Hangul Is Not a Phonetic Alphabet

Some Background on Hangul

The Korean alphabet gets a great deal of praise, and for good reason! Rather than continuing the prior tradition of using Literary Chinese[1] to write down documents or using other baroque methods to represent Korean sounds with Chinese characters[2], its inventor(s), King Sejong et fors. al., devised an elegant system that maps Korean morphophonology to characters in a compact and relatively simple way.

However, what irks me is frequently hearing comments that Hangul is a simple phonetic system. Some slightly misguided persons claim that it can mastered just by learning the sound of each individual character, and then stringing them together in order to read—if only if language were that easy! Whilst Hangul is simple in comparison to Korean's grammar system, Hangul does come with its own pitfalls and traps that may catch a first-time learner off-guard.

A Quick Aside About Linguistic Terms

There are a couple of terms that are helpful to know in discussions of spoken & written language. Humans can produce a wide variety of sounds that can be strung together to produce language, and we must give them good names and categories in order to talk about them. There are multiple names for 'sound' and groups of sounds in linguistics, mainly to be specific about what type of sound we're talking about.

Phones

To put it simply, a phone is a sound.

More specifically, a phone is a certain sound as distinguished by its very physical properties. These properties are a result of the way a person articulates a sound, that being derived from the location of their tongue, the rounding of their lips, the positioning of the teeth, the contraction of the vocal cords, the lowering of the palate so that air may pass through the nose, etc.

Phonemes

On the other hand, a phoneme is a unit of sound for a specific language.

To be more precise, a phoneme is a cluster of the multiple phones that a speaker of a language believes is the same 'sound'. Different languages (and even different dialects!) differ in whether they consider different phones to be different enough to constitute another phoneme.

If you're in disbelief that sounds get categorized differently based on the language you speak, I have an exercise for you! Place your hand in front of your mouth. Now, say the words 'crime' and then 'caught'. If your dialect is similar enough to mine, you should have felt a bit more air on your hand when saying 'caught' rather than 'crime'. I would wager that you consider both of those words to start with the same sound, a sort of 'c'. This is because aspiration—putting out a large amount of air when producing a consonant—does not distinguish phonemes in English. To put it simply, more air doesn't affect the type of consonant you're saying in English.

Phonemes are usually notated by putting a transcription between a pair of slashes, i.e. / /, whereas phones are usually notated by putting a transcription between a pair of square brackets, i.e. [ ]. Therefore, both 'c's in 'caught' and 'crime' would be notated[3] phonemically as /k/, however the 'c' would be notated phonetically as [k] and [k^h] for 'crime' and 'caught', respectively, with the ^h being used to indicate the aspiration—that extra breath of air.

This example carries us nicely over to our next category.

Allophones

An allophone is the specific phone that a phoneme is spoken as. In the previous example, [k] and [k^h] are allophones of /k/. Due to various phonetic rules that develop in languages, the specific allophone that is chosen when speaking is dependent on its surrounding phonemes, and choosing a different allophone from an expected one can lead to confusion.

Basically, if you choose the wrong variant of a sound you won't sound "native".

My favourite example of this is the word 'water'. In your average American accent, the /t/ is softened, becoming more of a light tap of the tongue upon the roof of the mouth. Try it for yourself! Pronounce 'water' as you normally would, and then again by over-emphasizing the /t/ sound to be a bit harder, as you might in the word 'tablet'. Sounds a bit weird, doesn't it? The usual allophone of /t/ in 'water' is that of [ɾ], a flap of the tongue, rather than of [t] or even [t^h], as you might find in the word 'tablet', 'stick', 'tick', 'time', etc.

For those Brits out there, your allophone of /t/ can show up as a [ʔ], a glottal stop, typically showing up in less posh accents which Americans make fun of by typing stuff out like "bo'oh'o'wa'er". Those über posh persons that speak in Received Pronunciation may have a stronger /t/, showing up as something like [t^h].

Morphemes

To speak normally, a morpheme is a unit of meaning for a language.

It is the smallest bit of language that gives some sort of meaning. For example, in English, 'cat' is a morpheme that means a temperamental and typically cute, fluffy animal. An '-s' is a morpheme that means that a word is plural, e.g. 'cats' being multiple cats. Some morphemes can stand by themselves, like 'cat', 'computer', or 'place', and so are called free morphemes, whereas other morphemes can only exist when attached to another morpheme, like the aforementioned '-s', 'con-', 'pre-', '-ed', or 'sub-'.

Some words are their own entire morpheme while others can be composed of many. 'fight' vs 'fire-fighter' is a one example.

You can string morphemes together to produce sentences.

Allomorphs

An allomorph is the specific realization of a morpheme, much like how an allophone is the specific realization of a phoneme, with an allomorph having a specific pronunciation, spelling, etc. For example, the plural morpheme can be written as '-s', '-es', or even nothing at all (represented using '-∅') depending on the morpheme that it attaches to (e.g. 'cats', 'boxes', or 'sheep', respectively).

Depending on the morphemes around it, the specific allomorph that a morpheme is concretely realized as can differ, whether that be at a phonemic or phonic level. The rules that govern this for a language are called its morphophonemic rules.

