Initials The Suzhou dialect has series of voiced, voiceless, and aspirated
stops, and voiceless and voiced
fricatives. Moreover,
palatalized initials also occur. Voiced obstruents are typologically partially voiced instead of fully voiced. These consonants are devoiced word/phrase-initially, but are fully voiced within a phrase. This is most apparent for the fricatives becoming . Because of this devoicing, in single-syllable forms the distinction is actually the tone contour. The glottal initials disappear if they are not at the beginning of a word/phrase, resulting in a smooth vocalic transition from the previous syllable. In this regard, it is possible to analyze both initials as a single phonological null onset when in this environment. : 蛙 ≠ 华 , but 青蛙 = 清华
Finals :
Syllabic continuants: Notes: • The Suzhou dialect has a rare contrast between "fricative vowels" and ordinary vowels . • phonetically, are regular plain vowels [i y], while have further constriction • an acoustic study found to be phonetically syllabic voiced laminal post-alveolar fricatives • is pronounced before rounded vowels. • is a true mid vowel, . May also be transcribed with the
Sinological symbol /ᴇ/. • In open syllables, is articulated close to a position for a close back vowel • Depending on the source, transcriptions differ: • may also be transcribed as • may also be transcribed as ; also applies to on-glide final rhymes • may also be transcribed as • Close vowels may be analyzed as diphthongs and transcribed as }
Historical Finals The Suzhou dialect allows a nasal coda but does not distinguish between them. As such, the
Middle Chinese nasal codas have largely either merged or been lost depending on the vowel it follows. Historical
rimes following certain vowels are distinguished as the nasalized vowels , but otherwise merge into modern . Historical and rimes are entirely merged and also result in modern , or are lost after certain vowels becoming modern . Monophthongization of the historical diphthong rime (as seen in Modern Standard ) resulted in a high-mid vowel /e/ distinct from the /ɛ/ that arose from the historical nasal rimes *-an/am (as seen in Modern Standard ). The distinction between /e ɛ/ is preserved within some Pingtan opera performers, but has merged together in general speech. Middle Chinese rimes have become glottal stops, . Like other Northern Wu varieties, syllables with an underlying glottal stop coda usually manifest as a
shortening of the vowel instead of an actual glottal stop , unless before a pause or at the end of an utterance.
Tones Suzhou is considered to have seven tones. However, since the tone split dating from Middle Chinese still depends on the voicing of the initial consonant. Yang tones are only found with voiced initials, namely [b d ɡ z v dʑ ʑ m n nʲ ŋ l ɦ], while the yin tones are only found with voiceless initials. These constitute just three phonemic tones:
ping,
shang, and
qu. (
Ru syllables are phonemically toneless.) In Suzhou, the Middle Chinese 阳上 tone and 阳去 tones have fully merged as (2)31. The original 阳去 313 tone possibly still occurs in tone sandhi patterns as the second element of a chain, following a 阴入 syllable (though it could be analyzed differently; see Tone Sandhi section below). Therefore, 买 and 卖 has exactly the same pronunciation in literary and colloquial readings
6ma , but can be distinguished in tone sandhi. 弗买 ≠ 弗卖 .
Tone Sandhi Tone in Suzhou dialect, like other Northern Wu varieties is generally grouped by phrasal tone pattern, also called
sandhi chains or
sandhi domains. An analysis by Wang (2011) describes Suzhou tone sandhi as rightward tone-spreading of the left-most (i.e. initial) syllable of a phrase. Such described "left-prominent" phrases with non-checked initial syllables of a given length have one of five possible contours, each equivalent to each of the five tones. While generally described as rightward tone-spreading of the initial syllable, it is also common for the phrasal tone pattern to not be the same as that of the initial tone. This is currently the system used on Wiktionary entries with Suzhou data. To distinguish the individual tone from the pattern expected from its tone spreading, the patterns themselves are referred to with the format of
tone number + X (1x, 2x, 3x, etc.). A tone level of 0 in the above chart indicates a syllable with a neutral tone (), functionally comparable to that of
