Jump to content
  • Sign Up

Tangent Constructed Chinese


Recommended Posts

Edit: So I see that the thread has been split. This is originally in response to a thread about Middle Chinese. The original post will be spoilered.


Edit: A few things mentioned in the following posts are outdated. Please check out the latest replies for current revisions. Current version includes voiced initials.


LATEST developments can be read here: http://tcchinese.tumblr.com/


So I created this Tangent Constructed Chinese, which is also used by a friend of mine.

(given in the video is the IPA representation of TCC, not the Romanization)


The Tangent Constructed Chinese (TCC) is meant to resemble a modern Chinese language (like Cantonese, the Hokkien languages, etc) that closely resembles Middle Chinese. In fact, I even prepared for a “future sound change of palatalization of consonants before /i/”, as with many other Asian languages, including Korean, Mandarin and Japanese.


I started to have this idea for TCC when I was learning about ancient Chinese culture at school, and was trying to tell my Western friends about them. I needed a sort of Romanization, for say, concepts like and .

Looking at online sources, they mostly only list the Mandarin readings, and sometimes also the Cantonese readings, although the Jyutping Romanization has been modified to Yale because of opposition (most likely). But the point is, I needed a Romanization for representing these Chinese concepts, but why not use Mandarin or Cantonese, the two Chinese languages I speak?

There are several reasons. I watched a video that was part of a Japanese-teaching series. In one of the videos, the Japanese teacher said that the Japanese kanji have a “Chinese reading”, which is meant to imitate the Chinese pronunciation. And he started with , which is pronounced ichi in modern Japanese “Chinese reading”. He comments, however, that it “does not actually sound like Chinese because in Chinese it is pronounced i (English approximation: ‘ee’)”.

Another thing is, I just do not feel comfortable when I tell my friends that say, is pronounced nian in “Chinese”. Because Chinese is a collection of languages (even though people would say that those languages are actually dialects because of political reasons), and the character, as far as I’m concerned, is pronounced with an –m ending in a majority of Chinese languages. So it seems like I’m claiming the character is pronounced “nian” when like 90% of “Chinese” isn’t pronounced this way. In TCC, the character is written as “niem~” in my Romanization.

So I think Mandarin and Cantonese are both not very good representations of “Chinese” when I need to refer to a concept/name/etc. Mandarin lacks one of the prominent “four tones” of Chinese: the checked tone. And Cantonese, the only other Chinese I can say I know well enough, lacks medials, which are also a prominent feature of Chinese that is found in many Chinese languages.

This TCC is designed to be used when I need to refer to Chinese things that aren’t things like placenames to Western people, for games that I make which need a Chinese pronunciation of things and for stories I or my friend Zsolt writes.

The TCC is not, however, a fully designed language. For each character, in constructing the TCC pronunciation, I need to refer to the characters' pronunciations in Cantonese, Mandarin, Hakka, Hokkien, Teochew, Japanese and Korean, and sometimes even Vietnamese and Ng (Wu Chinese), which, not very surprisingly, takes time. Although just Cantonese, Mandarin and the TCC Pronunciation Construction Rules would mostly do the trick, so it doesn’t always take THAT long.

Also, the tones have not been fully determined. Only the tone types have been: level, rising, departing or checked.

Since it is mostly just a Romanization that’s meant to be used when presenting Chinese-related things and used in games and stories, the grammar and vocab are a side dish, although they are meant to be based on Classical Chinese, and not using words that are unique to special language groups. One example I could give here, is “Yang Xiao Gui” in Mandarin, which would be written as “Yong Gwi Ds(ê)i” in TCC. (The “g” represents an unaspirated /k/ while the e-circumflex represents a schwa).


Hello. I'm new to this forum.

For that link up there, I don't really see any middle Chinese pronunciations. I see Cantonese, Mandarin, Japanese, Vietnamese and Korean though.

I myself have been trying to look up on Middle Chinese pronunciations.

And here's something that I did, which you might find useful. I created "Tangent Constructed Chinese."

