Wiktionary:About Chinese/Wu

Wu is a subdivision of Chinese, spoken by about 80 million people. It is spoken to southern Jiangsu and most of Zhejiang, an area sometimes referred to as Jiangnan. The historically wealthy Lower Yangtze Delta is home to the largest branch, Northern Wu, and is also the lect area that is represented on Wiktionary. Suzhounese (zh) was the cultural capital of the area, though due to Shanghai's relative economic status, Shanghainese (zh) has, to some extent, eclipsed Suzhounese's prestige status. Characteristically Wu varieties have voiced (濁／浊 (⁸zoq, “muddy”)) obstruents, and all except Oujiang varieties such as Wenzhounese have final glottal stops derived from the final stops of Middle Chinese. Together, these features reduce the number of tones that are actually distinctive or phonemic.

Although no official romanization system is endorsed by any government, on Wiktionary, both of these lects are notated in Wugniu, which is the most common standardised system used in online circles nowadays. Wugniu is also available in a wide variety of localities, and correspondences between different lects tend to have the same glyph (eg. Shanghainese /ɔ/ and Suzhounese /æ/ both being "au"). Slight modifications have been made regarding tone notation, as Wugniu does not officially have a way to notate tone sandhi. The legacy Wiktionary romanization for Shanghainese is accessible in the expanded pronunciation infobox, and its use is now not recommended.

As most Northern Wu lects share the same set of initials, the following will be a pan-Northern Wu table. Irregularities will be listed in the footnotes. Discrepancies between the legacy Wiktionary system and Wugniu for Shanghainese are shown in bold.

Finals in Northern Wu lects vary from lect to lect. The following are tables for Shanghainese, Suzhounese and Hangzhounese finals.

• eu is only found in the speech of elderly Hangzhounese speakers. When -eu is to be specified, -ei must also appear in the input and a note must be made of the difference. See 狗.
• -u- and -iu- glides become fricated (/zʷ/) after sibilants.
• MiniDict merges a few more rimes that are not merged on Wugniu. Please exercise caution.

Unlike other varieties such as Beijing Mandarin (right-prominent, limited but systematic), Guangzhou Cantonese (rare and non-systematic, but with certain patterns) or Taiwanese Hokkien (right-prominent, widespread and systematic), the tone sandhi rules in Northern Wu comprise two parts—a left-prominent word tone sandhi rule, and a right-prominent phrase tone sandhi rule. Words of fossilised nature follow the first rule, and analysable phrases (usually of verb + noun composition) follow the second rule.

For Northern Wu romanizations in the template {{zh-pron}}, the romanisation is made up of the sandhi chain's tone value (often the same as the tone category of the first character), followed by romanizations of the initial and final of each character (separated by spaces); 6zaon he would be the input for 上海 (⁶zaon-he) in Shanghainese. If the underlying tone category of a syllable is not the same as the sandhi chain's realization, the underlying tone is written to the right of the syllable; 6veq8 khu i would be the input for 勿可以 (⁶veq₈-khu-i) in Shanghainese. This is largely only used in Suzhounese, though note that in Shanghainese, the negator 勿 (⁸veq) uses tone 6's sandhi chain.

Wugniu's tone categories span from 1 to 8 for all lects, though not all are used in every lect. These numbers correspond to the four Middle Chinese tones—level, rising, departing and checked—and are further split in two based on voicing (voiceless—dark, voiced—light) of the initial of the character.

Right-prominent sandhi can be inputted through the use of +: 6ne+3hau for 倷好 in Suzhounese.

0 forces a null tone onto a syllable, and - forces a mid-tone onto a syllable. These are both useful in usage examples and idioms, where tone nullification often occurs based on syntax.

Shanghainese

There are five tone categories (or "tones") in Middle and New (中派與新派) Shanghainese, a reduction from the six to eight in Old (老派) Shanghainese. The tone category a character belongs to can essentially be inferred from its Middle Chinese pronunciation, or pronunciations from other Chinese varieties. Middle Chinese had four tones—level, rising, departing and checked. In the development to Shanghainese, each of the four Middle Chinese tones split in two, conditioned by the voicing (voiceless—dark, voiced—light) of the initial of the character. Three of the resultant eight tone categories then merged with other categories, producing five tone categories in total today.

Several of these tone categories are non-phonemic; that is, they are predictable from the voicing of the initial consonant and from whether the syllable is checked (ending in a glottal stop). Only tones 1 and 5 are contrastive: they both occur in syllables with voiceless initials and no final glottal stop.

