-1

Hangul has a number of doubled letters (e.g. ㄲ or ᆖ), digraphs (e.g. ㅔ or ㄳ), and trigraphs (e.g. ㅞ or ㅴ).

Is there a distinction between these multigraphs and a sequence of the individual letters that make them up?

This question is rather hypothetical since it appears that no current Korean font will display a sequence of the individual letters in a single syllable block (cf. Is there a Korean font that can display arbitrary jamo combinations?). So the question really is: Suppose there were a font that could display a sequence like ᆨᆺ (U+11A8, U+11BA) in a single syllable block, should the result be identical to ᆪ (U+11AA) or should it look different? Or, giving an entire syllable block example, should 목ᆺ (U+BAA9, U+11BA) look like 몫 (U+BAAB) if there were such a font? And if they should look different, then how should they look different and why?

Regarding the possibility of sequences such as ᆨᆺ (U+11A8, U+11BA) or 목ᆺ (U+BAA9, U+11BA), note that Unicode explicitly defines there should be no syllable break between them, see Unicode Text Segmentation § Hangul Syllable Boundary Determination. This is why computers treat these sequences as if they were a single syllable block when selecting text or moving the cursor (unlike e.g. the sequence 목ㅅ with standalone ㅅ U+3145, which is treated like two separate characters with regards to text selection or cursor movement – just try it). In other words, Unicode allows single-syllable sequences like 목ᆺ, and computer character processing takes it into account. There are just no fonts yet that display them as single syllables (as far as I know). So my question is: when such a font is written, how should it look?

Other information:

0

3 Answers 3

2

The digraphs and trigraphs and sequence of single characters are all bogus things. They can be written and even created on a computer but they are undefined as far as the language is concerned.

The five doubled consonants (ㄲㄸㅆㅉ) and eleven composite consonants (ㄳㄵㄶㄺㄻㄼㄽㄾㄿㅀㅄ) are treated as unbreakable single consonants just like any other. Same applies to vowels like ㅔ, ㅞ, etc. They might have been different hundreds of years ago, but they are now single unchangeable units (a sort of historical artifacts), and any other permutations (like two ㅊs or ㄱㅂ) are undefined and illegal. That is why Unicode doesn't support them.

It is true that you can commonly see people writing ㅋㅋㅋ or ㅇㅈ and things like that on the internet, but none of them are legitimate usage as defined by the official Korean language. Think of them as the equivalent of those weird interjections (e.g. "#*^!!!!!") you see in English cartoons.

Is there then any legitimate use of unit consonants and vowels? Yes, we need them to describe the language, for example in sentences like ㅅ은 이렇게 발음한다 (You pronounce ㅅ like this), ㄱ은 한글의 첫 번쨰 자음이다 (ㄱ is the first consonant of 한글), etc. But such usages preclude sequencing them. If you exclude frivolous plays, there really is no use for such sequences at all.

So, going back to your question, ㄱㅆ is not only undefined but there is no reason to ever write it, as it makes no sense and serves no useful purpose in modern Korean other than as a play or abusing the language.

As for not being able to selecting only the 목 part of 목ㅅ, I think it is just an accidental anomaly. I cannot imagine any reason why they shouldn't allow it because 목 is the legitimate part of the broken piece. If I try the same at a different site or any programs on my computer, I can select it with no trouble, which makes me think it's a StackExchange problem.

[EDIT]

Regarding 목ㅅ, I checked further and found this. Although the composite consonants are single units as defined by the language, unicode defines them as digraphs. This is so as to provide an easy way of entering ㄱ and ㅅ to create ㄳ, for example. So the interface sees two consonants and decides: if it's a two-consonant sequence matching one of the eleven composite ones I mentioned, it turns it into that composite (ㄳ). Otherwise, it will make it a sequence of two regular consonants (e.g. ㄱㅂ). So they must have one code point for normal ㄱ and another for the ㄱ in ㄳ. We can even see the difference in their size (which I had overlooked). In any case, the small ㄱ used for ㄳ exists purely for the computer interface.

