Myanmar character notes [Draft]

These notes are still in development. These notes are the result of my explorations into how the Myanmar script is used for the Burmese language in the context of the Unicode Myanmar block.

Each example has a superscript that links to a deconstruction of the characters involved in the lower part of the document. In major browsers, you should see the same information by mouseing over the example in the HTML version.

Script introduction

Myanmar is a tonal language and is syllable-based. The script is an abugida.

It runs from left to right.

There are a set of Myanmar numerals, which are used just like Latin digits. From 0 to 9 they are ၀၁၂၃၄၅၆၇၉.

Consonants

Unicode 5.1 lists 38 consonants (including 5 special forms). Native Myanmar words use a subset of the consonants that make up the traditional articulatory arrangement of indic scripts, however additional symbols are available for use in loan words, especially Indian loan words. These include the retroflex and voiced aspirated consonants. Yet more characters in the Myanmar Unicode block are used for variations for minority scripts based on myanmar.

Syllable-final consonants and asat. Most native words are monosyllabic. When there is a word final consonant at the end of a syllable, it carries a visible mark called 'asat' to indicate that the inherent vowel is killed, eg. see the small 'c' like mark over the last character in အင် øṅ̸ ʔɪ̃. The U+103A MYANMAR SIGN ASAT is a new character proposed for Unicode version 5.1 for this purpose. It is effectively a visible virama.

In native Myanmar, 9 characters (5 nasals င ဉ ည န မ ṅ ñ ñ- n m and 4 stops က စ တ ပ k c t p) appear in syllable final position. In final position nasals are pronounced as a nasalization of the previous vowel, eg. အန် øn̸ ʔã, and all stops are pronounced ʔ, eg. အတ် øt̸ ʔaʔ.

Multi-syllabic words. In multi-syllabic words, consonants that have no intervening inherent vowel are typically arranged such that the consonant cluster is stacked. The second consonant appears below the first, eg. မန္တလေး mn̸tle:, and ဗုဒ္ဓ bud̸dʰ. In some cases the lower character is abbreviated or reoriented. In the case of a doubled သ s the result is a special combination, သ္သ.

This effect is achieved in Unicode by using the character U+1039 MYANMAR SIGN VIRAMA between the consonants forming the cluster. Note that the virama is not visible.

Medial variants. ယ y, ဝ w, ရ r, and ဟ h have special medial variant forms that act like combining characters, eg. ချက် kʰy̱k͑ cʰɛʔ, မြန် mṟn͑ mjã, မွေး mw̱e: mwè, and မှ mẖ hmá, and ရှ rẖ ʃá. It is also possible to find two medials associated with a consonant, eg. လှျ lẖy̱ hljá or ʃá.

Dedicated medial signs exist in Unicode 5.1 for each of these uses.

Kinzi. When the first consonant in a consonant cluster is a non-word-final င ṅ it rises over the following letter and keeps its virama, rather than pushing the following consonant below it, eg. အင်္ဂာနေ့ øṅ̸͑gāne·. This is called 'kinzi'. To achieve this in Unicode 5.1, follow the ṅ with both U+103A MYANMAR SIGN ASAT and U+1039 MYANMAR SIGN VIRAMA, then continue with the next letter.

Consonant clusters and vowel signs. A consonant cluster is treated as a unit when it comes to vowel-signs, for example အငွေ øṅw̱e, where the left-combining e is displayed to the left of the ṅ although the character appears after the w̱in memory. On the other hand, vowel signs that would normally appear below a consonant may be displayed to the right, eg. in the second syllable of ပုဂ္ဂိုလ်များ pug̸giul͑my̱ā:, the i appears above the cluster, but the u has now to the right to make room for the subjoined g.

Changes in Unicode 5.1. In Unicode 5.0, U+103A MYANMAR SIGN ASAT did not exist, and U+1039 MYANMAR SIGN VIRAMA had to be used for both visible and non-visible viramas. This approach was problematic in that, since there are no spaces between words, it is not easy to automatically ascertain whether a virama should appear above a consonant or cause the stacking effect. For example, should my sequence of characters appear like this, အမ်မီတာ , or like this အမ္မီတာ? To get around this in Unicode 5.0 you needed to use a zero-width non-joiner (ZWNJ) after the virama if you want it to remain visible (ie. the first example above would have been transcribed as øm̸ˣʲmītā and the second as øm̸mītā). The non-joiner prevents stacking. In practice, this meant that there were very many ZWNJ characters in Burmese text, since there are many syllable-final consonants, and typing Myanmar was therefore much more time-consuming than it needed to be.

Unicode 5.1 also introduces dedicated medial consonants. This makes it easier to type myanmar text, but also allows for easy distinction of subjoined variants of these consonants rather than the usual medial forms.

Other phonetic information. Note that the combination of velar stop and r or y are pronounced as c, eg. ကြက် kṟk͑ cɛʔ, ကျပ် ky̱p͑ caʔ.

Sandhi affects the pronunciation of sounds at syllable boundaries in certain grammatical contexts, eg. လူထု lūtʰu is pronounced ludú, rather than lutʰú.

Some conventions exist for representing foreign sounds. f is ဖ pʰ, v is ဗ b or ဗွ bw̱, eg. တီဗွီ tībw̱ī tivi. A foriegn syllable final sound can be rendered by placing a second killed consonant after the syllable, sometimes in parentheses, eg. ဘတ်(စ်) bʰt͑(c͑) bas

Vowels

The Unicode 5.1 Myanmar block groups vowels into 8 independent vowels and 7 dependent vowels.

Inherent vowel. The inherent vowel is a. The independent form, used for syllable initial position, is အ, although this is classed as a consonant rather than a vowel by the Burmese.

