Cambodian

Plain vowels

Complex vowels

Phonetic syllables

Many native Cambodian words are monosyllabic. `Khmer also has many bisyllabic words, in which the first syllable is typically unstressed, and the vowel is rendered in colloquial speech as a schwa. Some bisyllabic words are compounds, however, and this may not apply.

Syllables start with one or more consonants or an independent vowel, however the latter represent a vowel sound after a glottal stop, so the syllable onset is always C(C)(C)V. The rhyme is composed of a short vowel in stressed syllables that is always followed by a consonant (VC), or a long vowel, that may not be.

Many monosyllabic words begin with consonant clusters, and some monosyllabic words end with clusters, although only one consonant is pronounced in syllable final position.

Word-internally, a syllable with a final consonant followed by another syllable creates a stack of consonants. The top item in the stack is the syllable final consonant, and the initial consonant of the next syllable is rendered in subjoined form.

Polysyllabic words are usually of Sanskrit, Pali or French origin. These words tend to alternate stress across their syllables, but may not.

Orthographic syllables

An orthographic syllable differs from a phonetic syllable in that it always begins with a consonant, or cluster of consonants. Where word-medial stacks occur, the orthographic syllable may begin with the final consonant of the previous phonetic syllable and continue through the onset(s) of the following syllable. Alternatively, an orthographic syllable may be just a final consonant (or consonant cluster) in a phonetic syllable.

    

An orthographic syllable includes all the combining characters associated with the consonants it contains.

Silent vowels & consonants

Inherent vowels are not pronounced after syllable final consonants.

Vowels are not pronounced between stacked consonants. The first character in the list below, 17D2, is used to create those consonant stacks. It is never visible when part of a cluster. (When rendered alone or without a following consonant, fonts do however apply a visible form.) See clusters.

Khmer also has other diacritics that silence vowels or parts of the text.

The following two diacritics are used to silence written characters.

17CD is used over a consonant, particularly in loan words, to silence it and any attached vowels or subscripts.

សប្ដាហ៍

រេហ៍ពល

កេរ្តិ៍

17CC is not very common and word-medially may produce rə before the consonant it sits on. When it appears word-finally, it may silence the final consonant (and may also affect vowel sounds).

បរិបូណ៌

17D1, the sanskrit virama, is sometimes used in Sanskrit words to indicate that a final consonant has no vowel sound.

អាត្មន៑

17DD, on the other hand, is a rarely used sign that indicates that a final consonant retains its inherent vowel sound.

(Check the detailed information for each of the above marks, since this section only looks at the use for silencing sounds.)

Plain vowels

The pronunciation of vowel signs depends on whether they follow an ɑ-class consonant or an ɔ-class consonant. In some cases the alternative may be a diphthong, so compare this also with the items in diphthongsz. Additional variations can arise from context or whether a syllable is open or closed. For more detailed information, click on the characters in the panel, or see vowel_mappings.

As an example of how the same vowel sign is pronounced differently after an ɑ-class and an ɔ-class consonant, check out the pronunciations for 17B7 in the following word.

កិរិយា

The following panel lists basic monophthongs following ɑ-class consonants.

The next panel shows vowels following ɔ-class consonants. In several cases the same vowel sign is used, but with a different pronunciation.

17C8 is a 20th century addition to the Khmer repertoire. It is used as a vowel after consonants that are pronounced as stressed syllables at the end of a word, or preceding an internal juncture in compounds.

17CF is used over just two words: ក៏ ដ៏

For 17D0 see vowel_mod.

Diphthongs

The following panel lists basic diphthongs following ɑ-class consonants. In some cases the alternative may be a diphthong, so compare this also with the items in monophthongs.

The next panel shows vowels following ɔ-class consonants. In several cases the same vowel sign is used, but with a different pronunciation.

Khmer diphthongs generally end with -j, -w, or ə, but there are a few others that don't. It is noteworthy, also, that not all diphthongs ending with -j or -w use យ or វ.

