Spelling in Gwoyeu Romatzyh

GR, like Pinyin, uses contrasting unvoiced/voiced pairs of consonants to represent aspirated and unaspirated sounds in Chinese. For example b and p represent IPA [p] and [pʰ] (p and p' in Wade-Giles). Another feature of GR surviving in Pinyin is the representation of words (usually of two syllables) as units: e.g. Gwoyeu rather than the Wade-Giles Kuo²-yü³.

The basic features of GR spelling are shown in the following tables of initials and finals, the latter referring to the basic T1 forms.[1] Many of the spelling features are the same as in Pinyin; differences are highlighted in the tables and discussed in detail after the second table. The rules of tonal spelling follow in a separate section.

In the tables Pinyin spellings are given only where they differ from GR, in which case they appear in (parentheses). The tables also give the pronunciation in [brackets].

Finals

		Coda
		∅			/i/		/u/		/n/		/ŋ/			/ɻ/
Medial	∅	-y [ɨ] ㄭ (-i)	e [ɤ] ㄜ	a [a] ㄚ	ei [ei] ㄟ	ai [ai] ㄞ	ou [ou] ㄡ	au [au] ㄠ (ao)	en [ən] ㄣ	an [an] ㄢ	ong [ʊŋ] ㄨㄥ	eng [əŋ] ㄥ	ang [aŋ] ㄤ	el [aɚ] ㄦ (er)
	/j/	i [i] ㄧ	ie [je] ㄧㄝ	ia [ja] ㄧㄚ		iai [jai] ㄧㄞ	iou [jou] ㄧㄡ (iu)	iau [jau] ㄧㄠ (iao)	in [in] ㄧㄣ	ian [jɛn] ㄧㄢ	iong [jʊŋ] ㄩㄥ	ing [iŋ] ㄧㄥ	iang [jaŋ] ㄧㄤ
	/w/	u [u] ㄨ	uo/o [wo] ㄨㄛ	ua [wa] ㄨㄚ	uei [wei] ㄨㄟ (ui)	uai [wai] ㄨㄞ			uen [wən] ㄨㄣ (un)	uan [wan] ㄨㄢ		ueng [wəŋ] ㄨㄥ	uang [waŋ] ㄨㄤ
	/ɥ/	iu [y] ㄩ (ü)	iue [ɥe] ㄩㄝ (üe)						iun [yn] ㄩㄣ (ün)	iuan [ɥɛn] ㄩㄢ (üan)

Key

GR differs from Pinyin

Note: In this section the word "tone" is abbreviated as "T": thus T1 stands for Tone 1, or first tone, etc.

Wherever possible GR indicates tones 2, 3 and 4 by respelling the basic T1 form of the syllable, replacing a vowel with another having a similar sound (i with y, for example, or u with w). But this concise procedure cannot be applied in every case, since the syllable may not contain a suitable vowel for modification. In such cases a letter (r or h) is added or inserted instead. The precise rule to be followed in any specific case is determined by the rules given below.[2]

A colour-coded rule of thumb is given below for each tone: the same colours are used below in a list of provinces. Each rule of thumb is then amplified by a comprehensive set of rules for that tone. These codes are used in the rules:

• V = a vowel
• NV = a non-vowel (either a consonant or zero in the case of an initial vowel)
• ⇏ = "but avoid forming [the specified combination]"

Pinyin equivalents are given in brackets after each set of examples. To illustrate the GR tonal rules in practice, a table comparing Pinyin and GR spellings of some Chinese provinces follows the detailed rules.

Tone 1: basic form

• Initial sonorants (l-/m-/n-/r-): insert -h- as second letter. rheng, mha (rēng, mā)
• Otherwise use the basic form.

Tone 2: i/u → y/w; or add -r

• Initial sonorants: use basic form. reng, ma (réng, má)
• NVi → NVy ( + -i if final). chyng, chyan, yng, yan, pyi (qíng, qián, yíng, yán, pí)
• NVu → NVw ( + -u if final). chwan, wang, hwo, chwu (chuán, wáng, huó, chú)
• Otherwise add r to vowel or diphthong. charng, bair (cháng, bái)

Tone 3: i/u → e/o; or double vowel

• Vi or iV → Ve or eV (⇏ee). chean, bae, sheau (qiǎn, bǎi, xiǎo), but not gee
• Vu or uV → Vo or oV (⇏oo). doan, dao, shoei (duǎn, dǎo, shuǐ), but not hoo
• When both i and u can be found, only the first one changes, i.e. jeau, goai, sheu (jiǎo, guǎi, xǔ), not jeao, goae, sheo
• For basic forms starting with i-/u-, change the starting i-/u- to e-/o- and add initial y-/w-. yean, woo, yeu (yǎn, wǒ, yǔ)
• Otherwise double the (main)[3] vowel. chiing, daa, geei, huoo, goou (qǐng, dǎ, gěi, huǒ, gǒu)

