My stance on the issue of Hakka romanisation is as follows:[phiên-siá | Kói ngièn-sṳ́-mâ]
- There is no official and authoritative Hakka romanisation, and the one used on this site should only be taken as a guide.
The reason for this, is, even though Hakka as a language is fairly homogenous, there are regional variations in pronunciation, lexicon and grammar, and no single romanisation can account for all the varied differences in pronunciation.
The romanisation used on this website to represent the sounds of Hakka is based upon a Missionary romanisation of 彭德修 Pang2 Det5Siu4, based in Toi2van1 台灣, who produced several works which aimed to help students of Hakka either learn the correct pronunciations, or to learn the language itslef. Hakka Wikipedia editors using the romanisation face several technical problems outlined below. Not everyone's speech accent is the same, and in different dialects, conflicts of reading for a particular Chinese character may be encountered. The following non-exhaustive list of publications all use different romanisation systems.
- 彭德修 Pang2 Tet5Siu4's dictionary of pronunciation 客家話發音字典 Hak5ga1va4 fat5yim1 su5dien3
- McIver's dictionary of Hakka 客英大詞典 Hak5-Yin1 Tai4Su4dien3
- 張維耿 Zong3 Vui3Guen3's Hakka dictionary 客家話詞典 Hak5Ga1va4 ci2dien3 uses 客家拼音方案 Hak5ga1 pin1yim1 fong1on4
- 劉鎮發 Liu2 Zin3Fad5's dictionary of Hakka pronunciation 客語拼音字彙 Hak5ngi1 pin1yim1 su4fui1.
- The Loeh-Foeng-Dialect by Simon Hartwich Schaank records the 陸豐 dialect of south eastern Guangdong.
- The Hakka Dialect by Mantaro J. Hashimoto uses a pseudo-phonetic system
- 梅縣方言詞 典 ed. 李榮 Li3 Yin2 and 黃雪貞 Vong2 Set5Zin1 uses IPA with tone letters.
As I speak a Hakka dialect that doesn't represent the medial u well, words like 光 guong1 and 江 gong1 are merged into gong1 etc in my dialect. Clearly then, my romanisation without -u- medials leads to those who have the medial reading such material with some puzzlement, if it leads to ambiguity. Moreover, my dialect differentiates between 高 -au and 歌 -o endings, whereas, some speakers have merged readings as go1, rendering romanised text liable to confusion. Hence, a dual text of Chinese characters and romanisation often helps to clarify an situation where confusion is likely to arise, but unfortunately, at present, most articles here are not dual written language.
I use a hybrid of Liu, Zong and MacIver's systems. The Zong and Liu romanisations are fairly similar, whilst MacIver's sytem is slightly superior to Pang's version in my opinion. My name ought to demonstrate the point.
- Hakka romanisation should be viewed as guide to pronunciation of Chinese characters, not a replacement. However, there are many Hakka words which have no fully recognised Chinese character equivalent. On this site, borrowed Chinese characters are used to fill these gaps. Only with practice can one distinguish the use of which pronunciation goes with the character in the context it is in.
Tone and Markup for copy and paste[phiên-siá | Kói ngièn-sṳ́-mâ]
I prefer using the markup instead of the accented characters. As my dialect of Hakka doesn't differentiate between the Qu tone in the Yin and Yang states, all my Qu tones will be represented by <sup>4<\sup>, without exception. However, below I've included the <sup>4'<\sup> as a way to distinguish the yang state if users should so choose it. The inclusion of the tone contours shows why in my dialect the qu tone is merged. The choice to use the IPA rendering of tone contours is available in this table.
