Simmons 2017 Gwoyeu Romatzyh 國語羅馬字
Simmons 2017 Gwoyeu Romatzyh 國語羅馬字
Volume 4
Shā–Z
Associate Editors
Wolfgang Behr
(University of Zurich)
Yueguo Gu
(Chinese Academy of Social Sciences)
Zev Handel
(University of Washington)
C.-T. James Huang
(Harvard University)
James Myers
(National Chung Cheng University)
General Editor
Rint Sybesma
Associate Editors
Wolfgang Behr
Yueguo Gu
Zev Handel
C.-T. James Huang
James Myers
LEIDEN • BOSTON
2017
All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval
system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or
otherwise, without prior written permission from the publisher. Authorization to photocopy items for
internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid
directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.
Fees are subject to change.
should be flexible and practical: j [ʨ], tc [ʨʻ] 1926; Chao 1976:14). But it was not until two years
(later ch), c [ɕ] (later sh) can represent pala- later, in September 1928, that the Institute of
tals before high front vowels [i] and [y] and Higher Education (Dàxuéyuàn 大學院) offfijicially
retroflexes elsewhere. announced that Gwoyeu Romatzyh should serve
21. The Romanization should use as many of the as an offfijicial form of transcription along with
26 letters of the alphabet as possible. the National Phonetic Alphabet—zhùyīn zìmǔ
22. It should use “voiced” consonants to rep- (Luó 1934:36).
resent unaspirated voiceless initials (for An outline of the fijinal version of GR and its
example b, d, g can represent [p, t, k])—in tonal spelling system is provided at the end of
part to achieve principle #21, and also to this article. In representing tonal distinctions by
increase readability by making distinctions spelling rules and avoiding diacritics altogether,
more visually prominent. GR “gives an individuality to the physiognomy
24. It should keep the morphemes/syllables of word, with which it is possible to associate
of words together as one unit-for example meaning in a way not possible with tone-signs
inyueh (yīnyuè) 音樂 ‘music’ not in yueh. added as an afterthought” (Chao 1964:11). The
25. Tones should be spelled as part of the word. system that Chao and his colleagues designed
does not simply represent each of the four tones
Notably, however, this fijirst draft was still by tacking on a letter that was otherwise silent
designed for the mixed lán-qīng Mandarin stan- (having no vowel or consonantal pronuncia-
dard: rù tone syllables were marked with a fijinal tion), as in Lín Yǔtáng’s earliest system noted
h plus an apostrophe—hʼ; and the dental sibilant above. The original insight that silent letters
initials z [ʦ] (later tz), ts [ʦʻ], s [s] were used could mark tone was retained for some sylla-
before fijinals with high-front vowels in jiānyīn bles in GR such as the -r following vowels to
尖音 syllables to preserve the jiān-tuán distinction. mark 2nd tone and the syllable fijinal -h to mark
But the mixed Mandarin standard was proving 4th tone. But the GR designers realized that
impossible to implement and meeting stifff resis- (a) tone did not always need to be marked with
tance. In 1920 Zhāng Shìyī 張士一 (1886–1969) letters dedicated to that single task alone, and
penned an influential call to use the language of (b) a given tone did not have to be marked the
Běijīng as the standard for Mandarin pronuncia- same way in all cases. They found that letters
tion; and in 1923 the Ministry of Education fijinally could do double duty, marking both consonant
offfijicially designated Běijīng as the national stan- and tone, or vowel and tone and that the tone
dard (Lí 1934:96, 171). Over the next few years, marking feature could be incorporated into the
successive drafts of GR were subject to much dis- initial in some cases.
cussion by Chao and other scholars who formed Thus in GR syllables that begin with an
a committee that was convened to study the unchanged medial or main vowel i or u (a form
matter: Guóyǔ Luómǎzì wěiyuánhuì 國語羅馬 of zero initial) are 1st tone; while syllables where
字委員會 ‘Committee on the National Roman- i and u are prefijixed by y and w (respectively)
ization’. This committee worked intensively on are 2nd tone (unless marked as 3rd or 4th tone).
