“3! W85 WWW” ”WWW-170325.” 223:. .‘j éfi .512 21': ..§.1 fi‘. ‘ J I ' o 4. "v a... -‘v— a. Siétz‘mlwliaj ”6 2%? 22} 04.14323.“ .; Rat. M “A - ‘ A «‘“H.3ou Manda-WM - “-- .-'T:V- en‘s- 3..-: slow-.- mur- W‘f‘ “—... .w' . ‘ 'Y ' 211,774.33. f" : . (egv q‘ch C8! .“ .._~..,..s»c- ‘ V ~ __ . .. - . . .‘ Ii" . o. —. . .- l 1.": r. ~“.‘,.‘I .. A. -O- “”2. q I". ’ . ”J”. ' - f7 1“ "1’. 1 ’1': ‘ .J‘" ‘ I. -’ 12; d “'m ‘L .1. _ f' .\ C. 1. ‘z.'-'V'J .. . “‘ fl -I , -..--.o-§.6Ifi -“-‘«~".“-"4":?1 1.13.2 22-521. "'2? "rut? ” 4223* > ' 33%;” ’2‘" ‘- “£32522 fif'2"'21.1"%;{33 bug RAW 3&2 a '233‘2‘~‘:}..§¢s°3 “"9392 .zit-‘tnzsw-‘d-lzir.'f1- “Iv-414.. .I.-I.I.I. _2I_~ L" I". ;?f‘$.]:;’- “‘59:; I. 41161-07"; ¢»—' 3 {.i’ - “51.5.“, 4;.- 1,, I 30")- V,. r "n-1,, - ‘wfilf w 1, ")f’y - .‘A In“ -"2;‘.....‘-- M ii.‘ "Q" A 6 4‘ .‘ . .. '." 1 _ 2 " ‘- . 3' '- 5 f” "”11, 11:2 ‘1;.';2'é.3" 3'12? ' 42;;‘3431‘:1.I--'._I ' '1'»'.22_‘, 23"} " , “:11:- 94‘5".» 3.: .‘ “23:. "VAX 2. 222:2 -—~22~2- 211 , _ . I o J . . - _ -1. .. .. ‘A h I . ’ . _ ‘. .1 1- - . .I‘ ‘ 4. . - M0 rut? 2”— “ “-('".““.I 'If “g? L' V I ip‘.‘ . km" (-2.: § 03539: I": J” ‘1'; h'vk‘.‘ )“° " ‘3 . " 1. 21‘? ‘_' "fl ‘9‘... I. ‘f .A'. .2 #1: $- \'r\‘L-a' if I Yd .J‘ ~_-.:~:.\": - ’ 2 17"” -"‘f=“03 i." 2‘1“" V 1'1. ‘2 7154;: . -y‘ ""5 31'5” ”Y‘s“ 3.31:1; ' -‘2- ,z'ni-fi“"‘ii-.“f‘ «VJ; a. . .- . ‘. 2 J o ‘3’“ ' ' . . ..€;’,‘f‘)& ,lc'r' 7},;‘;‘-‘L- 1-. .. ,‘_.’.q‘1-"f",fl" 31:: L3 , ‘ :14? Ail} 3‘ $3. r. .I.‘ ".4, i. \.P:'£. Jt .. .’~-. 2' 2- .11 .L 11%.. Bum -» 2 2 22-222- .. 2., 1.2 a,” r 1: ”'1’ \.. ‘ ,M,., g..-...- .e‘ -...1. I a 22- 20W s . . - - -!'1‘_‘ (xxx: ' -h\20 "" ““"‘"' '?"“" ' ' . " ‘~ . ‘1'. . (c o ~'I-." dd :4‘ J g ",:\.| - .u ”)4; :fl:_ (‘3 I-, o .. t... I J5‘3I’ fir_ h‘:'£2; 2 b," . W14 3‘ .' if. n ‘n- .L41’ .“ V 4 ‘ ‘0 ' ' 'p‘ ..hz “ "."'. " :4 v 4;. 2~ (3 ’21:" .unfl.’ _ “fun ’ :M‘ In; . a", :!;<‘ d, { ‘11.-I’.\ 4245-“ - fl“. .’:;‘1?LR¢3?‘::; '.‘F ‘ ‘q. 3‘ 3“...” '|.v.:2"'_‘ . . . . . I '. ‘ 2h ..'T' a b ‘ ‘ ’h‘;'q 2.232; "- H ‘, 'n w 9" 2" “ .‘ I ‘~". ‘ 5- 1 "' -,s.{7IV-.d‘..fi'\1 '1' t 4‘ fizsku‘."5 ":.':_ ‘£.§i\ ‘ tfl:g . ,‘v-uéoiva- . “3"; “§...‘- ‘ '5‘! m p. 2n... ‘ 1 A 2. CL? ’.1-;.:"J. 3'. . ‘. I . " 2 o .»’ 2’; {3:03 “"1 .5. ' 1-..”:fl t ' ‘ “1"" £"“-"""'I"‘ ' “" I; U‘y‘ n... ‘1' ' 3g} _ 9'1“ I \ :4“ . . fizrt «.us‘ ":1“ V-‘ S 2; 2'5; ”2 x 15:13.?" 1"“’::.‘2.. r. 222' 2.21.1.2: I‘SL ‘, g.._ CL" - ‘ fl. g“ O'L,’R.Ju~; 21:: via?» -ifz'z‘?‘ , .M . . “I ' ’ ”"’.: ‘1‘ ‘.J I!" d. ’5 ‘ A" '6 0' : W‘IL" .. £1.52; ’22 2’2 '12 g5. ‘En‘x’é‘a 22 22 33:22: 23.33,; “-3 1:!“ , ~l "2’: '1‘;\5‘: p.159... 1". ."_..... In ‘- b‘nq-‘A 2.: 1.22.132. 2.21;: - Ham 2 RM .1 .1. .2... '9. ’ Q“ o: 4, 5‘ . ‘ 0 A ‘WI. . ’- I ~‘2 . . 2. «mp-134' « . of; Oy‘i‘f a" aj‘hfiyéfl ‘2'“ U . ’ . 2 Ra‘s- k, . '0' _;;.:.I¢9.:.'. _. .‘n.~. . .~ *nt'..y§2 L:‘.""~"¥.£ ,‘1‘: 1:. - u .0 ICBM ”7:5... '- ‘ I 3:2LPhU‘-d o.- l3.\aq.. ~~ m 1‘4‘1"~- :i'.’ -v :I 2'. :%*f ~2~~ w- 22'» '22 "-2 .12 .2 2.22.232;.m&.- 2.30123. .. m .31“": 'tuh“;f~l . ' u ‘. A. $35,; fi?I-.;‘ "b (26‘. :u‘Iofi.‘§.§-8I" 0‘13: a" ' “'01“... .4; 1“: ‘. .' .':..:‘l:- $.rh’v - .. .5?) ‘ 1:1;‘4 :72“; £133! r..- ".1. «J. .41- wi. .1 : . ' ‘ r‘ 37:? .‘ "’ Lt: tug“ H1 \~ I" ‘ ' “ ‘ if“; .95.”... 3:2 “I".“I‘L. J. 221,.” u . 9 v ." . u 3 ‘ '1 5d" 6! 'u . b. C 4 . ’-. ‘f “4 , .. .. .1<’.2....9't._ -.' V;;:;w 217:4“.14J-1Q. .Ic. "I . M‘ ”My“. - .‘N_‘,z 0. w.- 2 1 .11... I... < ‘ 2; “V22 ‘ .4? .. I ’ 222m ‘1 2 .. ;. 3:132 m. «1:1,...‘2513 1. .2 II a . 2'. $24312 £35?“ 2’222$1’1;‘?8 ; .' -. .2 2. .' 51cm -1\ ‘2'” ‘ ‘ .9 . ‘.. ‘1‘ ¥:~‘p ’1'“! 317‘» {N .- o. . I. I y‘: :Ig’ L‘: ‘5‘. f.¢ .1...“ . 2 2‘22 2.11»-‘:-2:,,~'a.. : arc-22.222.112.22 2212" {€22 - . 22.12.22 ~. . .'~ 0 -¢ . . . . ~ . . 5‘35?" ‘ .r.I1-.:.'L.-.. 54.62312-“6'JI w H1“:j{.,_°;";_: , ‘3 ."v‘ “.3, " -‘ r ”2 213?; £23. 2 229.222 2.: 12.1.. «.21.; .. .. . n1.- 1“ ,. u... -.&‘ "‘6" 44‘“ ..’..ua. K‘ " .1." V a" ‘t ' 4.2 ’2 ":44“ . June ‘11-. In". . 'u' .V you...“ 1. ‘Q8 ' .“E'ufiowfi‘ “ ' 1" ‘ ‘ z W'P‘""" ““ u G» -’\\h~2du "‘3“. '2'“: “:f‘r" ’ '- ‘1‘ "'2" "' ‘ -, .ni‘r. .x.-‘ .mdv- “f2! ‘4‘ .1 N 2 Q r r‘ 'fi".'< , ‘ub‘ ~( .. 8 .u' _ 1:02.. - V;:: r. .:?§ g. 3.. 2 q', ;:.'~’| ~ I . 5. . I...“ v...};~.':i.‘;£‘ .. EV” _ 514*“... -, 1‘3 I. ‘. .. :g x .15 - U;:ngtq.2 L!" {2;11" - - a. 3:551, '. :‘1. d‘du’w ~ .- . *1 ”I I _ L ‘ : “*a’.‘;‘ 01“ §Q~R F‘L ' {2 i' ' ' ’ {F . 2w‘i‘ku‘a "saw-u . 3, .1.” 1:33.; :«vzo-O'M “‘2“. "(2.: he vxr a. E25}:I:'I» 1.3.3.4314 ." P ,1. . .u. _§,—m..uc1u:.¢ 4-2-19. 2.33”qu fl‘fi’ifie Isafiirb’x? W, jazz. Ak‘ttficc Kg: £5. {11.1%. 13‘. . ' ' 13:; ‘ .3. 27.:- ‘4 "5 up; .Jx'b 34129:”? Mum- Vt‘.‘ 2” . . .- » - 2 ‘ 2 ‘ - 2 ' - J“ (I 3... ‘f‘ 7 ' J .3 tntud 8‘ “Had 'h 2': . I; 1"" \ ‘ ‘3? .1 ‘ f': ”3"” 3': '51.." I: " "‘ K4 r”. l |.. .N' -O- n- .41.“) .‘ . . - . ." «uh; .MflQ I‘m. . ..' JhL 3:9! §".' 1‘ . gift: . 'f - "2’ 1”)“ x~"""‘2" "' " " 7-"1" ”W3; "1:; 24. .122-‘5‘7; «'3'. ”21'3“"? 4.21;: _.. “f." {.14 -.. .24 ”,3 o. "3 Ifi 5:93le ”,2 ..:‘?..>,..."m _.r;?."u~‘ 3:143 31:: 1‘" ,._ ' Pear“:'f'-:’r:tf‘.‘nI ‘ 5.’J§;22{:;2‘3“ '7 3;? '2‘" l 9:“ "Ev-Mk" : {19:2 2‘: 1ft“: - :12: L22“; 2;“, v -- ' 5‘ 6Q “.5 7;.“ 4' . “V154 .2 :.;'223'u;;r43_:u21.€.:u.12 ; L, :2 1. .Q. 2. "1.01:. :V. I—y‘fk“? 3,;.,J.,:2 : 1' . IV 1;“ “EL-I'vjx' “I - ’ “3' 22'~""- 2 :"fi k'” r‘.'3‘~“'2{:‘u'_u'1 H22" 1 Prv " . . ' ‘ ll. ~f‘ ( ‘3 0“. 'l n. 9 ‘ ‘ :. :Y?’ {‘ng J;;:14flug’”zflz* _ :1 w . . 222.1312: {gammy—22» "5' . 2; £3 2 . ~ -', VU'T‘JS‘ I“; ‘.,"§f v [14“ 3'.ij .‘26 ' ‘ r .“2 I? 1' ’2'." : h" :4". IJISOI “ 0- .3 .‘ ' I t . u ""‘a’ ' 22 2 N2 2 - M X'fzisa'1uw- we ‘9’»! has»: 0 , I _ Y 3:: . «a 9 ‘3‘ .. 5“"? .2.“ - - ‘12:; “"3333! “’2‘ ~'.-’-l:" J! 3 ’ ' 'J.é-:.:. ‘ ’ :3 -. ‘o 2|. IRE-3‘, ‘3».y ; - L ‘ . - ‘ I . 2' ‘ ' .ct I c» .1;3.£v ‘1': :I I 4 ‘ 3.. ~ ~‘ 12;- '1 '1‘,“ ‘ “‘ .f" “1";S‘: . :L4‘1 1‘ . :3 . :3”. . a. v ‘ -. ‘ g. _ ' 1 ‘ -.. . u . ' o‘ ._ ’. Hammond 2;- , {’rt‘ ". ," .14.,32‘?‘ . 12.. p.22: '8. ‘Jq! safi’h- . 3125‘ - . .3123 21. “J. .2 s.» - v 3.. a} .1 . . . .... . .. . . r . .M . . 2 ‘ g - 2 .» .-. r. - 1“ U 5-0. ‘ - .4 1' II" '1 ' . .,2. a I fl ; . if _'.‘ . B I. 4 's'? ‘ . ‘ . . -» . - . .- 2 ' . " ‘ ‘ E":.‘:‘2’..; 9. 24... .m 4~ 2. . .'.I‘ 9 j. 0. 2 -' - "'- d I 2" VI ‘ . _ ' r . . ..' . . . . V ,' ' ~ _ 21‘- . , , f‘ I ‘ n' .53: 9‘ J ‘W "I r. ' H. 4.0.: cc; ‘ ‘ - ~ . . : a o :6: 02' f k. 1"! e :72 '4 4'." . . «._ a.“ yum-1'21 . I o\'J2I.9.." fl» J‘u‘ - .22 t" I . l r 2 ‘l ..' ' _ . .r‘. .. f . - . + - a 20 I7"” \I-;- ' 5'2 , 3} = '. -’u I "‘5 ." < - J I ‘ .1.' z-,. "I" “I- 4 .-.'- ~ a, '< JJ’U' [[6. o.‘. ..;“ - r_.2':.: 1-4 . .., ,3” 2;. 12311213534 .51? 91.3.51 ; 2 . ”£722 .2 : .2- -;,. -.. o I 3 l . J , .. - ..L .2. I6 1 _ , _ , _ - . . f., . .9“. I .’ "1\ ‘ {if 3;. . f.,":‘ . .... J.,_ . .1 .u If“, . 7!: III.2<1'- '.’ .5" ' . ' . 2;} .2 ‘ )1 f'. a . :' 375:1,” s: l‘ I 1:. ,.. .. {¢.'_P"I':IQ'. v at. ., .,- I'l‘l.l"§0.,“".."'.‘—. 'I°)\'.. “.13; ,1,.‘. 'er’! Inf-‘- 11.; '2' L - ‘uofo‘rw' ‘$:- " new 1:22; .If ‘ '7‘ .15 L" "" ‘.‘J;.-‘§:Ioi.a“"¢ 'f '3" 'P} ”9:0 ('1 9"4 1' ' ' ' ‘°' .2 :I‘d.’ ‘2 _ ._.-¢ J m- .‘V ’ '2- .|,. 913; ‘y, -. .29; . .Mrw “Wm. -. . a 1. _ ‘v -:A,‘ _--.If,f¢¢' . ‘ ,1 -‘ “1.7.1." :I¢I.- 2 I u rJ (In 1 t . _ ' 1 2 . . v ‘g - I A . . .. 2.1 1 - " - _ ' ‘2 551+] JFK-9. ‘f! v.r'"".'1 ’3 2’"! “D .II .‘c 2‘! ‘ 3": .- . .1 I "fi't. o .- _ .- .. \- ~ .2 '~l.“~"fd--; I‘:*.I"" '-.’1J!‘ . .' ‘ ‘1 3' ”5".I1 '2 . 3:3-1 ‘2’J' -.".I 1 ‘4‘ .‘l .I'-’. v"- '2 ' . . . . . '.. .. ‘.' .. ,1 2. . . . . ' Irish”,- .' {XII‘Di'HJ ;:4‘ 0‘: a .3I5/‘k'1": ‘15:. if ‘ a) 1|: (”I { ‘5“-"f‘~:”.1 ' '. - '3’ E" ' . g ”s E' ‘I N I‘ '7- 1' .2. 'u‘t‘lr ’5 I. '52.. 2 '7'.” 2.:. ’2 ."Fny-‘Q: .2. rial-‘5‘; 1“. ,:A-;‘."., 12.1. 3-};- '61-. 2‘ ‘ xI- -' ’ ' '{' '--2" ('2' 1.20212: .2 w ‘M \2 7v" "' . 2. ‘39:}? .T;’-,o1-,-4.1-’-(2¢ ‘ ‘2 1w???" ‘ "“T1‘ri?’. - 1‘25.“ ’5‘ ‘. '..- L. .I- v... .. . ’0. .--|_- 4‘. . ' ‘. -1 '2. . .‘-:hd"l ‘ _ "Lyg’u :o'procu‘nr: t: .2. 7:24\~—' 1.“:-. _ .. 4-1‘32'0' . W 1;.” .v'.‘ 3.. v 2:! ”. z‘r’ . :I’o‘ u" I“ 1 ° ‘ 1" ‘ 1"!" ~"' ” ’ . “l ' '4’- . ' ' o ‘2! {iv} '2 ‘:; ( 2.1.4. 3“" A). “(314’ --' 7') ' 3‘ A" ‘- 3' '. ". . ,.' -‘ ., . q- 2:- tJt‘ ”r '92‘ .22 y'. -"1 . .‘ 2.. . -'.' -. Tut —" - '. ‘ ‘f: ".3215... ' t". "' ' I"? ' :1 9 ' ‘3' I ’ .:r:p :g‘?-.. ”12“”. ."."[.I9',;’ '- "9-3 fat-‘5'; 1:- :PIn -:D-.‘::.-J.:}tollmltl WW 0" '3“- “-1: . 2 . , .- a 2 . I . .' ' ‘ ' ' ‘ ‘4'. TV". F’ '1' ‘ J ' 2"“-3: 7”}. A:‘:-' on': o 7" ’ ’r " 2v 22222;,2-‘11: .1: ,., '1: ”35,141 .. .. . . ,... - »: 3 “." ‘ . q-vo ‘2 'f' ’.-~"n-' w 1.: ‘7' .rf ~orq'r-‘I 2‘ Jr'fi.‘ ' ' 92 ‘U " ‘ rug-1‘: s .‘vov oi.- -:IO:~V’~2I 3"?“ 3"‘11 _‘. 1!." r' ’ . .T." _. ‘O"..“. -2 '1‘2 2 2 22‘22' ' 2 farm 2 "‘ -" ' 2; 15.1.? .' ‘::.."."..‘="'J.,rx. ', " f. J1 -.": . ' -;' . . . 2 .- ._ 2' ‘ s .. . _ . , . .1,” .' . - - -- ¢ . . A, .55.... ‘1 2, v .11 I : ' ' 1‘." 32'1”! ;'.". flint"; ‘1'” £3114..- . 1.! .'.' ' (-‘r‘a .2'v-~-.I ~ %' Pk. '.- T'fI’: . '1 2"”: r!"'-:"". ‘:'::’AV;" " . . .‘ov 2. 1‘ I ‘I '.'1 r _ . . _ ”h, ‘31,: 112,0 -5 , "'{i 'r , 3. ‘f‘ .. "if 0. ' x1- 4" v ’ nT'” ‘ "“13,‘ ., «2n ,1. '1 1 222""‘-"1L‘ I ‘. - Iy.» '- u-v;. n”2o—o -.- 2 '1 .' . .. .. 0:! 4 5 -uf; ‘3‘; v-I. ”311'; 1. 3:4 .. - v : .v-. — t 04'.’: "'z‘W~-~‘. "frI: ‘9"- ‘.' ~.. - . 7 .'I 1&if:-.2.1-:, ‘l' *9 ! 1"!" f" " " ')" "'I. ”m :4 ":;o‘ .ng-‘(J‘rx 1r .2" """§'-' ;":l“ 4‘ I"{";“?;‘.':..: 'il‘a. .r.’}’.‘:' r“ a'—-.r",. ';.’”O I- ' ”£3151é: 2;§"I,’ 'Y.“ 1%“ 11:} . .- a I 2:! “1'12. wt; '1" . 7:; 5.00:1;z 7:;112...‘ ’20 'I: 4?: ‘ . ‘ ' ' v ' a . . {fiaq- Iva J20" 111231;: .r‘v 7:110 10:4 ‘222 - p. "V1" 3.1 .' o o my 0. "0' " . . pr.\."a-‘.. oro,.o:o«c ::—(ooo 1:2‘ 'I ". . '. a . . a . ‘I 2 - " . 90 " .‘ ‘42. r 1 "fv’f‘u ' ‘ - . ' ’ 4 00's 0'" .2, '{.O o',c..'Ir'..pf'I7-.‘| .. r . " A I . .| ' .. _‘ ,.'... K ., 'V.‘ . '. ...' .. ..-‘-I . - ~ - 01-. 'I or; . _ - - .. ’ n.1,.o39’ -j.u."'q . c....ov 2.. 1 - u, c - . v n. 1 ' l O I, V f' . 1 It‘lyol I . ..- f r ‘ I - (J _I." n '1‘...’ ‘ ', . .".o_ l- . 2 . a. Q! (91"“' _ Pry!“ ”'l' f” . . A I“ .1 . . no I ,_" . . x . 1 J . l,‘ ' 0; ‘Ir . I - .4 0 $ ' .1. . v. "’. . "Ofinno‘o-nt «or 5.0-- 0": '. 1' - v I 1 ' ‘ -'.'I' "' f' " :='-€"..(..~o:: L‘f"‘ '.0 1""'°"' "°‘ " 9' - .r I a\2"‘. . .n .u 'ALI .- "l '1 " ' ‘l "_ '41! -n.-‘n‘...-..‘ .I'O. '.‘,,,lo’,' ”tan-’2‘ ft a I"_ J M l! {ills lilllllllllllllllllll h 93 U0876 4890 it! This is to certify that the thesis entitled LANGUAGE VARIATION AND CHANGE IN MANDARIN: THE MAINLAND CHINESE SPEECH COMMUNITY IN THE UNITED STATES presented by LI LING has been accepted towards fulfillment of the requirements for Masters Linguistics degree in //7/2 7 4" ‘2; 67 . Major professor A -« Date/491K 3‘) y, /771 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution 44...- A HA ,.~.—_A‘a-.—-.~.r>“‘—A..—-’—< __A_ ,..-J\ ’. vkiir‘ - 4‘ a l‘ (1—— LIBRARY Mlchigan State Untverslty PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before due due. ll DATE DUE DATE DUE DATE DUE «3:35 29‘ II J "t L '11 1 - 0 qu‘v ‘1 "1} _ I! MSU Is An Arnnnntwe Action/Equal Opponunity Institution J cMMpma-m LANGUAGE VARIATION AND CHANGE IN MANDAFIIN: THE MAINLAND CHINESE SPEECH COMMUNITY IN THE UNITED STATES Li Ling A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Linguistics and Languages 1991 ABSTRACT LANGUAGE VARIATION AND CHANGE IN MANDARIN: THE MAINLAND CHINESE SPEECH COMMUNITY IN THE UNITED STATES By Li Ling This thesis aims to describe linguistic variation in correlation with social factors in the mainland Chinese speech community from a sociolinguistic perspective. The linguistic variety under examination is Mandarin spoken by mainland Chinese speakers. This thesis is organized into three parts. In part I, a subjective evaluation is carried out to find out how the mainland Chinese speakers unconsciously show their social attitudes towards their own speech and the speech used by other Mandarin speakers such as Taiwanese. In part II, the lexical variation of airen (wife/husband) is described. In part III, the syntactic variation of the aspect marker guo is studied. The data is collected by random sampling of the mainland Chinese speech community through interviews. The total informants are 128 in number. A quantitative analysis of the data is used. The findings show that the motivation for the synchronic variation is socially controlled. Acknowledgment The author wishes to acknowledge her indebtedness to Dennis Preston, Professor of English Department at Eastern Michigan University, for his help in research methods and statistical design of this study. TABLE OF CONTENTS Introduction Background Information Hypotheses 3.1 Hypotheses 1 3.2 Hypotheses 2 3.3 Hypotheses 3 Subjective Evaluation Study 4.1 Introduction 4.2 Method 4.3 Analysis and Results 4.4 Discussion Lexical Variation 5.1 Introduction 5.2 Background Information 5.3 Data Collection 5.4 Data Analysis 5.5 Results 5.5.1 Sex 5.5.2 Duration of Stay in USA. 5.5.3 Desire of Stay or Not Stay in USA. 5.5.4 Age 5.5.5 Attitudes Towards M and Duration of Stay in USA. 5.6 Discussion Syntactic Variation 11 11 12 12 14 14 14 17 33 36 36 37 39 41 43 43 45 46 47 49 51 57 6.1 Introduction 6.2 Background Information 6.3 Data Collection 6.4 Results 6.4.1 Age 6.4.2 Sex 6.4.3 Duration of Stay in USA. 6.4.4 Place of Origin 6.5 Discussion 7. Conclusion Appendix A. Voice Samples in the Subjective Evaluation Study Translation of Appendix A Appendix B. Rating Scales in the Subjective Evaluation Study Translation of Appendix B Appendix C. Questionnaires for the Lexical Variation Study Appendix D. Questionnaires for the Syntactic Variation Study List of References 57 57 62 63 63 64 65 65 66 70 71 72 73 74 75 76 77 1 Introduction Studies on. language variation and change have become well known through the sociolinguistic surveys of phonological and syntactic variability in urban centers. The systematic analysis of spoken forms reflecting such variations in statistically significant samples of different speech communities shows that social variables such as class, sex, age, ethnic origin, education, etc. directly affect language usage, and studies in different cities in the world have found that language variation is woven into the social fabric of the speech community. The fundamental aim of sociolinguistic research is the study of the covariation of social and linguistic variables. This type of study requires the determination of relevant factors in the social make-up of a speech community on the one hand, and the identification of socially significant linguistic variables on the other. The two are the necessary parts of the investigation. The serious investigation of linguistic change and variation in correlation with socially defined factors was begun by William Labov (1966, 1972). Labov selected the speakers of New York City as a speech community and studied phonological variables such as /r/ in words like car, floor, and fourth through a quantitative analysis. The frequency with which /r/ occurred varied with the speaker's social class, age, and speech styles. Trudgill (1974) did a similar study on /r/ in Norwich, England and reached the same conclusion that linguistic change and