Using the keys on the (computer) keyboard, you cannot create 목ᆺ. You have to enter the four Unicode values using the special escape sequence (like CTRL-SHFT-U on my linux machine). And as you said, 목ㅅ created using normal keys leave the parts selectable while the specially created 목ᆺ does not. And in my opinion, this is not a defect since the small ㅅ is nothing more than an artifact of computer software which doesn't have any use on its own.

In short, I don't think you can do anything useful with those special small consonants. If you need to display old consonants and vowels no longer used in modern Korean, you'll probably need specialized software that goes beyond what's available in unicode.

2
  • It is no anomaly, but a widely supported (I have not yet found any app that would not support it on Windows, Linux, or macOS) standard behaviour as defined in Unicode Text Segmentation § Hangul Syllable Boundary Determination. Of course, when you enter 목ㅅ (U+BAA9, U+3145), then there will be a syllable break. If you want to observe the behaviour supposed not to have a syllable break, you have to make sure you really have the characters 목ᆺ (U+BAA9, U+11BA). I wonder how it should really look, without syllable break.
    – mach
    Apr 12 at 5:39
  • You're right. I overlooked certain things in your original post. I updated my post.
    – Tony
    Apr 12 at 7:55
1

Update: after a long, long discussion, I think OP and I finally reached to a certain level of agreement at what this question is about. I'm updating my answer to the original question with some rephrasing of mine (in square brackets) based on the agreement, and cleanup of unnecessary discussive material.


Is there a distinction between these multigraphs and a sequence of the individual letters that make them up? [...] So the question really is: Suppose there were a font that could display a sequence like ᆨᆺ (U+11A8, U+11BA) in a single syllable block, should the result be identical to ᆪ (U+11AA) or should it look different [if the intention of sequencing (U+11A8, U+11BA) was to produce the composite jamo ᆪ ]? Or, giving an entire syllable block example, should 목ᆺ (U+BAA9, U+11BA) look like 몫 (U+BAAB) [if the intention of sequencing (U+BAA9 U+11BA) was to produce the syllaby 몫 and] if there were such a font?

When [] (or even [ ]) is used as an alternative sequence of code points for , I think they should be displayed identically to U+BAAB.

However, none of widely adopted USC/Hangul standards recognize (U+BAA9, U+11BA) or (U+1106, U+1169, U+11A8, U+A11BA) as an equivalence of U+BAAB (or (U+1106, U+1169, U+11AA).

  • KS X 1026-1 table 7 & 8 (EN translation) says there to be a syllable break between U+BAA9 and U+11BA. So (U+BAA9 U+11BA) is not allowed.
  • ISO 10646 says A complete syllable block is composed of a Choseong and a Jungseong, and optionally a Jongseong. and An incomplete syllable composed of a Jongseong alone shall be preceded by a CHOSEONG FILLER (115F) and JUNGSEONG FILLER (1160) So (U+BAA9, U+11BA) shall be normalized into two syllables (U+BAA9) . (U+115F U+1160 U+11BA).
  • Unicode recognizes (U+BAA9 U+11BA) as a legitimate sequence of code points to form a single syllable block accroding to UAX#29 (UAX#29#GB6, UAX#29#GB7, UAX#29#GB8). BUT,
    • According to Hangul syllable decomposition rule (Unicode 14.0 (latest) Chapter 3), canonical decomposition of U+BAAB is (U+1106 U+1169 U+11AA). Namely U+BAAB is not (U+BAA9 U+11BA).
    • According to Hangul syllable composition rules, which are only defined for <L, V>, <L, V, T>, and <LV, T> sequences, (U+BAA9/LVT U+11BA/T) is undefined, thus can't be further composed. Namely (U+BAA9 U+11BA) is not U+BAAB.
    • As U+BAAB != (U+BAA9 U+11BA) && (U+BAA9 U+11BA) != U+BAAB, we know that (U+BAA9 U+11BA) is a wrong way for representing U+BAAB.