It's a constructed form of Chinese that is meant to be an accurate representation of Chinese. For instance, for the character 念, it is read in Mandarin as "nian" and Cantonese as "nim". While Mandarin lacks the -m ending found in many other Chinese languages and the Korean reading, Cantonese lacks the "core vowel" from other Chinese variants. Mandarin also lacks entering tones, which is one of the "four tones" of classical Chinese. So I have taken the interest of creating my TCC, after studying the characters' pronunciations in Cantonese, Mandarin, Hakka, Hokkien, Teochew, Japanese and Korean. Vietnamese and Ng are sometimes referred, but not much. While my constructed Chinese is meant to be a descendant of Middle Chinese that closely resembles Middle Chinese, this was the result. So I think my TCC would be something you would be interested in if you're looking for Middle Chinese pronunciations.


Link to comment
Share on other sites

Wow...6 years later. In case a later reader wants a more complete answer, there is now a Wikipedia article on Middle Chinese with relevant information.


Anyway, if I'm understanding you correctly, you created a composite language out of several descendants of Middle Chinese (and Hokkien)? Sort of reminds me of General Chinese, except General Chinese is a romanization system that can notate Middle Chinese without ambiguity.


So you claim it's an accurate representation of Chinese. How did you choose reference varieties and what were your methods of aggregating them?


Actually, wouldn't this be better as a new thread?

Link to comment
Share on other sites

Hmm... Probably. Should I start a new thread to explain all this then? I do have some explaining to do, as well as answering your question.

I didn't realize this was way back in 2007 because I was simply looking up on Middle Chinese for my TCC and this popped up, and so, seeing how that link doesn't offer Middle Chinese pronunciation (seemingly), I wanted to mention my work, which might be of your/their interest.


I didn't know about General Chinese though. I'll read about it.

Link to comment
Share on other sites

OK, so I've restarted the topic with a proper introduction. The posts under mine were originally targeted at the post I had in spoilers.

I still have work on my TCC. But in the meantime, I do welcome comments. Just so that readers wouldn't be confused and think that I haven't been answering questions.

Link to comment
Share on other sites

Well I guess no one would be interested in this, or will comment. But I don't think the General Chinese mentioned will be a working diaphoneme, from what the article stated. For instance, it has no mention of the Cantonese phoneme oe. The characters 長, 棋 and 地 are shown to be voiced, but they are unvoiced in Sino-Japanese. 子 and 是 have no vowel.

Towards the beginning, the article states that anything yang has a voiced initial. That made it seem easy to construct voiced initials. But then we have 蠻, maan4 in Canto, man2 in Mandarin but ban in Japanese. And a large number of the voiced words turn out to be unvoiced in Japanese. And there's also no mention of what tone the characters are in when it is a checked tone.

After some analysis of my own, I think Chao's idea, which isn't clearly shown in the article, is that, for yang tones, if it starts with an aspirated plosive, it is voiced in GC if the tone is level or rising; and if it starts with an unaspirated plosive, it is voiced. But no mention of what that voiced phoneme is or what about non-plosives, like the case of 聞, which I have enough info to reconstruct GC, fortunately. For instance, 十 is ship in voiceless TCC, and it is zhip in GC, but  it is actually dzyip in Middle Chinese. So we know it is voiced. But what is it? Dz? Z? Dzr? We don't know either. And stranger is the character 怪...

 Though, languages have their own grammar and vocab, and it would only work to the creator's intention if everyone was speaking written Chinese.


re-EDIT: Did some more analysis today, after knowing about GC:

Click for full size.


Link to comment
Share on other sites

General Chinese doesn't claim to be able to predict with 100% accuracy how a descendent language will be pronounced because languages have exceptions to their patterns of evolution. That's why 長 (平) is voiceless in Sino-Japanese and the Cantonese oe (in Jyutping, I assume) is sometimes produced from 魚三等開 (iu) and 戈三等合 (iue) when it is usually 陽三等開 (iang).


After some analysis of my own, I think Chao's idea, which isn't clearly shown in the article, is that, for yang tones, if it starts with an aspirated plosive, it is voiced in GC if the tone is level or rising; and if it starts with an unaspirated plosive, it is voiced. But no mention of what that voiced phoneme is or what about non-plosives, like the case of 聞, which I have enough info to reconstruct GC, fortunately.

Are you talking about Mandarin? In any case, evolution to modern Chinese tones is summarized in a table in this article.


For instance, 十 is ship in voiceless TCC, and it is zhip in GC, but  it is actually dzyip in Middle Chinese.

OK I take back my statement that GC can write Middle Chinese without ambiguity. It seems some MC phonemes aren't preserved, such as the 常 initial.