|Tone Name||Vowels + diacritics||Nasal Syllabics||12/3/56/78 Tone||12/3/44'/56 Tone||Tone Contours|
|陰平 Yin Ping||Â â||Ê ê||Î î||Ô ô||Û û||Ṳ̂ ṳ̂||M̂ m̂||N̂g n̂g||1||<sup>1<\sup>||<sub>33<\sub>|
|陽平 Yang Ping||À à||È è||Ì ì||Ò ò||Ù ù||Ṳ̀ ṳ̀||M̀ m̀||Ǹg ǹg||2||<sup>2<\sup>||<sub>11<\sub>|
|上聲 Shang Sheng||Á á||É é||Í í||Ó ó||Ú ú||Ṳ́ ṳ́||Ḿ ḿ||Ńg ńg||3||<sup>3<\sup>||<sub>31<\sub>|
|陰去 Yin Qu||A a||E e||I i||O o||U u||Ṳ ṳ||M m||Ng ng||5||<sup>4<\sup>||<sub>53<\sub>|
|陽去 Yang Qu||Ā ā||Ē ē||Ī ī||Ō ō||Ū ū||Ṳ̅ ṳ̅||M̅ m̅||N̅g n̅g||6||<sup>4'<\sup>||<sub>53<\sub>|
|陰入 Yin Ru||A a||E e||i i||O o||U u||Ṳ ṳ||M m||Ng ng||7||<sup>5<\sup>||<sub>3<\sub>|
|陽入Yang Ru||A̍ a̍||E̍ e̍||I̍ i̍||O̍ o̍||I̍ u̍||Ṳ̍ ṳ̍||M̍ m̍||N̍g n̍g||8||<sup>6<\sup>||<sub>5<\sub>|
|IPA||a||e ɛ||i ɪ||o ɔ||u ɯ||ɿ||m m̩||ŋ ŋ̩|| Usage : |
Fonts[phiên-siá | Kói ngièn-sṳ́-mâ]
For the correct display of diacritic characters, this site recommends the use of the CharisSIL font. However, on inspection, the font also has characters situated in the Private Use area. Although it doesn't affect the symbols above, I should warn against using the PUA characters, since not everyone who views this Hakka Wikipedia will have the CharisSIL font, or want to install it. It is better to stick with characters grouped under Unicode areas.
Diacritics[phiên-siá | Kói ngièn-sṳ́-mâ]
One of the main problems with inputting text with diacritics is the lack of software to do this simply and easily. The table above is for my convenience, but, the diacritics used may not be the same as those found on this site. This leads to the problem of text compatibility in searching for articles since the software does not distinguish between a single precomposed character and a character composed of two unicode letters (i.e. a vowel + diacritic). As such, I favour a romanisation which dispenses with them, and uses numbers as tone labels instead.
IPA and Romanisation[phiên-siá | Kói ngièn-sṳ́-mâ]
The use of IPA (International Phonetic Association) characters is to enable the reader to see a faithful representation of a sound irrespective of the romanisation used. The only difficulty for Hakka users, is that their own local pronunciation differs to that of Hakka speakers from another region. For instance, Hakka speakers from the Longchuan 龍川/龙川 dialect of Heyuan 河源 Guangdong Province， China have vocabulary and pronunciations which seems almost as if it were heavily influenced by Yue 粵 dialects. Their pronunciation although having a regular correspondence to Middle Chinese differents somewhat from the 梅縣/梅县 Moi2yen4 paradigm. Or, the dialects of Hakka which have lost its 入 Ru Tone. Not even the 四縣 Si4yen4 is a perfect match to Meixian, on account of the difference in tone contours, which IPA can distinguish. A phonemic approach with a romanisation can disguise the tone pitch differences, since the majority of pronunciations of words occur using the same tone classes, therefore between the groups, the correspondence of tone classes for each different speaker is fairly regular. This pitch accent often creates great difficulty for listeners to a dialect whose accent is different to their own.
Using Ruby Text[phiên-siá | Kói ngièn-sṳ́-mâ]
Ruby mark-up allows Chinese characters to be appended with pronunciation. Currently, there are ruby versions for Mandarin and Japanese. In Hakka, it would require manual coding at present.
ruby : <ruby> <\ruby>
ruby body : <rb> <\rb>
ruby text : <rt> <\rt>
Ruby Text with Chinese Characters, and IPA[phiên-siá | Kói ngièn-sṳ́-mâ]
If the code blocks are sequential without spacing, they form a horizontal line of text. However, if you separate each code block with a blank line, they become a column of text instead.
Viewing Ruby text depends on each browser. Some supports the extension, others don't. For IE users, Ruby markup support is included in version 6 and later. For Mozilla Firefox, you need to use a plug-in. A suggested plugin was discussed here.