variation correlated closely with social factors. Empirical studies which also describe the correlation of language variation with extralinguistic factors have been done in different places and different linguistic systems. Wolfram (1969) did a study on phonological and syntactic variation in Detroit black English and found that social factors affected the variation pattern. The results show that final consonant cluster simplification such as the deletion of -t in the word “west” is determined by social classes. The syntactic feature such as the present tense of the third person singular -s in verbs is also controlled by social classes. Fasold (1972) did a similar study in black English in Washington, DC. Fasold concluded that some linguistic features correlated strongly with social factors. In Canada, Malcah Yaeger-Dror (1989) studied some phonological variables such as [ 8 z] and [ .9 z] in the Montreal French vernacular and found that the variation between the phonological variables were sensitive to age, sex, and social classes. Sankoff, Thibault and Berube's (1978) study on lexical and semantic variation in Montreal French revealed that lexical variation was to some extent affected by socioeconomic class. Susan Gal (1978) did a study on socially motivated linguistic variation and change between German and Hungarian in the Oberwart speech community. Gumperz (1964) studied variation between Hindi and the local dialect in correlation with social relationships in the Khalapur speech community in India, and the variation between Ranamal and Bokmal in the Hemnesberget speech community in Northern Norway. All these studies investigated relations between linguistic. behavior, dialectal variation or style-shifting, and hierarchically ranked social categories such as class, age, sex, social networks either in one linguistic system (Labov 1966, Trudgill 1974), between two languages (Gal 1978), or between two dialects (Blom & Gumperz 1986). Labov (1972) emphasized that socially defined factors played an important role in linguistic behavior: A review of the asymmetrical history and evolution of most linguistic changes, and of their remarkably uniform evaluation by society, shows that an asocial account would be incoherent. The histories that l have outlined would not exist if social differences were abstracted from the grammars involved, for the accounts of change would then be vacuous (319). Following Labov, most sociolinguistic studies, taking speech communities as the essential settings for understanding variation and change in language, are concerned with the theoretical questions of how to explain the linguistic variation which society manifests and with the methodological question of how to explore it. Linguists also find that variation within languages are on all levels: phonological, lexical and syntactical. Such variation can be studied along three synchronic dimensions: geographical, social and stylistic. In the past thirty years, many studies have been done on linguistic variation and change in the social dimension. A majority of these studies are confined to languages such as English (Labov 1966, 1972; Trudgill 1974), Montreal French (Sankoff et al 1978; Yaeger-Dror 1989), some European languages (Gal 1978: Mier 1987), Indian languages (Gumperz 1964 ), and African languages (Russell 1982). So far no published studies have described Mandarin spoken by the mainland Chinese speech community. This study aims to make a contribution to the literature on linguistic variation and change in Mandarin. Most previous studies have dealt with phonological variation, and there are many practical advantages in studies of phonological variation in terms of data collection. Many previous studies suggest improved methods for collecting data of phonological variables. In addition, it is easier and faster to obtain enough occurrences of phonological variables from the spontaneous speech in interviews. What the researcher should do is to create a relaxed environment, turn on the tape-recorder and get the informant to talk. However, lexical variation and syntactic variation is more controlled by the topic and context of talking. A carefully designed method is required in order to get enough occurrences of lexical and syntactic variables. If we want to further understand our language, we should extend the analysis of variation beyond the phonological level. The present study selected speakers from mainland China as a speech community and described the lexical variation of airen (wife/husband) and syntactic variation of the aspect marker guo from a sociolinguistic perspective. This thesis is organized into three major parts. In part one, a subjective evaluation study is carried out to find out how the mainland speakers unconsciously show their social attitudes towards their own speech and the speech with which they are in contact since such an attitude study may contribute a lot to understanding the mechanisms of linguistic change and variation in the mainland Chinese speech community. Part two outlines the lexical variation in correlation with social factors. The finding of the research is presented and the implication of the findings is discussed. Part three does the same for syntactic variation as part hue does. 2 Background Information The mainland Chinese people started to come to the United States in the late 1970s after Deng Xiao-ping's new open-door policy. The majority of these people were sent by the Chinese government to study and do research at various universities in the United States. Very few peeple came as private students at that time. In the last five years, more and more people have come as private students. They have enrolled in the universities either as visiting scholars or students. Each year the enrollment number of mainland Chinese students has increased. With Michigan State University as an example, the largest foreign student group is from Taiwan. The third largest foreign student group is from mainland China. The following table shows the enrolled students of both groups in the fall term since 1980:1 IaDELlL. . . I0 “3| a0. 0 03I 0“ 0". cl. Ile-IO II- lzuu: .Iauuu1 hhflnkuuLJHflna 1980 125 4 1981 133 11 1982 139 25 1983 170 65 1984 184 87 1985 191 108 1986 233 162 1987 272 212 1988 327 211 1989 361 235 1990 382 274 1This information was provided by the Office for lntemational Students and Scholars at Michigan State University. ‘ The number of enrolled students from both Taiwan and mainland China has been increasing rapidly since 1980, especially the number of mainland Chinese students. With the increasing number of both groups, the contact between these groups has become more extensive. The linguistic diversity on the phonological, lexical and syntactical levels is wide in the Taiwan speech and mainland China speech. This diversity is due to the long separation of these two groups. For about 30 years, communication was completely blocked between Taiwan and mainland China. Changes on the phonological, lexical and syntactical levels in Taiwan took place independent of those in mainland China. The diversity is also shown in the naming of their language. People from Taiwan and mainland China all speak the same language called Mandarin, but each group has its own term for this language. The Taiwanese people call it guoyu (national language), but the mainland Chinese people call it putonghua (common speech). Although guoyu and putonghua are under the same label - Mandarin, the differences at different linguistic levels are so salient that it would be more apprOpriate to consider guoyu and putonghua as being two dialects. The speakers of these two dialects belong to two different speech communities as defined by Labov (1989): The speech community has been defined as an aggregate of speakers who share a set of norms for the interpretation of language, as reflected in their treatment of linguistic variables: patterns of social stratification, style shifting, and subjective evalua- flons(2) In general, guoyu and putonghua are used to differentiate other regional dialects such as the Min dialect, the Cantonese dialect, and others, but guoyu and putonghua are all referred to as Mandarin out of Taiwan and mainland China. Mandarin is also a lingua franca among both Taiwanese and mainland Chinese. This is due to the unique language situation in China as a whole (including Taiwan). 90% of the Chinese population are the Han people speaking the Han language (Hanyu) which is referred to as the Chinese language in the broad sense. Within this language, there are two major sub- categories. One is Mandarin, the other includes eight regional dialects (Wu & Gordon 1987): the North Chinese dialect, the Wu dialect, the Xiang dialect, the Gan dialect, the Hakka dialect, the Northern Min dialect, the Southern Min dialect, and the Cantonese dialect. Most of these dialects are not mutually intelligible in speaking forms. Therefore, Mandarin is encouraged as a lingua franca for communication among different dialect speakers. For example, students come to the United States from different parts of China with different native dialects. Those who speak the Wu dialect do not understand those who speak the Cantonese dialect. They all use Mandarin as a lingua franca for communication. Mandarin becomes a lingua franca in the sense that it is a variety which is commonly used by peeple whose native varieties are different (Samarin 1972). Regardless of pronunciation, in Chinese dialects, there is little difference in syntax, but great difference in lexical items, especially in high-frequency general lexical items (Chao 1970). Considering guoyu and putonghua as two dialects, the distinct features of the two dialects show a parallel with those in British English and American English which are also regarded as two dialects (Trudgill 1986). As for the differences between British English and American English, there are few morphological differences, a number of important syntactic and phonological differences, but a very considerable number of lexical differences (Trudgill & Hannah 1982). The diversity between guoyu and putonghua is similar to that between British and American English. When the speakers of guoyu and putonghua come into contact, the distinct features from both dialects offer alternative ways of saying the same thing, that is, for each utterance in guoyu, there is a corresponding utterance in putonghua which provides the same referential information and both utterances are synonymous (Weinreich, Labov & Herzog 1975). Here are a few examples: guoyu putonghua English katongpian donghuapian cartoon diannao jisuanji computer From the perspective of current sociolinguistic theory, the study of languages and dialects in contact as against their study in isolation is important for our understanding of the mechanics of synchronic variation and motivation for synchronic variation (Rickford 1987). What is under examination in this study focuses on the contact-induced variation which has the merit of having a referent with high frequency in everyday speech, serves as salient 1O linguistic features, and carries social significance in the sense that the variation is constrained in terms of social factors such as sex, duration of stay in the USA, age, attitudes towards guoyu (hereafter T) and putonghua (hereafter M) and the education level of the speakers. 11 3 Hypotheses 3.1 Hypothesis 1 (subjective evaluation study) In the study of language change and variation, a generalization has been made by previous studies: dialect differences can be systematically associated with regional differences, social class differences, age group differences, stereotype of different dialects, and other social variables (Bouchard 1969; d'Anglejan & Tucker 1983; Lambert et al 1965; Williams 1983). Differences exist in the speaking form of M and T such as timing, the nature of intonation contours, accents, and others. For example, the speakers of M generally speak faster than the speakers of T. M speakers and T speakers use different modal particles such as ne, Ia, we, et cetera. The speakers of T often use we and ya at the end of sentences, but the speakers of M do not often use these two modal particles. The hypothesis is that these distinct features are sufficient to signal a distinction between M and T and serve as cues for the listeners to judge M speakers and T speakers by associating stereotypic semantic differential adjectives with M and T speakers, respectively. By using the word ”stereotypic", it means that the mainland Chinese speakers have different images of M and T speakers. Based on my own informal observations, those who speak M are thought of as being reliable, natural, and intelligent, but those who speak T are thought of as being soft, wealthy, or elegant. It is likely that the listeners evaluate the voices of T speakers and M speakers differently along these evaluative adjectives. In addition, the listener’s rating on these evaluative adjectives may be affected 12 by the social factors such as: age, sex, and duration of stay in the USA. ’ ' 3.2 Hypothesis 2 (Lexical variation study) In the subjective evaluation study on M and T, if the informants tend to associate the evaluative adjective wealthy with T, it implies that T community is more socioeconomically advanced than M community. A hypothesis can be derived from this association. T is more likely to be considered as being a prestige form. When T is endowed with prestige in a contact situation, the lexical items from T will be used as a means of displaying social status. Those who use T will be thought as such people who are from more socioeconomically advanced community and are more urbanized. However, the choice on the lexical variables may be constrained by attitudes towards M and T, sex, age, duration of stay in the U.S.A., and desire to stay or not to stay in the U.S.A. For example, it is likely that those who have been in the contact situation for more than two years will use T more than those who have been in the contact situation less than two years, and those who want to stay in the U.S.A. will use T more than those who want to go back to mainland China. 3.3 Hypothesis 3 (Syntactic variation study) For syntactic variation, the hypothesis is that an important factor influencing the choice among syntactic variables is the belief 13 that the syntax of M is the standard form and superior to other dialects. With this prescriptive norm, the informants may show a uniformity in their choice of the syntactic variables. Syntactic variables may not be likely to show great sensitivity to age, sex, duration of stay, and place of origin, but may be related strongly to the education level of the informants. 