Standard Chinese. The surface realization at the end of an utterance is a low akin to downstep, but in flowing speech is a mid/neutral pitch or may appear to copy the previous tone target. Additionally, Li (1998) describes the 5x chain such that the second syllable has a slight rise. Li also describes a higher mid/high-level for the second syllable of a 6x chain. Li's 1x chain describes the pitch declining after the second syllable. In phrases with checked initial syllables, the first two tones determine the overall contour. The resulting contour can be summarized as retaining the tone class (平上去) of the second syllable, but not the voicing class (陰陽). Both Tone 1 陰平 /44/ and Tone 2 陽平 /223/ will result in a Tone 2 contour (/223/). Both Tone 5 陰去 /523/ and Tone 6 陽去 /231/ will result in a Tone 5 contour (/523/). Ye 1988 describes additional patterns where • Tone 7 阴入 + Tone 1/3/5 retaining full tone, resulting in a /5ˀ 5/ pattern if Tone 7 阴入 is followed by Tone 1 阴平 • the original un-merged Yangshang 阳去 313 tone still occurs as the second element of a chain, following a 阴入 syllable (7.6 chain). • The second syllable of an 8x chain having a low-falling /21/ regardless of original tone However, Wang describes the same phrases differently, and so it is debatable whether these form distinct patterns:
Tone Category Shifts As mentioned above, the tone pattern of a phrase frequently does not match the expected pattern based on the initial syllable's underlying tone. Most frequently: • a phrase beginning with a Tone 3 syllable takes on the tone pattern expected of a Tone 5 syllable (in other words, a 5x chain) or a Tone 1 syllable (a 1x chain) • i.e. expected
3x >
5x or
1x • (5x) 短衫 : 5toe3-se1 • (1x) 暑假 : 1syu3-ka5 • a phrase beginning with a Tone 5 syllable frequently takes on the tone pattern expected of a Tone 1 syllable (a 1x chain) • i.e. expected
5x >
1x • (1x) 菜飯 : 1tshe5-ve6 • a phrase beginning with a Tone 6 syllable frequently takes on the tone pattern expected of a Tone 2 syllable (a 2x chain) • i.e. expected
6x >
2x • (2x) 大菜 : 2da6-tshe5, 2dou6-tshe5 • less frequently, the above shifts can happen in reverse • i.e. expected 5x > 3x • i.e. expected 2x > 6x • syllables following Tone 7 can also shift chains • Tone 7 + Tone 5/6 > (Tone 7 + Tone 1/2) > 7.2 • Tone 7 + Tone 6 > 7.3 • most non-checked syllables following Tone 8 collapse into a falling tone, equivalent to an 8.3 chain • Tone 8 + {Tone 1, 2, 3, 5, 6} > 8.3 Functionally, a Tone 3 pattern (3x chain) is the least common to occur and mostly surfaces when the initial syllable is a numeral phrase ( 3ci-zyu6 ) or reduplicated verb ( 3sia-sia3 ). Below is a chart with examples of the common tone patterns:
Tone reduction Wang (p. 50) additionally identifies a pattern where in certain constructions Tone 5 (/523/) followed by another syllable simplifies to [52] while the second syllable retains its full tone. This can be analyzed comparably to Shanghainese right-prominent sandhi that prioritizes the second syllable and reduces preceding syllables. This right-prominent sandhi pattern occurs commonly in
Verb + Object compounds. :做人 In addition to the above simplification of Tone 5 /523/ to [52], Li (p. 216) additionally describes Tone 2 /223/ and Tone 6 /231/ similarly simplifying to [23 ˨˧] in similar
Verb + Object, as well as
Adverb + Adjective structures : 穷大 dʑioŋ223-23 dou231 : 是鬼 zɿ231-23 tɕy52 : 过桥 kou523-52 dʑiæ223 Identified by Bu (2025) describing
Suzhou pingtan (but also applicable to Suzhou dialect normally), such tonal reduction generally occurs particularly for Tone 2 and Tone 6 syllables even when not in sandhi chains, and can further reduce to a simple mid/low tone. Because it can occur outside of
Verb + Object or
Adverb + Adjective syntactic conditions, Bu considers this tonal reduction to simply be a reduction of non-final syllables motivated by those tones (Tone 5 /523/, Tone 2 /223/, Tone 6 /231/) underlyingly being longer and having more tonal targets. In contrast, Wang (p. 348) treats this pronoun + copula construction as a single 6x phrase. {{Fs interlinear|{我 是} {老 三}|ŋəu‿zɿ1 læ‿sᴇ23|{1sg COP} {Lao San}|"I am Lao San"|indent=2|italics2=no}}
Casual Speech There can be additional variation in how reduced the tones can become based on how casual the sentence is spoken by the speaker. In the above sentence, the falling tone [˥˩] on 仔
tsy and 再
tse is reduced to a high-flat [˥] in casual speech, in addition to the Tone 6 /231 ˨˧˩/ (倷
ne, 飯
ve) and Tone 5 /523 ˥˩˧/ (再
tse, 去
chi) words already reducing to [23 ˨˧] and [52 ˥˩] even in slower speech. In the case of casual speech spoken quickly, Wang does describe a pattern where the preceding syllable takes a neutral tone. If the word (often a pronoun, adverb, or quantifier) precedes another phrase, it can reduce to a simple /3/ tone. This reduced pattern can apply across polysyllabic words or even multiple words. This can be considered as describing the same phenomenon as above but with less phonetic detail. : 交差 | : 搿人有点弗大适意 |
Stress The same phrase can take a different chain depending on which syllable or word is stressed. : 看戏 'to watch shows/movies" : (a 1x chain) ::generally without emphasis, it would be treated as a single concept and be a single sandhi chain. : ::emphasizing what is being watched—the verb is treated separately and reduced to either /52/ or /3/ == Writing ==