179A is usually silent in syllable-final position, but the combination 17D0 179A produces the sound ɔə.

ជ័រ

Nikahit and reahmuk

The descendants of the anusvara and the visarga, 17C6 called និគ្គហិត niʔkəhət, and 17C7, called រះមុខ reə̆hmuk, are regarded as vowels in Khmer, even though they represent the sounds -m and -h, respectively.

Used on their own, they can change the inherent vowel, affecting both the pronunciation and the meaning, eg. compare ដម ដំ

Nikahit can also follow vowel signs that would normally produce aː, and u/uː, and change their sound value. The following example shows its effect on the vowel sign 17B6u,643: ពាម ពាំ

The following panel lists rhymes following ɑ-class consonants.

The next panel shows rhymes following ɔ-class consonants. In several cases the same vowel sign is used, but with a different pronunciation.

The 2 following sequences are regarded as letters in the Khmer alphabet. They are not encoded separately in Unicode (but they are named sequences).

Vowel modifier marks

17D0 is used in some Pali and Sanskrit loan words (although alternative spellings exist) and indicates that the syllable has a particular vowel (click on the character in the box for more details). 

17CB is always placed above the final consonant, and basically shortens the preceding vowel.

ខាត់

Pre-base vowel signs

Three combining marks are displayed to the left of the onset consonant(s).

Relative to the consonants, these combining marks are always typed and stored in the order in which they are pronounced, not the order in which they are displayed. The positioning of the glyph when displayed is managed by the font and rendering software.

In fact, these vowel signs appear before the start of the orthographic syllable. When a syllable onset is a consonant cluster, the vowel sign appears to the left of the initial character in the consonant stack.

Circumgraphs

Five vowels are produced by a single combining character with visually separate parts, that appear on different (mostly opposite) sides of the consonant onset.

These are also combining marks that are always stored after the base consonant. The font places the glyphs in the right place relative to the base consonant.

Like the pre-base vowels, the components of a circumgraph vowel are arranged relative to a stack.

None of these circumgraphs decompose during normalisation. See also encoding.

Composite vowel signs

Vowels represented by combinations of the above characters (not including decomposed versions of the 2 circumgraphs described below):

Characters that don't appear in the combinations:

Vowel length

Vowel length is distinctive in Khmer. Short and long vowels are written using different vowel characters. For example, compare the long and short i vowel signs in the following word.

ហ្វ៊ីស៉ិក

Independent vowels

Khmer also has independent vowel letters that can be used to represent some of these sounds, but unlike most South Asian scripts, there are fewer independent vowels than vowel sounds, and some do not have direct correspondences with a simple vowel sign. For example, 17AA corresponds phonetically to the multipart vowel 17A2 17BC 179C. (See also vocalics.)

Whether a vowel sound is represented using an independent vowel letter or the glottal consonant plus vowel sign varies from word to word. In Cambodian orthography the two are not interchangeable. The independent vowel letters appear in relatively few words, but some of those words are quite common, eg. ឪពុក ឲ្យ ឮ

Other characters

Plain vowels

Diphthongs and rhymes

Stacking

Subscript consonant forms are called ជើងអក្សរ (or 'coeng', pronounced cəːŋ).

Cambodians see these subscripts as distinct letter forms, but, unlike Tibetan, they are produced in Unicode by inserting 17D2 before the consonant that will become a subscript. This character, which has no visual form in Cambodian, is called the coeng in Unicode, although it should rightly be called the coeng generator.

All the shapes are simplified and reduced in size compared to the non-subscript form. Many have significantly different shapes.

This list shows consonants in their normal and subjoined forms

179A produces a subjoined form that wraps to the left and under the preceding consonant. Several others wrap below and to the right of the consonant. 17A1 doesn't normally appear in subscript form.

Some subscripts change the sound of the preceding consonant.

Stacks commonly occur in multisyllabic words. A medial cluster will typically contain a final consonant for the first syllable and the initial consonant of the next syllable.

