Of making many books...

There Is No End

Cantonese and Mandarin

Most students of Chinese just learn Mandarin. A few just learn Cantonese (or other Chinese hua such as Shanghainese). But there must be a few who need to learn both Mandarin and Cantonese. Yet there is very little help available for this group.

In my case, I am living in Hong Kong (Cantonese speaking) and Shenzhen (Mandarin speaking), so any help for learning both languages would be useful. Of course, I could learn both hua separately. But there seems to be considerable similarity between them, and in fact both are descendants of Middle Chinese. So it should be possible to simplify the process to some extent.

There is at least one learning course, Yin-Ping Cream Lee's A Short Cut to Cantonese (Greenwood, Hong Kong, 1998) that addresses the grammar issues and the differences in vocabulary. This is designed for English-speakers who know some Mandarin. What the Short Cut does not address is the relationship between the forms of Mandarin words (or rather, of the pronunciations of characters) and their Cantonese equivalents, for the 70% or so of characters that are common to both hua. (The differences between traditional and simplified characters are irrelevant here, as both orthographies can be used for any hua).

So I have been doing my own research. I've found enough to think that there should be some simple help available. I'm not sure why it's not.

What follows is very much my own unchecked research, and no doubt full of errors, but enough to suggest that someone who knows more than me — someone who actually knows Mandarin and Cantonese for example — might be able to do the job properly.

Finals

Take for example the following table: ('ϕ' means 'nothing')

MandarinCantonese
ϕϕ
-n-n
-ng-ng
-n-m
-u-u
-i-i
ϕ-p
ϕ-t
ϕ-k

Now this covers at least 85% of all characters, at least of the 20,000 or so in Unicode that have both Mandarin and Cantonese pronunciations.

Tones

Tones are the most difficult part of spoken Chinese to learn, for native speakers of English and other non-tonal languages. But consider the following table:

ClassMandarinCantonese
closed/open11
open/open24
closed/open32
open/open35
closed/open43
open/open46
closed/closedanything1 or 3
open/closed46

Here 'open' means nasals ('m', 'n', 'ng') and vowels ('a' etc). 'Closed' means anything else ('p', 'f', etc). The first of these is the start of the syllable and the second is the end, so 中 (Mandarin zhong1, Cantonese zung1) is closed/open (and the tones are in fact as in the first line of the table: Mandarin 1, Cantonese 1). 牛 (Mandarin niu2, Cantonese ngau4) is open/open (and the tones are in fact as in the second line of the table: Mandarin 2, Cantonese 4). 八 (ba1/baat3) is an example of closed/closed (the final is only closed in Cantonese as Mandarin does not have syllables ending in -p/-t/-k). In this case, the Cantonese tone is 1 for short syllables and 3 for long syllables.

This table is less accurate, but represents the most common cases. Sometimes the 'open' or 'closed' is wrong, such as for 哀 (Mandarin ai1, Cantonese oi1) which should be open/open but has tones for closed/open.

Initials

It should be possible to create a similar table for the initial consonants. Sometimes these are straightforward eg 三 is Mandarin san1, Cantonese saam1. See the following values:

Mandarin (otherwise)Mandarin (before -i-)Cantonese
k-ci-k-
g-ji-g-
h-xi-s-
ch-ci-c-
zh-ji-z-
sh-xi-s-

Note that the Mandarin ci- can correspond to a Cantonese k- or c- (etc). Other changes include: Cantonese has lost the 'sh' sound, Cantonese s- can correspond to Mandarin sh-. Mandarin has no ng- initial, so Cantonese ng- can correspond to Mandarin ϕ. Sometimes there is a loss of aspiration eg Mandarin t- vs Cantonese d- or vice-versa.

What else?

What is not covered by these tables is the vowel or dipthong in the middle. This seems to be all over the place.

Historical basis

These tables are based on the fact that Mandarin and Cantonese originate from the same language Middle Chinese, roughly 700 CE. Some sound shifts occurred in both hua, and can be ignored for these purposes. Others gave rise to the patterns in these tables. test: [中]

'hua' (话)

I use 'hua' to mean language/dialect/topolect. See this Wikipedia article for an explanation.