑪 褞褊韋 ォ゚鍄糺 瑣裙韋 裝竟頽:

竟璢瑣顆褥韜 瑰裲サ. 22-24 褊碣 2011 . ツ琅韲頏, 2011 ( 褶瑣).

y pan Speech rate as reflection of speaker痴 Com social characteristics ing Svetlana B. Stepanova lish Pub This paper presents an analysis of speech rate in spontaneous conversations be tween Russian interlocutors conducted on the basis of recordings of 40 speakers and their interlocutors collected in the Speech Corpus of Russian Everyday In ins teraction 徹ne Day of Speech (ORD Corpus). The results allow us to compare the speed rate between speakers of different mother tongues using data drawn am from other researchers work. For example, Russians speak on average faster than Norwegians, but considerably slower than Spaniards and Brazilians.

The impact of different factors on the rate of speech is illustrated by the follow enj ing findings:

nB 1. There is a statistically valid difference between men痴 (m) and women痴 (f) speech rate: men speak substantially (from the statistical perspective) faster than women.

Joh 2. With age we start speaking more slowly.

3. Informants whose level of verbal competence was assessed at a high level by experts speak more slowly, while an articulation rate higher than the aver age is typical of speakers with a lower level of verbal competence.

4. Furthermore, we observed that the speech rate is dependent on a statisti cally significant way on the number of syllables of the utterance: the longer ofs the phrase, the faster the rate.

pro Keywords: speech corpus, speech rate, Russian ted Introduction 1.

rec The Russian speech rate has been studied by many Russian phoneticians. How ever, those studies are based either on recorded reading material (Krivnova 2007), cor on limited samples of colloquial speech (200 sec for 7 speakers in Rozanova 1983), or on taking the number of words per minute as the rate unit (Svetozarova 1988).

Un Svetlana B. Stepanova y Unfortunately, it is difficult to compare the results of those analyses with the re pan sults obtained on spontaneous speech material in other languages.

In the present study, we use material from the Speech Corpus of Russian Ev Com eryday Interaction (ORD corpus) created at Saint Petersburg State University痴 School of Philology. The main aim of the ORD speech corpus is to fix Russian spontaneous speech in natural communicative situations. A characteristic feature of the corpus is that speech is recorded not in a laboratory (or in the presence of ing the researcher or interviewer), but with the help of a technique designed by the developers of the British National Corpus (http://www.natcorp.ox.ac.uk/corpus/ lish creating.xml, latest access 19/11/2012): on an average day, volunteer informants carry a speech recorder on their neck all day long and record their own speech and the speech of their conversation partners.

Pub Currently, 350 hours of Russian spontaneous speech have been recorded in this manner. At present, the ORD corpus contains recordings made by a demo graphically balanced group of 46 subjects (23 men and 23 women) representing ins various social and age strata: students, military students, engineers, managers, scientists, doctors, IT specialist, a merchant, a builder, a psychologist, a photogra am pher, baby-sitter, a drawing teacher, etc. The subjects age ranges from 16 to years. Although the recordings were made while maintaining absolute anonymity, enj all subjects filled in sociological questionnaires, were given a psychological evalu ation and kept diaries of their 電ay of speech, noting the main conditions of com nB munication. In addition to the subjects speech, 600 interlocutors with whom the subjects conversed were also recorded. The interlocutors as well were of different Joh ages (from 3 to 68 years), professions, and occupations;

they were in both formal and informal relations with the subjects. About 40 hours of recordings have been transcribed (on average, one hour each from the 40 informants) with multi-level annotation using the professional tool ELAN (cf. Asinovskij et al. 2009: 250 for further information about the ORD speech corpus).

ofs 2. Research material pro As is well attested, the perception of the rate of speech is affected by both the ted tempo of articulating sounds and by the presence of pauses of certain duration.

At the first stage of the project, we decided to consider the 殿rticulation tempo, rec regardless of pauses.

In the ORD Corpus, we selected utterances without pauses, sighs, laughter, cor and or overlapping words from more than one speaker at a time.

The total number of selected utterances is 13,459, their total length is Un 19,298 seconds for a total of 5.36 hours. The rate was calculated as the number of Speech rate as reflection of speaker痴 social characteristics y syllables per second (cf. Pellegrino, Coupe and Marsico (2011: 545) discussing the pan advantages of this method). The syllable length was equated with the number of vowels. The material was processed using the STATISTICA program.

Com 3. Results ing General characteristics of the speech rate 3. lish Taking all the material into consideration, the average rate equaled 5.31 syllables per second (syll/s), see Table 1.

For different speakers the average rate changes from minimum 3.6 syll/s Pub (female-speaker, 63 years old) to maximum 6.7 syll/s (female-speaker, 20 years old) see Figure 1.

Our results concerning the average speech rate do not contradict the observa ins tions made by other experimental phoneticians: Krivnova (2007: 58), for example, claims that the average duration of a syllable in Russian is 150200 ms, which am corresponds to 4.766.67 syll/s.

enj Table 1. Average speech rate (m male, f female, k children) nB Descriptive Statistics (Tempo.sta) Valid N Mean (syll/s) Minimum Maximum Std. Dev.

Joh f+m+k 13459 5.31 0.63 13.19 1. Average speech rate 8. ofs 7. 6. pro 5. Speech rate 4. ted 3. 2. rec 1. 0. cor S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S S Speakers Un Figure 1. Average speech rate depending on the speaker 120 Svetlana B. Stepanova y In their review of research on speech rate Verhoeven, De Pauw and Kloots pan (2004: 298) list the average speech rate in several languages:

in British English:

Com 3.165.33 syll/s (Tauroza and Allison 1990) in French:

4.31 syll/s (Grosjean and Deschamps 1973) ing 5.73 syll/s (Malcot, Johnston and Kizziar 1972) in Northern Standard Dutch:

5.2 syll/s (Blaauw 1995) lish in Norwegian ranges:

3.54.5 syll/s (Almberg 2000) Pub In Brazilian Portuguese:

6.57 syll/s In Spanish:

ins 7.81 syll/s (Rebollo Couto 1997) am Compared with these data, Russian speech rate is situated in the middle of this list: Russians speak faster than Norwegians, but slower than Brazilians and enj Spaniards.

This is probably connected to phonetic differences. Here, we agree with nB Verhoeven, De Pauw and Kloots who argue that When speech rate is expressed as number of 層ords per minute or 壮yllables per Joh seconds, the measures will reflect these structural differences, and language with long words or syllables will necessarily have a lower speech rate measures then a language with short words or syllables.

(Verhoeven, De Pauw and Kloots 2004: 298) Further, the country the speaker lives in has a significant effect of on the speech ofs rate: Verhoeven, De Pauw and Kloots (2004: 306) mention that Dutch-people speaking in Belgium show a slower speech rate than those living in the Nether lands. Our data show similar results: speakers from Saint Petersburg speak slower pro than speakers from the provinces (see Figure 2).

ted The impact of gender on speech rate 3. rec The material also allowed us to check for statistically valid differences between men痴 (m) and women痴 (f) speech rate. Our analysis shows that men speak (from a cor statistical perspective) substantially faster than women (5.46 syll/s vs. 5.30 syll/s).

Children痴 speech rate (k) (neglecting gender) is the slowest 3.86 syll/s (see Ta Un ble 2). The statistical significance of these differences has been checked using the Speech rate as reflection of speaker痴 social characteristics y Box & whisker plot pan Non-residents vs. SPb-residents 6, Com 6, 6, 6, ing Mean Speech rate 6,0 Mean ア SE lish Mean ア 1,96* SE 5, Pub 5, 5,7 ins 5, am 5, Non-residents SPb-residents enj Figure 2. Significance of differences in average speech rate: Saint Petersburg (SPb) residents and non-residents nB STATISTICA program. The t-test showed a statistically significant difference be tween the mean values in these groups (p 0.001) (see Table 2).

