Learning different dialects by deduction and not memorization

[louiejac]

My goal is to learn as many different Chinese dialects as possible, including: Taishanese, Shanghainese, Teochew, etc., whilst minimising what I have to memorise from scratch. I already learned Cantonese and Mandarin, and have noticed strong relationships between the finals of Cantonese, Mandarin and Taishanese. Heck, there are even well-documented but imperfect relationships between Cantonese and Mandarin. 

 

I'll post the examples of the correlations I've noticed between Cantonese, Mandarin and Taishanese finals in a comment in this thread. So far, these relationships have worked for me such that I can sort of speak to my taishanese grandmother, without having to memorise the Taishanese reading for thousands of characters as if learning it from scratch.

 

Is there another dialect that I should learn so that I can easily learn a bunch of other dialects by deduction and not by rote memorisation? Should I invest my time to learn Middle Chinese phonology so I can use MC to deduce the sounds of differing dialects? Or is there no longer enough correlation between MC and the different dialects anymore? I'm asking only about how to guess the sound in a dialect; I will learn about the grammar differences some other time.

[louiejac]

Here are some examples I noticed between Cantonese, Mandarin and Taishanese. I don't know all the variants of Taishanese, just the one my grandmother uses  (for which I don't know the specific name).

 

If a word begins with a "y-" sound (j in jyutping) in Cantonese, and contains an "r" in Mandarin PinYin, then there's a high chance that the Taishanese reading begins with a "ng-" sound. I also noticed that these words have an initial of either 日 or  疑 in Middle Chinese.

 

If a cantonese jyutping final and mandarin pinyin final are both "-ing", then the Taishanese final is "-ang". If the same character has a separate Cantonese colloquial reading that ends in "-eng", then Taishanese final becomes "-iang". 

 

If a cantonese jyutping final is "-in", and the mandarin pinyin final is "-an" (eg. "-ian"), then the Taishanese final is "-an".

 

If the mandarin pinyin final is "-in", then the Taishanese final is "-in".

 

And then there's more documentation about the correlations between the tones and initials: http://www.stephen-li.com/TaishaneseLesson/TaishaneseLessons.html

[louiejac]

And after looking up some words in Stephen Li's Taishanese dictionary, I've been able to guess most other Taishanese words correctly without looking them up prior. I verify my guess by asking my grandmother how to pronounce Cantonese words (since she's fluent in Guangzhou Cantonese and Taishanese).

[Michaelyus]

Middle Chinese. A good knowledge of all the possible variations and how they relate to the common stock of recorded Tang dynasty Chinese is really helpful. Even if you don't want to know what a 重纽 is (and really, who does?), it's good to see how they all derive.

 

 

[louiejac]

So I guess now the question is, can I rely solely on a MC rhyme dictionary, learn how each MC initial and final sounds like in each of the modern target dialects, and then deduce the sound of any other character in any target dialect by looking up its MC initial and final? Is there enough documentation what each MC initial and final sounds like in the many modern dialects?

 

I also wonder, why did many words beginning in MC 日 and 疑 get redistributed into Cantonese Jyutping "j-" and "ng-"? Aside from Mandarin, do other dialects have these redistributions and are they predictable?

[Michaelyus]

Most Chinese philology seems to use MC rime dictionaries as a starting base, comparing modern speech varieties to the Guangyun or Jiyun.

 

Then they work out how the distribution occurred. There will usually be a lot of irregularity though.

 

Obviously for Min varieties, there is a lot of Old Chinese, pre-Tang phonology, so that can be a lot harder.

[Michaelyus]

As for the redistribution of MC日 and 疑 to Cantonese: 

 

I think MC 日 becomes Cantonese [j] as standard; there are of course exceptions (餌 goes to [n]; 扔 exceptionally goes to [w]). So this is pretty easy.

 

 

On the other hand, MC 疑 splits; palatalising finals go to [j], losing the nasal; non-palatalising finals preserve the [ŋ]. Compare what happened to Korean ㄴ (corresponding to /n/)  in the South in the initial position before /i/ or /j/.

 

What counts as palatalising of course can be hard to define; the changes in the medial and the final can easily influence which way it turned. In essence, MC ngi always palatalises (apart from, exceptionally, 銀; that's the legendary 重纽 [Division] III weirdness going on), nge usually does too; ngj- usually palatalises but not always; ngjw- is borderline and could go either way; nga- doesn't; ngo- doesn't; ngu- doesn't.

 

Stephen Li's Toisanese stuff has a good set of initial comparison tables! Useful.