Wugniu tone number	Legacy tone number	Tone name (tone category)	Voicing	Tone value	Example characters	Ancestral MC tones
1	1	dark level 陰平／阴平	voiceless	˥˧ 53	刀漿司東／刀浆司东	dark level
5	2	dark departing 陰去／阴去	voiceless	˧˦ 34	島到獎醬／岛到奖酱	dark rising, dark departing
6	3	light departing 陽去／阳去	voiced	˨˧ 23	桃導道牆／桃导道墙	light level, light rising, light departing
7	4	dark checked 陰入／阴入	voiceless	˥ʔ 55	雀削說足／雀削说足	dark checked
8	5	light checked 陽入／阳入	voiced	˩˨ʔ 12	嚼石曲食	light checked

Left-prominent tone sandhi (word sandhi)

Each of these five categories then has a tone sandhi pattern, depending on the number of syllables in the word. In Shanghainese, left-prominent sandhi patterns are always analysed to be entirely dependent on (ie. predictable from) the tone of the first syllable (except 勿 (⁸veq)), and as such, subsequent syllables' tones are unmarked.

Left-prominent word tone sandhi patterns

Tonal category	Monosyllabics	Disyllabics	Trisyllabics	Tetrasyllabics	Pentasyllabics
1. Dark level (1)	53	55＋21	55＋33＋21	55＋33＋33＋21	55＋33＋33＋33＋21
2. Dark departing (5)	34	33＋44	33＋55＋21	33＋55＋33＋21	33＋55＋33＋33＋21
3. Light departing (6)	23	22＋44	22＋55＋21	22＋55＋33＋21	22＋55＋33＋33＋21
4. Dark checked (7)	55	33＋44	33＋55＋21	33＋55＋33＋21	33＋55＋33＋33＋21
5. Light checked (8)	12	11＋23	11＋22＋23	11＋22＋22＋23 22＋55＋33＋21	22＋55＋33＋33＋21

Right-prominent tone sandhi (phrase sandhi)

When words combine form a phrase, the following right-prominent sandhi rules apply. In short, when the word A in appears non-finally in a phrase, its last syllable (A-x) changes to a flat (level) tone. The tone sandhi value that syllable A-x changes to is conditioned by three factors: (1) the tonal category of syllable A-x, (2) the number of syllables in word A, and if the number of syllables in A is 1 – (3) whether word A is "tightly associated" with the word preceding word A. The exact realization of the phenomenon varies from person to person, and as such, right-prominent sandhi is only visible in the IPA realization of the input.

Word undergoing sandhi (i.e. word A)	Tonal category of last syllable of word undergoing sandhi (i.e. syllable A-x)					Example
	Tone 1	Tone 5	Tone 6	Tone 7	Tone 8
Monosyllabic word unbound to any word preceding it	44		33	44	22	儂好／侬好 (6non+5hau) 炒魷魚／炒鱿鱼 (5tshau+6yeu ng)
Monosyllabic word tightly bound to a monosyllabic word preceding it						硬碰硬 (6ngan+-phan+6ngan)
Multisyllabic word	33			33		嘸著嘸落／呒着呒落 (6m zaq+6m loq)
Monosyllabic word tightly bound to a multisyllabic word preceding it						前世作孽 (6zhi sy+7tsoq+8gniq)

Conversion from MiniDict and Wugniu tone notation

• 平 (level), 上 (rising), 去 (departing), and 入 (checked) should be used for reference, not the numbers.

Suzhounese

Tone sandhi in Suzhounese is usually analysed as a system in which every underlying tone contributes to the chain's contour. However, some generalizations can be made. On Wiktionary, the system as described in A Study on Suzhounese Phonetics (蘇州方言語音研究) is used.

Disclaimer: as Suzhounese has many irregular sandhi chains, and this system drastically simplifies sandhi, many words are irregular. Please transcribe realised tones using the sandhi chain with the closest pitch contour.

It is widely agreed upon that Suzhounese has seven tones.