14 4Subjective Evaluation Study 4.1 Introduction The results of a number of empirical studies suggest that dialect variables may come to elicit socially conditioned value judgments which in turn affect the attitudes of listeners towards speakers from various backgrounds (Lambert et al 1965; Labov 1966). A speaker may be judged favorably or unfavorably by the dialect he speaks. It has been suggested that the distinct phonological, lexical, and syntactic features of different dialects are used by the listeners as a basis for evaluating the speakers (d'Anglejan & Tucker 1983; Graff et al 1989). Because the use of a particular dialect is a feature of all members of the group speaking that dialect, any reactions to the dialect should reflect the stereotyped characteristics of the group that speaks the dialect. This study aims to describe to what degree the informants can identify M and T speakers, how they evaluate these speakers, and to what degree the ratings on the semantic differential scales are related to the stereotype of T and M speakers. 4.2 Method Subjects. All subjects in this investigation were mainland Chinese who came to the United States after Deng Xiao-ping's new open-door policy since 1978. Most of them were students and visiting scholars, and a few were spouses of the students' and visiting scholars'. 60 subjects were approached with a stratified random sample. The sample was divided up in terms of three social 15 factors: age, sex, duration of stay in the U.S.A. The distribution of the sample is shown in Table 2. IIIZ'D'I'I' [ | I 'II A98 Male Eemale 125mm Wars Wm M 20-29 5 5 5 5 30-39 5 5 5 5 40- 5 5 5 5 Materials. Since this study is concerned with dialect variation, a sample of free speech is suggested as the most appropriate stimulus (d'Anglejan & Tucker 1983). The speakers used in the study were interviewed separately and informally on a specific topic. The topic permitted the speakers considerable freedom of expression (Appendix A). The topics varied according to the speaker's own interest in talking about something such as fishing, eating, cooking, looking for a job, seeing a movie, and learning another dialect. A 20-second sample of uninterrupted speech was selected from the output of each speaker and used as a stimulus for the subject's evaluation of the speaker. The final stimulus tape consisted of six voices: 2 were M speakers: 1 female and 1 male; and 4 were T speakers: 2 females and 2 males. The age of the six speakers ranged from thirty to forty, and all had college education. 16 The subjects were asked to make subjective judgments about each voice on a 7-point semantic differential scale which is the chief measurement technique in present attitudes studies. Typically a semantic differential scale involves the evaluation of a concept by rating it on scales comprised of adjectival opposites (Williams 1983y e.g. : uneducated _: _: _: _: _: _: _: educated On the 1 through 7 scale, 1 was low and 7 was high. For example, if 1 was checked, it meant that this speaker was rated as uneducated. If 7 was checked, it meant that this speaker was rated as well- educated. There were 12 evaluative adjectives used in this study. Some of them were selected from the previous attitude studies (Lambert et al 1965, Lambert et al 1960). Some of them were collected from a sociolinguistic interview (Labov 1984) conducted before the data collection started. In the sociolinguistic interview, three mainland Chinese students were interviewed together. This group interview was designed to obtain overt attitudes towards M and T. The interview focused on the differences in M and T. The interview was tape-recorded. The adjectives used by the three informants to describe M and T were used as evaluative adjectives in this subjective evaluation study such as: soft, elegant, emotive, wealthy, and others. Em. A sociolinguistic interview was carried out first to collect some evaluative adjectives. Then it took one month to finish the 60 individual ratings on the evaluation sheets in 60 interviews. 17 The 60 subjects were tested separately. They were given instructions on how to use the 7-point scales before they started. Since there were six speakers in the tape, each speaker had to be evaluated separately on one evaluation sheet which consisted of 12 semantic differential scales, and one multiple choice question. The multiple choice question involved was to identify where the speaker was from by choosing one of the given places such as Hong Kong, mainland China, Taiwan, Singapore, and other countries. (Appendix B). When they finished rating each speaker on the basis of the voice cues alone, they were asked to identify the speaker by marking one of the answers to the multiple choice question. Finally they were asked to identify their own age, sex, and duration of stay. 4.3 Analysis and Results A technique known as factor analysis was used to analyze the present data. This technique was suggested by some previous attitudes studies (Williams et al 1972; Shuy & Williams 1973). In general terms, factor analysis takes the interrelationships that have been found in the use of scales in rating the different types of speeches. This technique allows us to see whether underlying the subject's use of the individual scales, there are some basic dimensions or factors of judgmental behavior. Each dimension reflects a cluster of scales which are interrelated in their use. In other words, if the subjects are given the 12 semantic differential scales to evaluate T and M speakers, they will tend to subsume 18 individual scales into a few relatively general response dimensions. In a practical sense, these dimensions can be used as a basis for differentiating subjects' attitudes towards M and T. Results of the factor analysis on the present data led to the identification of two factors characterizing the interrelated use of the 12 semantic differential scales. Table 3 presents the correlation coefficients between the 12 semantic differential scales (see next page). Correlation coefficient is a measure of linear association between two variables, that is two semantic differential adjectives. A large r (that is a larger number) indicates that the two variables are strongly related to each other. For example, the correlation between the column variable 1 and the row variable 3 is .65 which indicates a relatively strong linear relationship between the two variables. However, a small r (r close to zero) indicates a weak relationship between two semantic differential adjectives. For example, the correlation between the column variable 4 and the row variable 9 is .13 which indicates a very weak relationship between the two variables. In Table 3, each entry represents the correlation coefficient between the row variables and the column variables. Since the correlation matrix is symmetric, only the lower triangle of the matrix needs to be presented. Since the correlation between the same variables (that is the column variable 1 and the row variable 1) is always equal to 1.00, the diagonal of the matrix is omitted. Below Table 3, " indicates that the correlation coefficient is significantly different from zero at a significance level of .05 and ** indicates that the correlation coefficient is significantly 19 different from zero at a significance level of .01. 2 .46" 3 .65".47" 4 .19 .42" .27' 5 .61".53" .67" .41" 6 .33' .44" .53" .40“ .56” 7 .64".39" .67“ .17 .55" .27" 8 .30' .42" .52" .44" .50“ .54“ .34" 9 .48".45“ .68" .13 .57" .43" .56" .39" 10 .31' .39“ .45" .39" .52" .57" .21 .73“ .42" 11 .49".45" .71" .36" .59" .60" .51" .64“ .57" .63” 12.43".23 .58" .41" .50" .55" .22 .64" .41" .71" .68" *-significant level .05 "-significant level .01 (2-tailed) The factors were given labels to reflect overall dimensions of the judgment of M and T speakers. The labels and their respective scales are shown in the following example. (See next page) 20 status (factor I) : good-sounding--bad-souriding soft--hard elegant-mot elegant intimate-mot intimate good-looking--ugly wealthy--poor solidarity (factor II): emotive--unemotive natural--unnatural kind-- unkind reliable-mot reliable educated-uneducated intelligent-mot intelligent The existence of the two-factor judgmental model was confirmed by a factor analysis of the interrelations among the semantic differential scales. Table 4 shows that factor analysis with principal component extraction and Varimax rotation is applied on the 12 evaluative adjectives (variables). (See next page) Principal components are linear combinations of variables which are orthogonal to each other and which account for, successively, maximum variance of the variables. The amount of variance accounted for by a particular principal component (factor) is called the eigenvalue of the principal component. For example, the first principal component accounts a variance of 6.2995 which is 52.5 21 percent of the total variance of 12 variables. , The larger an eigenvalue is, the more useful the corresponding principal component (factor) is in data reduction. Usually, principal components with eigenvalues greater than 1.00 should be retained for further analysis. As a result, two principal components (factors) are retained since only two factors have eigenvalues greater than 1.00. Table 4: [PW taster eigenvalue WW cumulatlmnsmm 1 6.2995 52.5 52.5 2 1 .5300 12.8 65.2 Varimax rotation in Table 5 is a technique which orthogonally rotates the retained factors such that the correlation between a variable and the factor to which the variable belongs is maximized, whereas the correlation between a variable and the factor to which the variable does not belong is minimized. Hence, Varimax rotation results in factors which are easier to interpret. Table 5 shows the correlations between the variables and the rotated factors. A variable is considered as belonging to factor I if its correlation with factor I is larger than its correlation with factor II. For example, the correlations of variable 1 with factor I and factor II are .81 and .15, respectively. Consequently, variable 1 is considered as belonging to factor I. Based on Table 5,. therefore, variables 22 1,2,3,5,7, and 9 belong to factor I while the remaining variables such as 4,6,8,10,11, and 12 belong to factor ll. 0 I O U .0: .0, :0 . e n. . o o;A : so. :. e :“.| W195 EaflQLs mania: I II 1 . good-sounding .81 .15 2. soft .53 .36 3. elegant .79 .39 4. emotive .07 .64 5. intimate .67 .47 6. natural .32 .69 7. good-looking .86 .03 8. kind .25 .80 9. wealthy .73 .26 10. reliable .19 .84 1 1. educated .54 .65 12. intelligent .25 .79 After a careful content analysis, it is believed that factor I can be interpreted as status and factor II can be interpreted as solidarity, respectively. In order to evaluate the reliability of the two factors, coefficient alphas are computed which are indicators 23 of factor reliability. Using SPSS, the coefficient alphas of factor I and factor II are .88 and .87, respectively, which suggest relatively high reliability of the two factors. A hypothesis of this study is that M speakers and T speakers are associated with different stereotypic semantic differential adjectives. In addition, the social factors may also show a difference in rating M speakers and T speakers on the semantic differential adjectives. A summary of the two factor mean scores in terms of age groups, sex, and duration of stay is given in Table 6 (see next page). Then factor score for status (factor I) is the average of the variables 1,2,3,5,7,and 9, while the factor score for solidarity (factor II) is the average of the remaining variables 4,6,8,10,11, and 12. Table 6 shows the factor means for the subgroups as defined by sex, age and duration of stay, which indicates each subgroup's average scores on the two factors termed as status and solidarity. (See next page) In order to get statistically confirmed results of the differences in rating M speakers and T speakers on the two factors by the three social characteristics such as age, sex and duration of stay, a test of this hypothesis was undertaken in the form of paired t-test (one-tailed test). The results of comparisons between the factor scores of M speakers and T speakers are presented in Table 7. (See page 26). -°3 0° “ll-l 24 mindsets headsets £39191 £89191 L__ll l_II sex male 4.37 5.01 4.15 4.73 female 4.45 5.20 4.22 4.89 age 20-29 4.03 4.79 4.17 4.83 30-39 4.67 5.27 4.29 4.92 40- 4.54 5.25 4.10 4.66 duration of stay below 2 years 4.47 5.16 4.34 4.96 above 2years 4.35 5.05 4.03 4.66 These comparisons are made by breaking down the data into subgroups such as sex, age, and duration of stay. Each entry in Table 7 is the paired-t statistics and the number in the bracket is the significant level. For example, the first entry's statistic result is 1.354, and the significant level is p <.10. A significant level smaller than .05 indicates a significant result, while the level greater than .10 indicates an insignificant result. In addition, a significant level between .05 and .10 suggests a marginally significant result. The results of the paired t-test reveal that male as a subgroup rate M speakers as marginally higher than T speakers on status (t- value 1.354, p<.10), but rate M speakers as significantly having more solidarity than T speakers (t-value 1.966, p<.05). Women as a subgroup rate M speakers as having more status than T speakers (t- 25 value 1.726, p<.05), but rate M speakers as significantly having more solidarity than T speakers (t-value 2.264, p<.025). The below two year subgroup do not rate M speakers as having more status than T speakers (t-value 1.003, p>.10), but rate M speakers as marginally having more solidarity than T speakers (t-vale 1.685, p<.10). The above two year subgroup rate M speakers as having more status than T speakers (t-value 1.969, p<.05), but rate M speakers as significantly having more solidarity than T speakers (t-value 2.484, p<.01). The youngest age group of 20-29 do not rate M speakers as higher than T speakers on status (t-value -0.870, p>.10), and do not rate M speakers as having more solidarity than T speakers either (t- value -0.367, p>.25). The 30-39 age group rate M speakers as having more status than T speakers (t-value 2.072, p<.05), and also rate M speakers as having more solidarity than T speakers (t-value 1.946, p <.05). The oldest age group rate M speakers as significantly having more status than T speakers (t-value 2.556, p<.01), but rate M speakers as significantly having much more solidarity than T speakers (t-value 3.383, p<.005). Table 7 shows that except the 20-29 age group, all the subgroups rate M speakers as significantly having more solidarity than T speakers, that is to say that in general the subjects rate T speakers as having less solidarity than M speakers. For status, the 20-29 age group and the below 2 year subgroup do not rate M speakers as having more status than T speakers, and men rate M speakers as marginally having more status than T speakers, but women as a subgroup, the above 2 year subgroup, and the older age groups all rate M speakers 26 as having more status than T speakers. Iabinli ..'