Joh A similar impact of gender on the speech rate has been observed in other languages as well (cf. for Chinese and American English by Yuan, Libermann and Cieri 2006 and Jacewicz, Fox and Wei 2010;

for Dutch by Verhoeven, De Pauw and Kloots 2004 or Quene 2008;

and for English, French, German, Italian, Japanese, Mandarin, Spanish, and Vietnamese by Pellegrino, Coupe and Marsico ofs 2011: 546). Characteristically, all authors observe that the difference between men痴 and women痴 average speech rate is not substantial but statistically consider pro able. Quene explains men痴 faster speech rate in terms of social dominance: 杜ale speakers may also express their social dominance by speaking somewhat faster than female speakers (Quene 2008: 1112).

ted Table 2. T-test for male/female/children痴 speech rate rec Mean Mean t-value Valid N Valid N p Group 1 Group 2 Group 1 Group 2 Variances cor m vs. k 5.46 3.86 16.63 5067 471 0. m vs. f 5.46 5.30 4.67 5067 7921 0. Un k vs. f 3.86 5.30 16.49 471 7921 0. 122 Svetlana B. Stepanova y The average speech rate depending on the utterance length pan 8, Speech rate (syll/s) 6, Com 4, 2, ing 0, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 The utterance length (syll) lish Figure 3. Average speech rate depending on the utterance length Pub However, all these researchers recognize that the speech rate is to a great ex tent determined by the length of the generated utterance. Possibly, all other factors affecting the tempo can be derived from the length of an utterance, as Liberman ins argues: 溺aybe region, sex and age don稚 really influence speaking rate after all, except indirectly via their influence on phrase length (Liberman 2006: 1). Fig am ure 3, indeed, shows that the longer the phrase, the faster the rate: all material considered, the average rate in single-syllable utterances is 2.59 syll/s, while in enj ten-syllable ones it is 6.25 syll/s.

However, in even lengthier utterances, the speech rate does not change sub nB stantially and remains within the range of average values of 6.27.2 syll/s. This may be explained by purely physiological reasons: a syllable has to be articulated Joh by the speech organs, and the movement rate of the latter cannot grow infinitely.

Further, speech perception also sets certain limitations subconsciously recog nized by the speaker:

when the rate exceeds maximum, syllable recognition may be impeded or even become impossible. These limitations may be explained by the facts that the pro ofs cess of interpretation requires time and that duration is one of the formal indica tors of a stream element necessary for defining the latter (a section shorter than a certain value cannot be an element. (istovi 1976: 14;

transl. S. B.) pro In order to take into account the phenomenon of anticipatory shortening, we checked the average syllable length of the analyzed fragments. Our material ted showed that the average number of syllables in men痴 utterances is slightly lower than that in women痴 utterances (8.1 vs. 8.3) (see Table 3), which means that Rus rec sian men on average pronounce shorter phrases, but at a higher rate.

Here, our results differ from the observations made by Fenk-Oczlon and cor Fenk (2010). According to them (Fenk-Oczlon and Fenk 2010: 1538), for Russian the number of syllables per clause is 5.68. This is substantially lower than our re Un sults. This divergence is probably due to the different character of the speech data Speech rate as reflection of speaker痴 social characteristics y Table 3. Average number of syllables and words in a speech fragment pan Descriptive Statistics (words/syll in phrase) Valid N Mean Minimum Maximum Std. Dev.

Com words-m 5065 4.4 1 44 4. syll-m 5065 8.1 1 82 7. words-f 7921 4.4 1 45 4. syll-f 7921 8.24 1 88 7. ing analyzed and the different counting procedure. Fenk-Oczlon and Fenk (2010) lish analyzed simple declarative sentences, whereas our analysis was based on speech fragments between physical pauses. These fragments could include complex sen Pub tences or a few simple phrases, spoken without pauses and irregularities.

Furthermore, a more detailed examination of the material comparing the av erage speech rate of men and women in utterances ranging from 1 to 20 syllables ins has been conducted.

Statistical analysis showed that in single-syllable utterances the speech rate am does not depend on the speaker痴 gender and is almost twice as slow as the aver age rate in the whole body of the material 2.6 syll/s (see Figure 4). Two-syl enj lable utterances were pronounced slightly faster by women than by men (4.09 vs.

3.97 syll/s). However, the difference is not statistically significant. In all other nB cases (utterances of more than two syllables), men痴 speech rate was higher than Joh 8, 7, 6, Average speech rate (syll/s) 5, ofs Female 4,00 Male pro 3, 2, ted 1, rec 0, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 cor Number of syllables Figure 4. Average speech rate depending on the utterance length (according to speaker痴 Un gender) 124 Svetlana B. Stepanova y women痴. In 5-, 7-, 10-, 11-, 14- and 20-syllable utterances the difference was sta pan tistically significant.

Com 8, ing 7, 6, lish 5, Speech rate Speech rate 4, St.Dev.

Pub 3, 2, 1, ins 0, am y Age enj Figure 5. Average speech rate and standard deviation depending on the informant痴 age nB Box & whisker plot 40 vs. Joh 6, 6, 6, ofs 5, pro 5, ted 5, rec cor 5, 40 Un Figure 6. Significance of average speech rate differences (older vs. younger than 40) Speech rate as reflection of speaker痴 social characteristics y The impact of age on speech rate 3. pan Unlike research based on other languages, our analysis does not point to a di Com rect influence of speaker痴 age on the speech rate. As mentioned in Section 3.1, children (in our corpus age 312) display a much slower speech rate than adults.

However, from the age of 16 on, age does not seem to influence the speech rate in a clearly patterned way (see Figure 5). There is a tendency towards a decrease ing on speech rate with speaker痴 age, but 菟eaks of male speakers aged 53, 55 and 56 disturb it.

lish Nevertheless, should we divide our speakers into two groups (younger vs.

older than 40), a significant difference concerning speech rate can be observed (see Figure 6): with age the informants start speaking more slowly, although never Pub reaching the slowness level of children痴 speech rate (see Table 2).

ins The impact of other social factors 3. am Statistical analyses did not reveal substantial differences concerning speech rate between speakers with higher, incomplete higher and secondary education. Het enj erogeneous factors may account for this observation: the ORD Corpus contains only a small number of recordings from speakers without higher education. Fur nB ther, there is a statistical spread of their average speech due to other factors.

However, the level of the informants verbal competence (LVC) has on aver age an impact on the speech rate. The level of verbal competence is determined by Joh a combination of the speaker痴 social characteristics: a speaker痴 level of education, as well as professional or non-professional use of speech, play a crucial role in as sessing his or her LVC, but a speaker痴 personal involvement in social matters is also to be considered. In determining a speaker痴 LVC, we relied on an expert痴 as sessment and interpretation of audio extracts from the corpus. These results have ofs subsequently been checked with the background information from the speaker痴 questionnaire in order to find correlations between the expert痴 interpretation of pro speech samples and the 喪eal social characteristics of the speakers (Bogdanova et al. 2008: 5761).

It turned out that the informants whose LVC was assessed by experts as ted 塗igh speak more slowly. An articulation rate higher than the average is typical for speakers with a 斗ower LVC (see Figure 7). Probably, an analysis that consid rec ers pauses and their length will yield different results.

It seems evident to point out that we speak differently with different interloc cor utors. Nevertheless, at this stage we can only conduct statistically reliable analyses of the average speech rate in conversations with colleagues and friends. These Un 126 Svetlana B. Stepanova y Categ. Box & whisker plot: speech rate pan 6, Com 6, 5, Mean ing Speech rate Mean ア SE 5,4 Mean ア 1,96* SE lish 5, Pub 5, ins 4, m h l am Verbal competence level Figure 7. Significance of differences in average speech rate values: Average (m) high enj (h) low (l) LVC nB Box & whisker plot Friends vs. Colleagues 5, Joh 5, 5, 5, Mean ofs Mean ア SE Mean ア 1,96* SE 5, pro 5, 5, ted 5, rec 5, cor Friends Colleagues Figure 8. Significance of differences in average speech rate values: With friends vs. with Un colleagues (all material considered) Speech rate as reflection of speaker痴 social characteristics y show that the analyzed speakers speak faster in a relaxed and friendly atmosphere pan than in the office with their colleagues (see Figure 8).