GR in 1925 and 1926, bringing it to its mature Similarly, syllables beginning with sonorant ini-
and fijinal form and making it strictly reflective tials m, n, l, and r are 2nd tone (unless otherwise
of the new Běijīng-based standard. They noti- marked), which incidentally has a basis in actual
fijied the Ministry of Education of the results of history in that 2nd tone naturally developed in
their effforts in September 1926 at which time words with those initials. Manipulation of the
Y.R. Chao issued a description of the polished vowels marks tone by changing the medial (as
version of GR and named the seven principals in from -ian to -ean for 3rd tone), by changing the
the discussions: Y.R. Chao, Lí Jǐnxī 黎錦熙, offf-glide (as from -iou to -iow for 4th tone) or the
Lín Yǔtáng, Líu Fù 劉復, Qián Xuántóng main vowel (as from -in to -yn for 2nd tone), or
錢玄同, Wāng Yí 汪怡, and Zhōu Biànmíng by doubling the main vowel (as from -a to -aa
周辨明 (Luó 1934:25–30, 35; Lí 1934:166–170; Chao for 3rd tone).
The result of their effforts created multiple represents the vowel [y]; and fijinal h marks 4th
ways of representing a tone depending on the tone in tzyh. Additionally, Roma (the Italian
specifijic syllable structure. 1st tone is variously spelling of “Rome”) illustrates how GR allows the
marked with zero initial, no change in the fijinal, unadjusted incorporation of foreign words that
and adding h after sonorant initials, as lh- and are in Roman orthography.
rh-; 2nd tone is marked with w or y at the onset Collectively, the various ways of represent-
of the fijinal, sonorant initials, and r following ing tone in GR lends the system an organic feel
the vowel(s); 3rd tone is marked by doubling the and one that is elegant though subtly complex.
vowel or changing the vowel that represents the The overall structure seems to reflect a natu-
medial or the offf-glide, as -aa, -ean, -ae; 4th tone ral, historical development, thus fulfijilling Chao’s
is marked by changes at the end of a syllable, 3rd principle: Strive to match natural language
such as fijinal q in -nq, doubling a fijinal -n, or a trends.
fijinal -w, -y or -h. The following are examples of GR’s naturalness is enhanced by following
difffering ways to write the tones depending on Chao’s 24th principle, which called for the mor-
the syllable: phemes and syllables of words to be spelled
together as one unit. Through Y.R. Chao’s care-
ful and systematic analysis, GR was a leading
1st 2nd 3rd 4th pioneer in the application of this principle, for
Tone Tone Tone Tone which he provided the following guide (Chao
貪 彈 毯 炭 1935–1936): (a) Mid-word vowels in multi-
GR tan tarn taantann syllable words belong to the syllable on the left-
Pīnyīn tān tán tǎn tàn side where possible and mid-word consonants
English corrupt pluck blanket
charcoal in multi-syllable words belong to the syllable
优 油 有 右 on the right where possible, as in jiaren ( jiārén)
GR iou you yeouyow 佳人 ‘beautiful woman’. If (a) fails to separate
Pīnyīn yōu yóu yǒu yòu the syllables correctly, then (b) insert a hyphen
English superior oil to have
right between syllables, as in fan-ann ( fān’àn) 翻案
(side) ‘overturn a verdict’, to distinguish it from fanann
輕 情 請 慶 ( fānàn) 發難 ‘revolt’. Chao also provided an
GR ching chyng chiing chinq additional 10 guidelines for cases that remained
Pīnyīn qīng qíng qǐng qìng ambiguous after applying the above formula.
English light sentiment invite celebrate GR was also pioneering in the use of “voiced”
consonant letters to represent unaspirated
voiceless initials—see Chao’s 22nd principle
Beyond simple indication of tone, there are above. The issues involved with this choice are
other orthographic advantages to the vowel thoroughly explored in Chao (1931:108). Though
and consonant changes used in GR. The vowel this practice is now commonly accepted, and
manipulation reduces the number of letters in was adopted in pīnyīn, it had not been com-
some syllables (as -ao instead of -aao for 3rd mon before Chao advocated it. For example, the
tone), as does the change in syllable endings (as widely used Wade-Giles system marked aspira-
fijinal w replacing u, and fijinal nq replacing ng, for tion with an apostrophe and otherwise wrote
4th tone). The multi-function role of initial y and the unaspirated series with the same letters: p’, t’,
w also helps to clarify morpheme boundaries. k’ versus p, t, k. Hence Chao’s method had many
Some of these advantages are illustrated by the detractors in the beginning. Bernhard Karlgren
name for the system itself, “Gwoyeu Romatzyh”, (1928:18), for instance, strongly objected to using
which is written in GR: change of medial u p, t, k for voiceless aspirates and b, d, g for voice-
to w marks 2nd tone in Gwo; y marks the syl- less unaspirated initials in teaching and learning
lable boundary in yeu, while the e marks 3rd Chinese, but conceded it was an acceptable solu-
tone—changed from the i in the digraph iu that tion for a native orthography.