._ .h. . . ._ u. dJ. status 521mm! sex male 29 1.354 1.966 (p < .10) (p <.05) female 29 1 .726 2.264 (p <.05) (p <.025) age 20-29 19 -0.870 -0.367 (p >.10) (p >25) 30-39 19 2.072 1.949 (p <.05) (p <.05) 40- 19 2.556 3.383 (p < .01) (p <.005) duration of stay below 2 years 29 1.003 1.685 (p >.10) (p <.10) above 2 years 29 1.969 2.484 (p <.05) (p <.01) As was mentioned early, social factors such as age, sex, and duration of stay may affect the rating of the same speaker. There may be a difference of rating M speakers among the different subgroups, for example, the youngest age group may rate M speakers 27 lower on status than the older age groups. There may also be a difference of rating T speakers among the different subgroups. A test of this hypothesis was given in the form of ANOVA. The following tables show the ANOVA results of rating M speakers on status and solidarity by the three social factors and of rating T speakers on status and solidarity by the same three social factors. Iahle8: WW S I 1! . l' E S' 'l' I E main effects 3.037 .026 sex .240 .626 duration of stay .579 .450 age groups 5.665 .006 2-way interactions .339 .886 sex,duration of stay .391 .535 sex, age groups .197 .822 duration of stay, age groups .456 .636 3-way interactions 1.267 .291 sex, age groups, duration of stay 1.267 .291 Explained 1 .489 .1 67 In Table 8 each entry shows the F statistics and significant level of F statistics. For example, the F statistics in the first entry is 3.037, and the significant level is .026. A significant level smaller 28 than .05 indicates a significant result, while the level greater than .10 indicates an insignificant result. A level between .05 and .10 suggests a marginally significant result. The results shown in Table 8 suggest that there are no significant 2-way and 3-way interactions. 2-way interaction is two variables added together such as sex x duration. 3-way interaction is three variables added together such as sex x duration x age. A significant 2-way interaction between sex and duration means that the effects of sex on status are different at different levels of duration. For example, in the below two year group, sex and status are positively related, and in the above two year group, sex and status are negatively related or unrelated. In absence of significant 2-way interaction, the effects of sex on status are the same at different level of duration. For example, in the below two year subgroup, sex and status are positively related, and in the above two year subgroup, sex and status are also positively related. A significant 3-way interaction (age x duration x sex) means that the effects of sex and duration on status are different at different levels of age such as 20-29 age group, 30-39 age group, and others. If there is no significant 3-way interaction, it means that the effects of sex and duration on status are the same at different levels of age. Table 8 shows no significant 2-way and 3-way interactions, therefore, there is no need to give further explanation. In the rating of M speakers, sex, as a social factor divided up by men and women, do not show a difference (p -.626). Nor does the 29 duration of stay social factor. The below two year subgroup and the above two year subgroup show no difference in rating M speakers on status(p -.450). However, the subgroups in age show a significant difference in rating M speakers on status(p -.006). The youngest age group rate M speakers as having a lower status than the older age groups (see Table 6). In Table 9 (see next page) analysis of variance of M speakers on solidarity by the three social factors is presented. Again there is no significant 2-way and 3-way interactions. Men and women do not show a difference in rating M speakers on solidarity (p -.309). The below two year subgroup and the above two year subgroup do not show a difference in rating M speakers on solidarity (p -.564). The three subgroups in age show a marginally significant difference in rating M speakers on solidarity (p -.068). The older age groups rate M speakers as having a higher solidarity than the youngest age group (see Table 6). The analysis of variance of T speakers on status by the three social factors is presented in Table 10. (See page 31) There is no significant 2-way and 3-way interactions. All the subgroups in sex, age, and duration of stay show no difference in rating T speakers on status (sex p -.710,duration of stay p c.121, age p -.735). As for the analysis of variance of T speakers on solidarity by the three social factors, the results are shown in Table 11. (See page 32) There is no significant 2-way and 3-way interactions. The subgroups in sex do not show a difference in rating T speakers on solidarity (p -.438). Nor do the subgroups in duration of stay (p 30 -.150). The subgroups in age also do not show a difference in rating T speakers on solidarity (p -.587). IIIB'EI' I . [II I Ill 5 I I! . l' E S' 'I' I E Main effects 1.771 .150 sex 1.059 .309 duration of stay .338 .564 age groups 2.843 .068 2-way interactions .452 .810 sex,duration of stay . .006 .941 sex,age groups 1.066 .353 duration of stay, age groups .061 .941 3-way interactions .637 .533 sex, duration of stay, age groups .637 .533 Explained .965 .490 The results in analysis of variance of M speakers on status and solidarity and of T speakers on status and solidarity by the three social groups suggest that, in general, all the members in the mainland Chinese speech community agree with each other in rating their own speech M, and in rating T. Only the subgroups in age show a difference in rating M speakers on status (p -.006), and a marginal difference in rating M speakers on solidarity (p =.068), but all the members agree with each other in rating T speakers on both status and solidarity. Though the results do not support the hypothesis that the subgroups in the speech community would rate the same speakers differently on status, and the same speakers differently on solidarity, the results statistically indicate that all the members in the speech community show a uniform attitude toward their own speech and T. 31 IlllIl' EI'I' [I I II S I I! . |' main effects sex duration of stay age groups 2-way interactions sex,duration of stay sex, age groups duration of stay , age groups 3-way interactions sex,duration of stay, age groups Explained E .814 .140 2.496 .31 0 .674 .062 .243 1.41 1 1 .723 1 .723 .916 S. I I E .523 .710 .121 .735 .645 .804 .789 .254 .189 .189 .533 32 S I I! . I' E S' I I E main effects .957 .440 sex .611 .438 duration of stay 2.139 .150 age groups .538 .587 2-way interactions .217 .954 sex, duration of stay .862 .358 sex, age groups .054 .948 duration of stay, age groups .057 .945 3-way interactions .436 .694 sex, duration of stay, age groups .436 .694 Explained .526 .876 As for the identification of the six speakers in the stimulus tape for the subjects to evaluate, the majority of the 60 subjects identify M speakers and T speakers correctly. Table 12 shows the percentage of correct and wrong identification of the six speakers: 33 The results show that the subjects can correctly identify M speakers from T speakers according to the linguistic cues in M and T. 4.4 Discussion The present analysis aims to determine the judgmental model yielded by the mainland Chinese subjects, and to determine the degree to which the judgments vary according to the subject's age, sex, duration of stay in the U.S.A. The results indicate that from a limited list of the 12 semantic differential scales, two dimensions of judgment are identified. They are labeled as status and solidarity. In comparing judgments, the following results are worth discussing: 1. In comparing M speakers and T speakers on status and solidarity, the results of the paired t-test show that for solidarity, the subgroups such as men, women, below two year subgroup, above two year subgroup, 30-39 age group and 40-above age group, except 20-29 age group, all rate M speakers as significantly having more solidarity than T speakers. The results indicate that the majority of the subjects agree with each other in describing M speakers with the judgmental model of solidarity. As for the judgmental model of status, 20-29 age group and the below two year subgroup do not rate M speakers as having more status than T speakers, but women, above two year subgroup, and the older age groups rate M speakers as having more status than T speakers. In another word, the subjects show a diversity in describing M speakers with status. There are two possible reasons for this diversity in judging M speakers on status: 34 (1) Different social groups in the mainland Chinese speech community evaluate M speakers differently on status. Some think that M speakers have more status than T speakers. Some think that T speakers have more status than M speakers. (2) Through extensive contact with T speakers in the United States, M speakers start to adjust themselves to T speech, and the differences in M and T speeches start to become more and more minimal, which makes M speakers also sound elegant, soft, and others. There is a tendency indicating that those subgroups such as women, the older age groups, et cetera. are leading a change in evaluating M speakers on status. They start to think that M speech sounds as statusful as T speech now. However, further research is needed to give more detailed explanation of this diversity. 2. The statistical results indicate that the cluster of evaluative adjectives used to judge M speakers and T speakers show a consistency with those adjectives collected from the sociolinguistic interview. The stereotypic evaluative adjectives used to judge M and T are statistically confirmed in the form of factor analysis. They use soft, wealthy, or elegant to judge T speakers, but use natural, reliable, or kind to judge M speakers. A differentiation of using these adjectives to describe M and T speakers reflects some level of psychological reality of stereotype attitudes of the subjects towards M and T speeches. 3. In rating only M speakers on solidarity and status, and in rating only T speakers on solidarity and status, the results of ANOVA suggest that the unconscious evaluation of M and T speeches is 35 remarkably uniform throughout the mainland Chinese speech community. Members of this community share their attitudes towards M speech because the results of ANOVA do not show a difference of their rating M speakers on both status and solidarity. They also share their attitudes towards T speech and the results of ANOVA do not show a difference of their rating T speakers on both status and solidarity. This finding supports Labov's (1989) definition of speech community and some previous attitude studies (Lamber 1967, Labov 1972) which suggest that attitude towards different languages and dialects is shared by the members of the same speech community. 4. In Table 12, the low percentage of misidentification of M speakers and T speakers indicates that the subjects are well aware of the differences of intonation, timing, accent, modal particles, etc. existing in M and T speeches. In turn, these differences serve as cues for the subjects to evaluate M speakers and T speakers. If a high percentage of misidentification of M and T speakers is shown, the results of this subjective evaluation study are not very convincing, and powerful community-wide stereotypes of M speakers and T speakers do not strongly exist. 36 5 Lexical Variation 5.1 Introduction This part of the study is concerned with the direct observation of a lexical variation in the context of the community life in a language contact situation. The previous studies have found that it can readily be observed that related, mutually intelligible dialects do have effects on one another in contact situations, with or without the development of individual bidialectalism. Very often, for example, when two speakers of different varieties of the same language which are completely mutually intelligible come into contact and converse, items may be transferred from one of the varieties to the other and very often it is at the lexical level that the transfer begins first because lexical differences are highly salient, and are readily apparent to all the speakers of the speech community without any linguistic training or analysis (Trudgill 1986; Weinreich 1968). This generalization is applicable to the mainland Chinese speech community. When they came into contact with T speakers, variation on lexical level, as being observed, took place first. They started to use less the lexical items that the community brought with them from mainland China, but used more and more the lexical items which disappeared long ago in mainland China but which are still used by T speakersz; that is, the lexical 2The lexical item airen is a typical mainland China lexical item, only used by the mainland Chinese. It is a marker of group identity. But the lexical item taitai and xiansheng are not only used by Taiwanese, but also by oversea Chinese from Hong Kong, Singapore, 910.. Since mainland Chinese speakers in the United States have extensive contact with people from Taiwan, they associate the differences in lexical items with 37 items referring to husband and wife: airen (M)(wife/husband), taitai (T)(wife), xiansheng (T)(husband). Airen is a typical mainland lexical item synonymous with taitai and xiansheng which are typical Taiwanese lexical items with taitai referring to wife and xiansheng referring to husband. By studying the frequency and distribution of these variants according to age groups, desire to stay in the U.S.A., sex, duration of stay, and others, the pattern of this lexical change can be constructed by correlating the linguistic pattern with differences in social structure. 5.2 Background Information In both T and M, quite a lot of lexical items can be used to refer to husband and wife, but the frequently used ones in speaking forms are not many. There is a distinct way of using these items between T speakers and M speakers. What is widely used by T speakers is not widely used by M speakers. For example, taitai (wife) and xiansheng (husband) are used by T speakers, but airen (referring to both husband and wife) is only used by M speakers. In order to have a better understanding of the lexical variation and change under examination, a brief history of these lexical item is reviewed here. Before 1949 when the PeOple's Republic of China was established, Taiwanese only, but not with those from other places. Evidence shows from the interviews in the subjective evaluation study that when the subjects were asked to identify the different speakers according to the multiple choice question, some of them said that they did not know anybody from Hong Kong and Singapore, and never heard of the Chinese language they spoke, but they could tell the differences in the Chinese language spoken by Taiwanese. 