Hakka Dialect and Other Topics

My name is Dylan Sung, I authored the first edition of this a long while ago, after a call for FAQ's at the forum. It has turned into an essay on Hakka and linguistics and there seems to be a number of loose ends to tie up yet. Meanwhile my email address should you want to contact me by email is mabr12@dial.pipex.com. If you don't hear from me for a while, please beare with me, I'll try to do the best I can, as quickly as I am able.

List of Contents
• 0.0 Introduction
  • 0.1 Who or what is Hakka as a language?
  • 0.2 What is special about the Hakka dialects
• 1.0 Hakka sound and Tones
  • 1.1 How do we write down Hakka Sounds?
  • 1.2 What are tones?
  • 1.3 Why does the tone depend on the ending of a word's sound?
  • 1.4 Why are there so many tones?
  • 1.5 How is a tone described
  • 1.6 What are initials and endings?
  • 1.7 How many initials and endings are there?
  • 1.8 Hakka Rhyme Tables
  • 1.9 Fan1 Qiad5 - Initials for Hakka
  • 1.A Fan1 Qiad5 - Rhymes for Hakka
  • 1.B Fan1 Qiad5 - In use in Hakka
• 2.0 Hakka Grammar
  • 2.1 Are there any grammars on Hakka?
• 3.0 Hakka Vocabulary
  • 3.1 What about Hakka Vocabulary?
  Top
• A.0 Appendix
  • A.1 Fan1 Qiad5 / FanQie Method of Initial and Ending
  • A.2 Pin2, Song4, Hi4, Ngib6 - Si4 Sang1 Four Tone Categories
  • A.3 If there are only four tone types, how come Hakka has six tones altogether?
  • A.4 Why does my Hakka dialect have falling tone contour for the the Song4 Sang1, when this phrase means suggest a rising tone, and that a rising tone contour is given for my Hi3 Sang when it suggests a falling tone?
  • A.5 Some Hakka Tone Contours
  • A.6 Remembering The Tone Numbers
• B.0 Lau's Pinfa System
  • B.1 Some initial concepts
• C.0 Appendix C
  • C.1 Books and References consulted
  Top

• 0.0 Introduction

  ---

• 0.1 What is Hakka as a language?

  Hakka is a dialect of Chinese. It has many regional dialects, that are strewn across southern China, Taiwan, Hong Kong, and many of the Pacific Rim countries. With the migration of the last century, its speakers can now be found on all the great continents, except Antarctica, but we are working on that one. There are approximately 37 million plus speakers worldwide. Meixian (Moi Yan) is often considered to be the 'standard' Hakka subdialect, though all are equally valid in their own right. The only thing which hinders a linguistic unity, is the lack of a common romanised standard. We aim to provide a method whereby this last point can be overcome in Lau's Pinfa method in section Appendix B.

  List of Contents : To Appendix Index : Top

  ---

• 0.2 What is special about the Hakka dialects?

  Within the Hakka group of dialects, there is a core vocabulary that is common between them. Many Hakka words can not be found in other Chinese dialects. For instance soup in Hakka is called 'sun4' but 'tong1' in other dialects. Often Hakka retains endings that are not present in the Mandarin Chinese dialect and said therefore to be more 'conservative'. Amongst these is the endings -t, -k, -p, -m. Also, the use of the 'ng-' initial sound, although found in other conservative dialects such as Cantonese seems be better preserved as an initial in Hakka [sll]. It has spawn much research into its linguistic origins.

  List of Contents : To Appendix Index : Top

  ---

• 1.0 Hakka Sound and Tones

  ---

• 1.1 How do we write down Hakka sounds?

  For the purposes of recording the sound, we can use a sort of alphabet to transcribe the sounds. There are many types used today. The reason for this is that various authors have developed, independently, their own ways of rendering a sound syllable into a written transcription.

  There seems to be no set standard since the various dialects may not have exactly equal sound values for each 'letter' of the alphabet they employ. See Appendix B for a proposal for a new Standard Romanisation.

  A sound of Hakka can be written in a romanised form. Next, to fix the tone of the word, a number (or some graphical element) is added on to the word. So, in the above words for soup, 'sun4' has the tone number 4, and 'tong1' has tone number 1.