Though the National Phonetic Alphabetic (Chao 1964:11)—though McGinnis (1997) has
and later pīnyīn both maintain separate repre- provided credible empirical evidence to dispute
sentations for Mandarin voiceless retroflex and that claim.
palatal afffricates and fricatives, both groups are The earliest practical use made of GR was in
represented with the same set of initials in GR: j, the revised dictionaries issued by the Ministry of
ch, sh. According with Chao’s 4th and 12th prin- Education after the adoption of the Běijīng stan-
ciples, as these initial sets are in complemen- dard, such as the 1932 Guóyīn chángyòng zìhuì
tary distribution and never occur with the same 國音常用字彙, the fijirst offfijicial dictionary in
fijinals, the distinction between them in GR is which GR was used together with the National
determined by the following vowel. Before fijinals Phonetic Alphabet (by then renamed zhùyīn
that begin with the high-front vowels [i] and fúhào 注音符號 ‘phonetic symbols’) to gloss pro-
[y]—i.e., all the fijinals in Rows -i and -iu in Table nunciation. The National Language dictionaries
1 below—they represent palatals: afffricates [ʨ] (Guóyǔ cídiǎn 國語詞典) continued to include
by j and [ʨʻ] by ch, and the fricative [ɕ] by sh. GR until the Republic of China government in
Before all other fijinals—those of Rows -a and -u Táiwān dropped it as a standard in the mid 1980’s
in Table 1—they represent the retroflexes: afffri- (Ministry of Education 1986). Other dictionar-
cates [tʂ] by j and [tʂʻ] by ch, and the fricative [ʂ] ies that used GR are Walter Simon’s Gwoyeu
by sh. The voiced retroflex fricative [ʐ] is spelled Romatzyh Chinese-English vocabulary (Simon
with r in GR and also only occurs with Row -a 1944, a page from which is shown in Figure 1)
and Row -u fijinals. and Y.R. Chao’s Concise Dictionary of Spoken
Aside from tonal spelling, one of the most Chinese (Chao and Yang 1947). Lin Yutang’s 1972
characteristic features of GR is the use of y to dictionary is sometimes cited; but it used a mod-
represent the alveolar apical and retroflex apical ifijied tonal spelling that is markedly less complex
vowels, [ɿ] and [ʅ] respectively. Both are thus than GR. In textbooks, Walter Simon enthu-
easily distinguished from [i], which is written i. siastically adopted GR in England for the Chi-
The distinction between the two is determined nese Sentence Series (Simon and Lu 1944), as did
by the initial: y is [ɿ] following the dental sibi- Y.C. Liu (1960). In the United States, Y.R. Chao
lants tz [ʦ], ts [ʦʻ], and s [s]; it is [ʅ] following the of course also used it in his Mandarin Primer
retroflex j, ch, sh, and r. Another characteristic (1948) and his Sayable Chinese volumes (1968);
innovation is the use of the digraph iu to repre- and GR was long used at Princeton University,
sent the vowel [y] (pīnyīn ü). Though criticized, where Ta-tuan Ch’en, et al. also compiled a Chi-
for example Karlgren thought it a “phonetic lie” nese textbook that used GR even in the late 20th
(1928:20), it proved to be neatly adaptable to century (1994). Beyond textbooks, a few scholars
tonal spelling rules (see Row -iu in Table 1). have occasionally adopted Gwoyeu Romatzyh
Champions of GR had hoped that it would not in their writing on Chinese, for example Harry
merely be used as a transcription for the citation F. Simon (1958), Chao (1968), Simmons (1999),
of Chinese names and words in foreign texts, and Branner (2000). GR also found sporadic use
but that it would also help to promote a single in the Romanization of personal names; those of
national language and pronunciation, serve to the mathematician Shiing-Shen Chern 陳省身
advance literacy, and even develop into a liv- (Chén Xǐngshēn) and the linguist Gong Hwang-
ing orthography that could eventually replace cherng 龔煌城 (Gōng Huángchéng) are two
the Chinese characters (DeFrancis 1950:82; Chao well-known examples.