38 taitai and xiansheng were widely used in mainland China. When the Kuomintang Party led by Chiang Kai-shek went to Taiwan, they brought with them not only a political system which was different from that of the Chinese Communist Party led by Mao zedong, but also the language used before 1949 in mainland China. This language has been developed independently in Taiwan, and given the name guoyu since then. The year of 1949 is the dividing line between the new China and the old China by the Chinese Communist Party. Some lexical items associated with the old China were criticized by the Chinese Communist Party. The salient example was the lexical items taitai and xiansheng which were associated with the exploiting class. These lexical items started to disappear in mainland China after 1949, and a new one was encouraged in use. That lexical item was airen. Criticism about using these old China lexical items was carried out again during the Cultural Revolution. Those who happened to use these lexical items risked being persecuted by the Red Guards.3 As a result of the political and social change, taitai and xiansheng disappeared from the verbal repertoire of the mainland Chinese speakers, but airen came into being with a literal translation of the "loving person" in terms of the morpheme structure: ai- means "love", and -ren means ”person". Its use in practice is encouraged by the Chinese Communist Party with the explanation that airen is associated with the idea that men and 3Red Guards refer to the young generation who were the vanguard in launching the Cultural Revolution. They were the generation ranged from 16 years old to 30 years old including the students at middle school, high school, and universities. 39 women are equal so that wife and husband should do the four ”each others": love each other, help each other, show respect to each other, and be polite to each other. This lexical item has become popular in mainland China since 1949, not because it is simple in the sense that it covers both sexes (wife and husband), but because it is associated with the four “each others" which give both men and women an equal social status. However, this lexical item is “locked” in mainland China. Elsewhere out of mainland China this item is not borrowed by other Mandarin speaking people; thus, it remains as a representative lexical feature of M. When the M speakers came into contact with T speakers, variation between these lexical variables occurred. 5.3 Data Collection The data was collected by randomly sampling the mainland Chinese speech community through individual interviews. 68 informants were approached. Among them, there were 39 males, and 29 females, divided into three age groups. The body of the interview consists of questions designed to get the informants to talk as freely as possible (Appendix C), for example, "Which lexical item 00 you choose to use?". Some of the questions were open-ended questions for the purpose of getting more informative and accurate feedback from the informants, including collecting direct attitudes towards M and T and free remarks made on M and T. Such questions are ”Why do you choose this lexical item?" ”What do you think of the lexical item you choose and the lexical item you do not choose?” 40 Many early studies of language in its social context used tape recorded interviews as data since these studies were all confined to phonological variation. However, the analysis of lexical variation imposes some specific constraints on the fieldwork method( Cheshire 1982). The main problem is that very large quantities of data are needed in order to ensure an adequate number of occurrences of lexical variables, and to identify the lexical variables from a number of potential lexical variables. Thus a combined method of observation and questionnaires was used in the present study. Before the study on the lexical variation was carried out, there was a series of preliminary observations. These included informal interviews and a number of anonymous observations in public places. These preliminary observations led to the definition of ten lexical variables which were studied here, including the typical lexical items from M -airen, from T -taitai, xiansheng, from English - wife, husband, and including some items which were neutral such as: qizi (wife), furen (wife), zhangfu (husband), Iaogong (husband), and name. These items are neutral in the sense that they are not associated with either M and T. The purpose of giving these neutral lexical items and English items is that: (a) in practice those lexical items are sometimes used by the mainland Chinese speakers, and (b) not to force the informants to make a choice only between taitai,xiansheng and airen, but to empirically support the preliminary observation that taitai,xiansheng, and airen 41 appeared to be sensitive to any measure of social factors, and the choice on English words related to other interpretations. The method used to collect the data on these lexical variables is very simple, and similar to the methods used by Mier (1987), Anshen & Aronoff (1989), and Ljung (1989). The ten lexical items were written on a card which was presented to the informants during the interview. The informants were asked to make a choice among these variables and answer some questions. Their answers to these questions were written down, and their choice on the variables was sorted out when calculating the frequency of different variables. 5.4 Data Analysis A most important aspect of the study of linguistic variation and change is the careful and precise description of what features of language are in variation and change, and in what ways (Baugh & Sherzer 1984). Since there is no easy way of measuring and characterizing the total impact of one language on another in the speech of the mainland Chinese community, the only possible procedure is to describe the various forms of lexical items and to tabulate their frequency of occurrences, then to account for the different frequency of occurrences (Cheshire 1982; Labov 1966, 1972; Trudgill 1974). In order to ensure that a representative sample of speakers of M as a speech community are included in the study, two different kinds of samples are used: a stratified sample with detailed social characterization in the subjective evaluation study and the syntactic variation study, and a random sample through the speech 42 community in the lexical variation study. In doing so, a large portion of the speech community is approached. The two different sources of the available data from the same speech community provide a better laboratory for studying the mechanism and motivation for synchronic variation. The data obtained from the 68 informants formed the basis for the study of the ten lexical items in two patterns: the single choice on the lexical items from M, T , and English (hereafter E); the dual choice on M and T, M and E, and T and E. The two patterns are shown in Table 13. (See next page) The frequency of choices was tabulated and converted to percentages based on the Labovian framework of analyzing linguistic variables. Four tables were made to show the statistical results in correlation with four social factors: sex, age, duration of stay in the U.S.A., and desire to stay or not to stay in the U.S.A. One table was made to show a direct measure of attitude towards M and T. The different attitudes towards M and T were strongly related to the duration of stay in contact with T which in turn affected the informant's choice on the lexical items. 43 single choice M airen husband/wife T taitai wife xiansheng husband E husband wife dual choice M/T airen husband/wife taitai wife xiansheng husband M/E airen husband/wife husband wife T/E taitai wife xiansheng husband husband wife 5.5 Results 5.5.1 Sex A number of sociolinguistic studies show that variable features are distributed differently for men and women (Labov 1966, Ljung 1989, Trudgill 1974, Wolfram 1969). This observation is applicable to the lexical variation studied in the mainland Chinese speech community. Among the 68 informants there are 29 females and 39 males. The statistical result on lexical variables by sex is presented in Table 14. (See next page) 44 temals. male single choice M 10 21 28 3 8 34 8 dual choice M/T 3 1 5 T/E 17 3 WE 0 3 As is shown in Table 14, both men and women show a higher percentage of single choice on T (38% and 28%), rather than on M (21% and 10%). However, women show the highest percentage of single choice on E (34%) and the lowest one on M (10%). Men show the highest percentage of single choice on T (38%), but the lowest one on E (8%). Men use M more than women do, and men also use T more than women do. But women use E significantly more than men do. For the dual choice pattern, men use more MIT than women do, but women use more T/E than men do. The differences of percentage indicate that both men and women favor T more than they favor M. Women, as a subgroup in the speech community, share a different prestige norm towards E from men. This difference is also supported by the women's higher percentage of dual choice on TIE. Table 14 suggests that the distribution of sex accounts for differences in speech behavior which, in turn, accounts 45 for lexical variation conditioned by social factors. 5.5.2 Duration of stay in the U.S.A. Duration of stay as a social factor also affects linguistic variation (Forschungsprojekt 1978, Weinreich 1968). In Table 15 the differences in lexical variation in correlation with duration of stay are measured. Among the 68 informants, 20 have been in the U.S.A. (is. in the contact situation with T) for less than two years, and 48 for more than two years. Those with a shorter duration of stay show the highest percentage of single choice on M (35%), the next highest percentage on E (15%), but the lowest percentage on T (5%). As for the dual choice pattern, they only make choice on M/T. Those with a longer duration of stay show the highest percentage of single choice on T (46%), the next highest percentage of single choice on E (19%) but the lowest percentage of single choice on M (8%). For the dual choice pattern, they show the highest percentage on T/E, the next highest percentage on M/T, but the lowest percentage on M/E. Those who have stayed a shorter time use M much more than those who have stayed longer. mm; W W single choice M 3 5 8 T 5 4 6 E 1 5 1 9 dual choice M/T 2 0 6 T/E 0 1 3 M/E 0 2 46 For those who have stayed longer use T much more than those who have stayed shorter. There is no great difference between the two subgroups on the lexical choice of E. For the dual choice pattern, the below two year subgroup use more MIT than the above two year subgroup. No choice is made on TIE and WE by the below two year subgroup. The differences of percentage show that duration of stay as a social factor affects the lexical variation. In general, those who have stayed longer favor T more than M, but those who have stayed shorter favor M more than T. 5.5.3 The desire to stay or not to stay in the U.S.A. Table 16 shows the desire to stay or not stay in the U.S.A. as a social factor affecting the informant's choices on the lexical variables. IBM. '- -I o- o .0 - 0| - o- o e- .. e o I. MAMA. 513! may single choice M 8 26 T 46 19 E 1 6 19 dual choice M/T 8 1 3 T/E 14 3 47 Among the 68 informants, 37 want to stay, and 31 do not want to stay. Those who do not want to stay use more M (26%) than those who want to stay (8%). However, those who want to stay use much more T (46%) than those who do not want to stay (19%). The percentages on E are very close in both subgroups. Those who do not want to stay give the highest percentage of lexical choice on M, but those who want to stay give the highest percentage on T. For the dual choice patterns, those who do not want to stay use more MIT but less T/E than those who want to stay. Those who want to stay use more TIE than those who do not want to stay. The differences in statistical results on the lexical variables show that those who want to go back to mainland China intend to keep their own typical lexical item, but those who want to stay intend to get rid of the typical mainland lexical item. Those who want to go back to mainland China will use more M than T, but those who want to stay will use more T than M. 5.5.4 Age The age of the 68 informants is within ten-year intervals: 20-29, 30-39, and 40-49. Among the 68 informants, there are 25 in the 20- 29 age group, 36 in the 30-39 age group, and only 7 in the 40-49 age group. The differences of the three groups' choices on the lexical variables are presented in Table 17. (See next page) There is a great difference between percentage of choices on M and T among the 20- 29 age group, and this group use M much less than the older age groups. For the older age groups, the differences between 48 percentage of choice on M and T are not that great. single choice M 8 22 29 32 3 6 43 E 20 1 7 14 dual choice M/T 8 8 14 T/ E 1 6 6 0 WE 0 0 0 It seems that the older age groups intend to keep a balance of the old prestige form and the new prestige form, but the youngest age group are beginning to adopt the new prestige form only. The percentages of single choice on E are pretty close among all the three age groups. However, the youngest age group show the second highest percentage of choice on E, but the older age groups show the second highest percentage of choice on M. Since T and E are all associated with prestige forms, the youngest age group show the highest percentage of dual choice patterns on TIE, but the oldest age group show the highest percentage of dual choice patterns on MIT. As a whole, all the age groups favor T more than other lexical variables. 49 5.5.5 Attitudes towards M and Duration of stay in the U.S.A. Dialect variables may come to elicit socially conditioned value judgements which in turn affect the attitudes of informants (d'Anglejan & Tucker 1983). In this study, the different attitudes towards M are derived from the duration of stay in the contact situation. Table 18 presents a direct measure of attitudes towards M in correlation with duration of stay. The number of the informants who have positive and negative attitudes towards M is calculated and converted into percentage for interpretation of the data. 5131 W W positive 45 17 negafive 10 50 neutral 45 33 Positive in Table 18 covers the overt prestige associated with M, but overt stigma associated with T, while negative covers the overt prestige associated with T, but overt stigma associated with M. Since the informants were asked to give comments on M and T, they showed their overt attitudes towards M and T such as: "I like to use airen because it sounds comfortable to me. I do not like to use taitai and xiansheng. They remind me of the old 50 China.” ”I hate using airen. It sounds disgusting. I always use taitai and xiansheng because they sound statusful." The different attitudes towards M are strongly influenced by the duration of stay in the U.S.A. Those who have stayed in the contact situation with T longer show a much higher percentage of negative attitudes towards M than those who have stayed shorter. Those who have stayed less than two years show a much higher percentage of positive attitudes towards M than those who have stayed longer. In addition, the differences of percentage of positive and negative attitudes are great in each subgroup. For example, the percentage of positive attitudes in the less than two year subgroup is 45, but the percentage of negative attitudes is 10, and the percentage of positive attitudes in the above two year subgroup is 17, but the percentage of negative attitudes is 50. The statistical differences shown in Table 18 are significant. In general, the longer the informants stay in the contact situation, the more they favor T and the less they favor M, but the shorter they stay in the contact situation, the more they favor M and the less they favor T. The change of attitude in Table 18 is consistent with the choice on M and T shown in Table 15. Those who have stayed longer use T significantly more than those who have stayed shorter, but those who have stayed shorter use M significantly more than those who have stayed longer. The correlation is obvious: the duration of stay affects their attitude and attitude in turn affects their choice on the lexical variables. 