Quote
Changgi
unfortunately Hong Kongers themselves are not familiar with any Romanization system other than Basterdized (please excuse the term as that is the most accurate analog to the official Wade-Giles equivalent) and Modified Eitel.
CPS would indeed require them to learn 38 new symbols (37 in Taiwan ROC). Considering the thousands of complex characters necessary for literacy, 38 simple ones are a laughable matter.
Quote
Changgi
I believe I have told you my personal preference of a Romanization system, what foreign learners will think of the system and the flaws with the tone markers
I've found that binary elevation (i.e. either upper-half or lower-half) is sufficiently distinct in handwriting. It isn't common for us to confuse 'P' and 'b' in reasonably clear handwriting, after all. Even with messy handwriting, one can mentally quantize the analog elevation into a binary 'up' or 'down' category.
Quote
Changgi
my computer has not been able to display a number of Cantonese characters and the Hokkien and Hakka specific zhuyin fuhao, so that may cause problems.
Indeed, but as long as they fall within the Unicode standard (U3100 & U31A0), any number of new fonts may be created, and a precious few already exist.
Quote
Changgi
More on the tones, I do not think entering tones should be given separate symbols because first of all there is no phonemic distinction between them and the yim pieng, yim khio and yang khio tones
My tonal glyphs are not strictly based on pitch levels and contours, but rather classical tone categories. By including the entering tone, I have significantly reduced the number of required phonetic glyphs (14 fewer).
Quote
Changgi
and secondly there exists entering tones for say, tone 2 as well, as in tsaak2, dip2. Despite being colloquial pronunciations resulted from tone shifts, they are as real as tones can get and should have status as a tone, and be properly notated. Same goes for Jyutping “eu” and perhaps (non-Jyutping) “diq”
CPS primarily focuses on phonetically representing Chinese characters based on their literary pronunciations. Of course, it is equipped to handle most colloquial sounds as well, but in the case of the 'rising-entering' tone, I consider this an instance of colloquial variation and therefore unnecessary to record (I treat tone sandhi the same way, as does Mandarin Zhuyin). I briefly mention in my document, however, that such tonal precision is still possible with the use of additional clarification marks in dictionaries, textbooks, etc. It is in a way similar to the Hebrew script, wherein vowels are not typically written, yet the option remains for educational and reference materials by adding dots below the consonants.
I know what you might be thinking: "If CPS is not as accurate as Romanisation for transcribing the true sounds of colloquial Cantonese, then what is the point?". Well, this is why I stress that CPS isn't intended to entirely replace Romanisation. So what is its use? CPS is primarily a means of phonetically identifying existing Chinese characters, and most of them have a literary reading, or its reading is otherwise the same in both colloquial and formal speech. Unlike colloquial pronunciations, literary readings are less subject to evolution and drift, and thus offer a more reliable means of identifying specific characters. While most Romanisation schemes aim to visually describe the actual modern tones to varying degrees of explicitness, CPS tonal glyphs represent the original eight-tone classification because this is a more etymologically consistent means of identification. Correct me if I'm wrong, but each and every character used in both Standard Chinese and Colloquial Cantonese possesses an identity in one of these eight tones, even if the Dark-Entering tone can be further divided into high-level and mid-level pitches, and even if some Light-Entering tone characters can become 'rising-entering' or 'tone 2' as you've pointed out.
Quote
Changgi
One thing I haven’t liked about zhuyin is how it represents i and y with the same symbol. While they are semivowels in Mandarin, I do not think that analog would work well in Cantonese.
I don't see why not. /j/ and /i/, along with /w/ and /u/, need not have distinct symbols if they pose no conflict or ambiguity. In fact, having separate symbols for each creates redundancies found even in Hanyu Pinyin like 'ji' and 'wu'. According to the IPA values, the situations are the same in Mandarin and Cantonese.
Quote
Changgi
Another thing I don’t like is how an and ang have separate symbols, but in uses two symbols instead of one, thus requiring learners to learn more symbols than necessary.
Why wouldn't 'an' and 'ang' have separate symbols? They are different sounds. If you meant 'aan' and 'aang', then they each have two symbols because they have long vowels, while 'an' and 'ang' each have one symbol because they have short vowels. Finals with short vowels have their own group of symbols because the short vowels cannot occur without codas.
