Thread by @ZevHandel

This is part of a sign in a station on Line #4 of the Seoul subway system.

It’s trilingual.

Um, it is trilingual, right? Or … is it?

Take a moment. Think about it.

It’s in three different scripts, that’s for sure. But is it in three different languages?

When I lived in Seoul I spent a lot of time on the subway. And I spent a lot of time looking at signs in the subway stations. When I looked at signs like this, I thought:

Is this sign in Korean, English, and Chinese ❓

Or maybe … is it in Korean, English, and Japanese ⁉️

And then I thought: Why am I assuming English? Why couldn't "Dongdaemun" be … Dutch? Indonesian? Turkish?

Or even … Korean? 🤔

This raised a thornier question: Is there any way to know for sure how many different languages are being written here?

Maybe the sign is entirely in Korean, represented in three different written forms: in the Roman alphabet, the Hangul alphabet, and Hantcha (i.e. Chinese characters).

So maybe it’s not trilingual at all, but monolingual and *tri-scriptal*! 🤯

A Korean word or proper name made up entirely of Sino-Korean components can be written in all three scripts. You can’t necessarily tell by looking at the outward form what spoken language is represented by it.

Let me illustrate a bit more what I mean. If we have a full sentence or even just a phrase, we can identify the language because we can inspect the words and grammar.

1. 사과를 먹어
2. eat apples
3. 吃蘋果

There’s no doubt that those sentences are Korean, English, and Chinese.

On the other hand,

“sagwa-reul meogeo”

has to be Korean, even though it’s written in the Roman alphabet.

So when we see:

동대문 / Dongdaemun / 東大門

It could be Korean-English-Chinese, or Korean-Korean-Korean, or Korean-English-Korean, or Korean-Korean-Japanese, …

And it got me thinking … is there any way I can determine which languages are being written here?

Is it even plausible that the signage is entirely Korean? What would be the point of writing the same Korean name three times in a row, in different scripts?

I think it is plausible. After all, adding romanized Korean allows users of languages written in the Roman alphabet to recognize and identify stop names. And adding Korean written in hantcha allows users of the Chinese and Japanese scripts to recognize and identify stop names.

It's also plausible that the signs would be in Korean, and also in English (the most commonly recognized language in the world), and either Japanese or Chinese (because so many visitors to Seoul are from China, Taiwan, and Japan).

But how to decide which of these two plausible possibilities is actually the case?

I started inspecting more signs to see if I could resolve my question, and the answer soon became clear.

Here are a few more signs. See if you can figure it out.

Did you figure it out?

Look at the first photo. Take another look at this:

미아삼거리 / Miasamgeori / 彌阿삼거리

"彌阿삼거리" is clearly neither Chinese nor Japanese, since neither of those languages is ever written in Hangul.

It’s gotta be Korean!

It's mixed-script Korean.

The Sino-Korean components are written in hantcha. The native Korean components like geori ‘road’ have to be written in Hangeul because modern mixed-script Korean only permits Chinese components to be written in Chinese characters.

So the third script isn’t Chinese or Japanese. It’s a mixed hantcha-hangeul written form of Korean.

Kind of crazy, right? Why write Korean twice?

We'll come back to that question in a bit.

Although we've resolved the question of the the third language on the signs, we still can’t tell if the second language is English or romanized Korean. “Miasamgeori” could be either.

But this sign seals the deal. We've finally found a location name that disambiguates the languages for us.

동서울터미널 / Dongseoul Bus Terminal / 東서울터미널

The second language is clearly English, and the third is clearly Korean. The Korean name “Dong-Seoul Teomineol” has only one Sino-Korean component, namely dong ‘east’. So that’s the only part that can get written in hantcha.

(Seoul is very unusual. As far as I know it's the only modern city name in Korea that is native Korean instead of Sino-Korean.)

Maybe you're asking this question:

Does the written form “東서울터미널” help a Chinese or Japanese tourist find their way around?

It's a fair question. My answer is:

Probably not.

But 동서울터미널 is an outlier; the majority of Korean place names are mostly or entirely made up of Sino-Korean roots, so the Seoul subway signage system works pretty well overall for Chinese and Japanese tourists.

Here, for example, is a sign that works really well for users of a Chinese-character script, because the mixed hantcha-hangeul written form is entirely in hantcha:

주변지역 안내도 / Neighborhood Map / 周邊地域案內圖

The whole map is fun to look at and compare the three ways of writing each location.

So now, assuming consistency in signage, we can go back to the sign we started with and say confidently:

It's bilingual (Korean-English-Korean)

and

tri-scriptal.

It’s yet another reminder that script and language are not the same, and can intersect in surprising ways, as I illustrated in this language quiz from a few months ago:

https://twitter.com/ZevHandel/status/1339459027831455744

/end

I should have made this clear from the beginning: These photos are from 2007.

The signage today has changed, as @dailykoreanfact helpfully pointed out.

https://twitter.com/dailykoreanfact/status/1370635349395836929

Now there are unambiguously four languages in four scripts:

1. Korean in Hangul
2. English in Roman alphabet
3. Mandarin in simplified Chinese characters
4. Japanese in Katakana

There's a fine line between translation and transliteration when it comes to proper names.

The English translation of Tōkyō 東京 is "Tokyo".

The normal way to translate "Tōkyō-ni sundeiru 東京に住んでいる" into English is "I live in Tokyo", not "I live in the Eastern Capital".

Translations of proper names are typically sound-based borrowings or transliterations, not calques.

Two more examples: Máo Zédōng 毛泽东 in English is "Mao Zedong", not "Shining-East Fur". Düsseldorf in English is "Dusseldorf", not "Dussel Village" or "Dusselthorp".

For place names with transparent etymologies and no widely accepted English form, whether a proper translation should be a calque or a transcription is an open question.

Is the English translation of Dongdaemun 동대문 "Dongdaemun" or "East Gate"? I can imagine either one.

But I find the choice made in this sign odd. I would probably have gone with "East Seoul" not "Dongseoul".

There's an added twist when the place name is made of Sinitic elements with known Chinese characters, and the languages involved are Chinese, Korean, Japanese, and/or Vietnamese.

This is because sinogram-based orthographic transliteration is more or less the same as etymological calquing, and is a viable alternative to strict translation and to strict transliteration.

To give an example. Suppose you want to translate Korean anneido 안내도 "guide map" into Standard Written Chinese. You have three choices, unlike the two choices available if you were to translate into English.

1. You could *translate* it into the Mandarin word with the most similar meaning, dǎolǎntú 导览图.

2. You could *transliterate* it into Mandarin syllables approximating Korean pronunciation: ānnèidǒu 安内斗.

3. You could identify the Chinese characters that represent its etymological origin (案內圖), adjust them into the simplified script (案內图), and then pronounce them in their conventional Mandarin readings (ànnèitú).

Sometimes this yields a word that exists already in the target language, sometimes it doesn't. But because the morphemes have recognizable semantics, the process usually "works" for language users, with an intelligible result.

This third option is kind of like if you were to translate German Düsseldorf into English "Dusselthorp". We just don't do this sort of thing with any regularity outside of the Sinographosphere.