To follow up on my post on the Taishanese spoken by my parents (the 端芬 Duanfen dialect), here’s a look into the initials and finals of this dialect.
Inspired by Stephen Li’s excellent work on documenting the phonology of the 臺城 Taicheng dialect of Taishanese. Kudos to TypeIt for making inputting these IPA characters easier.
Similar to other Chinese languages like Cantonese and Mandarin, a Taishanese syllable can be analysed as an onset/initial + nucleus + coda. A nucleus and a coda combined is called a final. A tone is across the entire syllable.
About the romanisation
While I was designing a database to investigate how Duanfen Taishanese corresponds to Middle Chinese, I realised it would probably be a good idea to make a romanisation of it due to the high number of characters in IPA untypeable with a regular QWERTY keyboard.
This will be the romanisation I’ll be using in that database, as well as any projects relating to Duanfen Taishanese I do from now on.
November 2022 update: A lot has changed in the five years(!) since I’ve made this post.
I don’t expect anyone to use this romanisation, due to it being designed for internal use only. In fact, there are many areas of the romanisation that I would do differently if I had the opportunity to start from scratch:
- The idea of using the tones of the basic numbers as the basis for the tone numbering has since gained traction in the online Taishanese enthusiast communities. The idea is that tones 0, 1, 2 and 3 are defined as whatever tones the Taishanese pronunciations of 零, 一, 二 and 三 have. The only tone remaining (the low falling tone, or the tone in 以) is defined as tone 5.
- /eɐ̯/ and /oɐ̯/ would be better represented as <ie> and <uo> to prevent confusion between, say, <guai> and <gwai> (which would become <guoi> and <gwai> respectively).
- <eo> and <a> may be better represented as <a> and <aa>, due to these two sounds mapping to Cantonese <a> and <aa> very well.
As for the phonology itself, expect this to be a living document – I’m still learning new things about this 方言, and so I’ll still be updating this (as well as the site as a whole) periodically.
Initials
My father and mother
| Labials | Alveolar / Dental | Palatal | Velar | Glottal | ||
|---|---|---|---|---|---|---|
| Stop | Unaspirated | p 背 | t 總 | k 夠 | ʔ 答 | |
| Aspirated | pʰ 旁 | tʰ 速 | kʰ 求 | |||
| Nasal | m 馬 | n 你 | ŋ 牙 | |||
| Fricative | Central | f 放 | s 色 | h 吃 | ||
| Lateral | ɬ 三 | |||||
| Affricate | Unaspirated | t͡ʃ 裝 | ||||
| Aspirated | t͡ʃʰ 床 | |||||
| Approximant | w~ʋ 話 | l 律 | j~ʝ 養葉 | |||
| Other |
Initials in romanisation
| Labials | Alveolar / Dental | Palatal | Velar | Glottal | ||
|---|---|---|---|---|---|---|
| Stop | Unaspirated | b | d | g | ' | |
| Aspirated | p | t | k | |||
| Nasal | m | n | ng | |||
| Fricative | Central | f | s | h | ||
| Lateral | hl | |||||
| Affricate | Unaspirated | dx (t͡ʃ) / dz (ts) | ||||
| Aspirated | tx (t͡ʃʰ) / ts (tsʰ) | |||||
| Approximant | w/v | l | j |
Finals
Columns represent nuclei, with rows representing codas (which can be semivowels, unreleased consonants, nasal consonants, or simply nothing).
My father and mother
| /a/ | /i/ | /ə/ | */wə/ | /ɛ/ | /eɐ̯/ | /ɔ/ | /oɐ̯/ | /u/ | other | |
|---|---|---|---|---|---|---|---|---|---|---|
| -∅ | a 牙 | i 你 | eɐ̯ 寫 | oɐ̯ 我 | u 肚 | |||||
| -j | aj 危 | əj 使 | *wəj 規 | ej 笑 | oɐ̯j 愛 | uj 碎 | *waj 乖 | |||
| -w | iw 少 (see note) | ɪw 收 | ɔw 掃 | |||||||
| -m | am 啱 | im 鹽 | əm 冚 | ɛm 點 | m̩ 五 | |||||
| -n | an 散 | in 現 | ən 緊 | *wən 群 | ɛn 面 | oɐ̯n 幹 | un 管 | |||
| -ŋ | aŋ 生 | eʲŋ 另 | eɐ̯ŋ 將 | ɔŋ 狼 | oʷŋ 龍 | |||||
| -p̚ | ap̚ 鴨 | ip̚ 葉 | əp̚ 急 | ɛp̚ 接 | ||||||
| -t̚ | at̚ 八 | it̚ 折 (e.g. 折本) | ət̚ 實 | *wət̚ 骨 | ɛt̚ 跌 | oɐ̯t̚ 撥 | ut̚ 律 | *wat̚ 刮 | ||
| -k̚ | ak̚ 得 | eʲk̚ 滴 | eɐ̯k̚ 弱 | ɔk̚ 落 | oʷk̚ 叔 |
Finals in romanisation
“eo” is used to represent the schwa [ə].