Com 4. Conclusion The present study based on the Speech Corpus of Russian Everyday Interaction ing (ORD Corpus) has yielded new data on the rate of spontaneous Russian speech.

Although the ORD Corpus is not as large as the corpora used by researchers of lish Dutch, English, Chinese or Japanese, the results obtained confirm many of their conclusions: speech rate depends on such factors as gender, age, level of verbal competence, as well as on the length of the utterance. Our analyses show that Pub 1. Men speak substantially faster than women.

2. Speech rate decreases with age.

ins 3. Informants with a high level of verbal competence (determined by experts assessment) speak more slowly, while an articulation rate higher than the av am erage is typical of speakers with a lower level of verbal competence.

4. Finally, there is a statistically significant relationship between the speech rate enj and the number of syllables in the spoken phrase: the longer the phrase, the faster the rate.

nB However, other factors concerning the speaker (e.g. the communicative situation, the kind of relation with the interlocutor, the emotional condition) can also sig Joh nificantly influence the speech rate. The ORD corpus contains several genres and different styles of speech. There are casual conversations at home with relatives (talk during meals, party talk, family chats, etc.) as well as professional and infor mal conversations with colleagues, communication in academia (lectures, practi cal lessons, and students informal conversations), communication with friends ofs in different situations and at different places, consultations with doctors, talk with shop assistants, telephone conversations, etc. Recordings were made at home, in pro the office, in public transport, on the streets, at universities, in a military college, in coffee bars and restaurants, in shops, in amusement parks, and so on. Further analyses of these data will allow us to check how the speech rate is affected by ted communicative situations and, subsequently, by the communicative role of the speaker.

rec cor Un 128 Svetlana B. Stepanova y Acknowledgments pan The study was supported by the Ministry of Education and Science of the Rus Com sian Federation, project 8554 鉄peech Corpus of the Russian Language: Complex Analysis of Oral Speech. My sincere thanks to all colleagues working on the cre ation of the ORD Corpus and especially to Tatiana Sherstinova for her continuous help in the automatic processing of annotated materials.

ing lish References Almberg, Jrn. 2000. 適or fort snakkar vi eigentleg? Nordlyd 28: 6073.

Pub Asinovsky, Alexander S. et al. 2009. 典he ORD Speech Corpus of Russian Everyday Com munication 徹ne Speaker痴 Day: Creation Principles and Annotation. In メ蛉t, Speech and Dialogue. TSD-2009 [LNCS/LNAI 5729]. Matouek Vaclav and Mautner Pavel (eds.), ins 250257. Berlin/Heidelberg: Springer.

Bogdanova, Natalia V. et al. 2008. 徹 ォkorpuseサ ivoj rei: principy formirovania i vozmonosti am opisania. In Komp男uternaja lingvistika i intellektual地ye texnologii: Trudy medunarodnoj konferencii ォDialog 2008サ (7/14), Aleksandr E. Kibrik (ed.), 5761. Moscow: MGU.

Blaauw, Eleonora. 1995. On the Perceptual Classification of Spontaneous and Read Speech [OTS enj dissertations series]. Utrecht: Led.

British National Corpus: http://www.natcorp.ox.ac.uk/corpus/creating.xml.

nB istovi, Ludmila A. (ed.). 1976. Fiziologija rei. Vosprijatie rei elovekom. Leningrad: Nauka.

Grosjean, Franois and Deschamps, Alain. 1973. 鄭nalyse des variables temporelles du franais spontan II Comparaison du franais oral dans la description avec l anglais description et Joh avec le franais interview radiophonique. Phonetica 28 (34): 191226.

Fenk-Oczlon, Gertraud and Fenk, August. 2010. 溺easuring Basic Tempo across Languages and some Implications for Speech Rhythm. In International Conference on Spoken Lan guage Processing (Interspeech-2010), Makuhari, Chiba, Japan, 15371540.

Jacewicz, Eva, Fox, Robert A., and Wei, Li. 2010. 釘etween-speaker and within-speaker varia tion in speech tempo of American English. Journal of the Acoustical Society of America ofs 128 (2): 839850.

Krivnova, Ol暖a F. 2007. 迭itmizacija i intonacionnoe lenenie teksta v 叢rocesse rei-mysli.

Abstract

of PhD thesis. Moscow: MGU.

pro Liberman, Mark. 2006. 敵uys are a bit gabbier in Dutch, too. http://itre.cis.upenn.edu/~myl/ languagelog/archives/003682.html (latest access 13.11.2012).

Malecot, Andre, Johnston, R and Kizziar, P. (1972). 鉄yllabic rate and utterance length in ted French. Phonetica 26: 235251.

Quene, Hugo. 2008. 溺ultilevel modeling of between-speaker and within-speaker varia rec tion in spontaneous speech tempo. Journal of the Acoustical Society of America 123 (2):

11041113.

Pellegrino, Francois, Coupe, Christophe, and Marsico, Eggidio. 2011. 鄭 cross-language per cor spective on speech information rate. Language 87 (3): 539558.

Rebollo Couto, Letcia. 1997: 鏑e rhythme en espagnol et en portugais: syllabique ou accentuel. Un Travaux de l棚nstitut de Phontique de Strasbourg 27: 6390.

Speech rate as reflection of speaker痴 social characteristics y Rozanova, Natalia. N. 1983. 鉄upersegmentnaja fonetika. In Russkaja razgovornaja re: Fone pan tika. Morfologija. Leksika. est, Elena A. Zemskaja (ed.), 579. Moskow. Nauka.

Svetozarova, Natalia D. (ed.). 1988. Fonetika spontannoj rei. Leningrad: LGU.

Tauroza, Steve and Allison, Desmond. 1990. 鉄peech rates in British English. Applied Linguis Com tics 11 (1): 90-105.

Verhoeven, Jo, De Pauw, Guy, and Kloots, Hanne. 2004. 鉄peech rate in a pluicentric language.

A comparison between Dutch in Belgium and in Netherlands. Language and Speech (3): 297308.

ing Yuan, Jiahong, Liberman, Mark, and Cieri, Cristopher. 2006. 典owards an Integrated Under standing of Speaking Rate in Conversation. In International Conference on Spoken Lan guage Processing (Interspeech-2006).

lish Pub ins am enj nB Joh ofs pro ted rec cor Un Un cor rec ted pro ofs Joh nB enj am ins Pub lish ing Com pan y y pan Russian everyday utterances Com The top lists and some statistics ing Tatiana Sherstinova lish Pub This paper presents lists of the most frequently used Russian utterances in ev eryday interaction and gives a statistical description of Russian spontaneous speech;

including, frequency distribution of utterance length measured in words ins and syllables and temporal statistics of utterances. Audio data for this research were taken from the speech corpus of Russian everyday communication known am as the Speech Corpus of Russian Everyday Interaction 徹ne Day of Speech (ORD Corpus). Special investigation was conducted to analyze average utter ance duration and syllable rate of speech relative to utterance length in syllables.

enj Additionally, we propose the existence of two temporal registers in Russian everyday speech.

nB Keywords: Russian spoken language, everyday utterances, utterance length, Joh ORD speech corpus, everyday speech communication Introduction 1.

This paper continues the statistical description of Russian spoken interaction that ofs was started in (Sherstinova 2010).

Audio data for this research were taken from the ORD speech corpus of Rus pro sian everyday communication. All recordings were made under natural condi tions, with digital voice recorders hung around the neck of subjects, and captured ted natural spoken language communication during a single 24 hour period. A de tailed description of the ORD corpus is presented in Asinovsky et al. (2009).

To date, transcripts and multi-level linguistic annotation have been made for rec 40 hours of the ORD recordings and represents speech from 20 male and 20 female subjects and their interlocutors. For this research we have excluded from investi cor gation all utterances containing overlapping speech, unintelligible speech, unfin ished (broken) remarks, utterances with inner pauses and/or hesitations, and thus, Un 106 Tatiana Sherstinova y we our corpus consists of roughly 15,000 utterances produced by approximately pan 200 different people. These utterances formed the basis for our investigation.