កម្មករ

Some clusters involve two subscripts. These are, with 3 exceptions, composed of a final nasal, followed by a stop and r.

កន្ត្រៃ

កញ្ជ្រេង

The 3 exceptions are loan words.

អង្គ្លេស

សងស្ក្រិត

សាស្ត្រាចារ្យ

Subscript consonants that appear at the end of a word, are silent.

ពេទ្យ

រដ្ឋ

Stacks and registers

Where the two consonants involved in the cluster are in different classes or registers, the pronunciation of any following vowel is normally determined by the register of the subscript consonant. For the following exceptions, however, the vowel pronunciation is determined by the register of the first consonant (all ɔ-class):

Combining characters & stacks

Most vowels that are pronounced after a consonant stack are displayed as if they were attached to the consonant at the top of the stack (including pre-base vowel signs and circumgraphs).

Click on the examples to see the underlying sequence more clearly.

ក្លិន

ឆ្វេង

ភ្លៀង

This can also lead to ligation between the top consonant in the stack and a vowel sign, unless the subjoined consonant is right-spacing.

ក្ដារ

ក្រាក់

ក្បាល

The exceptions are the vowel signs that appear below the consonant; these appear below the last subjoined form, unless the subjoined form is left-spacing.

ខ្ពុរ

ត្រុំ

Other diacritics are treated in a similar fashion. This can lead to ambiguity if a diacritic could modify either consonant in a stack.

Subscript consonants after vowels

It is rare but possible to find subscripts used after independent vowels. One common word spelled this way is ឲ្យ

It is also possible to find subscript forms of independent vowels. Four of these are named sequences in Unicode.

Glyph order

Although all combining characters follow the base in memory, the visual order of syllable components may not follow a linear progression from left to right. In the following example the order in which the glyphs are pronounced is far left, far right, down, left, left: កន្ត្រៃ In the word ច្រៀង the spoken order of the separate visible parts, numbered left to right, is 3,2, 1+4, 5, Some vowel signs span two or three sides of the base consonant or cluster.

Joining forms for AA

Some small joining features occur in relation to 17B6 and similarly shaped vowels. Unicode provides the following list of common forms:

1. ក + ា = កា
2. ប +  ា = បា (avoids confusion with ហ )
3. ប +  ៅ = បៅ
4.  ្ស +  ា = ្សា

Register-shifter position

The function of the register-shifter marks is described in register_change.

When 17C9 museʔkətoə̯n or 17CA trəisaɓ appears with a vowel sign above the consonant, the ក្បៀសក្រោម kɓiəhkraom form may be used. This looks exactly like 17BB. See fig_posn_muse.

fig_posn_triisap provides some examples of TRIISAP.

The Unicode Standardu,647 gives the impression that both of these diacritics are moved below the consonant any time a vowel appears over that consonant. However, in reality only certain consonants cause this behaviour. The behaviour varies a little by font, but in general the MUUSIKATOAN diacritic is lowered for the following letters:

TRIISAP is lowered for these characters.

For more details, see this SIL document.sil

If needed, this behaviour can be modified using 200C, if the font recognises it, to prevent the low form appearing. The ZWNJ must be typed and stored between the base consonant and the register-shifter. See fig_zwnj_triisap for an example using ប្រតឺងអ‌៊ឹះ.

 

Observation: Textarea entry fields on the Mac don't seem to recognise the ZWNJ character (mentioned below) and incorrectly display the combining characters that follow it. The text looks ok if the browser displays the sequence as part of the HTML.

NYO with subscripts

Another common feature is that ញ drops the swash below the baseline when followed by a subscript consonant.

Also, when it appears as a subscript under itself it uses a special full form subscript (see fig_nyo_subj).

Vowel ligatures

In the âksâr mul style, some vowel signs ligate with the consonant characters to which they are applied. See the word វិទូ in fig_vowel_ligatures.

Observation: This behaviour is font dependent. For example, the Khmer Mool font (used for the figure), and Khmer OS Muol produce the ligation, but the Khmer Ratanakiri font does not.