  Sounds themselves can be thought of as two parts, the initial letter which is often a consonant, and the ending, or rhyme of the sound. This method is known as the FanQie (in Mandarin) or in Hakka, the Fan1 Chiad5 method. See Appendix A.1 for more on this.

  List of Contents : To Appendix Index : Top

  ---

• 1.2 What are tones?
  In Hakka, like many languages of the world, tones are used. It is said to be a 'tonal language'. Tones are the variations in pitch of a sound, which are so regular that it can be assigned a mark or number to represent it. As we are writing in English, an example would be:
  • I have.
  • I have?
  The difference between these, is that in english a question can be implied tonally by raising the pitch of the word at the end of the sentence. In writing, we see that english uses the question mark (?) as an indicator of a question, so when it is read aloud, the reader can ajust her pronunciation so that the listeners can realise the added information. But in reality, English is not a tonal language, so there are no tone indicators on any of the words written.

  In Hakka, however, there are several tones. Often the tone also depends on the ending of the sound. There are generally six tones to Hakka, though in one particular, Hailu or Hoi Liuk (HoiLiug), there are seven. The tones are numbered 1 through to 6, and added to the end of a written sound. Alternatively, some sort of accenting can be used, such as -\^/.' though these are just hypothetical.

  List of Contents : To Appendix Index : Top

  ---

• 1.3 Why does the tone depend on the ending of a word's sound?
  There are two main categories of sound in Hakka, called by linguists, "legato" and "staccato". For our purposes we can call them long sounds and short sounds. A long sound is created by a glide. This is the lengthening of a sound which can occur at the begining (on glide) or at the end (off glide). Mostly, Hakka sounds have a small off glide to vowel endings so they appear longer. This also occurs with the three nasal consonant endings -m, -n and -ng in Hakka. This resultant long sound will need a tone to pin its final sound value. The tones associated with these long or legato sounds are 1, 2, 3, and 4. This leaves the short or staccato sounds and they are numbered 5 and 6. They are often associated with the (airstream closing during the articulation or saying of the) occlusive endings -p, -t, -k (or -b, -d, -g).

  List of Contents : To Appendix Index : Top

  ---

• 1.4 Why are there so many tones?
  There are this many to account for the various sounds present. If we were to write "giang" down as a sound, without its tone, we would not be totally sure what its meaning is from all the other words of this sound. By placing the tone number 4 against it, we may recognise it as meaning 'mirror' as opposed to "giang3" for 'neck'. So tones play an integral part in distinguishing, and narrowing down the possible number of meanings that a romanised sound has.

  List of Contents : To Appendix Index : Top

  ---

• 1.5 How is a tone described?
  Tones are described using a arbitrary scale of five levels. A tone can be level, rising, falling, long, short, or anything between. We can think of these five arbitrary levels to be numbered 1 through to 5 from the lowest to the highest rung of a musical stave repectively. The lowest pitch (1) like in music, occurs at the lowest part of the stave, and the highest on the top (5)

  This enables us to write what is known as a tone contour. It is the numerical equivalent of using the stave as a visual representation.

  Tone contours are written within two forward slanting slashes /~/. 5 is the highest level, and 1 is the lowest. A tone contour /23/ describes a tone rising from level 2 to 3. Similary, /423/ is a contour that begins at level 4, decends to level 2 and rises to level 3. If we had /5/, then this shows a short tone (possibly a staccato tone) of level 5.

  You may like to read a suggestion on how to remember the tone numbers if you have a command of Hakka in section A.6.

  List of Contents : To Appendix Index : Top

  ---

• 1.6 What are initials and endings?
  Chinese words are for the best part monosyllabic (having just one syllable). This enables us to split the sound into two arbitrary, but consistent, parts. They are known as the initial and the ending. Initials are single letters, dipthongs, and (in some transcriptions) trithongs (has 3 letters). The remaining piece of the sound is known as the ending.

  For example, taking the word phrase for spider, la2 kia2 the sound of the sounds are "la" and "kia". 'l-' and 'k-' are initials, and '-a' and '-ia' are endings.

  List of Contents : To Appendix Index : Top

  ---

• 1.7 How many initial and endings are there?
  This depends on which method of transcription you use.