And stranger is the character 怪

What's so strange? With a 見 initial and 皆二等合 final, it is "cuay" if I'm getting the system right.

But hey, I'm interested in your system, if you can explain the rules used to make it. For example, your 十 is "ship" but 何 is "ĝo". Why is one voiceless and the other voiced? Also why do 何 "ĝo" and 他 "ta" have different finals when they're both 哥一等開?

Link to comment
Share on other sites

Well, Cantonese, about the tones. Sorry I overlooked that. I was talking more about constructing voiced initials though, which includes turning Cantonese and Mandarin /m/ into Japanese /b/


OK so I guess GC is more of a diaphoneme for most current Chinese languages instead of 100% all Chinese languages AND Middle Chinese. Still, it's great.


Oh I think I misread for then. I don't see a problem now. Although I have been wondering what GC <ae> and <e> actually stands for... I used /ɛ/ and /æ/ simply out of guess.


But technically my TCC has a lot of rules, and I probably cannot explain all of them, but I'll list out some of the more important ones in my opinion. Thanks for your interest.


About ship, it was back at the time when I was clueless about how to construct voiced stop initials, because Mandarin and Cantonese both lost them, and Korean doesn't help. Now I will probably reconstruct it as zhip, as with GC, the reason being GC simply voices the initial without turning say, sh into dz, only to zh.


Teochew and Japanese takes precedence, in a way, because they are said to preserve Middle Chinese more. So if there's any contradicting info, usually I will go with those.

For my Romanization, b/d/g would stand for unaspirated, p/t/k for aspirated, and bh/dh/gh for voiced.

If a character is pronounced "m" in Cantonese, "w" in Mandarin, and "b" in Japanese, it got reconstructed as bh, after 今晚:

Old Canto: gom maan, Old Mandarin: gin wan, Japanese: konban 

But now that I know GC, it seems that GC uses <v> for it, which is a separate phoneme, so I'll probably need to work on that.

If in both Canto and Man, it starts with "m", but it starts with "b" in Japanese, it gets reconstructed as "bh":

(Canto: mok, Man: mo, Japanese: baku, TCC: bhak) and (Canto: maan, Man: man, Jp: ban, TCC: bhan)

If a word is pronounced -eng in Mainland Mandarin but -ong in Taiwan Mandarin, it gets reconstructed as -ung or -ông.

If a word is pronounced -uu or -ou in Sino Japanese and ends with ng in Cantonese, it becomes -ung or -ong/-ông in TCC.

e.g. (Cantonese: lung; Mandarin: long; Japanese: ryuu; Korean: ryong; TCC: liung {the i here represents [j], the medial form of /i/.})

If a word is pronounced "ng-" in Cantonese but "y-" in Mandarin, it's reconstructed as "gn". e.g. : Canto: ngaa, Man: ya, TCC: gna

If a word is pronounced "y-" in Cantonese but "r-" in Mandarin, it's reconstructed as "gn"

e.g. : Canto: yan, Mandarin: ren, Japanese: nin/jin, Middle Chinese: nyin (ng represents a palatal nasal, while I use "gn"), TCC: gnin; : Cantonese: yap; Mandarin: ru; TCC: gnup

The Mandarin palatal series "j q x" was the result of merging historical "g k h" with "z c s". Refer to Cantonese for which to use. If Cantonese uses neither, use "g k h".

e.g. : Canto: hei, Man: qi, Japanese and Korean: ki, TCC: ki

Cantonese eu [œ] mostly becomes "io/yo", after Hokkien, although it seems that it is “ya” more, in Middle Chinese.

The tone types are the same as Cantonese, as reflected by Baxter's Transcription for Middle Chinese. They can be omitted, but when written, I usually use superscript / for rising and ~ for departing.

Cantonese "short a" becomes "i", "u" or "o" depending on Japanese and Mandarin pronunciations if it's not "e" in Mandarin:

(Canto: can; Man: chen; Jp: chin; TCC: chin); (Canto: mat; Man: wu; Jp: butsu; TCC: bhut); (modern Canto: gam; Jp: kon; TCC: gom), distinct from (Canto: gam, Kr: kim; Jp: kin; TCC: gim)

It becomes “ê” before –ng when it is –eng in Mandarin

Use Cantonese final consonants, except syllabic ng

Mandarin medial yu becomes –iu-

Cantonese yu and Mandarin ru becomes gnu

Cantonese –ei and –ai will mostly become –i. To be exact, the character’s pronunciation in other languages will need to be referred to.