51 5.6 Discussion A breakdown of the data by social factors reveals that these lexical variables reflect great sensitivity to social factors. The selection among these variables carries social significance in different subgroups in the mainland Chinese speech community. Although individuals sometimes use M and sometimes use T, or other lexical items, the average percentage for each group falls into a quite predictable pattern because the members of the same speech community share rules but differ in their degree of realization of the rules (Preston 1989). A few generalizations can be made as follows: 1. For the whole speech community, in general, T is regarded as a prestige form because it is associated with a social group which is one step above the mainland Chinese group socioeconomically. Evidence is found from the individual interviews. Some informants said that they liked to use T because T carried the association of being "open" and ”economically developed”. Some said that they did not like to use M because M sounded "not urbanized", ”revolutionary", and “from a developing country”. A rise of T in the mainland Chinese speech community is promoted by this favorable association with T and this unfavorable association with M. These associations are shown in all the subgroups in terms of social factors. 2. In addition, different norms are shared by different subgroups in this speech community. Women, as a subgroup, not only associate T with a prestige form, but also associate E with a prestige norm that men seem not to share. Women show a much higher percentage 52 of choice on E than men do because E is associated with 'well- educated" and “westernized”. Since most of the women are spouses who came to the United States because their husbands are studying in the American universities, most of the women stay at home, but only a small number of them go to universities. Some of them are working on a master's degree, but very few are working on a Ph.D. degree. However, the majority of the men are in Ph. D. programs. In this sense, men are considered as being more well-educated than women, and at the same time men are considered as being more westernized. Being well-educated and westernized is associated with a higher social status in this speech community. Some previous studies report that, generally speaking, women are more status- conscious than men. Trudgill (1983) suggests that since women have lower social status than men they are more aware of the value of linguistic indicators of status. Women in this mainland Chinese speech community use E significantly more than men do when women are expected to use less E than men are expected. This hypercorrection of the women's use of E suggests that women in this speech community are more sensitive to the social significance of status-related linguistic variables such as E since speaking E is associated with ”being westernized” and ”advanced education”. In a speech community where women favor a prestige form, a linguistic variable that shows women in the lead may very well be a new change (Labov 1981). 3. The choices on M and T vary significantly between those who 53 want to stay and those who do not want to stay. Those who want to stay use more T than those who do not want to stay in order to separate themselves from mainland China, but those who do not want to stay use more M in order to signal their desire of going back to mainland China and to keep their group identity as being someone from mainland China. This finding shows a parallel to Labov's (1972) famous study on some phonological variables in Martha's Vineyard. Those who want to leave the island intend to get rid of the typical island phonological features, but those who want to stay use the typical island phonological features significantly. 4. In addition, it is important to discuss is the direct measure of the subjective attitude towards M and T. This direct attitude measure helps to explain why there is a rise of T and a diminution of M in the mainland Chinese speech community. An important factor causing the change of attitude is the length of duration of stay in the contact situation with T. The longer they stay, they more they favor T and the less they favor M. It is not a simple thing to explain. The longer they stay, the more demerits they see with the political and economical systems in mainland China. It happens that the use of airen is encouraged by the Chinese Communist Party. Their dissatisfaction with the systems led by the government is also reflected in their attitude toward M which in turn affects their choice on lexical variables. Evidence from the interviews shows an explicit attitude such as: "What the Chinese Communist Party encourages is not good, airen is not scientific and it confuses people in social interactions.” 54 M is not scientific in the sense that M lacks sex differentiation. A semantic componential analysis on M and T explains why: airen: [+married] taitai: [+married, +female] xiansheng: [+married, +male] Airen refers to both husband and wife and covers both sexes, but taitai only refers to wife and xiansheng only refers to husband, and they differentiate sexes. Evidence also shows that the morpheme structure of M affects the choice on M which consists of two morphemes: ai- means 'love", and -ren means "person”. The literal meaning of M is ”loving person” which is considered as being a synonym of the English word 'lover" which contains a bad semantic connotation.Some informants give remarks on M as follows: "I don't use airen here because I don't want to be laughed at, airen is associated with 'lover'”. The lack of sex differentiation of M and bad semantic connotation appear to be as linguistic factors which constrain the informant's choice on lexical variables. As a matter of fact, these remarks are attitude-induced and contact-induced. These two so-called linguistic factors are not true in mainland China where airen has been widely used by the speakers who have a different attitude toward M and have no contact with other Mandarin speakers from other places out of mainland China, and are not true for those who show a positive attitude towards M and do use M in a contact 55 situation as shown in Table 18. In general, the informants favor T because T is associated with a social group which is more economically advanced, thus the linguistic variety used by this group is endowed with prestige. Their lack of favor of M is complicated for a few reasons particularly because M is associated with a less economically advanced social group, thus bearing no prestige; M is associated with a political and economic system that is disliked by those who have the opportunity to come into contact long enough with other social groups from different political and economical systems; and M is associated with the contact-induced remarks on the bad semantic connotation of M. What has been mentioned in this paragraph forms a base for the attitude toward M and T. When the informants make their choice among the lexical variables, they all make their remarks on their choice and in their remarks they show overt attitude toward M and T, in turn their attitude directly affects their choice on the lexical variables. 5. Age, as a factor, is also worth discussing. Some empirical findings suggest that if a linguistic variety is favored as a prestige form, a rise in this prestige form will be expected in all age groups. This generalization is supported by the present study. All age groups show a much higher percentage of choice of T. The previous studies also suggest that differences in generational speech may be indicators of language change and one social group in the speech community has been found to lead in the development of a linguistic change (Preston 1989). The age group which leads the change from M 56 to T in the mainland Chinese speech community is the 40-49 age group which shows the highest percentage on T than the other two age groups. These oldest informants are the ones who underwent the Cultural Revolution when T was criticized and disappeared. They underwent this transitional period from some use of T to a complete use of M, and they had had contacted T before they came to the United States. When they came into contact with T again, probably they intend to get back the 'old" prestige form - the lexical items that they were exposed to before. In summary, the picture which emerges clearly from the data suggests an increasing favor of T in the mainland Chinese speech community. The alternation between M and T carries social meanings, and the variation pattern is socially motivated. The speakers of the mainland Chinese speech community make lexical choice as part of a verbal strategy to identify themselves with social categories, and their choice symbolizes their attitude. The findings support the hypothesis mentioned early in Hypothesis Two (Lexical variation), and indicate that airen, xiansheng,andtaitai are sociolinguistic variables as defined by Labov (1972). They provide the option of saying the same thing in several different ways, and they are identical in referential or truth value, but oppose in their social significance. 57 6 Syntactic Variation 6.1 Introduction Sociolinguistic studies are concerned with documenting and understanding the relationships between language and various social forces. The relationships between language and various forces are studied on different linguistic levels: phonological level (Labov 1966, 1972, 1989, Trudgill 1974, 1984), lexical level (Sankoff et al 1978), and syntactic level (Fasold 1972, Kroch & Small 1978, Rickford 1983). Empirical findings from previous studies suggest that variation at all linguistic levels can be correlated with social factors such as age, sex, social economic class ethnicity, social ideology, et cetera. In this part, linguistic behavior at the syntactic level is studied in the mainland Chinese speech community. The syntactic feature under examination is the aspect marker: guo in M and you in T. These two aspect markers serve as variables in this syntactic variation study, and what controls the variation pattern is discussed below. 6.2 Background Information Communication was completely blocked for more than thirty years between the Mandarin speakers in Taiwan and the Mandarin speakers in mainland China. Distinct linguistic features developed independently in both Taiwan and mainland China. A salient syntactic feature is the different use of the aspect marker in spoken Mandarin in Taiwan and mainland China. 58 In M, there are three aspect markers: zhe, Ia, and guo. They are all put either after verbs or adjectives in sentences to indicate the change of an action or a state. The differences between these aspect markers are: zhe indicates the action or state in a continuous aspect. Ie indicates the completion of action or state. guo indicates the completion of action or state, but with an emphasis on gained experiences (Huang & Liao 1985, Wang 1987). In this study, only guo is under examination based on the preliminary observation that guo has a variant you. In spoken T, a frequent use of you has been observed and you replaces guo in a different position in sentences, but in M, you is not an aspect marker and not used in a syntactic status as an aspect marker. The following examples give a comparison between M and T. The following gives the literal meaning of each word in the sentences: M 1. Biaojie shuo guo zhe jian shi. cousin say this measure word thing/matter My cousin has talked about it. T 2. Biaojie you shuo zhe jian shi. cousin say this measure word thing/matter My cousin has talked about it. 59 Another example shows that you is added in a sentence. The literal meaning of each word is given as follows: M 3. Ta dalai guo dianhua. he call/make phone call He has called. T 4. Ta you da guo dianhua. he call/make phone call He has called. Probably a tree diagram will give a better picture of the relationships between the constituents in the above four sentences. M1: / I biaojie (cousin) shuo guo zhe jian shi . (talk) (this matter) 60 T2: /3\ NP VP biaojie ap/ NP (cousin) l l I you shuo zhe jian shi (talk) (this matter) M3: / n: vp _ | ta (he) V' p NP l l l dalai guo dianhua (call) (telephone call) T4: S N P l ta (he) ap V p NP you da guo dianhua (call) (telephone call) 61 In M1 the aspect marker guo is put after the verb shuo, but in T2 the aspect marker guo is replaced by you which precedes the verb shuo. In M3 the aspect marker guo is following the verb dala, but in T4 the aspect marker guo remains after the verb, but you is added preceding the verb da. The aspect marker you is a salient syntactic feature In T. The alternation between you and guo gives a base for syntactic variables and offers an alternative way of saying the same thing for the M speakers. In the study of syntactic variation, the nature of being a syntactic variant has to be talked about. As is known, syntactic variants are more difficult to identify and compare, qualitatively and quantitatively, than morphological, phonological, and lexical variants because they have to meet the definition of syntactic variation. Labov (1972) gives synonymy as the criterion for determining the set of variants of a particular syntactic variable. The notion of synonymy specifies that the alternative forms should refer to the same state of affairs and have the same truth value. Jacobson (1989) suggests that the most obvious way to define syntactic variation is to say that syntactic variation involves the availability of two or more constructions that differ in form but share the same meaning, and he also suggests that different researchers should be free to define it in different ways. In doing so, this makes syntactic variation a field of study that is both interesting and challenging. In the present study of syntactic variation, guo and you meet the criterion of syntactic variables 62 proposed by sociolinguists. Although guo and you differ in structural forms, they do share the same meaning and the same syntactic status as aspect markers. 