The principles of division of speech into utterances were described in Com Sherstinova, Stepanova and Ryko (2009).

All statistics presented here have been obtained for the whole subset of 15, utterances regardless of age, gender and other social or psychological character istics of speakers. Another analysis of speech rate was conducted on this same ing data (Stepanova 2011). However, it took into account different social factors of speakers.

lish 2. Frequency distribution of utterance length in words Pub As an initial step, the frequency distribution of utterance length in words was ob tained (see Figure 1). Based on our data, the average utterance length is 4.35 words, ins and standard deviation is 4.02 words. The longest utterance in the ORD corpus has 71 words. It turned out that the majority of all utterances consist of a single am word or word-like particle (25.26%). Two-word utterances make 15.58% of the whole data, three-word utterances has the third rank making 12.45%. Four-word enj utterances make 10,98%, five-word utterances 8.74%, etc.

nB Joh Absolute frequency ofs pro ted rec 1 4 7 10 13 16 19 22 25 28 31 35 44 Utterance length in words cor Figure 1. Frequency distribution of utterance length in words Un Russian everyday utterances y In general, the shorter the utterance length in words, the higher its frequency pan in real communication. Subsequently, utterances consisting of up to three words make more than a half of all speech interaction.

Com 3. The top lists of Russian everyday utterances ing Frequency lists were further obtained for utterances of the same length in words.

In this article we present high-frequency zones of these lists for most common lish utterances consisting up to 4 words (see Tables 14). Making these lists we try to keep reference to functional type of utterance marked by the following cor respondent symbols: declarative utterance (//), interrogative utterance (?) and Pub exclamatory, greeting, vocative or imperative utterances (!). The percentage was calculated within each group of utterances. Whenever possible, we tried to save the transcript痴 division into rhythmic groups (or syntagmas) marked by a for ins ward slash (/) (e.g., 砺s / ja ponjal 奏hat痴 enough, I got it). However, if multiple division into rhythmic groups was possible for the same word string (e.g., 電a da am and 電a / da 惣eah yeah and 惣eah / yeah), we left in the list just one variant summing all relevant frequencies.

enj It should be noted as well that when making these lists we excluded from them all phrases that only occur in the speech of a single speaker. This left only nB those which are more common. For example, we excluded from the lists an utter ance of a triple occurrence 店oma, razdevajsja! (銭homa, take off your clothes!) Joh when a mother was trying to get her three-year old son to change his clothes after walking, or a favorite quote of a song of another subject 鍍a enina v okoke (禅his lady behind the window), which he repeated several times, as well as other non-common utterances.

As shown in the previous section, more than a quarter of all Russian speech ofs interaction is performed using one-word utterances. The top list of one-word ut terances obtained on our material is given in Table 1. The total number of one word utterances in our ORD transcripts is 3,648 (the percentages shown in the pro right column were calculated out of this number).

We may see that many of them are formed by discourse particles, which have ted mainly the pragmatic function of providing successful interaction: e.g., 砥gu, 殿ga, 鍍ak, 電a (all four meaning 惣ep / yeah), 砺ot (層ell/ 蘇ere is), 渡u (層ell), rec 杜 (a hesitation marker), 菟onjatno and 屠asno (both meaning 訴t痴 clear), 徒oneno (壮ure), 渡ormal地o (訴t痴 all right), and many others.

cor Among interrogative utterances the highest rank has 殿?, 杜?, 鍍o? (all used in the meaning of 層hat are you speaking about? or asking for repetition), 電a? Un 108 Tatiana Sherstinova y Table 1. Most frequent one-word Russian utterances pan Rank Utterance Count Percent Rank Utterance Count Percent 1 ugu // 595 16.31% 30.5 sejas // 11 0.30% Com 2 da // 436 11.95% 30.5 aj ! 11 0.30% 3 tak // 142 3.89% 34 molodec ! 10 0.27% 4 aga // 104 2.


85% 34 pojdm ! 10 0.27% 5 vot // 100 2.74% 34 poalujsta ! 10 0.27% ing 6 a? 93 2.55% 36 net ? 9 0.25% 7 net // 74 2.03% 38 ja ? 8 0.22% 8 m? 73 2.00% 38 al ! 8 0.22% lish 9 da ? 65 1.78% 38 bljad ! 8 0.22% 10.5 oj ! 55 1.51% 40.5 kak ? 7 0.19% Pub 10.5 m // 55 1.51% 40.5 gospodi ! 7 0.19% 12 a // 50 1.37% 44 na ! 6 0.16% 13 to ? 47 1.29% 44 netu // 6 0.16% 15 nu // 42 1.15% 44 normal地o // 6 0.16% ins 15 xoroo // 42 1.15% 44 o! 6 0.16% 15 vs // 42 1.15% 44 jasno // 6 0.16% am 17 nu ? 29 0.79% 51 vo ! 5 0.14% 18 davaj ! 26 0.71% 51 vs ? 5 0.14% enj 19 zdravstvujte ! 23 0.63% 51 est // 5 0.14% 20.5 ponjatno // 22 0.60% 51 zaem ? 5 0.14% nB 20.5 poemu ? 22 0.60% 51 kakie ? 5 0.14% 22 spasibo // 21 0.58% 51 net ! 5 0.14% 23 koneno // 18 0.49% 51 opa // 5 0.14% Joh 24 privet ! 17 0.47% 51 xm // 5 0.14% 26.5 a! 12 0.33% 51 xoroo ? 5 0.14% 26.5 gde ? 12 0.33% 62.5 allo ! 4 0.11% 26.5 kto ? 12 0.33% 62.5 mama ! 4 0.11% 26.5 tak ? 12 0.33% 62.5 kakoj ? 4 0.11% 30.5 kuda ? 11 0.30% 62.5 zameatel地o // 4 0.11% ofs 30.5 ego ? 11 0.30% 62.5 prikol地o // 4 0.11% pro (惣es?), which is often used to demonstrate astonishment, 渡u? (層ell?), stimu lating interlocutors to continue speaking, explaining or performing some other ted action. This list is continued by standard question utterances: 菟oemu? (層hy?), 堵de? (層here?), 徒to? (層ho?), 鍍ak? (壮o? or 前K?), 徒uda? (層hereto?), and many others.

rec The most common one-word interjections are the following: 登j! and 殿j! (similar to English 双h! and 疎h!), 殿! (疎h!), 登! (双h!), 杜olodec! (層ell done!), cor 堵ospodi! (銑ord!), and others.

Un Russian everyday utterances y Table 2. Most frequent two-word Russian utterances pan Rank Utterance Count Percent Rank Utterance Count Percent 1 nu da // 54 2.39% 24 a zaem ? 4 0.18% Com 2 nu vot // 33 1.46% 24 a poemu ? 4 0.18% 3 ne znaju // 29 1.28% 24 v smysle ? 4 0.18% 4.5 nu ladno // 15 0.66% 24 ne xou // 4 0.18% 4.5 m da // 15 0.66% 24 nu-nu // 4 0.18% ing 6 nu davaj // 12 0.53% 24 tak / sejas // 4 0.18% 7 da da // 10 0.44% 24 tak to... 4 0.18% 8.5 i vs // 8 0.35% 24 ja ponjal // 4 0.18% lish 8.5 nu ponjatno // 8 0.35% 37 a kuda ? 3 0.13% 10 o gospodi ! 7 0.31% 37 a ty ? 3 0.13% Pub 11.5 spasibo bol弛e // 6 0.27% 37 a to ? 3 0.13% 11.5 ugu / ugu // 6 0.27% 37 vot to ? 3 0.13% 16 a ego ? 5 0.22% 37 vs / spasibo // 3 0.13% 16 vot tak // 5 0.22% 37 vs ravno // 3 0.13% ins 16 da ladno // 5 0.22% 37 da / koneno // 3 0.13% 16 da net // 5 0.22% 37 da u // 3 0.13% am 16 ne nado // 5 0.22% 37 ili net ? 3 0.13% 16 do svidanija ! 5 0.22% 37 ne ponjal // 3 0.13% enj 16 nu to ? 5 0.22% 37 niego sebe // 3 0.13% nB Regarding greetings (salutations), there were no surprises here. At the top are 奴dravstvujte! (蘇ello!), 菟rivet! (蘇ello! or 蘇i!), and the telephone greetings Joh 殿l! (蘇ello!), and 殿llo! (蘇ello!).