Independent/initial vowels. The symbol အ is used as a support for vowel signs, and the combination of that and the vowel sign is the normal native way of showing independent/initial vowels.

Some independent/initial vowels have an alternative form that is used in some words only - typically Indian loan words, eg. ဧရာဝတီ Erāwtī 'Irawaddy river', ဩဂုတ် Ogut͑ 'August', and ဤ Ī 'this'. There are normally different forms for different tones, and normally only one or two vowel+tone combinations have these forms.

Long vs. short vowels. The 'primary' vowels have 'short' and 'long' written forms that hark back to the earlier Indic script origins, but the distinction is used nowadays for indicating different tones.

Tall AA. In order to avoid visual confusion, there are two forms of the long -aa vowel sign in Burmese. ဝာ wā looks very much like တ t, so a taller form of -aa is used, ie. ဝါ wā. This form is also used after the following: ခဂငဒပ. Whereas in Unicode 5.0 the choice of appropriate form was left to the font or implementation during rendering, such contextual decisions are not appropriate for Sgaw Karen and other minority scripts, which only use the tall form, so U+102B MYANMAR VOWEL SIGN TALL AA was added to Unicode 5.1 as a separate character. It's not clear, but I'm assuming that for Burmese, you don't use that character, but continue to rely on the font or other rendering algorithms to choose the appropriate glyph - however, the Myanmar2 font doesn't seem to automatically do the right thing, so that may be incorrect.

Tones. There are three tones in Burmese, creaky, low and high. A vowel plus tone combination is called a rhyme. The tone of a syllable is indicated either by the vowel used, or by a combination of vowel and diacritic.

There are 7 main vowel sounds in open syllables. The following lists those sounds and their different representations for the three tones in Burmese, creaky, low and high. (Combining symbols are shown with အ.)

a Primary central vowel. As the inherent vowel, this has the creaky tone. With low tone: အာ. With high tone: အား.

i Primary front vowel. With creaky tone, 'short' symbol အိ (alternative initial is ဣ). With low tone, 'long' symbol အီ (alternative initial is ဤ). With high tone အီး.

u Primary back vowel. With creaky tone, 'short' symbol အု (alternative initial is ဥ). With low tone, 'long' symbol အူ (alternative initial is ဦ). With high tone အူး.

e High front mid vowel. With creaky tone အေ့ (alternative initial is ဧ). With low tone အေ. With high tone အေး.

o High back mid vowel. With creaky tone အို့ုိ. With low tone အုိ. With high tone အိုး.

ɛ Low front mid vowel. With creaky tone အဲ့. With low tone, killed y အယ်. With high tone အဲ.

ɔ Low back mid vowel. With creaky tone အော့. With low tone, adds asat အော် (alternative initial is ဪ). With high tone အော (alternative initial is ဩ).

The following table summarises the above in a way that allows you to see how the various ways tones are applied to open syllables using the native Myanmar characters.

	creaky	low	high
a	inherent vowel	long form	visarga
i	short form	long form	visarga
u	short form	long form	visarga
e	dot below	inherent	visarga
o	dot below	inherent	visarga
ɛ	dot below	killed-y form	inherent
ɔ	dot below	asat	inherent

Vowels in closed syllables. Vowels in 'closed' syllables end in a glottal stop or nasalisation. Historically, however, they ended in one of four nasals or four stops, and this is still reflected in the orthography. The vowel quality has also evolved in these syllables, typically producing diphthongs.

The following list shows all the main combinations of vowels and finals, and the phonemic changes involved.

i becomes eɪ: အိန် øin͑ ʔeɪ̃; အိမ် øim͑ ʔeɪ̃; အိတ် it͑ ʔeɪʔ; အိပ် ip͑ ʔeɪʔ.

u becomes oʊ: အုန် un͑ ʔoʊ̃; အုမ် um͑ ʔoʊ̃; အုတ် ut͑ ʔoʊʔ; အုပ် up͑ ʔoʊʔ.

ɔ becomes aʊ: အောင် eāṅ͑ ʔaʊ̃; အောက် eāk͑ ʔaʊʔ.

o becomes aɪ: အိုင် iuṅ͑ ʔaɪ̃; အိုက် iuk͑ ʔaɪʔ.

a is a lot more complicated.

-ṅ	-ñ	-ñ	-n	-m
အင် øṅ͑ ʔɪ̃	အည် øñ-͑ ʔɪ̃	အဉ် øñ͑ ʔi	အန် øn͑ ʔã	အမ် øm͑ ʔã

-k	-c	-t	-p
အက် øk͑ ʔɛʔ	အစ် øc͑ ʔiʔ	အတ် øt͑ ʔaʔ	အပ် øp͑ ʔaʔ

There are other combinations of vowel and final consonant found in Burmese words of Indian origin, which often stick to the original Indian spelling, however, they tend to follow Myanmar pronunciation, eg. ဓာတ် dʰāt͑ daʔ, ဗိုလ် biul͑ bo, ဥယ္ယာဉ် uy̸yāñ͑ úyĩ.

Rhymes in closed syllables ending in nasalization are inherently low toned. Do they still need any special marks? What about closed syllables ending in plosives?

Vocalic weakening. A process called vocalic weakening affects the first syllables of certain words (mostly nouns and adverbs), eg. ထမင် tʰmṅ͑ is pronounced tʰəmĩ́, not tʰámí̃; ဘုရား bʰurā: is pronounced pʰəyà, not pʰúyà.

Word boundaries, punctuation, etc.

Spaces are used to separate phrases, rather than words.

The Unicode 4.1 Myanmar block lists 2 punctuation characters. Punctuation is commonly limited to ၊ and ။, with significance close to comma and full stop, respectively.