To prevent a ligature forming, use 200C (ZWNJ) between the consonant and the vowel. To cause a ligature to form when there isn't one, if the font has the appropriate rules, use 200D (ZWJ) instead.u,647

Managing glyph shaping

Content authors can use 200C and 200D to change the position of register shifters (see consonant_shift_posn), and to manage ligatures in mul fonts (see vowel_ligatures).

Names

- (called សហសញ្ញា sɑːhɑː sɑːɲɲiə) is used between the parts of a person's name. Typically the family name (written first) and following names, but often all names for Chinese Cambodians, eg.

ញ៉ុក-ថែម ɲok tʰaem

លី-ធាម-តេង liː tʰiəm teiŋ

Grapheme clusters

Base ZW(N)J? Combining_mark*

This fixed ordering makes it easier to search for and collate text. For more details, see this SIL document.sil

Combining marks may include zero or more of the following types of character.

1. Consonant shifters [2] (see register_change) Only one per grapheme cluster, typed and stored immediately after the base consonant.
2. Dependent vowels [19] (see combiningV) Always a single code point, even for circumgraphs (see circumgraphs).
3. Final consonant marks [2] (see nikahit_reahmuk) One of 2 possible combining marks, at the end of a grapheme cluster sequence. (Cambodians consider these to be vowels.)
4. Coeng generator [1] (see subscript_consonants) Normally occurs immediately after a consonant at the beginning of a cluster. It sometimes also appears after an independent vowel.
5. Other diacritics [6] (see vowel_mod, novowel, and emphasis) Normally occurs at the end of a syllable.

In some cases, a ZWNJ is inserted between a consonant and a consonant shifter in order to specify special shaping rules for the shifter (see consonant_shift_posn).

To prevent or encourage the formation of ligatures (especially for mul fonts) a ZWNJ or ZWJ can be inserted between a consonant and following dependent vowel (see vowel_ligatures).

The following examples show a variety of grapheme clusters:

Click on the text version of these words to see more detail about the composition.

	បរិបូណ៌
	ប៉ាន
	នុ៎ះន៎
	ចង្អូរ
	កញ្ជ្រេង
	ប្រតឺងអ‌៊ឹះ

The last syllable of the last item in the list contains a ZWNJ (to keep the register shifter above the base).

Note how grapheme clusters segment the parts of a conjunct after the virama. This is not always desirable (see orthographicS just below).

Larger typographic units

(Consonant Coeng)* Grapheme_cluster

Khmer commonly stacks glyphs, to form conjuncts (see clusters). These conjuncts represent consonant clusters or gemination. The stacking does not occur across word boundaries; only within words.

Grapheme clusters terminate after a sequence of marks that includes a coeng indicator code point, but editorial operations that change the visual appearance of the text, such as letter-spacing, first-letter highlighting, line-breaking, and justification, should never split stacked forms apart. For this reason, an alternative way of segmenting graphemes is needed. This may not apply, however, for some other operations such as cursor movement or backwards delete.

Where stacks appear, a typographic unit contains multiple grapheme clusters. The non-final grapheme clusters all end with 17D2, which is never visible, and the final grapheme cluster begins with a consonant.

The following are examples. Some of these examples were shown in the previous section: here the stack is treated as a single typographic unit.

Click on the text version of these words to see more detail about the composition.

	ចង្អូរ
	កញ្ជ្រេង
	អភិវឌ្ឍ
	ភ្លៀង

Browser behaviour

Cursor movement. Move the cursor through the text.
Gecko and Blink step through the text using grapheme clusters. It takes 2 or more steps (depending on the number of GCs) to get through the stacks, one grapheme cluster at a time. In some cases the cursor may look as if it is not moving. WebKit steps through all words using the orthographic syllables as described here (ie. they step over a stack and all associated combining characters in one jump), except that WebKit treats the ZWNJ as the start of a new grapheme cluster, requiring an extra jump to get past it.