6.3 Data Collection The informants in this part are the same as those described in the subjective evaluation study. The total informants are 60 in number and distributed in four social dimensions: 1. sex: female (30), male (30) 2. age: 20-29 years old (20) 30-39 years old (20) 40-above (20) 3. duration of stay: below Mo years (30) above two years (30) 4. place of origin: south (39) north (21) The 60 informants are all selected from the mainland Chinese speech community, and they were interviewed individually. In the interview, they were presented with a card on which a few sentences were written. These sentences covered the aspect marker guo and you, respectively. The informants were asked to make a choice on these sentences and make comments freely on the 63 sentence which they did not choose (Appendix D). Such questions are "Which sentence do you choose to use?” "Why do you choose this sentence?” "What do you think of the sentence you choose, and the sentence you do not choose?” The field method used in the present study is similar to that used in Kachru's (1982) study on syntactic variation in Hindi. A pilot study was carried out to determine the sample sentences, and the appropriateness of the questions before the quantitative data collection began. It took more than one month to finish collecting the data. The sentences containing you were collected from a conversation with some native speakers of T, and also justified by several other native speakers of T. The sentences containing guo were intuitively judged by five native speakers of M. 6.4 Results 6.4.1 Age Previous studies suggest that there are social constraints on the selection of the syntactic alternative forms (Cheshire 1982, Labov1972, Montgomery 1989). The social constraints on the selection of syntactic variation are shown through a quantitative analysis. For sentence M1 and M3, the label M is given, but for sentence T2 and T4, the label T is given, in order to differentiate the distinct syntactic features from M and T. Table 19 shows the percentage of choices on M and T by age. (See next page) 64 M . I 20-29 85 15 30-39 100 0 40- 95 5 Among the 60 informants, there are three age groups: 20-29, 30- 39, and 40-above, with 20 informants in each age group. The youngest age group use T much more than the other two age groups. The choice of the 30-39 age group on T is categorical in the sense that it is a none or all phenomenon. All the informants in 30-39 age group use M, and no one uses T. For the oldest age group, the percentage of choice on M is 95, but only 5 on T. What is obvious is that all the informants use M significantly more than they use T. 6.4.2 Sex Among the 60 informants, 30 are males and 30 are females. The differences of their choices on M and T are shown in Table 20 : mum—WWW M female 97 3 male 90 10 65 Women choose M more than men do. Men choose more T than women do. However, both men and women all choose M significantly more than they choose T. 6.4.3 Duration of stay in the U.S.A. Table 21 shows the differences of choices on M and T by those who have stayed in the U.S.A. for more than two years, and those who have stayed for less than two years. .1: ' - 3| 30:0 le‘ 01“ .10 p o .010 M below two years 97 3 above two years 90 10 Among the 60 informants, 30 are in the below two year subgroup, and 30 in the above two year subgroup. Those who have stayed longer choose more T than those who have stayed shorter. Those who have stayed shorter use more M than those who have stayed longer. Still, both of the two subgroups use M significantly more than they use T. 6.4.4 Place of origin Place of origin is another social characteristics of the informants among whom 39 are from the north part of mainland China, and 21 from the south part of mainland China. The differences of their choices on M and T are shown in Table 22. (See next page) 66 .0: ' : 31-030 l0: 0.1):10 0 0: :0 001 M north 100 0 south 90 10 Those from the north show a categorical pattern with all the choice made on M, and on one from the north selects T. Those from the south use some T. Again both subgroups significantly favor M more than T. 6.5 Discussion The statistical results reveal the following generalizations: 1. All the subgroups show a significantly high percentage of choices on M, and a significantly low percentage of choices on T. In addition, there are a couple of categorical choices, with all the choices made on M and no choices made on T. Those who did not select T commented on T as follows: "I do not use T because T is not the standard Mandarin." "I do not use T because T is not grammatical." "T sounds like using English auxiliary: have/ has. In standard Mandarin you is not used as that in T." Table 23 shows the percentage of those who do not select T and make comments on T as as above. (See next page) 67 M ": :n_.o: o to := to .10 nul- I: do : 01111:: not selecting T and making the above comments 83 not selecting T and not making the above comments 10 Such evidence shows that this phenomenon is better explained in terms of syntactic variation in correlation with education level rather than in correlation with age, sex, and others. An important factor influencing the informant's choices among syntactic variables is the idea that the forms of the standard dialect is superior to other regional dialects. To them, M is considered as being the standard dialect among all the dialects. This kind of belief is suggested as prescriptive norms (Kroch & Small 1978, Lefebvre 1989). This norm is strongly expressed in the comments on T by the informants. It is clear that those who do not select T associate T with a non-standard form, and they also think that T is influenced by the English auxiliary have. The foundation of the prescriptive norm is the education level of the members in the mainland Chinese speech community. Nearly all the members in this speech community had received college education before they came to study in the United States. It is difficult to find a group of mainland Chinese speakers who have not received college education before they come to the United States, therefore, education is not listed as a social factor under investigation in the syntactic variation study ~because every one in 68 this speech community has received college education. Since putonghua is the standard form, it is used in all publications and taught in schools. As a result, those who have a college education have been exposed to this standard form since elementary schools. In mainland China, speaking the standard form is related strongly to the education level of the speaker, and the degree of education is a better indicator of linguistic behavior (Carlock & Wolck 1981). However, the prescriptive norm derived from one's education level is not present in every speech community, but it is highly developed in literate communities (Denison 1988), such as this group of mainland Chinese speech community. This norm derived from the education level controls the syntactic variation. 2. Evidence also shows that the informants are well aware of the distinct syntactic feature in M and T. In the interview questions, the majority of the informants identified T with guoyu speakers. Only a few said that they were not sure who used T. 3. In order to check possible regional dialect influences on the informant's choice of the syntactic variables, the place of origin of the informant's is asked about. The percentage scores obtained from the data in correlation with the place of origin indicate that the informants from the north part of mainland China do not choose T at all, but those from the south use some T. An interpretation of this difference could be that you used as an aspect marker is a syntactic feature in many southern dialects such as the Min dialect, and the Cantonese dialect. Further research will be needed for a better account for the dialectal differences. 69 In summary, the syntactic variables involved in the present study do not show a great sensitivity to social variables such as age, sex, et cetera, but the selection on the variants indicates their education level which influences the linguistic behavior of the mainland Chinese speech community. 70 7 Conclusion This present study consists of three parts. In part one, an attitude study on M and T among the mainland Chinese speech community is carried out. A general picture emerges in a community-wide stereotype of M speakers and T speakers based on the distinct linguistic features in the voice samples. In part two, analysis of lexical variation between M and T is listed in correlation with some social factors. The generalization shows that lexical variables are sensitive to different social factors. A general tendency is found in favor of T due to its status of prestige in the restricted association with social economic advancement of T speakers as a social group. In part three, syntactic variation between M and T is studied. A general finding is that syntactic variables, unlike lexical variables, is related to the education level of the informants. A consistent and strong tendency is found in favor of M due to the association with standardness of M. To put the studies of subjective evaluation, lexical variation, and syntactic variation together, several generalizations can be made through an overall discussion of the three studies. 1. Two different measures of language attitude are used, with an indirect measure of attitude towards M and T in the subjective evaluation study by asking the subjects to evaluate the speakers on the semantic differential scales, but with a direct measure of attitude towards M and T in the lexical variation study and the syntactic variation study by asking the informants direct questions 71 such as ”What do you think of the lexical items you choose?". The attitudes towards M and T, measured directly and indirectly, show a correlation in subjective evaluation study, lexical variation study, and syntactic variation study. In the subjective evaluation study, status and solidarity serve as two important judgmental dimensions of social attitude. In the subject's ratings along these dimensions, some of them tend to associate the relative status with T speakers, and some tend to associate status with M speakers, but nearly all of the subjects relate solidarity with M speakers who are in-group members of the subjects. This association of T with status is also shown in the lexical variation study. T is associated with a more socio- economically advanced group, thus T is considered as being prestigious restricted to socio-economic advancement. As a result, the members of the Chinese speech community show a shift in the direction of T in their lexicon. The association of M with solidarity is also consistently shown in the lexical variation study. Those who relate themselves to the mainland China life tend to keep their own typical lexical items to show solidarity with the mainland Chinese speakers. However, in the syntactic variation a different picture emerges in a favor of M rather than T due to the informant's comments that M is the correct Mandarin. It seems that the phonological and lexical features of T are far more accessible than the grammatical features of T to the members of the mainland Chinese speech community. Linguistic variables at 72 different linguistic levels are acquired with great differences in this particular speech community. This finding agrees with the study done by Ash and Myhill (1986). They find that in Philadelphia those whites who have had extensive contact with blacks show a shift in the direction of phonological and lexical variables of Black English Vernacular (BEV), but show no shift to the BEV-marked syntactic variables. They also find that those blacks who have had extensive contact with whites show a major shift in the direction of syntactic variables of standard English, but lesser shift in phonological and lexical variables of standard English. Ash and Myhill suggest that the differential acquisition of linguistic variables at different linguistic levels in a speech community involves prestige, but they have left the description about the divergence for their future research. My own research leads me to suggest that linguistic variables at different linguistic levels reflect different kinds of prestige. An important finding from the present study indicates that linguistic variables at different linguistic levels are judged differently by members of the mainland Chinese speech community. Evidence from the statistical results reveals that at the phonological level (voice samples), mainland Chinese speakers associate T with a statusful variety; at the lexical level they also favor T and give an overt association of T with prestige; but at the syntactic level they favor M and associate M with a standard variety. Further studies will be needed in order to account for this divergence in detail. 73 2. In many previous sociolinguistic studies done in the urban areas in the western countries such as the United States, Canada, and Britain, the standard variety is always associated with a higher social status group such as upper class, upper middle class, but a non-standard variety is associated with a lower social status group such as lower working class (Labov 1966, 1972; Trudgill 1974; Yaeger-Dror 1989). Linguistic variation between the standard variety and the non-standard variety is strongly correlated with social economic classes. This generalization is not applicable to mainland China where there are no social economic classes. The standard variety is associated with education level. My assumption is that the more one is educated, the more he is exposed to the standard variety, and the better he can differentiate the standard from the other varieties since the medium of education is in the standard variety. Using standard variety reflects the education level and reflects a higher social status. The prescriptive norm derived from education level seems to be the main influence on syntactic variation in the mainland Chinese speech community where nearly all the members have received college education. A different picture may emerge if the same research is conducted in a mainland Chinese speech community which is not homogenious in education level. Some studies done in the non-western world have found that education level controls the verbal repertoires of speakers, and it is the major factor influencing the speakers' linguistic choices (Jahangiri & Hudson 1982). This finding is consistent with the present syntactic variation which 74 shows little sensitivity to social factors such as age, sex, et cetera, but education level appears to be the best indicator of the mainland Chinese speakers' probable linguistic behavior. 3. In both subjective evaluation study and lexical variation study, T is regarded as a statusful variety, however, in the syntactic variation study M is regarded as a statusful variety. Linguistic variables at different linguistic levels are judged differently in the speech community. T is favored at phonological level and lexical level, but not favored at the syntactic level. This generalization is supported by the women's choice on linguistic variables. As it is well-known, women are more status-conscious than men, and women tend to use linguistic variables associated with status (Trudgill 1983). In lexical variation, T is associated with status, women use more T than M. In syntactic variation, M is considered as being statusful, women use more M, but less T. Women as a social factor are a good indicator of a status-related variety. 