Cantonese/Hakka linguistics

[96981339r@polyu.edu.hk (cflau)]

Gog5 vui4 pen2 yiu1 (Dear Friends),

I would like to comment on the relationship between Hakka, Cantonese and
Mandarin.

Before I studied Chinese dialectology in detail, I just considered Hakka a
Northern dialect coming to the South, and Cantonese came so much earlier
that it is thought that it is almost indigenous to Guangdong. According to
this belief, then hakka should have more deep-rooted northern
characteristics and its similarity to Cantonese should be superficial.

However, this does not fit the facts dug out by me in the past years.
Ironically, Hakka and Cantonese shared deep-rooted features such as
development of tonal classes. From last year onward, I changed my
hypothesis: Hakka and Cantonese were more related to each other than any of
them to Mandarin. The evidences are:

1. Characters with voiced initials of the second tone in Mediveal Chinese
(Zhonggu Zhuoshang zi, ZGZSZ) are divided into two groups. One group (Group
B2a) has about one-third of the total and is changed to Yangshang in Hong
Kong Cantonese, and to Yinping in Hakka. The remaining two third (Group B2b)
are merged to Qu. In Mandarin almost all of these ZGZSZ are merged into Qu.
The most interesting point is that the characters contained in the Groups
B2a and B2b are quite uniform throughtout most dialects in Cantonese and
Hakka, but totally different in Min dilects (e.g. Chaozhou, Xiamen, Hainan),
Wu dialects and Mandarin. This characteristic is so fundamental in dialect
grouping that I can only believe that Hakka and Cantonese are very closely
related from the point of linguistics.

2. Historical linguistics of Cantonese also shows that the Cantonese accent
spoken in Canton and Hong Kong is relatively new. According to the records
of the missionaries in the past three centuries, Cantonese has undergone a
lot of changes:
i. diphthongization of simple vowels - i to ei, u to ou, y (vowel of fish in
C.) to oey (vowel of push in C.), Therefore, for the sentence "I took an
aeroplane to USA" was "Ngo ts'o figi hy Migwok" instead of "Ngo ts'o feigei
hoey Meigwok". The former is still heard in some suburbs near Canton. 
ii. Loss of medials which are present in most other Chinese dialects. For
example, the vowel of the first syllable of "Hong Kong" is now "oe", but it
was "io" centuries ago. Hakka still retain this combination.
iii. Loss of contrast between zh, ch, sh and z, c, s. intersting, almost all
Cantonese dialect in Guangdong has lost the contrast, but in Guangxi some
dialects still distinguish between Zhong (middle) and zong (ancestor). In
Hakka, the Mexian dialect also loss the contract, but dialects in counties
around Meizhou e.g. Wuhua, Dabu still retain tis.
iv. Loss of the apical vowel such as si (four) in Mandarin. In Hong Kong,
some places was spelled into English 150 years ago and has fossilized this
sound. There is a place called "Tsz Wan Shan" which can cause troubles for
foreigners to guess the pronunciation, in which the "Tsz" represented the
apical vowel, which is the same as ci (mercy) in Mandarin. However, we all
pronounce it as ts'i in Hong Kong and is not distinguished from the pond,
chi. In Meixian Hakka, however, even the pond is pronounced as ci for mercy.
But in Hong Kong Hakka, the contrast is still there (mercy is ts'u and pond
is ts'i), but the apical vowel is lost. Interestingly, some Cantonese
dialects still has the apical vowel, such as Doumen near Zhuhai.
v. Gradual loss of contrast between n and l as initials, and the most recent
loss of "ng" as initals.

Combining these facts together, we can find that modern Cantonese appears to
be much more different from Hakka and Mandarin than its variant spoken 300
years ago. Cantonese should have been more similar to Hakka by the beginning
of the Qing Dynasty, but because of language contact with foreigners and
other minoriity races, it is diverging away.

3. Deep-rooted structure such as grammar and pronounciation methods also
shows than Cantonese are more similar to Hakka. This is almost a common
sense for Hakka-Cantonese bilinguals when they encounter Mandarin.

4. Historically, Cantonese-speaking people did not admit that their are
natives here since the Qin dynasty as most other people suggest. They talk
of a unaminous geneology starting from the Zhujixiang of Nanxiong (Northern
Guangdong) near the end of Song Dynasty (ca. 1200 AD). They also insist or
even record that they are Tang-ren and have northern origin, like Hakkas.
While Hakkas also show a similar date of immigration to Guangdong (e.g. My
ancestor Liu Kaiqi came to Guangdong from the year 1235). Therefore,
Cantonese are no "Early comers", or according to the definition of Luo
Xianglin, all Cantonese are "Hakkas". If Cantonese and Hakka were just
Tangren coming to the South at about the same time, then their similarity in
language can be well interpreted.

I am still studying the linguistics between Hakka and Cantonese, which
suggest that Hakka and Cantonese are more related than expected. I will keep
you informed on my new findings.

Yours sincerely,
Liu Zinfad (Lau Chunfat).