Quote
Changgi
If I have not interpreted incorrectly, you have chosen to represent p/t/k with m/n/ng? I do think that adds to unnecessary confusion…
I actually think that writing p/t/k as a coda adds to unnecessary confusion because they are not the same sounds as the initial p/t/k anyway; for example, foreigners could easily be mislead (as I was) into thinking that 'saap' should be pronounced as it would be in English, whereby the final consonant has a release of air. Conversely, pairing p/t/k with m/n/ng stresses that the Entering tone demands the coda be abruptly and sharply cut short, but in the same oral positions as m/n/ng respectively. It also has etymological value, as p/t/k morphemes were once considered a different tone of m/n/ng morphemes; writing a final p/t/k is actually a very Occidental concept, and Cantonese is an intrinsically Oriental language.
Quote
Changgi
And you chose to distinguish between the two yim entering tones by vowel length?
But there exists things like gap7 gap8 gap9 and sik7, which is a short vowel, but it has tone 7 instead of 8.
Yes, there are a few exceptions, but I've deemed them a worthy trade-off for having a simplified tonal system. The exceptions, in addition to being comparatively few in number, are mostly for colloquial morphemes anyway, and many of those have alternate tones that fall within the vowel-length rule. As I've stated above, even these differences may be explicitly clarified with additional marks wherever necessary (context usually clears it up, though). Perhaps one instance of common confusion could be the standard morpheme 必, but rare instances like these may be taught individually.
Although I don't express this personal sentiment in my document, my hope is that this tonal system will encourage the high-level and mid-level entering tones to either merge, become allotonic, or at least more rigidly enrich the vowel-length dichotomy. According to a couple of articles I've read, this merger is already starting to take place, albeit slowly. I would liken the reunion of 陰入 to the recent reunion of 陰平, whereby the 'high-level' and 'high-falling' tones rejoined into a single tone. Both of these reunions reflect a reversion to a more classical structure.
You might be interested in an informal survey I did on the matter here: [www.cantonese.sheik.co.uk] (I collected all of the exceptions to the vowel-length rule I could find, however obscure)
Quote
Changgi
I don’t think writing 一 differently depending on the orientation is a good idea… It is supposed to be a character and not a punctuation, after all…
The orientation of 一 is not a strict requirement, but rather an encouragement for more compact writing. Indeed, even in Mandarin Zhuyin, this symbol is not always correctly or consistently orientated, despite having exactly the same orientation rule as CPS. Since 一 is the simplest shape you can really make (aside from a dot), the orientation would likely not confuse many readers (except those literate in Korean perhaps?).
Quote
Changgi
Zhuyin uses separate symbols for diphthongs but I see that you chose to use the Jyutzyu for “j” and “w”, but not very consistently because you have separate jyutzyu for some of them like ai (correct me if I’m wrong)
This has everything to do with vowel length. The rule is that finals with a long vowel and coda always have two symbols, while finals with a short vowel (which require a coda) always have one symbol. If I made all finals into individual symbols, I would have to add 18 new symbols (and that's not even considering p/t/k codas). The clear visual distinction of long and short vowels helps readers very quickly distinguish between high-level and mid-level entering tones (with the exceptions I addressed above).
Quote
Changgi
For the z/c/s things, they aren’t really postalveolar, but more like alveolo-palatal.
According to professionals with whom I've consulted, it really depends on not only the dialect, but on the following vowel as well. They are mostly alveolar (usually before A/E), but can be postalveolar (usually before OE/EO/O/U) or alveolo-palatal (usually before I/YU). In the 19th century, they were distinguished lexically as well, which can be seen today in HK Romanisation.
Quote
Changgi
“CPS explicitly indicates vowel length, which is necessary for the complementary
distribution of the high-level and mid-level forms of the Dark-Entering tone”
Wouldn’t that be as confusing as having a Romanization that indicates im/yang tones of Cantonese by indicating voicing of the corresponding initial consonant in MC?
I actually considered doing just that, but not only would too many additional symbols become necessary, it would mislead learners into thinking that these initials were indeed voiced when they are in fact no longer voiced.
Quote
Changgi
In Cantonese, vowel length is not phonemic.
Unless it's a difference between 'A' and 'AA'... In any case, I mostly differentiate vowel length to help readers distinguish which form of the Dark-Entering tone to use. It also allows for 'kwing/kwik' and 'gwing/gwik' while prohibiting a following 'i' in other cases. It also reminds learners that the short vowels are, while related, indeed different vowel sounds from their long counterparts.
Quote
Changgi
As I have probably stated, I know there exists a number of languages that use “i” for both a consonantal sound and a vowel sound and all, but using that for Jyutzyu, I can only see unnecessary confusion.
'j/i' and 'w/u' are not really pure consonants anyway though, just semivowels that happen to occur as initials. As I had mentioned previously, adding new symbols for them added nothing beneficial to the script and only created redundancies (although I did consider it).