Additionally, there were a considerable amount of utterances expressing eval uation, like 度oroo (前K!), 渡ormal地o (訴t痴 all right!), 奴ameatel地o (組reat!), 菟rikol地o (祖ool!), etc.

Examining at frequent two-word utterances (see Table 2), we noticed that many of them have a similar structure, where each begins with a particle. The ofs most 殿ctive particle here is 渡u (層ell), which initiates the most frequent utter ances: 渡u da (層ell, yes), 渡u vot (an utterance making a boundary in speech pro or interaction), 渡u ladno (前K!), 渡u davaj (層ell, (let痴) do it), 渡u ponjatno (層ell, I see), 渡u-nu (層ell, well), etc.

ted Particle 電a (惣eah) is used in a similar function, forming utterances such as 電a da (惣eah yeah), 電a ladno (前K!), 電a u (層ell, yeah), 電a / koneno rec (惣es, sure), 電a net (although a literal translation of this utterance is 惣es (well) no, meaning 創o!), etc.

About 10% of all two-word questions begin with conjunctions: 殿 (疎nd, 礎ut):

cor 殿 ego? (疎nd what?), 殿 gde? (疎nd where?), 殿 zaem? (疎nd what for?), 殿 poemu? (疎nd why?), 殿 ty? (疎nd you?), 殿 to? (疎nd what?), 殿 kuda? Un 110 Tatiana Sherstinova y Table 3. Most frequent three-word Russian utterances pan Rank Utterance Count Percent Rank Utterance Count Percent 1 da da da // 14 0.78% 24.5 a da da // 2 0.11% Com 2.5 nu i to ? 6 0.33% 24.5 a nu da // 2 0.11% 2.5 nu ne znaju // 6 0.33% 24.5 da ty to ! 2 0.11% 4.5 vot i vs // 4 0.22% 24.5 da ty to ? 2 0.11% 4.5 vot tak vot // 4 0.22% 24.5 i to teper ? 2 0.11% ing 7.5 vs / ja ponjal // 3 0.17% 24.5 kak vy moete ? 2 0.11% 7.5 nu i vs // 3 0.17% 24.5 ne figa sebe ! 2 0.11% 7.5 ja ne znaju // 3 0.17% 24.5 nu / spasibo bol弛e // 2 0.11% lish 7.5 ja ne mogu // 3 0.17% 24.5 nu vs / poka // 2 0.11% 24.5 a / nu da // 2 0.11% 24.5 nu davaj / poka // 2 0.11% Pub 24.5 a / nu ponjatno // 2 0.11% 24.5 nu i xoroo // 2 0.11% (疎nd where to?), etc. The most typical negative utterances are the following: 渡e ins znaju (選 do not know), 渡e nado (層e don稚 need it or 租on稚 do it!), 渡e xou (選 don稚 want it), 渡e ponjal (選 don稚 understand).

am The only complete utterance (defined here as having both a subject and a predicate) presented in this list of most frequent two word utterances is 屠a pon enj jal (選 understood). The total number of two-word utterances in our corpus is 2262.

nB Three-word utterances are less numerous and their frequencies are rather low (see Table 3). Thus, the most frequent utterance, consisting of triple repetition of Joh the particle 電a (惣es/惣eah) occurs only 14 times. In this corpus of 1804 utter ances, it represents less than one percent of the total number of three-word utter ances. We see here mainly the same words and particles that found to be the most frequent for one and two-syllable utterances as well: 電a (惣es / 惣eah), 砺ot (層ell), 渡u (層ell), 砺s (層ell / 奏hat痴 enough), 鍍ak (惣es), 鍍o (層hat), 電avaj (前K! / 鼠et痴 ), 度oroo (前K!) and others. Many of these utterances are ofs very close in meaning with shorter forms.

Moreover, Table 3 shows that some of three-word utterances are a combi pro nation of the most frequent one and two-word elementary utterances presented above. For example: 渡u ne znaju (層ell, I don稚 know), 砺s / ja ponjal (奏hat痴 enough / I got it), 殿 / nu da (疎h, I壇 rather agree), 殿 / nu ponjatno (層ell, I see), ted 渡u / spasibo bol弛e (層ell, thanks a lot), 渡u vs / poka (層ell, let痴 stop, bye!).

However, in contrast to utterances consisting of one or two words, three-word rec utterances use the personal pronouns 屠a (選), 鍍y (惣ou), 砺y (惣ou) rather regularly.

cor Table 4 presents four four-word utterances that occur in the speech of differ ent speakers: 徒ak u tebja dela? (蘇ow are you?), 殿 to to takoe? (層ell, what is Un Russian everyday utterances y Table 4. Most frequent four-word Russian utterances pan Rank Utterance Count Percent 1.5 kak u tebja dela ? 3 0.19% Com 1.5 a to to takoe ? 3 0.19% 3.5 tak / tak tak tak // 2 0.13% 3.5 ja voobe ne ponimaju // 2 0.13% ing it?), 鍍ak / tak tak tak (壮o so so so), 屠a voobe ne ponimaju (選 absolutely do not understand). All the other four-word utterances occur just once in our data lish or are peculiar only to individual speakers. The total number of four-word utter ances investigated is 1565.

Pub Interestingly, we in our subset of the ORD corpus there were no utteranc es of five or more words that occurred at least twice in the speech of different speakers.

ins am 4. Frequency distribution of utterance length in syllables enj Frequency distribution of utterance length in syllables is shown on Figure 2. The utterances classified here are according to the ideal (or 吐ull-style1) number of nB syllables.

Utterance length in our data ranges from 0 syllable (as in 杜 // or 杜 ?) to 132 syllables. The average utterance length is 8.01 syllables and standard deviation Joh is 7.71 syllables.

The most frequent Russian utterances consist of one or two syllables and rep resent 11.0% and 11.7% of all utterances respectively. Three-syllable utterances are ranked third (8.7%), four-syllable utterances are ranked fourth (7.7%), five syllable represent 6.88% and six-syllable utterances represent 6.50%. Utterances ofs longer than 20 or more syllables occur less than 1% of the time. Thus, more than half of all spoken communication consists of short utterances with a length up to pro 6 syllables.

ted 1. According to Russian phonetic tradition, 吐ull pronunciation style is a well-articulated rec pronunciation without reduction, simplification or omission of sounds and syllables. This style is inherent (is it inherent or simply strongly correlated with?) to slow rate of speech, and is typi cal for public (mainly, declamatory) speaking. It may be also called 鍍he ideal style, as opposed cor to that of real spontaneous speech with multiple reductions. The ideal number of syllables in textual transcripts is automatically calculated basing on rules of the full pronunciation style and Un may slightly differ from the real number of syllables in the original speech.

Tatiana Sherstinova y pan Com Absolute frequency ing lish Pub 200 ins 0 5 10 15 20 25 30 35 40 45 50 55 60 79 Utterance length in syllables am Figure 2. Frequency distribution of utterance length in syllables enj 5. Average duration of utterances depending on their length in syllables nB Special investigation was made to analyze average duration of utterances depend Joh ing on their length in syllables (Sherstinova 2011). The main statistics of this dis tribution for utterances under 30 syllables are given in Table 5. Here we did not analyze longer utterances, as their number in the ORD corpus is not sufficiently significant.

One may notice that one- and two-syllable utterances, being the base of spo ken communication, have the minimal variation when compared to other types ofs of utterances. Thus, the average utterance duration was 453 ms for one-syllable utterances (n = 1608, SD = 208), while it was 563 ms for two-syllable utterances pro (n = 1710, SD = 242). That means that their durations are the most predictable.