1964:11; Chen 1999:183). Ultimately these hopes Though GR survived as a dictionary tran-
were not realized. But GR did have a fairly sig- scription in Táiwān until dropped in 1986, it
nifijicant life in the classroom as a tool for foreign was essentially ignored on the mainland after
learners of Chinese, where Y.R. Chao believed the founding of the People’s Republic of China,
that tonal spelling “proved in practice to be where Hànyǔ pīnyīn 漢語拼音 was offfijicially
a most powerful aid enabling the student to adopted in 1958 (Zhōu 2003:103–104) (→ Tran-
grasp the material with precision and clearness” scription Systems: Hànyǔ Pīnyīn). Gwoyeu
Romatzyh thus never gained a foothold anywhere when pīnyīn transcriptions are spelled without
other than for the limited uses described above tone marks, which has become the common
and today has been abandoned for all practical practice. Thus the Chinese province of Shǎnxī
purposes. Surely the complexity of the GR tonal is spelled Shaanxi to distinguish it from Shānxī
spelling system—the learning curve for which in diacritic-free pīnyīn. In another example, the
met with much criticism and resistance, and no pre-Qín state of Wèi 衛 is often spelled with
doubt also caused much confusion—rendered the GR 4th tone spelling—Wey—to distinguish
the efffort to widely promulgate the system an it from the Wèi 魏 state with which it coex-
uphill battle. Yet, the primary cause of GR’s isted in the 3rd and 4th centuries BCE (though
failure is no doubt because, as Jerry Norman here the ambiguity exists even in pīnyīn with
concluded: “it never received the offfijicial sup- the tone marks, as it would if GR were used
port it would have required to succeed; per- for both).
haps more importantly it was viewed by many Beyond the Romanization of Mandarin,
as the product of a group of elitist enthusiasts variations of GR-type tonal spelling have been
and lacked any real popular base of support” used for the transcription of Chinese dialects.
(Norman 1988:259–260). Y.R. Chao included examples of tonal spelling
Tonal spelling now only survives in the odd for six dialects besides Běijīng in his 1922 article
instance where unacceptable ambiguity arises “Studies on National Language Romanization”:
Row-a ᴀ a a ar aa ah
Finals ɤ e e er ee eh
z̩ (ɿ), ɹ̩ (ʅ) -i y yr yy yh
ɚ er el erl eel ell
ai ai ai air ae ay
ei ei ei eir eei ey
ɑu ao au aur ao aw
ou ou ou our oou ow
an an an arn aan ann
ən en en ern een enn
ɑŋ ang ang arng aang anq
ʌŋ eng eng erng eeng enq
ɷŋ ong ong orng oong onq
Row-i i i yi i yi ii yii ih yih
Finals iᴀ ia ya ia ya ea yea iah yah
iai (yai)
iɑu iao yao iau yau eau yeau iaw yaw
iɛ ie ye ie ye iee yee ieh yeh
iou iu you iou you eou yeou iow yow
iɛn ian yan ian yan ean yean iann yann
in in yin in yn iin yiin inn yinn
iɑŋ iang yang iang yang eang yeang ianq yanq
iŋ ing ying ing yng iing yiing inq yinq
iɷŋ iong yong iong yong eong yeong ionq yonq
Row-u u u wu u wu uu wuu uh wuh
Finals uᴀ ua wa ua wa oa woa uah wah
uɤ uo wo uo wo uoo woo uoh woh
labial+uɤ o o or oo oh
uai uai wai uai wai oai woai uay way
uei ui wei uei wei oei woei uey wey
uan uan wan uan wan oan woan uann wann
uən un wen uen wen oen woen uenn wenn
uɑŋ uang wang uang wang oang woang uanq wanq
uʌŋ ueng weng ueng (woeng) wenq
Row-iu y ü yu iu yu eu yeu iuh yuh
Finals yɛ üe yue iue yue eue iueh yueh
yan üan yuan iuan yuan euan yeuan iuann yuann
yn ün yun iun yun eun yeun iunn yunn
Alphabet], Shànghǎi 上海: Shāngwù 商務印書館, the National Language’, Guóyǔ luómǎzì 國語羅
1934. 馬字), formerly offfijicial transcription systems for
McGinnis, Scott, “Tonal Spelling versus Diacritics for
Teaching Pronunciation of Mandarin Chinese”, The
the National Language Guóyǔ 國語 of Repub-
Modern Language Journal 81/2, 1997, 228–236. lican China, became obsolete in the People’s
Norman, Jerry, Chinese, Cambridge: Cambridge Uni- Republic of China.