Quote
Changgi
I do not see using “ng” to represent the velar nasal sound as a problem in itself, so I wasn’t feeling very comfortable to hear that as a flaw, but “ng” initials have caused mispronunciations to non-learners, yes, although I’d imagine that isn’t the main goal.
Why write a single phoneme with two letters, irrespective of whether it is an initial or coda? It's a flat-out waste of time and space, and I'd rather the English language adopt the letter 'ŋ' to be honest...
Quote
Changgi
But the underlying things, having to learn something new just for the pronunciation of one’s own native language seems a bit extra and foreigners consider something non-Latin based to be “useless”.
It's this kind of Eurocentrism that plagues linguistic instruction. There's nothing about Latin-based alphabets that renders them inherently superior as a means of transliteration, it's merely how history played out. I'm not accusing you of sharing these sentiments yourself, but foreign students who hold such views are ill-informed. Students of Japanese and Korean, for example, only use Romanisation as a means of introduction to Kana and Hangeul respectively, and those systems are more symbol-laden and complex than CPS (even Mandarin Zhuyin is more complex than CPS). As I mentioned previously, if a foreigner is prepared to learn a language with several thousands of complex characters, an extra 38 simple ones is nothing.
As for the native speakers, CPS would be ideally taught from the moment they enter the education system as a means of making sense of the sounds they've been hearing and imitating since infancy. They would be ideally learned before a Latin letter is even a concept.
Quote
Changgi
But with Jyutzyu being there, will people still bother with Jyutping? Wouldn’t they just use A Romanization when writing English essays and similar things instead of actually using them to learn Cantonese?
It would work much like Japanese and Korean instruction for foreign students: you become briefly acquainted with Jyutping (although IPA is certainly preferable), and use it only in the earliest stages of learning. Once the learner has a sufficient grasp of the core phonology, you assign these phonemes to the CPS symbols and use them exclusively to describe Chinese characters for the remainder of the course. Certainly for English prose (or those in other European languages), the use of Romanisation is ideal to ensure phonetic familiarity. Also, in the rare case that CPS fails to adequately describe the true sound of the character, Romanisation can be used in parentheses to clarify. My biggest problem with full reliance on Romanisation to transliterate the sounds of any non-European language is simple: foreign learners will confuse the letters with their own language's phoneme-letter assignments. If Romanisation must be used, IPA is the only 'one size fits all' answer that I can envision, since other Romanisation schemes tend to be slanted in favour of a given European language (usually English).
Quote
Changgi
I am currently using the Jyutping input system and I’d imagine I’d have to scroll through a lot of candidate words even with Jyutzyu anyway. The input system I’m using includes l- characters for n- characters, requires you to input literary pronunciations (so ge3 does not exist) and requires hak1 for 黑 and a large number of similar characters which I’ve only ever heard being pronounced with a long vowel instead of a short one. I had to scroll through 50 candidate words before I could get to 立 when I typed laap6, and the first 30 or so were characters with an n- initial (typing lap6 would get me there faster but I've only heard that pronunciation a few times).
These troubles you're describing perfectly illustrate the need for a standardised and simplified Cantonese orthography and set of character readings, which is only possible through improved and more rigidly structured Cantonese education (I'm counting on you, LSHK). I programmed a Jyutzyu IME myself, so I had the displeasure of experiencing first-hand how many candidates are possible for a given syllable.
Believe it or not, I actually seriously considered reviving the lexical distinction between Z/C/S and J/CH/SH because it would greatly reduce the number of candidates for syllables with Z/C/S initials. The only issue is, unlike the N/L distinction, very few people would know which characters are Z/C/S and which are J/CH/SH (I guess we're unfortunately too late to save that one). If the N/L merger were allowed to become official, the candidate list would become completely unwieldy (the same goes for other products of so-called 'lazy sounds', such as the disappearing NG, GW, and KW initials, along with the merging N/T and NG/K codas).
For Cantonese to thrive in an increasingly Mandarin-dominated Sinosphere, it needs to be tamed and whipped into shape by government decree. Pronunciations need to be rigidly standardised and unified, and orthography needs to be set in stone, regardless of whether it is CPS or Romanisation. This kind of tough love is what made Putonghua and Guoyu possible, after all. Cantonese can keep as many dialects as it wants, as Mandarin does, but it needs a solid standard (even if artificial) to survive in today's world.
One of the unintended benefits of CPS, actually, is that while you must write the symbol for initial 'N' wherever etymologically appropriate, certain speakers may pronounce it as 'L' if they so wish. By analogy, 'W' and 'WH' are identically pronounced in today's English; they maintained distinct pronunciations until very recently, yet the orthography remains consistent.