4. There are, however, some limitations in the lexical variation study and the syntactic variation study. For a systematic study on linguistic variation in the mainland Chinese speech community from a sociolinguistic perspective, the following should be taken into consideration for future research: A. A larger sample size should be approached. B. A stratified sample covers a group of speakers without college educaflon. C. Methodology should be improved in obtaining lexical and 75 syntactic variation data. We should try to develop a spontaneous speech which will stimulate the occurrences of lexical and syntactic variants as many as possible. In doing so, it will cost a great amount of time and energy, but will provide a better sociolinguistic data. This present study with an emphasis on linguistic variation describes the linguistic behavior of the mainland Chinese speakers from a sociolinguistic perspective. Since no published studies have been done on the linguistic behavior of this particular speech community, the results of this study have no source to be compared. It is reasonable to say that the results of this study may not be representative of any other mainland Chinese speech communities in the United States, but the results of this study as a whole will not be invalidated. APPENDIX 76 Appendix A Voice samples in the subjective evaluation study Voice sample 1 nuns, sweatshirt, me. maiEIITfio ° V ' I Z fiicitiiné‘gaeato assist-re. ran-z. magnesium. LEI EWWE, mm. W. Voive sample 3 WEEK MWWR WWW: IIWF-i? 954% $71 ‘ 8313, fiTéfiliI? Wit?! .. naawtm L... V ' 4 8:71 We items. was 3mm- :flifl’rflfiiléo iii/immersing. Voice sample 5 ma, armaments. Mfififfifif’fflifi. aria—earn gum sea. seams. lWéEWUWaEM 0 Voice sample 6 SEER—“HERE, flfiifiwig—Wo litiiltliii fiMl-‘t‘iléiiffiél 13, fiWlfio Wififlifi, WTfliBBo Wfiifim Ho fixfih‘flfifififlifio Translation of Appendix A Voice sample 1 For cooking a meal,if you have five minutes, you can cook egg with tomato. According to the servings, you may put in one or two eggs. Generally, I put in two eggs and one tomato. I slice the tomato first, then beat the egg, then put a little bit of salt in the pan. Voice sample 2. I have tasted Thai food in New York. Thai food is very delicious. It is similar to Chinese food. But in the central part of the United States such as Michigan, sea food is expensive and not fresh. Voice sample 3 That friend of mine went to apply for a job at a school. He was asked such questions: Do you have any kids? How old are they? If you teach here, will it be ok? What about your husband? Will he come here with you? Voice sample 4 When you shake the fishing pole, the end of the pole is shaking . You can feel its force. That kind of pole is good. The shaking should be strong. If the shaking is weak, it is not good. When the pole is not strong, you can not get the fish out of water because the pole will bent. Voice sample 5 I think if you want to communicate, you do not have to use that spoken language to communicate. But you are sure that you can communicate. I think so. When you communicate with that person, you will not use his language to express yourself. That is it. Voice sample 6 Another reason for me to see a movie is that seeing a movie is a kind of adjustment. For example, when you see a horrible movie, you get excited. When you get excited,you forget yourself. After that, you will feel comfortable. I do not know how to say that. 78 Appendix B Rating scales in the subjective evaluation study WIPE-n: WEIR; _: _: _: _: _: _:tf|!f 2132.. _ _. _ _. _ 4m 3m_: ______ 15R 4AIIM_; ______ ARM: 545st]; ______ at)! 68A. _. _. _. _. _. 4458A 74965.; ______ Iii 8§a_. _. _. _. _. _ .5556. 94988.; ______ 315 108158;.- _: _- __; __; _; Lawn-tit transit; _: _; _: _: _; .4883: ”“4444444mw 1 3fliii79iiifiifilltflil/ki _in JR _ai! _aima _E:alEfiI J13 14831868389138 MI: 5 1C *iflflsfifiui fi$UT 459220-29 30--39 40-- 79 Translation of Appendix 8 Please use the following scales to evaluate the speaker 1 not good- -sounding._: __:good-sounding 2 soft _: _: _: _: _: _: _: not soft 3 not elegant __ _ _ _ _: elegant 4 emotive _: _: _: _: _: _: _:not emotive 5 not intimate _: _: _: _: _: _: _: intimate 6 natural _: _: _: _: _: _: _: not natural 7 ugly _: _: _: _: _: _: _: good-looking 8 kind _: _: _: _: _: _: _: not kind 9 poor _: _: _: _: _: _: _: rich 10 reliable _: _: _: _: _: _: _: not reliable 11 not educated _: _: _: _: _: _: _: educated 12 intelligent _ _ _ _ _: not intelligent 13 Where do you think this speaker is from? _ Hong Kong _mainland China _Taiwan __Singapore _Malaysia __other 14 Check the following according to your own situation sex : M F duration of stay: below 2 years above 2 years age: 20--29 3039 40»- 80 Appendix C Questionnaires for the lexical variation study 1. Which lexical items do you choose to use? (Show the informants the card with the ten lexical items written on it) 2. Why do you choose this one? 3. What do you think of the lexical item you choose and the lexical item you do not choose? 4. How long have you been in the United States? 5. How old are you? 6. Would you like to stay in this country if it is possible? Lexical items written on the card: 1A M 96A £2} 3? 91351:. 5‘? wife SUE husband 81 Appendix D Questionnaires for the syntactic variation study I. Here are two sentences under column A and two are under column B. Which one do you choose to use? (Present the informant with the card with the sentences written on it) 2. Why do you use this sentence? 3. What do you think of the sentence you choose and the sentence you do not choose? 4. Who do you think use the sentence that you do not use? Sentences written on the card: A B lufifl 18‘an unseen. iflfiiiiifliio annexes. LIST OF REFERENCES 82 List of References Anshen,F., and Aronoff, M1989. Morphological productivity, word frequency and the Oxford English Dictionary.ln Ralph W. Fasold and Deborah Schiffrin (Eds), Language Change and Variation. Philadelphia: Amsterdam. Ash,S.,and Myhill, J. 1986. Linguistic correlates of inter-ethnic contactln David Sankoff (Ed.), Diversity and Diachrony. Philadelphia: Amsterdam. Baugh,J., and Sherzer,J.1984. Social bases of language change .In John Baugh and Joel Sherzer (Eds), Language in use: Readings in Sociolinguistics. New Jersey: Prentice-Hall, INC. Blom, J.,and Gumperz. J. J. 1986. Social Meaning in linguistic structure: code-switching in Nonlvay.ln John J. Gumperz and Dell Hymes (Eds.), Directions in Sociolinguistics. New York: Basil Blackwell INC. Bouchard, Ellen L.1969. Psycholinguistic attitude study. Studies in Language and Language Behavior. Ann Arbor: University of Michigan Press. Carlock,E.,and Wolck, W.1981. A method for isolating diagnostic linguistic variables: the Buffalo ethnolects experimentln David Sankoff and Henrietta Cedergren (Eds), Variation Omnibus. Canada :Edmonton. ' Chao, Yuen Fien.1976. Aspects of Chinese Sociolinguistics. Stanford: Stanford University Press. 83 Cheshire, Jenny, 1982. Linguistic variation and social function.ln - Suzanne Romaine (Ed.), Sociolinguistic Variation in Speech Communities. London: Edward Arnold. Variation in an English Dialect: A Sociolinguistic Study. Cambridge: Cambridge University Press. d'Anglejan, A., and Tucker, G.R.1983. Sociolinguistic Correlates of Speech style in Quebec .In: Ralph W. Fasold (Ed), Variation in the Form and Use of Language:A sociolinguistic Reader. Washington, DC. : Georgetown University Press. Denison, Norman. 1988. Language contact and language norm. Folia Linguistica, Tomus XXII, 12-35. Fasold,Ralph. 1972. Tense Marking in Black English: A Linguistic and Social Analysis. Virginia: the Center for Applied Linguistics. Forschungsprojekt, Heidelberger.1978. The acquisition of German syntax by foreign migrant workers In David Sankoff (Ed.), Linguistic Variation: Models and Methods. New York: Academic Press. Gal, Susan. 1978. Variation and change in patterns of speaking: language shift in Austria.ln David Sankoff (Ed.), Linguistic Variation: Models and Methods. New York: Academic Press Graft, D., Labov,W., and Harris, W.A.1986. Testing listeners' reactions to phonological markers of ethnic identity: 3 new method for sociolinguistic research.ln David Sankoff (Ed.),Diversity and Diachrony. Philadelphia: Amsterdam. Gumperz, John J.1964. Linguistic and social interaction in two communities. American Anthropologist 66 (6), 137-153. 84 Huang, borong ,and Liao Xudong.1985. Contemporary Chinese Language. Gansu: People's Press. Jacobson,Sven.1989. Some approaches to syntactic variation. In Ralph W. Fasold and Deborah Schiffrin (Eds.), Language Change and Variation. Philadelphia: Amsterdam. Jahangiri, N., and Hudson, R. A. 1982. Patterns of variation in Tehrani Persian.ln Romaine, S. (Ed), Sociolinguistic Variation in Speech Communities. London: Arnold. Kachru, Yamuna. 1982. Syntactic variation and language change: Eastern and Western Hindi. Studies in the Linguistic Sciences Volume 12, Number 2, Fall 1982. Kroch, A.,and Small, c.1978. Grammatical ideology and its affection speech.ln David Sankoff (Ed), Linguistic Variation: Models and Methods. New York: Academic Press. Labov, William.1966. The Social Stratification of English In New York City. Washington,D.C.: Center for Applied Linguistics. 1969. Contraction, deletion and inherent variability of the Language, 45: 715-62. __1972. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. __1972. Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press. 1981. What can be learned about change in progress from synchronic description.ln David Sankoff and Henrietta Cedergren (Eds.),Variation Omnibus. Canada: Edmonton. 85 _1984. Field methods of the project on linguistic change and variatioan John Bauch and Joel Sherzer (Eds), Language in Use: Readings in Sociolinguistics. New Jersey: Prentice-Hall, INC. 1989. The exact description of a speech community: short a in Philadelphialn Ralph W. Fasold and Deborah Schiffrin (Eds.),Language Change and Variation. Philadelphia: Amsterdam. Lambert, W. E., Anisfeld ,M.,and Yeni-komshian, 6.1965. Evaluation reactions of Jewish and Arab dolescents to dialect and language variation. Journal of Personality and Social Psychology 2, 84-90. Lambert, W. E. 1967. A social psychology of bilingualism. Special Issue of the Journal of Social Issues 23 (2). Lambert, W.E., R.C. Hodgson,and R.C. Gardner.1960. Evaluations to spokenlanguages. Journal of Abnormal and Social Psychology Vol. 60, No.1, 45-51. Lefebvre, Claire.1989. Some problems in defining syntactic variables: the case of wh-questions in Montreal French.ln Ralph W. Fasold and Deborah Schiffrin (Eds), Language Change and Variation. Philadelphia: Amsterdam. Ljung, Magnus. 1989. Social determinants of the use of English In Sweden. In Ralph W. Fasold and Deborah Schiffrin (Eds), Language Change and Variation. Philadelphia: Amsterdam. Mier, Jeanne Zang. 1987. A case of lexical diffusion in a language contact situation. Studia Linguistica 41. 1, 72-82. 86 Montgomery, Michael. 1989. Choosing between that and it.ln Ralph W. Fasold and Deborah Schiffrin (ed), Language Change and Variation. Philadelphia: Amsterdam. Preston, Dennis R.,1989. Sociolinguistics and Second Language Acquisition. ‘ New York: Basil Blackwell INC. Richford, John.1983. Carrying the new wave into syntaxJn Ralph W. Fasold (Ed.), Variation in the Form and Use of Language. Washington, D.C.: Georgetown University Press. 1987. language contact, variation and diffusion: microlevel community perspectivesGeorgetown University Round Table on Languages and Linguistics. Washington,D.C.: Georgetown University Press. Russell, Joan. 1982. Networks and sociolinguistic variation in an African urban setting.ln Suzanne Romaine (Ed.), Sociolinguistic Variation In speech Communities. London: Edward Arnold. Samarin, William J.1972. Lingua francas of the world.ln Joshva A. Fishman (Ed), Reads in the Sociology of Language. Paris: Mouton. Sankoff, D.,Thibault,P.,and Bembe, H.1978. Semantic Field Variabilitan David Sankoff (Ed.), Linguistic Variation: Models and Methods. New York: Academic Press. Shuy, R.,and W1lliams,F.1973. Stereotyped attitudes of selected English dialect communitiesln Roger W. Shuy and Ralph W Fasold (Eds), Language Attitudes: Current Trends and Prospects. Washington, D.C.: Georgetown University Press. Sobin, Nicholas1982. Texas Spanish and lexical borrowing.ln J. Amastae and Elias- Olivares (Eds), Spanish in the United States: Sociolinguistic Aspects. Cambridge: Cambridge University Press. 87 Tousignant, C.,and Sankoff, 0.1989. New results on Montreal French /r/.ln Ralph W. Fasold and Deborah Schiffrin (Eds), Language Change and Variation. Philadelphia: Amesterdam. Trudgill, Peter, 1974. The Social Differentiation of English in Norwich. London: Cambridge University Press. 1983. On Dialect: Social and Geographical Perspectives. Oxford: Basil Blackwell. 1984. Sex and Covert Prestige: Linguistic Change in the Urban Dialect of Norwich. In: John Bauch and Joel Sherzer (Eds), Language in the Use: Readings in Sociolinguistics. New Jersey: Prentice-HaII,INC. 1986. Dialects in Contact. New York: Basil Blackwell INC. degill,P., and Hannah, J., 1982. International English: a Guide to Varieties of Standard English. London: Edward Arnold. Wang, Honglian.1987. Basic Knowledge on Chinese Language. Beijing: Beijing Teacher's College Press. Weinreich, U., Labov ,W.,and Herzog, M.l.1975. Empirical foundations for a theory of language change. In W.P. Lehmann and Yakov Malkiel (Eds), Directions for Historical Linguistics. Austin: University of Texas Press. Weinreich, Uriel. 1968. Languages in Contact. Mouton: the Hague. \Mlliams, F., Whitehead ,J.L.,and Miller, L.1972. Relations between language attitudes and teacher expectancy. American Educational Research Journal 9, 263-277. Williams, Frederick1970. Psychological correlates of speech characteristics: on sounding " Disadvantaged”. Journal of Speech and Hearing Research 13, 472-488. 88 1983. Some research notes on dialect attitudes and stereotypesln Ralph W. Fasold (Ed.), Variation in the Form and Use of Language: A sociolinguistic Reader. Washington,D.C.: Georgetown University Press. Wolfram, Walter. 1969. A Sociolinguistic Description of Detroit Negro Speech. Washington, DC. :Center for Applied Linguistics. Wu, Zhiqian and Gordon, E.1987. Language variation in Shanghai. Working Papers in Language and Linguistics July, 21, 54-65. Yaeger-Dror, Malcah.1989. Patterned symmetry of shifting and lengthened vowels in the Montreal French Vernacular (MFV). In Ralph W. Fasold and Deborah Schiffrin (Eds), Language Change and Variation. Philadelphia: Amsterdam. I‘IICHIGRN STATE UNIV. LIBRRRIES lllll III III llll III III Ill ll llll |||| llllll ll lllll lllll Illll 31293008764890