Tone 4: change/double final letter; or add -h

• Vi → Vy. day, suey (dài, suì)
• Vu → Vw (⇏iw). daw, gow (dào, gòu), but not chiw
• -n → -nn. duann (duàn)
• -l → -ll. ell (èr)
• -ng → -nq. binq (bìng)
• Otherwise add h. dah, chiuh, dih (dà, qù, dì)
• For basic forms starting with i-/u-, replace initial i-/u- with y-/w-, in addition to the necessary tonal change. yaw, wuh (yào, wù)

Neutral tone (轻声 Chingsheng / qīngshēng)

A dot (usually written as a period or full stop) may be placed before neutral tone (unstressed) syllables, which appear in their original tonal spelling: perng.yeou, dih.fang (péngyou, dìfang). Y.R. Chao used this device in the first eight chapters of the Mandarin Primer, restricting it thereafter to new words on their first appearance. In A Grammar of Spoken Chinese he introduced a subscript circle (_o) to indicate an optional neutral tone, as in bujy_odaw, "don't know" (Pinyin pronunciation bùzhīdào or bùzhīdao).

It is important to note that any GR syllables beginning u- or i- must be T1: in T2, T3 and T4 these syllables all begin with w- or y- respectively. An example in all four tones is the following: ing, yng, yiing, yinq (Pinyin ying).

Erhua (兒化), or the rhotacized or retroflex[5] ending, is indicated in GR by -el rather than -r, which is already used as a T2 marker. The appropriate tonal modification is then applied to the rhotacized form: for example shell (shìr) and ideal (yìdiǎnr).[6]

Most other romanization systems preserve the underlying form, but GR transcribes the surface form as pronounced. These are the main principles followed when a syllable is rhotacized in GR:[7]

1. -el replaces final -y
2. -l replaces final -i (in -ai and -ei) and -n, but -in becomes -iel
3. -el is added to -i and -iu
4. -l is added to all other finals, but -ing acquires an additional -e- to become -iengl.

As a consequence the one-to-one correspondence between GR and Pinyin is broken, since one GR rhotacized form may correspond to several Pinyin forms.[8] For example:

• jiel corresponds to both jīr and jīnr (both pronounced [t͡ɕjɚ˥]).
• jial corresponds to both jiār and jiānr (both pronounced [t͡ɕjaɚ̯˥]).

The most important manifestation of tone sandhi in Mandarin is the change of a T3 syllable to T2 when followed by another T3 syllable (T3 + T3 → T2 + T3). GR does not reflect this change in the spelling: the word for "fruit" is written shoeiguoo, even though the pronunciation is closer to shweiguoo.[9] Four common words with more complicated tone sandhi (also ignored in the spelling) are mentioned below under Exceptions.

A number of frequently-occurring morphemes have abbreviated spellings in GR.[10] The commonest of these, followed by their Pinyin equivalents, are:

• -g (-ge)
• -j (-zhe)
• -m (-me)

occurs in sherm (shénme), jemm/tzemm (zhème) and tzeem (zěnme)

• sh (shi)

also in compounds such as jiowsh (jiùshi), dannsh (dànshi), etc.

• -tz (-zi)

In its original form GR used the two "spare" letters of the alphabet, v and x, to indicate reduplication. This mimicked the method by which the Chinese script indicates repeated characters with an iteration mark (々). In GR the letter x indicates that the preceding syllable is repeated (shieh.x = shieh.shieh, "thank you"), vx being used when the preceding two syllables are repeated (haoshuo vx! = haoshuo haoshuo! "you're too kind!").[11]

This concise but completely unphonetic, and hence unintuitive, device appears in Chao's Mandarin Primer and all W. Simon's texts (including his Chinese-English Dictionary). Eventually, however, it was silently discarded even by its inventor: in Chao's Grammar as well as his Sayable Chinese all reduplicated syllables are written out in full in their GR transcription.

The following words and characters do not follow the rules of GR:

• The name Romatzyh (which strictly speaking should be "Luomaatzyh") follows international usage (Roma).
• The characters 一 ("one"), 七 ("seven"), 八 ("eight"), and 不 ("no/not") are always written i, chi, ba, and bu, respectively, regardless of the tone in which they are pronounced. In other words changes due to tone sandhi are not reflected in GR.

• Chao, Yuen Ren (1948). Mandarin Primer: an Intensive Course in Spoken Chinese. Harvard University Press.
• Chao, Yuen Ren (1968a). A Grammar of Spoken Chinese. University of California Press. ISBN 0-520-00219-9.
• Simon, Walter (1947). A Beginners' [sic] Chinese-English Dictionary. Lund Humphries & Co. Ltd.