Why Does This Matter?

Hangul is neither a phonetic nor phonemic alphabet. It is a morphophonemic alphabet. Rather than representing the raw phones or phonemes that produce a word, Hangul instead tries to maintain the raw forms of a word's morphemes when possible.

This is simple for words that constitute of basically one morpheme, such as 바다, 'sea', which is pronounced and spelled as you might understand with a basic Hangul education: something like /bada/. ㅂ + ㅏ makes /ba/ and ㄷ + ㅏ makes /da/. Jamo correspond well to morphemes; morphemes correspond well to phonemes; morphemes correspond well with phones; and all is well in the world.

Things become slightly more difficult when it comes to more complicated words. Take this verb that means 'to get cold'. Its root form is usually given as 식다 /sikd͈a/ in Korean dictionaries, composed of its root 식- /sik/ and the suffix -다 /da/[4]. Korean is an agglutinating language[5], so we'll conjugate this verb by adding onto the root (식) an -어 /ʌ/ to indicate that we're talking like normal people and an -요 /jo/ to indicate that we're polite. Writing it in IPA, we'd get something like /sikʌjo/, and blindly writing that in Hangul by representing the sounds as jamo we'd get *시거요. This does produce the correct sound we're looking for.

However, according to modern orthography, it's wrong! It would be like spelling 'eat' as 'eet' in English. The root form must be preserved, and so we spell it as 식어요 and pronounce it as [시거요]. When reading, the ㄱ takes the empty spot that is an initial ㅇ.

A Quick Introduction to 받침법 (Final Consonant Rules)

Each Korean consonant has a different pronunciation at the beginning and the end of a block. See the Wikipedia page on Hangul to see all the sound values. Because of this, anyone with more than a month's worth of training in Hangul can tell you that there are only 7 effective phonemes that can be produced at the end of a block in Korean if one is to represent them with Hangul: ㄱ, ㄴ, ㄷ, ㄹ, ㅁ, ㅂ, ㅇ. All other consonants effectively become one of these consonants.

This is why the full names of consonant jamos are the way they are. They are composed of an initial block that demonstrates the initial sound of the jamo and then a second block that reveals the final sound of the jamo.

This may not sound so bad. But now, we must get into sound changes that occur when words are combined.

Assorted Trickeries

A prominent cause of allomorphs when reading Hangul are the nasalization of sounds. For example, the word for soup broth, 국물, is a compound word composed of the word/morpheme 국, soup, and the word/morpheme 물, water. However, because the ㄱ at the bottom of 국 meets with the nasalized ㅁ in 물, so too does the ㄱ become nasalized into an ㅇ and we end up with a pronunciation of [궁물] /kuŋmuɭ/ instead of *[국물] */kukmuɭ/.

For the uninitiated, how do you think the word 같아 is pronounced? If you were to say something like [가타] /kata/, you would be correct!

How about 같이? If you were to say something like *[가티] */kati/, you would be wrong! Instead, it's [가치] /kat͡ɕ^hi/! The 같- root must be preserved when written, but when a ㅌ at the bottom of a block meets an -이 (or [j] sound[7]), the ㅌ disappears and the -이 is replaced with a -치 when spoken. Same for when a ㄷ meets an [i] or [j]—it becomes a ㅈ attached to the next vowel.

A particularly egregious case is that of the reading of 깻잎, the humble sesame/perilla leaf. It's composed of the morpheme for sesame/perilla, 깨; the saitsori morpheme[8], -ㅅ-; and the morpheme for leaf, 잎. Rather than the naïve reading of something like *[깨싶] */k͈ɛsip/ or the slightly less naïve reading of something like *[깨집] /k͈ɛt͡ɕib/ by following the aforementioned rule, the proper reading of it is [깬닙] /k͈ɛnnip/. As far as I can tell, this is a result of the word leaf formerly being 닢 in all cases in Early Modern Korean but becoming /잎/ at the beginning of a word and /닢/ in compound words[9] like 꽃잎, 깾잎, and 찻잎. However, because of wanting to spell it like 잎 in all places, some batchim rules were invented so that it still came to be pronounced as /닢/ when read in the middle of a compound word. Simple[10].

Sino-Korean Vocabulary And Its Perils

Korean takes many words from Chinese and Japanese by reading the Chinese characters using a Korean pronunciation[11], much like how English has many borrowed Latin, French, and other foreign-coined Latinisms by taking the Latin and reading it as if it were English. Likewise, these Chinese characters, when combined into compound words (of which the vast majority of them are) must also follow Korean morphophonemic rules. However, current Korean orthography also wants to preserve their original forms (except for when it doesn't[12]).

This leads to the following scenario: every character is a morpheme, thus each character's spelling must be preserved. However, when these characters meet, these morphophonemic changes much occur, and we get the following situations:

• 법원 is pronounced like [버붠].
• 연락 is pronounced like [열락].
• 학년 is pronounced like [항년].
• 십륙 is pronounced like [심뉵].

Finding the exact reasons for these changes is left as an exercise for the reader.