| /a/ | /i/ | /ə/ | /wə/ | /ɛ/ | /eɐ̯/ | /ɔ/ | /oɐ̯/ | /u/ | other | |
|---|---|---|---|---|---|---|---|---|---|---|
| -∅ | a | i | ia | ua | u | |||||
| -j | ai | eoi | *weoi | ei | uai | ui | *wai | |||
| -w | iu | eu | ou | |||||||
| -m | am | im | eom | em | m | |||||
| -n | an | in | eon | *weon | en | uan | un | |||
| -ŋ | ang | ing | iang | ong | ung | |||||
| -p̚ | ap | ip | eop | ep | ||||||
| -t̚ | at | it | eot | *weot | et | uat | ut | *wat | ||
| -k̚ | ak | ik | iak | ok | uk |
Tones
Note that I’ve written a more thorough write-up about the tones over on my blog, which you can read here.
With Chao tone numbers, 5 is the highest relative pitch level and 1 is the lowest. The numbers used to represent each tone in the romanisation are based on that of Jyutping.
Note that unlike Cantonese, Taishanese has the 陰平 and 陰去 tone categories (represented as 1 and 3 in Jyutping) merged, hence both tone categories are represented as 3 in this romanisation and 1 is not used.
Other sources such as Wang and Qian (1950), as well as Chinese Wikipedia, include a separate 陽入 tone with the value ˨˩ 21 to address characters like 翼 having the low falling tone. I haven’t included this in the table below, as I consider this to be part of the same low falling tone change phenomenon that affects both 入聲 and non-入聲 syllables alike. This is discussed later in the article, in the Tone change section.
| Tone names | Chao tone numbers | Tone number in romanisation | Example |
|---|---|---|---|
| 陰平／阴平 | ˧˧ 33 (mid level) | 3 | 衣 [ji33] ji3 |
| 陰上／阴上 | ˥˥ 55 (high level) | 2 | 椅 [ji55] ji2 |
| 陰去／阴去 | ˧˧ 33 (mid level) | 3 | 意 [ji˧˧] ji3 |
| 陽平／阳平 | ˨˨ 22 (mid-low level) | 4 | 移 [ji22] ji4 |
| 陽上／阳上 | ˨˩ 21 (low falling) | 5 | 以 [ji21] ji5 |
| 陽去／阳去 | ˧˨ 32 (mid falling) | 6 | 易 [ji32] ji6 |
| 上陰入／上阴入 | ˥ 5 (high stopped) | 2 | 益 [jek̚5] jik2 |
| 下陰入／下阴入 | ˧ 3 (mid stopped) | 3 | 八 [pat̚˧] pat3 |
| 陽入／阳入 | ˧˨ 32 (mid falling, stopped) | 6 | 譯／译 [jek̚˧˨] jik6 |
Tone change
Words commonly undergo the rising changed tone phenomenon in Taishanese. Whereas in Cantonese they become the mid-low rising tone (e.g. 錢 cin4 becomes cin2 as represented in Jyutping), in Taishanese a rising contour is simply added to the end of the existing tone. This is represented in my romanisation by a star (*) at the very end.
Another type of tone change, where the tone is changed to the low falling tone (tone 5), is indicated by writing the original tone number, then putting a hyphen and the tone number of the tone it changes to (in this case tone 5). This may be used in conjunction with the rising changed tone phenomenon.
Interestingly enough, for some characters, the low falling tone is the only pronunciation for a character. This includes 街 and 翼. In this case, it’s not unreasonable to consider the low falling tone as being the ‘original’, lexical tone for these characters.
Here are some examples, as well as some more unusual sound changes:
| Example | Original tone | Changed tone |
|---|---|---|
| 掃／扫 (noun meaning “brush”) | [ɬɔw˧˧] hlou3 | [ɬɔw˨˩˥] hlou3-5* |
| 一味 | [mi˧˨] mi6 | [mi˧˨˥] mi6* |
| 餸 | [ɬoŋ˧˧] hlung3 (as 送) | [ɬoŋ˨˩] hlung3-5 or hlung5 |
| 發毛／发毛 | [mɔw˨˨] mou4 | [mɔw˥˥] mou4-2 |
| 牙刷 | [tɕʰat̚˧] txhat3 | [tɕʰat˨˩˥] txhat3-5* or (fast speech) [ɕat˨˩˥] xat3-5* |
Tones that are not covered in the romanisation are written using Chao tone numbers after a tilde (~) symbol. This replaces a hyphen if one was added to reflect a change of tone.
For example, 「确！」 (expression of disbelief, pronounced [kʰɔk̚˥˧]), would be written as khok~53 or alternatively khok3~53 (to show that 确 originally had tone 3), but not khok3-~53.
References and external sources mentioned
- Chinese Wikipedia article for Taishanese.
- 台山方音 (1950) by Li Wang (王力) and Songsheng Qian (錢淞生).