Cantonese o and e become "a" and “ia” mostly, except after velar consonants, when o remains o, as seen in other Chinese languages and old Mandarin.

Mandarin un, ui, iu and empty i become true un, wi, iu/yu and i respectively, as seen in Japanese and Korean. e.g. (Korean: wi, Japanese: i, Mandarin: wei, Canto: wai, TCC: wi) and (Japanese: ki, Mandarin: gui, TCC: gwi)

Mandarin -iao and Cantonese -iu becomes –ieu

Japanese o, if pronounced in Cantonese as w-, has w- as an initial. e.g. (Canto: wong; Jp: ou; TCC: wong). Although on this –ong vs –ang thing, Zsolt told me that Japanese au was historically merged with ou, so –ang and –ong become indistinguishable, so “” is probably “wang” in Middle Chinese.

- Because of the historical loss of medial -w- in Japanese, "w" is added when it's present in Mandarin, e.g. (Man: guai; Jp: kai; TCC: gwai)

- Cantonese "f" could become "hw", depending on Mandarin and Japanese

- Canto -eng/-ing becomes ieng


And if what I’ve interpreted about GC is true, for yong tones in Cantonese, if it begins with an aspirated plosive for level and rising tones, the plosive should be voiced, and if it begins with an unaspirated plosives for departing and checked tones, it should also be voiced. However, the problem is that , which is pronounced with an unaspirated initial in the yong departing tone, has a voiceless “ti” in Japanese. So either Cantonese tones will need to take precedence to Japanese, or that couldn’t accurately predict what should be voiced and what shouldn’t.

Link to comment
Share on other sites

Well, this is not to say that your method won't work, but if I were doing something like that, I'd rather just make a romanization system based on Middle Chinese instead of taking different samples of modern Chinese, because as you attempt to get more and more accurate, you'll just get closer and closer to Middle Chinese.

And then you'll be doing the same thing as Bernhard Karlgren and Li Fang-Kuei, William Baxter, and this guy. I actually prefer that last one as it's easiest to type. (Now that I've read the General Chinese more closely, it seems pretty incomplete.)

Link to comment
Share on other sites

I would make a Romanization of Middle Chinese if I had the pronunciations of the characters in Middle Chinese. There are many attempts at constructing Middle Chinese and they yield slightly different results.

Although I'm not so sure about Middle Chinese, because clustering things like kjuw (9 in Baxter) and a lot of those things seems a bit weird to say that it is pronounced that way in Chinese. I probably started work on this before I knew about Middle Chinese.

Because I AM trying to use a Chinese for games and explaining Chinese things and stories.

I think probably "a simplified version of middle Chinese" would be a good term to describe what I'm trying to go for.


But I guess even with all the pronunciations in Middle Chinese provided, doing this would still be interesting.

I will also happily take requests, although probably no one other than Zsolt is interested... He asked me to give him TCC readings of characters in his story.


Although I wonder if voiced initials are still "neutral". I mean, many Chinese languages lost them.


That Romanization from Polyhedron looks promising. If only I understand what those terms mean and refer to...

Link to comment
Share on other sites

Well, nobody knows how Middle Chinese is pronounced, although we're pretty sure about the initials.


Roughly, confidence level about

initials is 90%

finals is 70%

tones is 40%


and maybe if you read about it in more detail it won't seem so weird.

Polybiohedron also made recordings, although I think is pronunciation isn't very careful and some characters are just wrong.

Link to comment
Share on other sites

I am referring to how an individual character is pronounced, actually. As in finding the pronunciation of any character I want.


By weird I was more comparing it with modern languages and how an average Westerner who doesn't know about Chinese would think, after only seeing things like "Gung Hey Fat Choi" and "Kung Fu". I wouldn't want to make Chinese so repelling to them by clustering too many consonants together...


Because of the limitations I will not be able to watch videos in quite some time. So he pronounced things in Middle Chinese? I wonder how he deals with the checked tones, and whether he has distinguished yin from yang tones then...


Though from the thumbnails he seems to be using a notation I'm not familiar with... Perhaps his own?

Link to comment
Share on other sites

Characters' pronunciations can be found in 廣韻 or any dictionary that has 反切, like this one.


