Length-based Cantonese Romanisation

[ParkeNYU]

For those of you who don't know, I developed a Cantonese version of Zhuyin/Bopomofo because when I began studying the language, I wasn't satisfied with any of the existing Cantonese Romanisation schemes. Although I still use my system (CPS) religiously, its limited compatibility has led me to seek an alternative system–a Romanised one. I thought about why Zhuyin/Bopomofo felt like a more natural fit for the Cantonese language than the other systems; although perhaps the biggest (and most unavoidable) reason is that Latin letters have numerous phonetic interpretations, another crucial reason is the dissonance between orthographic and phonetic length (e.g. the short 'deoi' and long 'doi'). I realised that I would only be satisfied with a system of Cantonese Romanisation that found harmony between orthographic and phonetic length, and although Penkyamp admirably begins to approach this goal, it still left more to be desired.

And so, after much consideration and consultation, I made my own:

INITIALS

b / p / m / f

d / t / n / l

g / k / ŋ / h

j / c / s

y / w / gw / kw

NUCLEI (long)

aa [aː]

oa [ɔː]

oe [œː]

ae [ɛː]

ee [iː]

oo [uː]

uu [yː]

NUCLEI (short)

a [ɐ]+[i/u/m/p/n/t/ŋ/k]

o [o]+, [ʊ]+[ŋ/k]

u [ɵ]+[y/n/t]

e [e]+[ i ], [ɪ]+[ŋ/k]

CODAS

i [ i ] / u / (u)i [y]

m / p / n / t / ŋ / k

TONES

高¹ gôu

起² héi

半³ bōon

沉⁴ càm

仰⁵ yõeŋ

下⁶ hăa

SYLLABIC NASALS

m / ng

https://dl.dropboxusercontent.com/u/15757362/PCR.pdf

I attempted some impossible compromises: this system had to be (1) phonemically systematic, (2) phonetically intuitive (for native English speakers like me), and (3) easy to type and display. Minor sacrifices were necessary in each department (notably the inclusion of the letter Eng), but I feel that I have distilled a happy medium.

Perhaps some of you Cantonese learners out there have also struggled in finding a system to suit your tastes. Since the end goal of literacy in any Chinese language is the use of Chinese characters, I've arrived at the belief that everyone should use their own preferred means of phonetic transcription, as long as these systems correspond with a common set of phonemes.

What do you all think of this approach?

[Angelina]

Could you please copy the file content and post it, or maybe attach a screen shot? It seems like I can't download the file, I'm in China. 

 

Good idea! I am not fully satisfied with any existing Romanization schemes for Cantonese either. I would actually prefer a Cyrillization system because of aesthetic reasons. Maybe I can make my own. 

 

Angelina

[ParkeNYU]

Certainly, sorry about that. I'll update the first post.

And here is a sample text.

"Ya̖n ya̖n sāaŋ cūt la̖i ja̱u ha̱i je̱e ya̖u gae, hái jūun ye̖em to̖ŋ ku̖un le̱i so̱eŋ yāt lu̱t pe̖ŋ dáŋ. Ku̗i de̱i gu̱i ya̗u le̗i sēŋ to̖ŋ lo̗eŋ sām, ye̖e cáe yēŋ gōai yo̱ŋ hēŋ da̱i gāan gae gwāan ha̱i la̖i wo̱o sōeŋ dui do̱ai." -Sai Gaai Ya̖n Ku̖un Sūun Ye̖en (Lu̖un Ha̱p Gwoak)

[Angelina]

Sure. I can't offer you any (professional- I am linguist after all) advice until I see it. 

[ParkeNYU]

I've added it above, and I'm looking for a Cantonese Cyrillisation too.

[imron]

and (3) easy to type and display

How is ŋ easy to type?

 

I'm not a linguist, but I can tell you that any latin based romanisation scheme that uses 'funny letters' e.g. ŋ is never going to work and will be bastardised by users, see for example people who type pin1yin1 rather than pīnyīn, and how official things (passports/street signs) in China use v instead of ü.

[ParkeNYU]

imron, hence my concession: "Minor sacrifices were necessary in each department (notably the inclusion of the letter Eng)"

 

For compatibility's sake, users may use 'ng' if they must. After all, 'ŋ' is just a merger of 'n' and 'g' anyway, and the spelling 'ng' is used instead when it occurs as an autonomous syllable. In my IME, I mapped the 'r' key to 'ŋ' because 'r' is visually identical to the left half of 'ŋ' and easily reachable on the keyboard). The problem with the digraph 'ng' is that it takes up too much space, and specifically more space than its sibling phonemes 'm' and 'n'. This all ties into the syllable-length issue.

 

In any case, I plan to use it mostly for reading, and the IME (which is ultimately used to render characters anyway) takes care of the extra letter. I'd have liked to use the similarly velar 'q' for this sound, but it just appears out of place.

[imron]

hence my concession: "Minor sacrifices were necessary in each department (notably the inclusion of the letter Eng)"

Yeah, I read that, but it's like saying "This Ming vase is in perfect condition, except for the huge gaping crack down the side."  What you said above is saying it's easy to type, except for this one commonly used letter, which is impossible to type on a regular keyboard.