Selection. Place the cursor next to a character and hold down shift while pressing an arrow key.
The behaviour is the same as for cursor movement.

Deletion. Forward deletion works in the same way as cursor movement. The backspace key deletes code point by code point, for all browsers.

Line-break. See this test. The CSS sets the value of the line-break property to anywhere. Change the size of the box to slowly move the line break point.
Gecko and Blink set break opportunities based on orthographic syllables. WebKit wraps on grapheme cluster boundaries, but gets rather confused by the stacks, typically wrapping half the stack plus one adjoining (non stack) character together.

Bracketed text

Khmer commonly uses ASCII parentheses to insert parenthetical information into text.

Ellipsis

The following word (pronounced laʔ) is used as the equivalent of 'etc.' ។ល។

A character exists that represents that sequence, ៘, but the Unicode Standard deprecates it, and recommends the use of the three separate characters instead.u

Other spellings for et cetera also exist. These include:

• ។បេ។
• –បេ–
• –ល–

Repetition

It is common to repeat words or sometimes phrases in Khmer, particularly to provide emphasis. ៗ (called លេខទោ leːktoː) can be used for this, eg. ខ្លាំង ៗ klaŋklaŋ very strong គាត់មានផ្ទះថ្មី ៗ kaːtʰ miən pʰteə̯h tʰməitʰməi he has a brand new house

Sometimes this sign repeats a phrase rather than a word, eg. បន្តិចម្ដង ៗ ɓɑntecmɗɑːŋ ɓɑntecmɗɑːŋ little by little

It is also occasionally used to repeat the word at the end of a sentence for the beginning of a new sentenceh, eg. ខ្ញុំទៅផ្ទះខ្ញុំ ៗ នៅជិតផ្សារ kɲomtɨwpteə̆hkɲom pteə̆hkɲom nɨwcɨtpsaː

The sign is usually separated from the text by a space.

Zero-width space & word-joiner

Because Khmer doesn't separate words, applications typically look up word boundaries in a dictionary, however, such lookup doesn't always produce the needed result, especially when dealing with compound words and proper names. To counteract these deficiencies, authors may use one of the following invisible characters.

In order to manually fine-tune word-boundary detection, the invisible character 200B (ZWSP) can be used to create breaks.u,625

To prevent a break between syllables, 2060 (WJ) can be used.

It is also important to bear in mind that Khmer may be used to write various languages, in particular minority languages for which different dictionaries are needed. Since such dictionaries may not available in a given browser or other application, there is a tendency to use ZWSP in order to compensate.

Large-scale manual entry of ZWSP and WJ has potential downsides because the user cannot see them; this leads to problems with ZWSP being inserted in the wrong position, or multiple times. However, these don't set a state, so it doesn't create major issues. It would be useful, however, if an editor showed the location of these characters.

Care should also be taken when trying to match text, eg. for searching in a page. WJ should be ignored. ZWSP may or may not be ignored, depending on whether word boundaries are significant for the search.

Line-edge rules

As in almost all writing systems, certain punctuation characters should not appear at the end or the start of a line. The Unicode line-break properties help applications decide whether a character should appear at the start or end of a line.

Show (default) line-breaking properties for characters in the modern Khmer orthography.

The following list gives examples of typical behaviours for characters used in modern Cambodian. Context may affect the behaviour of some of these and other characters.

Click on the Khmer characters to show what they are.

• “ ‘ (   should not be the last character on a line
• ” ’ ) ? ! ។ ៕ ៚ %   should not begin a new line
• ៛   should be kept with any number, even if separated by a space or parenthesis.

The following character should not produce a line-break when they appear inside or alongside a word: ៙.

Numeric

The khmer numeric style is decimal-based and uses these digits.rmcs

Examples:

Alphabetic

The cambodian-consonant alphabetic style for the Khmer language uses these letters.

Examples:

Prefixes and suffixes

List counters typically use a full stop+space as a suffix.