What concerns the dependency of average utterance duration on their length ted in syllables, for our empiric data it is well described by the following linear func tion: y = 133,28x + 367,5, where x is a number of syllables, and y is an average duration of such utterances in milliseconds. Figure 3 presents both empirical data rec (marked by dots) and the theoretical line. Our proposed model fits the data well.

cor Un Russian everyday utterances y Table 5. Descriptive statistics of utterance length pan Average utterance Average utterance Utterance length Utterance length Average syllable Average syllable duration, ms duration, ms duration, ms duration, ms Com in syllables in syllables deviation deviation Standard Standard Count Count ing 0 188 641 1001 16 221 2484 562 1 1608 453 208 453 17 238 2647 537 2 1710 563 242 282 18 197 2715 631 lish 3 1263 744 303 248 19 152 2896 539 4 1118 903 329 226 20 114 3023 658 Pub 5 1001 1008 306 202 21 112 3121 661 6 943 1151 349 192 22 70 3353 735 7 858 1312 418 187 23 85 3342 602 8 752 1429 434 179 24 68 3572 725 ins 9 711 1554 414 173 25 77 3782 765 10 636 1674 434 167 26 67 3974 743 am 11 485 1816 450 165 27 47 3981 702 12 448 1986 548 166 28 45 4061 832 enj 13 395 2055 529 158 29 33 4280 594 14 319 2236 567 160 30 34 4343 1085 15 259 2268 537 nB Joh ms Average utterance duration ofs pro ted rec cor 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Utterance length in syllables Un Figure 3. Average utterances duration and utterance length in syllables. Empirical data and the theoretical model 114 Tatiana Sherstinova y 6. Dependence of average syllable duration on utterance length pan in syllables Com This section is concerned with average syllable duration for utterances of different length in syllables. The data we analyze here were presented above in Table 5 in two columns (Utterancelengthinsyllables and Averagesyllableduration, ms). The correspondent dependency curve is shown in Figure 4.

ing Differences between neighboring averages from Table 5 were tested for sig nificance by means of T-test (the normality of distribution was confirmed by the lish Shapiro-Wilk W-test). For example, we tested the differences between: (1) one syllable and two-syllable utterances, (2) two-syllable and three-syllable utteranc es, (3) three-syllable and four-syllable utterances, and so on.

Pub The result is quite convincing. It turned out that for utterances under 15 syl lables, the average syllable duration significantly differs between any two neigh boring utterance types, at a significance level of 95%. Thus, we may assume strong ins dependency of average syllable duration and utterance length in syllables.2 Based on our data, the average syllable duration in utterances (whose length is more am than 15 syllables) does not differ significantly and is equal to approximately to 150 ms.

enj nB Joh Average syllable duration Regular (dialogue) register Speedy (monologue) register ofs pro 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Utterance length in syllables ted Figure 4. Average syllable duration relative to utterance length in syllables.

A hypothetical division into two registers rec cor 2. The dependence of average syllable duration on utterance length in syllables is integrative part of the so called Menzerath痴-Altmann law, according to which the increase of a linguistic Un construct results in a decrease of its constituents.

Russian everyday utterances y An important conclusion drawn from our research might be summarized as pan follows: We can assume the existence of two registers of speech:

1. The 途egular (or dialogue) register is used for producing utterances whose Com length does not exceed 15 syllables. Its distinctive feature is a strong depen dence between average syllable duration (syllable rate) of utterances and their length in syllable. Therefore, in this case the average syllable duration ing is a function from utterance length in syllables, ranging approximately from 450 ms to 150 ms.

2. The 都peedy (or monologue) register is used for producing long utterances lish (exceeding 15 syllables). In contrast to a dialogue register, the average syllable duration (or an average utterance rate) of a 都peedy register does not depend Pub on utterance length in syllables and is equal to approximately 150 ms.

We may suppose that 150 ms functions as a sort of constant, around which the ins average duration of long Russian spontaneous utterances tends to occur. Finally, we note that the average syllable duration for the whole analyzed part of the ORD am corpus is 160 ms, which is only 10 ms longer.

enj Conclusion 7.

nB The main results of this research are the following:

Joh 1. The majority of Russian spoken interaction consists of one or a few word ut terances, which contain one or few syllables.

2. The top lists of the most common Russian utterances are obtained. Discourse particles occur in, or at times entirely constitute the most common Russian utterances.

3. The dependency of both the average duration of utterances and their average ofs syllable duration on utterance length in syllables was studied.

4. The hypothesis on existence of two temporal registers of speech was pro proposed.

Further investigations in this area may include studies on functional, rhythmi ted cal and syntactic typology of the most frequent Russian elementary utterances.

Quantitative analysis of utterance duration should be made on the basis of the rec real number of pronounced syllables in each concrete case (e.g., the common Russian greeting 纏dravstvujte! (official 滴ello!) having in full-style pronuncia cor tion 3 syllables in real spontaneous speech is often reduced to 2-syllable utterance like 纏dras稚e! or a completely reduced one-syllable form like 泥rast!). Such Un 116 Tatiana Sherstinova y investigation allows us to prove or reject the proposed hypothesis concerning two pan different temporal speech registers. Finally, it seems that interesting results may be obtained by conducting similar research for different groups of speakers (men Com and women, young men and adults, etc.).

Acknowledgments ing The study was supported by the Ministry of Education and Science of the Rus lish sian Federation, project 8554 鉄peech Corpus of the Russian Language: Complex Analysis of Oral Speech.

The author would like to thank Michael Furman for carefully proof-read Pub ing the paper and correcting her English, all remaining errors are the author痴 responsibility.

ins References am Asinovsky, Alexander S. et al. 2009. 典he ORD Speech Corpus of Russian Everyday Com enj munication 徹ne Speaker痴 Day: Creation Principles and Annotation. In メ蛉t, Speech andDialogue.TSD-2009[LNCS/LNAI 5729]. Matouek Vaclav and Mautner Pavel (eds.), nB 250257. Berlin/Heidelberg: Springer.

Sherstinova, Tatiana Y., Stepanova, Svetlana B., and Ryko, Anastassia. 2009. 鉄istema an notirovanija v zvukovom korpuse russkogo jazyka. In MaterialyXXXVIIImedunarodnoj Joh filologieskojkonferencii,SectionォFormal地yemetodyanalizarusskojreiサ, Pavel, A. Skrelin (ed.), 6675. St. Petersburg: St. Petersburg State University.

Sherstinova, Tatiana. 2010. 轍uantitative Data Processing in the ORD Speech Corpus of Rus sian Everyday Communication. In TextandLanguage.Structures芳unctions蜂nterrela tions鵬uantitativePerspectives, Peter Grzybek, Emmerich Kelih and Jn Mautek (eds.), Advisory Editor: Eric S. Wheeler, 195206. Wien: Praesens.

ofs Sherstinova, Tatiana. 2011. 迭epliki povsednevnoj russkoj rei v slogovom izmerenii. In Materialy XXXX medunarodnoj filologieskoj konferencii. Section ォPolevaja lingvistika.

Integral地oemodelirovaniezvukovojformyestestvennyxjazykovサ, Alexander S. Asinovsky pro and Natalija V. Bogdanova (eds.), 229239. St. Petersburg: St. Petersburg State University.

Stepanova, Svetlana B. 2011. 迭ussian spontaneous speech rate (based on the speech corpus of Russian everyday interaction). In Proceedingsofthe17thInternationalCongressofPho ted neticScience, Way-Sum Li and Eric Zee (eds.), 19021905. Hong Kong: City University of Hong Kong.

rec cor Un モトハ 81-42 + 811.1/2 + 811.161.1 チチハ (リ)81.2.ミ ハホヘムメミモハヨネ゚ () ムハタニナフ () ツ マホツムナトヘナツヘホノ ミモムムハホノ ミナラネ (瑣褞鞨 籵 蒻鞣 裝竟頽) ヘ. ツ. チ聿瑙籵-チ裙瑩 ツ 瑣 碵趾瑯 褄褥碣珸 鈕瑙 籵 蒻鞣 裝竟頽 糅褌褊 褶, 韲 珸粢 裼頸瑣鞣 韋, 粽鴦粢 褶襃趾褊. ハ褪 瑣褞鞨 蓁 瑙琿韈 琿 蓖 韈 瑕頷 裝竟頽, 裝珞褊 璧裨 褶 碚韲 顆褥粽 籵鞨: 琥褌, 瑕 琥褌, 琥褌 瑕, 瑟 琥褌, 琥褌 . マ鞣 瑕 顆褥粢 瑩瑕褞頌韭 褶裘胛 瑣褞鞨 鍄糺 萵, 瑕 蒡瑣 裲 襄 韜, 裝竟頽 糺褪 璧裨 褶.