 

users may use 'ng' if they must

Assuming you want this romanisation scheme to be used rather than just serving as an academic exercise, they will, not because they must but because it's the path of least resistance.  Be prepared for people to start using numbers instead of tone marks also.

 

This all ties into the syllable-length issue.

You'll find that users care much more about ease of typing than consistent syllable lengths.

[Hofmann]

Here's what I think. Vowel quality determines what kind of nucleus is perceived, not length. It seems like you have made the representation of vowel quality less intuitive by prioritizing the representation of length.

[ParkeNYU]

@imron

 

You're ultimately right, and so I think this system is better suited for reading (like Zhuyin).

 

@Hofmann

 

I made an effort to maintain intuitive nuclei (for English speakers), but because it is a Romanisation scheme, many aspects will be frustratingly subjective. For example, I think that 'ae' is a very natural assignment for [ɛː], but others may feel that it implies [æː] or something else. This kind of issue is sadly unavoidable when you use the Latin alphabet to represent an unrelated language.

 

I don't wish to imply that this is the best Romanisation possible for Cantonese, but it's what works best for me, and I thought perhaps there were others who might find this system more natural and intuitive than existing systems. Like I said in my first post, I ultimately want to encourage people to craft their own systems. In the end, we all use characters and speech to communicate in Chinese anyway. I think of these systems as learning tools.

[Angelina]

When I said Cyrillic, I meant Macedonian Cyrillic not Russian Cyrillic. 

 

We have the letter њ in Macedonian, so њ for /ŋ/ . All devices and platforms already have Macedonian input. 

 

INITIAL (+medial 'w')    

b б  /     p  п /      m м  /   f ф

d д /      t т/         n  н /     l л

g  г /      k к /       ŋ њ /    h х

gw ѓ /    kw ќ /     w в

j ѕ /        c  ц /        s с  /    y ј

 

I will have to check the tonal system for the rest.  You should type w for њ if you have a QWERTY keyboard. 

[Angelina]

It will probably look like this 

 

"Јàн јàн сāaњ цōт лài ѕau хai ѕee јàu гâe, хái ѕūuн јèeм тòњ кùuн лei сoeњ јāт лoт пèњ дáњ. Кǒј дei гoј јǎu лěi сêњ тòњ лǒeњ сāм, јèe цáe јēњ гōai јoњ хēњ дai гāaн гâe ѓāaн хai лài вoo сōeњ дôј дoai." -Сâi Гâai Јàn Кùun Сūun Јèen (Лùuн Хaп Ѓôaк)

[ParkeNYU]

I think that 'c' might be better written as 'ц' rather than 'ч', since it sounds more like /ts/ than /tʃ/ in Cantonese.

[imron]

I ultimately want to encourage people to craft their own systems

The great thing about standards is there are so many to choose from.

[ParkeNYU]

In this case, I'm not trying to make the standard, but rather a standard for myself and whoever else would like to adopt it (not competing). Even if I'm the only one in the world who uses it, it's no less helpful for learning if I adapt educational materials to this system. The reason that it exists in the first place is that I find it easier to use than other standards, after all.

[imron]

it's no less helpful for learning if I adapt educational materials to this system

Arguably it is, because the time you spend adapting existing educational materials could be spent instead on learning.  It's less helpful because you are unable to make ready use of pre-existing materials.

 

Note: Don't get me wrong though.  I'm sure there is plenty of value in developing your own systems as an educational tool to gain a thorough understanding of the phonetics of a language.  I just don't think they are that practical for actually learning a language.

[Angelina]

I think that 'c' might be better written as 'ц' rather than 'ч', since it sounds more like /ts/ than /tʃ/ in Cantonese.

 

Yes it does. Thank you!

[Angelina]

ѓ, ќ  and в are not used to represent the same phonemes in Macedonian, but it is convenient for me to use them now. 

[ParkeNYU]

the time you spend adapting existing educational materials could be spent instead on learning.

 

In the age of physical textbooks and dictionaries, sure. Fortunately, many educational resources are now digital, which allows me to apply conversion software; this effectively does all the work for me in an instant. In the age of digital technology, it's easier than ever for students to have it their way.

[Angelina]

Arguably it is, because the time you spend adapting existing educational materials could be spent instead on learning.  It's less helpful because you are unable to make ready use of pre-existing materials.

 

Note: Don't get me wrong though.  I'm sure there is plenty of value in developing your own systems as an educational tool to gain a thorough understanding of the phonetics of a language.  I just don't think they are that practical for actually learning a language.

 

There is no standard Romanization/Cyrillization system used as widely as 拼音 for Mandarin, we might as well make our own systems.

 

Most pre-existing materials are in Yale, and as I can see, it is not that good. It's like using Wade-Giles for learning Mandarin. Jyutping is acceptable, but it is not perfect and it is better to improve it before more materials are published. 

 

Also, my writing system is extremely phonetic, I will be forced to transcribe Cantonese personal names into Macedonian anyway.