  Pinfa was developed by Liu ZinFad [cfl], which uses the initials closely resembling the initials in Mandarin Chinese. It may become a standard if we all work towards it. Appendix B.0 gives information about its basic concepts. A copy of the most recent input can be obtained at :

  http://www.geocities.com/Tokyo/Pagoda/3847/sapienti/hagfa99b.htm for viewing first
  http://www.geocities.com/Tokyo/Pagoda/3847/sapienti/misc/hagahk.zip and download, respectively.

  In the Sa Tdiu Gok dialect, as given by Dylan Sung (mabr12@dial.pipex.com), at
  http://www.geocities.com/Tokyo/Pagoda/3847/sapienti/haksound.htm
  he has created his own romanisation.

  In the Taiwanese transcription given in the link in the Taiwanese Hakka Association of USA (scchang@softidea.com)
  http://www.softidea.com/twhakkausa/hktlpa.html
  shows considerably more initial and endings.

  Also, you can view in BIG5 (redding@oes.itri.org.tw) another Taiwanese Hakka transcription table at
  http://www.geocities.com/Tokyo/Ginza/2168/pinin.htm
  just to compare, giving also information about Hailu (Hoi Liuk) words, formation of sounds and its seven tones.

  List of Contents : To Appendix Index : Top

  ---

• 1.8 Hakka Rhyme Tables
  In the past, Chinese scholars created dictionaries according to their sounds and tones, and generally called Rhyme Dictionaries (such as the Qieyun of 601AD).

  The characters were classed according to their initial and end rhymes set out in accordance with the four basic tones which are called in Mandarin Ping, Shang, Qu and Ru. These rhyme dictionaries were used for poetry often, and so they form an integral part in our knowledge about the sounds of older chinese dialects. So, the ideas of sounds and tones underly the "recent" body of poetical works dating from the Six Dynasties period after the break up of the Later Han dynasty until today, having reached a peak of excellence in all the forms in the Tang Dynasty.

  The idea of using rhymes has lead the scholars in the past to use one character in each of the four basic tone types to represent a rhyme sound that could be compared to all words which had the same rhyme and tone.

  In the next three sections, 20 consonant initials are introduced, then a set of rhymes attributed to the auhor's own particular brand of Hakka, and then its application in Fanqie usage.

  List of Contents : To Appendix Index : Top

  ---

• 1.9 Fan1 Qiad5 - Initials for Hakka
  There are twenty consonant initials used in the Hakka FanCed method.

  

  Fan1Ciad5 requires two charaters for correct representation. In the case of vowels, where there is no consonant initial, the method substitutes a zero' initial, using the character 'ya' as the first character.

  List of Contents : To Appendix Index : Top

  ---

• 1.A Fan1 Qiad5 - Rhymes for Hakka
  The following is proposed by Dylan W.H. Sung as representative of his own dialect. The sounds are just a subset of the whole inventory of Hakka sounds, which Dr. Lau [cfl] has gathered.

  He uses the six tones according to the rhymes of his dialect as follows:

  

  Liu ZinFad's romanisation replaces the set -p -t -k with -b -d -g in most romanisations of Hakka. The vowel dipthong -ia- has mutated in SaTdiuGok to a single vowel -e- in some cases only. There is no -u- (-w-) medial in STG Hakka so they are not represented here.

  Yin and Yang represent the traditional linguistic assignments of high and low tone pitch registers, and P=Ping S=Shang Q=Qu R=Ru tones repectively. It may be seen that due to the differences in one Hakka dialect to another, not all the rhymes are represented here. Also, due to the difference in the tone contours from one dialect to another, the high and low registers may not logically be high and low meaning the tonal pitches can be switched in real life experience. Added to that, the tradtional meaning of Ping(level), Shang (rising), Qu(going/falling), and Ru(entering) are now misnomers - a hangup of the original uses which they were designed for. That is why the numerical representations are necessary here, though they can equally well be referred to in writing by these four names as a general indication of their tonal groups. I.E., Ru tones will always end in -p -t -k (or Lau's -b -d -g).

  So Hakka rhymes can be used to represent a sound which cannot be found in Character inputs using one of the above as a reference.

  List of Contents : To Appendix Index : Top

  ---

• 1.B Fan1 Qiad5 - In use in Hakka
  In order to show that fan1ciad5 was in use, dictionaries such as the Kangxi dictionary would have the word qiad5 displayed after the two characters. Note that the rhyme also embodies the final tone of the sound also.

  e.g. For the word sky, tian1, we have the initial t~, and the rhyme ~ian1. Fan1qiad5 using chinese characters - 2 methods:

  • Square brackets separate the three characters from the body of the text and this indicates the use of fanqie.