Whatever. I would rather present Middle Chinese as it was in order to facilitate explanation about evolution.


I don't think he distinguishes 陰 and 陽, since they developed from the loss of voicing.


He uses his own notation.

Link to comment
Share on other sites

I cannot open .rar for reasons I had better not state.

I might do a video on my another TCC bunch out of interest, since I've had some progress after you showed me the videos. Although I will have to edit the thing.

Does 廣韻 really give the exact tone contours and not just the tone types, level, rising, departing or checked? Because one could pronounce it only if they are given. Although one could still guess...

Link to comment
Share on other sites

7-zip is not an option. In fact, nothing is, which is why I'm stuck. Better not talk about the reason. Basically anything that requires downloading isn't a good option.


Well, that is why... I would want the contours too. The exact pitches, if you will. Although that guy did assume all the tones were yin tones. Not sure if the Gwangyun is something I'd look for though... If it gives me like for instance, 秦 is pronounced "dzin with a level tone", that will be good. I've checked out some Chinese articles on Wikipedia but I don't understand descriptions of the tones...


So I checked out an online version. Does it cluster a bunch of characters with similar/the same rhyme or characters with the same initial together? It doesn't seem to be the case... 多 and 水 are in the same column.


Meanwhile, Baxter says it is pronounced Dzin SyiX Hwang while the Wikipedia article on Qin Shi Hwang gives SiB. I've asked a friend to correct it and hopefully it will be and stay corrected.

Because from my TCC, it cannot be pronounced as an apical consonant /s/. It is retroflex in Mandarin, which means that it is either palatal or retroflex in Middle Chinese, and Baxter proved me right.

Btw, that is part of the thought process I go through when constructing TCC pronunciations.

Link to comment
Share on other sites

That's unfortunate. It's quite a useful spreadsheet. (but aren't you downloading these pages if you're viewing these forums?)


I'll just give you a screenshot so you can see what I'm talking about.



Nobody will give you exact pitches, and if they do, it's their guess. When I read aloud I just use 33 for 平, 15 for 上, 51 for 去, 33C (or 3C) for 入. It's a simple scheme that works for differentiating the tones but I don't know if that's really how they were pronounced.


What this spreadsheet gives is the 反切 (column C), initial (column H), final (columns F, I, J, K), tone (column L), Biopolyhedron's romanization (column M), and Blankego's romanization (column N). Characters are grouped by rhyme. As you can see, all characters in the screenshot rhyme with 東.


As you observed 始 starts with a voiceless alveolo-palatal fricative.

Link to comment
Share on other sites

Looks a bit complicated but I think I get it. Although it seems like there are some differences in the reconstructions, like iung vs hiung.

In modern Chinese, 入 isn't a tone itself, but rather, just any characters with a -p/-t/-k ending, so there could just be as many checked tones as there are for normal tones. What I mean is, for instance, Cantonese has 6 non-checked tones, and as I have observed, it has at least 5 checked tones, which are equivalent to the 6 non-checked tones, but checked.

Therefore, I'm not sure if pronouncing a checked tone as the level tone was really the case. Maybe it had checked level, checked falling and checked rising? Or maybe it was actually a seperate tone itself in Middle Chinese, so it was listed as one of the four tones? Like maybe level was 22, rising was 35, departing was 51 and checked was 3 (analyses use only one number for level-checked tones)? I guess we may never know...

Link to comment
Share on other sites

Linguists who study this stuff are pretty certain that 入聲 was not a true tone in a linguistic sense, but any syllable with an unreleased plosive at the end, just like in modern Chinese. Characters with 入聲 might have been pronounced with different pitch contours, but as you said we may never know. In any case, it did not make a phonemic difference, so reconstructions don't create any difference.


Cantonese has 6 non-checked tones, and as I have observed, it has at least 5 checked tones, which are equivalent to the 6 non-checked tones, but checked.

What are these at least 5 checked tones that you have observed? I recognize only 3.


陽 tones in Cantonese are a more reliable indicator of voicing in Middle Chinese, but it doesn't always work. For example, , whose only readings are (in Jyutping) lap1 and nap1 where one would expect lap6, must be voiced in Middle Chinese as it starts with a /l/.

Link to comment
Share on other sites

Join the conversation

You can post now and select your username and password later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Click here to reply. Select text to quote.

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...