ハ裘 籵: 瑙 褶, ヌ糒粽 , 蒻鞣 裝竟頽, 瑪鈞 裼頸璋韋, 裼頸璋韶 , 籵 蒻鞣 裝竟頽.

By discursive units we mean hesitative constructions which are very common in Russian speech. This report is dedicated to the Dictionary of Discursive Units (hesitative constructions). The research is based on actual realizations of one of such units: 琥褌, 瑕 琥褌, 琥褌 瑕, 瑟 琥褌, 琥褌 and so on. Basically all of them may be translated as Let痴 say. We analyze both quantitative characteristics and functional specificity of these units in Russian speech.

Key words: Russian Spontaneous Speech, Hesitations Pauses, Discourse Units, the Dictionary of Discourse Units, Hesitation Construction, Russian Speech Corpus.

フ瑣褞鞨 瑙 褶, 珞韜 裝裹 糅褌 磅裲 粹韲瑙 竟聹頌 瑟 珸 瑜珞褊韜, 珸顆 粽銕跫 裙 頌瑙, 頌 裲韭胙瑶顆褥 褊韋. マ裝裹 裝珞褪 褊裹 籵跫 鈿璞韲, 褌, 瑜韲褞, 頌瑙韃 褪韭 齏 胙瑟瑣韭 糂裝裘 褶. ヒ裲韭胙瑶顆褥韜 蔟 砒褶鞣瑯 韭璋 糺 (裲褌), 頷 裲韭 糅褌褊胛 頸褄 鍄, 糺 鈿璞褊韜 (褌瑙褌) 瑩 , 齏頌顆褥韃 葢鞳, 粽鈿韭粢韃 裲顆褥頷 裝竟頽 糺 璋韜, 韈褊褊韃 褊 瑩瑕褞頌韭, . 蓿. フ跫, -粨蒻, 胛粽頸 褥頷 瑜珞褊 裲韭胙瑶顆褥胛 頌瑙 瑙 褶: 瑕 碼褌 (. 裲 ム籵 磊粽 珸胛粽 褶 [チ聿瑙籵, ホ瑕 2011;

ホ瑕 2012, 2012畩), 瑕 籵, 裲 褻鞨韈頏籵 (瑰 籵 褥褌, 裝頏籵 蒻鞣 裝竟頽 .

[チ聿瑙籵, ホ瑕 2011]). ネ褊 蔟萵 鈕瑙 瑣褞鞨 籵 蒻鞣 裝竟頽 (トナ) 糅褌褊 褶 褊 瑰 瑣.

ム裝 顆竟 碣瓊褊 瑣褞鞨 蒻鞣 裝竟頽 糂裝裘 褶 跫 珸籵, 瑜韲褞, 瑕, 糂 裝珞 碚 褞糒 褞裝 粽褓碣珸 裼頸璋韶 韋 (ユハ), 鈞韃 瑪鍄 裔瑙. マ裝韃 粱 韵顆, 褓褌褌 粽鴦粽 瑙 褶, 趾瑯 粨 糅褌褊胛 蒟頽頸 (褄粢 糺趾褊 胛粽頸 砌籵 ト頌鞣 裝竟頽 粽 褶 蓖糅褌褊). 蒻 韈 瑰瑙褊 碚 粢矜胛 糺琥褊 裔瑙韜 胛粽裙 褥 褶襃趾褊. チ竟粽 瑕頷 韜 琺 瑰瑙褊 璧裨 褶, 瑕趺 籵鞨鞣 頡韶琿, 顏韜 珸 砒趾瑯 褓磆蒻 頷 裲韭胙瑶顆褥胛 頌瑙.

ト頌鞣 裝竟頽 跫 褥 瑕 珸籵褌 褶裘 珞瑣韈瑟, ォ糂褶 蓖 糺珸籵韋 褊裹 葢 珸 糺鍄籵 跖鈿 珞瑣顆褥, 竟蒻瑣 裙 趺 跖, 韲 裙, 糂褶瑯 萵 瑕趺 糺珸籵 蓿肛 竟瑙篏 [ツ褞褪籵 2010: 9-10]. モ粢 トナ 鈞韲瑯 褂 趺韃 褂蔘 褞褶裘 鈔瑟 (褌褊瑟) (-, -) 鈿璞韲 裲韭 (褶裘 褌褊瑟) (.

頌. 1).

ミ頌. 1. モ粢 褶裘 珞瑣韈 趾褊 瑙胛 裲 ハ 褶裘 珞瑣韈瑟 磊 瑰頽 (粽, 瑕, 韲褊, , , .);

鈿 裝籵 (, );

粐蓖 籵 韋 (趺 磊, 珞褞, 赳).

ホ碣瓊褊韃 瑙琿韈 瑙 褶 鈔褪 頸 褞褶褊 蒻鞣 裝竟頽瑟 (裼頸璋韶 ).

メ瑕韃 褌褊, 瑕 珸琿 瑙琿韈 胛 瑣褞鞨, 粢 珸碣珸 () 瑟, () 鈿, () () 蔘 (), () 瑟 蒟, (胛粽), 碵粢 (胛粽), 碚 (), () 琥褌 () . 胚 磊 頌褌瑣韈頏籵 銜褊 韜, 糺褌 韲 瑙 褶.

フ瑣褞鞨 蓁 瑰裙 頌裝籵 琿 葢 胛 鍄: ヘ璋韶琿 (粹 蓐) (ヘハミ゚) ヌ糒粽 (矜瑙頏籵 瑙頏籵 裲裲 (ムタメ) 硴 ォホ蒻 褶裘 蒟サ ホミト) (ヌハミ゚)1. ヘ 瑣褞鞨 褪 褥 蒻鞣 裝竟頽 裝 粹裨 趾褊 糺珸籵 跫, -粨蒻, 胛粽頸 葢 韵瑾 トナ:

1) 粽韃 裼璞韲 璢褊瑟 裲 (籵 瑩珸頸瑟/褞褶裘 褌褊瑟) 2:

マ蓿硼裹 ヌ糒粽 胛 鍄 ., 瑜韲褞: [チ聿瑙籵 蓿. 2011].

ム. 砠 粢糒頷 褞竟 : [ト瑩璢瑙 2000] [チ聿瑙籵 2011].

() () 瑟:

o 鈿瑯 / (:) 瑟 / *ヘ 鞴 褥 / 鈿璞頸 琅 裔籵 / 萵 / 赳磊 頌瑣 (ホミト)3;

o 瑟 / 蒡矜ク ? (ホミト);

粽 () 粽:

o 鈿璞頸 / 矜 瑕 蒻 / 矜 蒻 竟聿 / 瑰 竟聿 蒻 // / / / 粽 瑕 粽 / 褥 褥 糅褌 / / 胚褌 褥 / 蒟裘 (ムタメ);

o 襄琿 鈿璞頸 瑟 粽 粽 瑙瑣 蒡肄 / 珞萵 瑙瑣 蒡聰 / 襃裝粢 (ムタメ);

o 粽 / (@*ヘ.. 韲 瑕趺 韆 / 鈿瑯 / 粽 瑕 粽 // 褊 (ホミト);

o 萵趺 ... 韲 竟鞴褊 / 琥褪 鈿瑯 粽 /*マ 鈿 / 瑪頸 # 裴頸褄 # 萵 / 瑪頸 頏籵 粽 蒟鴦粨 / 瑪頸 跫 蒟瑣 糀褞ク / 粽 粽 蟶 *ヘ # 褄粢 ... 趺 / 褪 / 韲瑯 褪 # / 顆ク 瑕 粢褊 (...) 砒裼 (ホミト);

2) 粽韃 鈿璞韲 璢褊瑟 裲/褶裘 褌褊瑟 (蒟褌瑙韈璋):

() 鈿:

o 胛 褶 褥 韲褪 // 鈿 / - / 鈞 胛蒡 (ムタメ);

() () 蔘 ():

o - 胛蔘 赳褊 / 蔘 胚 襄瑣 萵-鞦蕈 / 瑕 褓磊 褥 / 竟褞褥 // 褌 褊裹 瑕 粽銕跫 褥糒褪 趺 磊 褞裼 胛 萵-鞦蕈 襌ク 裝 (ムタメ);

碚 ():

o [ヒ褊, 趺 / 52 / 瑰琥頌濔 メ 蓿胛 粽蔘 / 萵?