    

  • We get rid of the qiad5 character and substitute a \ backslash between the initial and rhyme.

    

  One more example: xiang3 =

  Zero Initial words.

  For sounds which begin with a vowel, it is said to have a zero initial. The character is used together with the rhyme to indicate the sound and tone of the word in question.

  • a hundred million, yid6 =

  List of Contents : To Appendix Index : Top

  ---

• 2.0 Hakka Grammar

  ---

• 2.1 Are there any grammars on Hakka?
  Grammars have been written mainly in book form. Try :

  
  Author    : Mantaro J. Hashimoto 
  
  Title     : The Hakka Dialect; A Linguistic Study of Its Phonology, 
  
  Syntax and Lexicon Princeton-Cambridge Studies in Chinese Linguistics, V
  
  Publisher : Cambridge at the University Press, 1973. 
  
  ISBN 0 521 20037 7
  
  Library of Congress Catalogue Card Number 72-85438.
  
  

  C.F.Lau's site at http://www.acad.polyu.edu.hk/~96981339r/enghom~1.htm gives further information about specifically Hakka sounds and a general overview of Hakka grammar.

  List of Contents : To Appendix Index : Top

  ---

• 3.0 Hakka Vocabulary

  ---

• 3.1 How about Hakka vocabulary?
  There are many internet sites with Hakka vocabulary. If we include some on-line dictionaries, then, some of these can be found at:
  
  • Sa Tdiu Gok Hakka
    http://www.geocities.com/Tokyo/Pagoda/3847/sapienti/hdindexc.htm
    
  • Taiwanese Hakka
    http://www.geocities.com/Tokyo/Ginza/2168/diction.htm
    
  • Comparisons between the sounds in one dialect to another of Chinese
    http://www.asiawind.com/hakka/language.html
    
  • Thomas Chin's pages
    http://132.229.12.115/Chin/Hakka/Tables/index.html

  List of Contents : To Appendix Index : Top

  ---

• A.0 Appendix

  ---

• A.1 fan1 ced5 / fanqie Method of Initial and Ending
  The method of splitting a sound into two parts, the initial and ending is not a Chinese invention. It comes from Indic sources, and applied to Chinese when Buddhism arrived in China at the end of the Han Dynasty, and required a way of translitterating the Sanskrit names of deities and the various theological concepts within it. This method is transcribed in Mandarin Pinyin as 'fan qie' - which can be written in Hakka as fan1 ced5 \ fan1 chet5 \ fan1 tet5 \ fan1 chiad5 etc, depending on the romanisation. Hence, if there was a character needed to be looked up, the dictionary user must first know the sound of the word. The user splits it into the initial consonant, and the ending rhyme. Reference characters representing these two classes are consulted in the dictionaries; one to represent the intial and the other the ending rhyme.

  When these two are used, any word can be looked up. For instance, say we had a word, "sim", we first split this up into the two parts, s- and -im. Then we look in the dictionary for two index characters which represents these two separate pieces of information. If the dictionary used the word 'sa' for its 's' initial, and "lim" for its end rhyme, then a combination of these two surrogate characters would represent 'sim'.

  List of Contents : To Appendix Index : Top

  ---

• A.2 Pin2, Song4, Hi4, Ngip6 - Si4 Sang1 Four Tone Categories
  The introduction of fanqie meant the new dictionaries could use this method to order their entries. Chim2 Yok1 (M. Shen Yue) invented the Si4 Sang1 or 4 Sounds/Tones method, which enabled a finer grading of the characters in the lexicons of the day. He proposed 4 tone types, a level tone called Pin2 (M. Ping), a rising tone which he named Song4 (M. Shang), a falling tone Hi3 (M. Qu), and a tone based one the words which possessed stops -p (-b), -t (-d), and -k (-g), known as Ngip6 (M. Ru) meaning "entering". The ending of Ngip is a -p, so it represents one of the three types of endings of its tone class. Historically, the high pitch tone subtype is named as Yim1 (M. Yin), whilst the low pitch tone subtype is called Yong2 (M. Yang).