Despite the standards, some people might still want to use (U+BAA9 U+11BA) or alikes to represent U+BAAB or alikes. I think that is reasonable thinking based on Hangul compositionality. And maybe less importantly not conforming to the international/national standards doesn't send someone to jail.

That being said, at the moment, as OP pointed out;

There are just no fonts yet that display them as single syllables

Concretely, the current situation is that

  1. Fonts render 목ᆺ into a horizontally segmentable single graphical unit.
  2. The resulting grapheme of (U+BAA9 U+11BA) behaves as a single syllable block when selecting text or moving the cursor (as OP found out).
  3. Fonts render 목ᆺ differently from 몫.
  4. Some fonts render U+11BA part shorter, place it in the bottom side, and make it clear that the part is from one of Unicode jongseong codepoints (but, for example, macOS default Hangul font does not do this).
  5. Fonts render [baa9 11ba] syllable quite differently - twice wider - than other regular syllables.

Based on the first two observations, it seems that fonts are "kind of" rendering (U+BAA9 U+11BA) as a single syllable grapheme, at least behaviorally. But from the last observation, we know that they do it in a very irregular (and somewhat weird and incomplete) way. Namely, render it double-syllable wide, hurting the cosmetic beauty and readability of Hangul 모아쓰기 writing style. (The third observation is maybe because those characters are canonically different characters accordingly to Unicode canonicalization rules? I can't be sure.)

It is obvious to me the current situation is not the best to neatly display all "possible" Hangul syllables. But considering real-world usage and usability of sequences like (U+BAA9 U+11BA), which are practically non-characters, I think I can live with the current situation as long as most fonts are capable of displaying all "in-use" Hangul syllables properly.

3
  • Comments are not for extended discussion; this conversation has been moved to chat.
    – user17915
    Apr 17 at 10:57
  • Sorry, I did not see that while you are now admitting that sequences like 목ᆺ are units, you are still denying that they should be displayed as hangul syllable blocks (which they should be according to UAX #29 § Hangul Syllable Boundary Determination). Regarding the withdrawal, cf. N3257 and subsequent Unicode versions, which contain only the new characters, but not the proposed ban on sequences like 목ᆺ (yes, I know it is mentioned, but nothing more).
    – mach
    Apr 17 at 14:53
  • The logic by which you arrive at the conclusion “that (U+BAA9 U+11BA) is a wrong way for representing U+BAAB” is fallacious. Sure, 목ᆺ is not a canonical or compatibility intermediate decomposition step or full decomposition of 몫, nor vice versa. But does this allow the conclusion that 목ᆺ must not be displayed like 몫? Not really.
    – mach
    Apr 18 at 19:11
0

Abbreviations

For simplicities sake, I will use the following abbreviations:

  • LV is any single-character Hangul syllable block composed of a leading consonant and a vowel, e.g. 모
  • LVT is any single-character Hangul syllable block composed of a leading consonant, a vowel, and a trailing consonant, e.g. 목.
  • L is a any single-character Hangul leading consonant, e.g. ᄆ
  • V is any single-character Hangul vowel, e.g. ᅩ
  • T is any single-character Hangul trailing consonant, e.g. ᆨ

Standards for encoding Hangul syllable blocks

There are two different standards for encoding Hangul syllable blocks:

  1. According to KS X 1026-1 (English translation: KS X 1026-1), a Hangul syllable block must be encoded as LV or LVT. If no Lv or Lvt is available, (that is, when using older letters), a complete single Hangul syllable block must be encoded by L V T? (a sequence of one L followed by one V followed zero or one T). So there are only three options: Lv, Lvt, L V T?
  2. According to UAX #29: Unicode Text Segmentation § Hangul Syllable Boundary Determination, a Hangul syllable block can contain sequences of L, V, or T, and there can be combinations of Lv or Lvt with either L, V or T. The possible ways to encode a single Hangul syllable block are thus the following: L+ V+ T* (a sequence of one or more L followed by one or more V followed by zero or more T), L* Lv V* T* (a sequence of zero or more L followed by one Lv followed by zero or more V followed by zero or more T), or L* Lvt T* (a sequence of zero or more L followed by one Lvt followed by zero or more T.