[ネ褞紆褞, 趺 / ? / ?] ト / 碚 [チ褥裝 瑰琥頌 // ネ 瑣褞鞨 ム瑙-マ褪褞碯胝胛 鞣褞頸褪, 2006] (ヘハミ゚);

ホ 砒 胙瑶顆褥胛 裝珞褊 瑣褞鞨 ホミト .: [リ褞竟籵 蓿. 2009].

() 鈿瑯 ():

o *ツ (:) / 鈿瑯 / 褥 褊 / 跫 ...

頸瑣 / 璞竟 蒟 (ホミト);

o 顏 瑕- 糅褌褊 齏 瑕 *ム 鈿瑯 聿 糂 *マ *ム 聰 // () 瑟 瑰 糅褌褊 褊 (ホミト);

碵粢 (胛粽):

o 碵粢 胛粽 / 糺蓖 蒟 - 褊 褪 糺蓖胛 蓖 // 裼瑣 // 粽 糂ク / 胚 珸瑣 / 糺蓖 蒟 (ムタメ);

o [ニ褊, 趺, 33] ゚ 襌ク 蔘 / 褞ク鈿 / / 碵粢 胛粽 / 珞褞 / 璧 磊 琿 [ト趺 碼褊韃, 珸胛粽 蒟 // トツテモ, チ珸 萵 ォミ褶 萵裘韭篏, 2009] (ヘハミ゚);

(胙碚) 胛粽:

o 粽 鈿瑯 / c裲褪瑩 糂ク-瑕 / 磊籵褪 瑕 裲褪瑩-襁褞褊 瑟 / 鈿 瑟 / / 胙碚 胛粽 韈粨 褊 / 裲褪瑩 瑰 萵趺 蒟瑯 蒻裲瑟 / 璞ク 胛 / 瑜韲褞 頷蒻 瑟 裔 蒟瑯 / 趺 瑕 磊 / 韈粨 (ホミト);

o 糂ク 韈 聿 齏頌 / 胙碚 胛粽 / 籵 顏 // *ヘ 鈿 褌 / 璞琿 瑕- 糅褌 襌ク 顆裙 (ホミト);

(胛粽):

o / 琺 / 褊 褪... / 琺 趺 鈕 / 胛粽 / 萵籵 瑣 (ホミト);

() 琥褌 ():

o 褥 琥褌 瑟 褥 褥裘 蓿裨籵 / 聿 瑟 頌 萵 / *マ 韃鉞瑯 / 瑯 裔 璧竟 (ホミト);

o 踟竟 鞴頸 鈞 / 萵磊 粲頸 / 鴈琿 // 琥褌 瑕 磊 裙 粡蒡 鈿瑙韋 / 琿 - 粽 鴈瑙 磊 / 蟶 粲糺 璞褥 (ムタメ).

マ裝 琿 褪 裝褪 瑙琿韈 瑰裨 珮. ツ 糺 籵 胛 鍄 琥褌 竟瑯 瑕 蓖 韈 鈿璞褊韜 肭璢 珸瑣: ォ1 . . . 碯. 糅.

琥褌 鈿璞. 粐蓖. . ヘ瑜韲褞, 韲褞. ミ珸胛粽 鈞褄 瑟 璢韈, 聿 褄粢 鈿瑯, , 琥褌, 瑰 矮 褥 糺裹 瑰, 砒 萵 褌. ツ褞褥瑯. ォト 鈕珞糒褪 粢 頏!サ [ム籵 胛 鍄 1984: 101]. ハ韋 琥褌 瑕, 瑕 琥褌 . 粽銕跫 籵鞨 鈞韭頏籵 鞳蒟, 褥糒褪 襌 *マ 珸瑣/珸瑣 粐蓖 粽褪瑙 鈿璞褊韃 宅粢, 褄頽褌褞, , 粹, 籵 (瑟 趺).

テ瑟瑣韭 胛 鍄 琥褌 粐蓖 鈿 , 糺琥 裲 褊 褞褶頌褊, 珸籵 韃 磊 褊 裨, 頷 錮 裝籵褄 頷 韈趺: 粽-褞糺, 粽-糘, 瑕褻, 鈿璞頸, 頸瑕, 瑜鞣, 瑜韲褞, 瑣 珸瑣, 蓖 , 蓿胛 , 琥褌, 裝籵褄 蓿. マ顆褌 瑟 粽 璞褥粢 齏璋韋 萵 胙 糂褶瑯 琺 裝.

タ琿韈 韶頏籵 頌裝褌 韋 珸 粨萵 褶 鈔齏 蒟瑣 蒡粽 竟褞褥 珮蒟.

メ瑕, 萵 粹胛 蓐 ヘハミ゚ (頌褊 褶), 璞褥粢 肭璢-裝韭瑣 琥褌 糂褪齏 顏 葢琥蕘 (5,7 % 磅褌 琺 糺碚 35 褞糺 裲):

ツ 鈞頸 琥褌 顏, ム珸 琥褌: 鈕瑣 琅, 裔褪 , 裘銕跫!

ム褥粢 碚 硼瑩跖 裔褊韜 琥褌 璞褥粢 蒻鞣 裝竟頽 (33 , 94,3 %). ツ 萵粱褌 碚竟粢 裲 (90,9 %) トナ 頌褊 褶 糺瑯 瑕 蓖粹 (. 頌. 2).

マ 萵 胛 蓐 ヘハミ゚ (粽 胛 瑪/硴顆 褶), 琥褌 璞褥粢 肭璢-裝韭瑣 糂褪齏 瑕趺 琺 裝 3 裔褊 (5,2 % 磅褌 琺 糺碚 58 褞糺 裲), 粽 糂襄 頷 鈿璞褊韃 萵 珸琿 硴韈韲 鈿璞褊 粐蓖胛 籵 蒡韲:

ム琥褌 / 跫 糺竟 / 琥褌 / 鈞 ヘ / 琥褌 / 瑕 ツ 萵粱褌 碚竟粢 裔褊韜 (94,8 %) 萵 瑣褞鞨 韶頏褪 瑕 トナ, 瑕趺 瓊 蓖粹, 韲褊 瑣褞鞨 糺粱褊 蒡粽 碚 顆褥粽 裔褊韜 韋 琥褌 瑕 (27,3 %, . 頌. 3).

ツ 褞韲褊琿 褶 (萵 ムタメ, 糺碚 韈 60 裲) 琥褌 肭璢-裝韭瑣 糂褪齏 珸, 顆褥粽 籵鞨 萵 トナ 襌 粽銜, 頌 ( 碼褌 褓硴琅瑙韋 蓖粹 琥褌) 琥褌 瑕 (30,0 %) 瑕 琥褌 (13,3 %) (. 頌. 4).



Pages:     | 1 |   ...   | 30 | 31 || 33 |
 





 
ゥ 2013 www.libed.ru - ォチ褥瑣 礪硴韶裲 瑪-瑕顆褥頷 褞褊韜サ

フ瑣褞鞨 胛 琺 珸襌褊 蓁 鈿瑕褊, 糂 珞 竟琅褂瑣 頷 珞瑟.
ナ ツ 肭瑰 褌, ツ璧 瑣褞鞨 珸襌ク 琺, 赳鴦, 瑜顏頸 瑟, 褶褊韋 1-2 珮頷 蓖裨 萵韲 裙.