  List of Contents : To Appendix Index : Top

  ---

• A.3 If there are only four tone types, how come Hakka has six tones altogether?
  This is a good question, as it requires one to hear the sounds that are represented in these categories. In Hakka, the first class of tones, Pin2, or level tone, has words which have a level tone in both a high and a low pitch, hence Pin tones have two subtypes, a high level tone, and a low level tone. They are given the numbers 1 and 2 respectively. The other class of tones in Hakka which has both low pitch and a high pitch pronunciations of a sound, is the Ngip or entering class. They are given the tone numbers 5 and 6 repectively. That leaves the Song4 Sang1, and the Hi4 Sang1 which are given tone numbers 3 and 4 repectively.

  List of Contents : To Appendix Index : Top

  ---

• A.4 Why does my Hakka dialect have falling tone contour for the the Song4 Sang1, when this phrase means suggest a rising tone, and that a rising tone contour is given for my Hi3 Sang when it suggests a falling tone?
  The answer to the apparent reversal of the tones is based somewhat in the history of the tones themselves and is too complex to answer here.

  List of Contents : To Appendix Index : Top

  ---

• A.5 Some Hakka Tone Contours
  Here are the tone contours of a few types of Hakka dialects, taken from Jerry Norman's book, "Chinese" Chp.9 page 226, and Mantaro J. Hashimoto.

  More tone contours can be found in Thomas Chin's pages.

  Table of Tone Contours I
  Name of Dialects		Ping		Shang	Qu	Ru
  Mandarin	Hakka	1	2	3	4	5	6
  MeiXian	Moi Yan	/44/	/12/	/31/	/42/	/21/	/44/
  HuaYang	Fa Yong	/55/	/12/	/31/	/53/	/32/	/55/
  ShaTouJiao	Sa Teu Gog	/33/	/11/	/32/	/53/	/33/	/55/
  Shi-yen (Miao-li)	Shi-Yen (Miao-Li)	/3333/	/1111/	/3222/	/5543/	/32/	/55/

  Table of Tone Contours II
  Dialect		Ping		Shang	Qu		Ru
  Mandarin	Hakka	1	2	3	4	5	6	7
  Hai Lu	Hoi Liuk	/53/	/55/	/13/	/31/	/22/	/55/	/32/
  The Hoi Liuk dialect has seven tones, the extra tone resulting from the high and low pitch distinction in the Qu tone.

  Sherman Cheung [sc] points out that the Wuhua dialect has been exported to Huayang - SiChuan, so they have the same tone patterns. I hope to put more tone contours up as soon as I have the relevant information.

  List of Contents : To Appendix Index : Top

  ---

• A.6 Remembering The Tone Numbers
  There are two possible suggestions on how to remember how to apply the tone numbers. Both of these really requires you to already have a knowledge of Hakka sounds, and hence some fluency.
  • Utilising the Rhyme Table.

    The rhyme table given above in section 1.8 gives the six tones found in the majority of Hakka dialects. Where possible the many rhyming endings are varied according to the tones. The first four legato tones are the ones which shall interest us. (The tones numbered 5 and 6 are legato tones, and it is easy to differentiate high and low pitch between the two.) By ordering the tone numbers into a high to low sequence, you can remember how to apply the tone numbers easily.

    The author remembers his tones as the sequence 4, 1, 3, 2. This is because tone 4 is the highest in pitch and tone 2 is the lowest, and the variation on tones can be remembered as 4132.

  • Utilising Tone Contours

    The tone contours can be used also since they give the variation of pitch of each tone. The tones contours given above in section A.5 shows that the Fa Yong (HuaYang) and Meixian dialects follow similar high to low tone order, 4132. The Shatoujiao tone is 4132.

    The principle can also be extended to Hoiliuk (HaiLu) and LiukFung (LuFeng) dialects. This has 5 tones are then ordered 21453.

    As a rough guide, the beginning and ending levels of the tone contours are a reasonable guide to find the scale of the pitch order.

    The Shi-Yen (Miao Li) dialect would have 4312 as the order of high to low.

  List of Contents : To Appendix Index : Top

  ---

• B.0 Appendix B - Proposed Standardisation Using Lau's Pinfa Method.