versity Press, 1988. Before this decision was taken, there had been
Simmons, Richard VanNess, Chinese Dialect Classifiji-
discussions for several years about which kind
cation: A Comparative Approach to Harngjou, Old
Jintarn, and Common Northern Wu, Amsterdam: of system should be selected from among the
John Benjamins, 1999. so-called systems of national form like Zhùyīn
Simon, Harry F., “Some Remarks on the Structure of zìmǔ (systems whose graphs were derived from
the Verb Complex in Standard Chinese”, Bulletin of Chinese characters or from their graphical
the School of Oriental and African Studies 21, 1958,
553–577. elements) and systems based on the Latin or
Simon, Walter, Gwoyeu Romatzyh Chinese-English even Cyrillic alphabet. But after 1955, reform-
Vocabulary: Being a Glossary to the Chinese Sentence ers concentrated on several drafts based on
Series, London: Arthur Probsthain, 1944. the Latin alphabet and fijinally presented the
Simon, Walter and C.H. Lu, Chinese Sentence Series:
First Fifty Lessons, part I: Text in Gwoyeu Romatzyh Hànyǔ pīnyīn fāng’àn 漢語拼音方案 (‘Chinese
with Translation, London: Arthur Probsthain, 1942. Spelling Scheme’), which is the offfijicial name for
Wú Zōngjì 吳宗济 and Zhào Xīnnà 赵新那, eds., this system (Wú 1978:151–155; Zhōu 1979:49–53;
Zhào Yuánrèn yǔyán lùnwén jí 赵元任语言学论文 Chappel 1980; Zhōu 1995:1–17; DeFrancis 2006,
集 [Linguistic essays by Yuenren Chao], essays in
Chinese, Běijīng 北京: Shāngwù 商务印书馆, 2002. Sū 2010:252–270).
Wu Zongji and Zhao Xinna, eds., Linguistic Essays by The reasons for the selection of a Latin-based
Yuenren Chao, essays in English, Beijing: Shangwu, alphabet instead of one of national form are
2006. reported as the following (Wú 1978:154–155,
Zhōu Yǒuguāng 周有光, The Historical Evolution of
Chinese Languages and Scripts, Zhōngguó yǔwén de
Zhōu 1979:82–89, 146–150, Zhōu 1995:12–14, Sū
shídài yǎnbiàn 中国语文的时代演进, Columbus: 2010:252–257, 266–267):
National East Asian Languages Resource Center,
2003.
• After Zhùyīn zìmǔ was devised (1913) and
offfijicially promulgated (1918) in early Repub-
Richard VanNess Simmons lican China, more Chinese had become
accustomed to the Latin alphabet in subse-
quent decades. This resulted from the use of
Transcription Systems: Gwoyeu Romatzyh and Latinxua Sin Wenz
Hànyǔ pīnyīn 漢語拼音 (‘New Latinized Script’, Lādīnghuà xīn wénzì
拉丁化新文字) in the 1930s and 1940s and
Hànyǔ pīnyīn (‘Chinese Spelling’) romanization increasing study of English and other West-
was drafted in the People’s Republic of China ern languages, and because of the common
and was offfijicially introduced by a resolution of usage of Latin letters as symbols in scientifijic
the People’s Congress on February 11, 1958. Its and mathematical notation;
implementation was situated in the context of
language and script reforms during the 1950s
• The use of a Latin-based system would facili-
tate international communication;
devised by the new Chinese government. The
development of a new system for the represen-
• The Latin alphabet was a system whose efffec-
tiveness was evident from experience during
tation of the sounds of standard Chinese was long historical periods and in many countries
one of the main tasks of those reforms, besides all over the world. It was an internationally
the reform (simplifijication) of Chinese charac- recognized instrument, which like the Arabic
ters and popularization of the standard Chinese numerals was applicable in any nation of the
language (Pǔtōnghuà 普通話). world, and therefore its adoption would not
With the promulgation of Hànyǔ pīnyīn, the damage the patriotic feelings of the Chinese;
“Phonetic Alphabets”, Zhùyīn zìmǔ 注音字母
and Gwoyeu Romatzyh (‘Roman Alphabet for
• By the selection of a Latin-based offfijicial system,
the necessity of putting two diffferent systems