The Unicode standard has been widely adopted with regards to how applications identify what sequences behave like a single Korean syllable block, e.g. when selecting text or moving the cursor. KS X 1026-1, on the other hand, has apparently not even been adopted by Samsung Android, cf. Peter Constable’s second 2020-09-04 comment to “‘Ancient Hangul’ is an exaggeration”. With regards to fonts, there is probably no font yet that displays any sequence behond KS X 1026-1 as a single Korean syllable block. If such a font were written, the wide adoption of the Unicode standard would ensure that it behaves correctly in most applications.

Consonant or vowel sequences in the different standards

KS X 1026-1 does not allow sequences of consonants or vowels For instance, there can be no T T sequences like ᆨᆺ or Lvt T sequences like 목ᆺ.

Unicode allows sequences of consonants or vowels. T T sequences like ᆨᆺ or Lvt T sequences like 목ᆺ are perfectly acceptable and behave as if they formed single Hangul syllable blocks – there is just no font yet to display them as such.

Hangul multigraphs

The question is whether single Hangul characters such as ㅔ or ㄳ should be considered atomic letters or multigraphs composed of other Hangul characters, in this case, of ㅓ and ㅣ or of ㄱ and ㅅ respectively. There are a number of reasons to consider them composed multigraphs:

  • The original account of the Hangul, the Hunminjeongeum, describes them as compositions.
  • They look like compositions, with the individual elements being easily identifiable and sometimes not even touching.
  • Their names are composed of the names of the individual elements.
  • The individual elements often represent different phonemes, especially in the case of characters composed of two different consonant elements such as ㄳ.
  • KS X 1026-1 recognizes them as complex letters.
  • Common keyboard layouts allow typing the individual letters seperately.

On the other hand, those characters have no Unicode decomposition rules, canonical or compatibility, for being split into their individual elements. This cannot change because of Unicode’s Normalization Stability. However, there is a ticket accepted for Unicode’s Common Locale Data Repository to decompose those characters for the purpose of correct collation, cf. CLDR-10246 • Updated Hangul collation tailoring (probably a follow-up on L2/17-078).

So there are strong reasons to think of characters such as ㅔ or ㄳ as multigraphs.

Display of consonant or vowel sequences

Sequences of consonant or vowel characters are allowed by Unicode (though not by KS X 1026-1). According to the Unicode Text Segmenation rules, they should be displayed in a single syllable block. How should that look visually?

There are well-defined rules for the arrangement of Hangul letters into syllable block. Basically, those rules have been layed out in the Hunminjeongeum:

  • Consonant sequences are arranged from left to right (except for ㅇ, which is placed below).
  • Leading consonants signs are placed above the horizontal vowel signs or to the left of vertical vowel signs.
  • Trailing consonants are placed below.

By these rules, the only way to display the sequence 목ᆺ in a single syllable block is to make it look like 몫.

Is there a difference between hangul digraphs or trigraphs and sequences of the respective single letters?

Visually, I think there cannot be any difference between Hangul multigraphs and the corresponding sequences of consonant or vowel letters – if those sequences are displayed in a single Hangul syllable block as per the Unicode Text Segmentation rules. So 몫, which contains the digraph ㄳ, will look like 목ᆺ, which contains a sequence of ㄱ and ㅅ.

By the logic of the Hangul script, I think there cannot be a difference either. The character 몫, containing the digraph ㄳ, can be analyzed as a sequence of ㅁㅗㄱㅅ – exactly like the sequence 목ᆺ.

The only way by which 몫 and 목ᆺ displayed as 몫 would remain different is the Unicode normalization. The character 몫 decomposes to ㅁㅗㄳ and normalizes back to itself. The sequence 목ᆺ decomposes to ㅁㅗㄱㅅ and also normalizes back to itself. This is a potential security problem with no easy answer.

See also

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.