  ---

• B.1 Some basic ideas
  If we approach Pinfa as a romanisation using english as the medium to transmit this idea, perhaps some english words may serve to demonstrate the initials of Lau's method, for those who are unfamiliar with Mandarin sounds.

  Initials:

  There are 17 consonant initials, the capital letters in the english are the approximate values of the consonants in Pinfa.

  b, p, m, f,
  Bone, Pineapple, Meal, Food,

  v (does not appear in Mandarin, nearest being w) Vandle

  n, l, t, d,
  Never, Lime, sTar, Docker

  g, k, h,
  Grey, Kite, Happy

  z, c, s
  aDZe, Talking, Soap

  q, x have the sounds ch (CHurch) and sh (SHout)

  Medials: There are two Hakka medials, i and u, used in Lau's input.

  i (=y) as y in 'yellow'. Can be used as a glide as in ngiad6 for moon.

  u (=w) as u in 'queen'. This is used as a glide where those speakers have it.
  E.g. guong1 (bright) but others who don't have this glide will read it as gong1.

  Mandarin has no occlusive endings -b -d -g (or -p -t -k) so there is no tone 5 and 6. Dependending on the tone contour of your own dialect, tone 5 should be (according to Meixian) a low pitch, whilst 6 is a high pitch, contrast /hab6/ (together) /sad5/ (to murder), so /hab/ is higher in tonal pitch than /sad/.

  In some cases, the mapping Mandarin -iang > -iong can be seen, though in the case of M.Jiang1 (river) we have H. Gong1

  Hakka -ia- can be thought of as -e- as well, rhyming with the english "bed", so ngiad4 can be nged4 (though there seems to be a slight glide before the e, hence ngied4), and ngian2 (year) is ngen2 or ngien2, tian2 (paddy field) is ten2 or tien2 etc....

  Short vowels length (with the middle of a written sound):
  a = bat
  e = get
  i = bit
  o = got
  u = put
  

  Long Vowel length (vowel appears as an end letter):
  a = rhymes with car
  e = rhymes with air
  i = rhymes with been
  o = rhymes with core
  u = rhymes with loo
  

  Vowel Dipthongs:
  These comprise two or more vowels used as a representation of a sound in Hakka. There are two rules to observe in the transcriptions.

  • Rule 1.
    Where there is a vowel at the end of the dipthongised sound, as in /la2 kia2/ (Spider), the second sound kia2 has a long -a ending, but a short glide -i-.

  • Rule 2.
    Where a vowel dipthong occurs within the sound, and is enclosed either by another vowel, or by a consonant ending, then the sound is short. If the enclosing letter is a vowel, hence creating a tripthong, then rule 1. is observed.

  • Example a:
    -iau is a tripthong as in kiau2 (bridge), -ia- are short vowels (rule 2), but u is long (rule 1).

  • Example b:
    -iang and in giang4 (mirror) -ia- short (rule 2) because a consonant ending encloses the sound.

  Differentiation between some of the vowel dipthongs.

  Dr. Lau [cfl]has pointed out the difference between using -iu and -eu in an email communication to [dwhs].

  Notice that you have a typo in my pinyin system, which is a bit hard for you because you cannot speak the Meixian tongue. "Ngid5 tiu2" in our dialect is "Ngid5 teu2" in Mexian, the difference being the lowering of the main vowel i to e. This will not cause communication problem in most of the cases, but they are distinctly two different sounds. Contrasts between "iu" and "eu" do occur after d, n, l, z, c, s but those pronounced with iu after z, c and s are changed to j, q and x. For the most circumstances, "eu" occurs more often than "iu". However, my family name, Liu2, is pronounced with "iu" in all subdialects. You and I make no difference between my surname and a building in our native tongue, both pronounced as liu2. But many other Hakkas pronounce the latter as leu2.

  It is then worth bearing in mind, that though this gravitates towards the Meixian dialect, it nontheless is valid as Standard Romanisation because the sound values of the vowels have been fixed above and hence there shouldn't be too much misunderstanding between one speaker's romanised work and another.

  We also would like to point out the following similarities between MeiXian and other dialects:

  
  Meixian > Other Mapping > Further Change
  
  -------->--------------->----------------
  
  -ian    >     -ien      >    -en 
  
  -eu     >     -iu
  
  -ieu    >     -iau 
  
  -o      >     -au
  
  

  Wu Hua as mentioned by Sherman Cheung [sc] and also SaTeuGog [dwhs] agrees with -au as the rhyme in words where they appear as -o in Meixian.