Wugniu tone number	Tone name (tone category)	Voicing	Tone value	Example characters	Ancestral MC tones
1	dark level 陰平／阴平	voiceless	˦ 44	廳千芭篩／厅千芭筛	dark level
2	light level 陽平／阳平	voiceless	˨˨˧ 223	排尋同廊／排寻同廊	light level
3	rising 上聲／上声	voiceless	˥˩ 51	耍總朵孔／耍总朵孔	dark rising
5	dark departing 陰去／阴去	voiceless	˥˨˧ 523	誘鈵贊介／诱𰽥赞介	dark departing
6	light departing 陽去／阳去	voiced	˨˧˩ 231	戀晚稻共／恋晚稻共	light rising, light departing
7	dark checked 陰入／阴入	voiceless	˦˧ʔ 43	戚鉑掬客／戚铂掬客	dark checked
8	light checked 陽入／阳入	voiced	˨˧ʔ 23	十薄達曆／十薄达历	light checked

Left-prominent tone sandhi (word sandhi)

The left-prominent sandhi system used on Wiktionary works quite similarly to Shanghainese. For tones 1-6, only the first syllable is taken into account, whereas in tones 7 and 8, the first two syllables determine the chain. The 0's listed below are not displayed in the IPA as it is considered non-standard.

Left-prominent word tone sandhi patterns in unchecked syllables

Tone category	Monosyllabics	Disyllabics	Trisyllabics	Tetrasyllabics
Dark level (1)	44	44 0	44 44 0	44 44 44 0
Light level (2)	223	22 33	22 33 0	22 33 44 0
Rising (3)	51	55 11	55 11 0	55 11 11 0
Dark departing (5)	523	52 33	52 33 0	52 33 44 0
Light departing (6)	231	23 11	23 11 0	23 11 11 0

Left-prominent word tone sandhi patterns in checked syllables

Tone category		Disyllabics	Trisyllabics	Tetrasyllabics
1st syl	2nd syl
Dark checked (7)	Level (1, 2)	44 23	44 23 0	44 44 23 44 0
	Rising (3)	55 51	55 51 0	55 51 11 0
	Departing (5, 6)	55 523	55 52 33	55 52 22 33
	Checked (7, 8)	44 44	44 44 0	44 44 44 22
Light checked (8)	Level (1, 2)	22 33	22 33 0	22 33 44 0
	Rising (3)	22 51	22 51 0	22 51 11 0
	Departing (5, 6)	22 523	22 52 33	22 52 22 33
	Checked (7, 8)	33 44	33 44 0	33 44 22 0

Right-prominent tone sandhi (phrase sandhi)

Suzhounese only has Shanghai-style right-prominent sandhi in monosyllabic dark departing terms. The dark departing syllabe mutates from 523 to 51: 笑別人 is pronounced as /siæ^523>51 bəʔ²² ȵin³³/ replace ȵ with n̠ʲ, invalid IPA characters (523>51<>22</>ȵ33).

In example sentences, all Northern Wu lects with Wugniu romanization are permitted. Other Wu lects are to be further investigated as to the potential for their inclusion. The Northern Wu localities with Wugniu notation are:

• Shanghai
  • Urban Shanghai, Baoshan, Jiading, Putuo, Qingpu, Songjiang, Jinshan, Fengxian, Chuansha, Chongming
• Jiangsu
  • Eastern Nantong (通東)
  • Urban Suzhou, Kunshan, Wujiang, Changshu
  • Urban Wuxi, Yixing
  • Urban Changzhou
  • Jingjiang (Old Coast 老岸話)
• Zhejiang
  • Urban Jiaxing, Jiashan, Haiyan, Haining, Tongxiang
  • Urban Huzhou, Changxing, Anji, Deqing
  • Urban Hangzhou, Yuhang, Linping, Fuyang, Tonglu, Xiaoshan
  • Urban Shaoxing, Keqiao, Shangyu, Shengzhou, Xinchang
  • Urban Ningbo, Cixi, Yuyao, Zhenhai, Beilun, Yinzhou, Fenghua, Xiangshan
  • Urban Zhoushan, Daishan

• 中華／中华: zh-pron|w=sh:1tson gho;sz:1tson gho2

• For checking the pronunciation of words, use The Comprehensive Dictionary of Shanghainese (《上海话大词典》), which uses IPA notations throughout the book. Wugniu also has a Shanghainese dictionary listed under 松江 Songjiang. However, note that Wugniu's dictionary spills into lect areas outside of Puxi
• For checking the pronunciation of characters, use Wugniu or Wu Chinese MiniDict

  Note that MiniDict uses a different romanisation system than the one implemented here. Their romanisation scheme is as listed on their website

See also

• Category:Wu language
• Category:Northern Wu
• Category:Shanghainese Wu
• Category:Suzhounese Wu
• Wiktionary:About Chinese/Wenzhounese another Wu variety

If unsure, hassle User:ND381 or User:Musetta6729.