  You can probably get away with leaving the tones out, but in general, the tones 1 to 4 in Mandarin agrees with Hakka in most cases. Where you know that there is an occlusive ending, then apply the tone 5 or 6 depending on the particular pitch contour of your brand of Hakka. Otherwise, use the Meixian usage above. (you could draw a guess from Mandarin, if you know that it is tone 2 for example, then the occlusive ending may be the low pitch tone 5 in Hakka - though this may be overstretching it a little).

  More information can be obtained from C.F.Lau's site at
  http://www.acad.polyu.edu.hk/~96981339r/enghom~1.htm

  List of Contents : To Appendix Index : Top

  ---

• C.0 Appendix C

  ---

• C.1 Books and References of Interest
  • Mantaro J. Hashimoto. The Hakka Dialect; A Linguistic Study of its Phonology, Syntax and Lexicon. Cambridge University Press. (Princeton University) 1973 hbk. ISBN 0-521-20037-7
    • Hashimoto has surveyed the materials written on Hakka upto 1973. There is a introduction concerning the origins of Hakka people, with two maps. Also there is a lot of linguistic information on tones and sounds, and discusses the development of today's dialects from a proto-Hakka dialect.
  • Si Jiao Hao Ma Xin Ci Dian - Four Corner Index New Dictionary. Shang W Yin Shu Guan - Commercial Press Ltd, H.K. 1995 April 38th Edition. ISBN 962-07-0024-4 / ISBN 962-07-0025-2
    • Once learnt, the Four Corner method is easy to use, and fast. No need to look up radicals, number of strokes, sound etc. Simplified Characters.
  • Shang Wu Yin Xin Zi Dian - Commercial Press New Character Dictionary. Shang Wu Yin Shu Guan, HK. 1992 December (4th printing). ISBN 962-07-0140-2
    • This is a based on the 214 Kangxi radicals. There are about 10,000 characters and their simplified doubles are also give.
  • Roy T. Cowles. A Pocket Dictionary of Cantonese. Hong Kong University Press 1992 (1914,1949,1986). ISBN 962-209-122-9
    • Cantonese using the Wemper-Meyer romanisation. Has a radicals index, plus an English-Cantonese wordlist. Characters are referenced by numbers.
  • Henry Henne. Sathewkok Hakka Phonology. Reprinted from Norsk Tidskrift For Sprogvidenskap. 1964 BIND XX. ISBN ?
    • Henne lived on the outskirts of Shataukok between 1949-1950. Gives a linguistic account of the dialect.
  • Lau Chun Fat. Hag5 Ngi1 Pin1 Yim1 Si4 Fui4. (Hakka Pinyin Dictionary - in Chinese) The Chinese University Press, Hong Kong. 1997 ISBN 962-201-750-9
    • Lau has surveyed the sounds of Hakka and has adapted the use of Pinyim to cover most of the subdialects. He gives information on specific Hakka words, 2 maps, a character listing (approx. 6000) by number of strokes, by Hakka Pinyim sounds, and a diskette with his input.
  • KeJiaHua CiDian. (Hakka Dialect Dictionary) Zhang1 Wei2geng3. Guangdong Renmin Chubansha. 1993 ISBN 7-218-01504-2
    • N.B. This is a precursor to Pinfa, as it only incorporates the Meixian Dialect, whilst Lau's Pinfa system embraces multi-sub-dialectal differences.

  List of Contents : To Appendix Index : Top

  ---

• Changes made since last posting

• Revised link to reflect the website change. 5th.08.1999
• Revised Link in section 3.1. 3rd.10.98
• New sections 1.9 , 1.A and 1.B added, with new gif representation of chinese characters. Added more link flexibility at the top of the page. 16th.07.1998
• Removed old message. 28th.06.1998
• Addition of information in section C.1, with information about two dictionaries on Hakka, and Dr. Lau's site in section 2.1. 14th.05.1998
• Slight errors corrected, addition of section A.6, and addition of some links. 1st.04.1998
  
• BIG5 Coding removed and the use of a GIF file substituted in section 1.8 23rd.03.1998

  ---

  That's all there is right now, updates will appear when I get more information.

  ---