Chapter 1. An Apple Is Not an Apple — The Work Called

The content we make is now sold in 169 countries. It is translated into thirty-nine languages as standard, and sometimes as many as seventy. Every day, somewhere, someone encounters a character we created in Korean through their own native language. We take responsibility for that work.

Let us introduce ourselves briefly. MEJE Works Corp. opened in 2021. At first, we worked with companies such as LG Electronics, LG Household & Health Care, and Gravity to build IP libraries. The larger the company, the more it wants its characters, worldbuilding, and terminology to remain consistent in other people's languages as well. We helped preserve that consistency. From 2024 onward, we began creating worldbuilding IP centered on K-pop ourselves, and have been producing characters and daily content based on that IP. We call this content “fandom activity.”

In this work, an encounter with foreign languages is unavoidable. A line I write in Korean is carried into English, Japanese, Spanish, and Indonesian. In one language, that line may arrive wrongly. More precisely, when a small misalignment accumulates 169 times, a character's face hardens into something different. For that reason, we do not see foreign languages simply as “things that get translated.” We treat them as distinct, independent keywords.

An Apple Is Not an Apple

A spoon and sujeo are different. A spoon is Western tableware; sujeo is Korean tableware. Their forms may look alike, but their use, their place at the table, and the feeling of holding them differ. A printing press and a printer are the same way. “Printing press” brings a newspaper rotary press to mind, while “printer” brings an inkjet on a desk to mind.

What, then, of an apple? An apple is not an apple. The Korean sagwa and the English apple are merely dictionary words that point to each other. The red, firm, thin-skinned fruit that comes to mind when we say “an apple” does not perfectly match the image carried by English apple. The color, fragrance, and family feeling an American imagines in “apple pie” differ from what a Korean imagines in sagwa pai.

One word makes this most evident: pear.

The Korean dictionary renders pear as bae. Yet the large, crisp, intensely sweet Korean pear is not what pear carries. A pear in an English garden is slightly astringent, smaller than a Korean pear, and often used in jam or dessert. A pear is not as delicious as a Korean pear. Thus, for a Korean reader, “He picked up a pear from the table” can become a lie as soon as it is translated carelessly.

Similar cases abound beyond the table. Translate kimchi as “pickled cabbage” and the time of fermentation and the craft of the hand disappear. Translate soju as “Korean vodka” and its alcohol content doubles while the mood of the meal changes. Translate Japanese tatami as “straw mat” and the family seated formally on it vanishes. Translate sarangbang as “guest room” and it loses its double nature: a place for receiving visitors and for reading and discussing with them.

Every word carries the time, table, people, and customs of its own country. A dictionary does not mark what it carries. It shows only that two words occupy the same position; occupying the same position does not mean they have the same weight.

Our approach to foreign languages can be compressed into one sentence: Every word has its own weight, and when we carry a word across, we must carry its weight with it.

The Time We Have Reduced

This attitude is good, but labor-intensive. To inspect the weight of every word requires materials; materials require people; and people require time.

When we first began this work—more precisely, when members of our team began doing this work five to ten years ago—it took more than six months to prepare the materials needed to carry a book or an IP set coherently into another language. People read a book line by line, extracted terms, searched dictionaries, researched an era, and wrote documents by hand.

That time shrank. For a while, we used a thirteen-week process, the result of standardizing procedures, refining division of labor, and introducing some tools. Thirteen weeks was still long.

Our standard today is two weeks, and with a simplified process we can finish in under two hours. The simplified process applies only to variations on eras and genres we have already worked on; a work opened for the first time, such as Pride and Prejudice, still has a two-week standard.

You may be surprised to hear that six months became two weeks. Frankly, we were surprised at first as well. This shortening was not achieved by removing people and running only machines. I want to make that clear. We still place human judgment first; the time became possible because we separated precisely the points at which that judgment must arrive.

We intend to disclose part of that reduction. That is the purpose of this text.

What This Work Is — Not Translation, but Data Construction

Strictly speaking, what we do is more accurately described not as “translation,” but as data work that builds a world and precisely refines the meanings of its keywords within it.

A translator carries an already created text into another language. We work differently. Before moving the text, we rebuild the world in which that text lives. We document who is who, which events occurred in what order, what time people of that era ate lunch, and precisely what weight the word “civility” carried in Britain in 1813. Translation becomes coherent only when those materials exist.

We gave this preparatory work a name:

(書筵閣).

Seoyeon (書筵) is an old word for the setting in which a king read books deeply with scholars. Gak (閣) is a pavilion for keeping materials. A working place where books are read closely together and what is learned is kept: that is the place we enter before we handle content in another language. Every preparatory task before the main work—translation, adaptation, and localization—takes place here.

This text records the entire process of applying the work called to a single book.

The Book We Will Open Together — Pride and Prejudice

This text will follow one work from beginning to end and show the full process. Our case study is Jane Austen’s Pride and Prejudice (1813), commonly known in Korean as Omangwa Pyeongyeon.

You may download this book free of charge from Project Gutenberg (www.gutenberg.org). Search for “Pride and Prejudice,” and it appears as the first result. Its PG number is 1342. You may read this text with a copy open beside you, or without one; either way, you can follow its course. With the book open, however, each line comes more vividly alive.

The same procedure can be applied to other works: Shakespeare’s Romeo and Juliet, Melville’s Moby-Dick, Dostoevsky’s Crime and Punishment, Mary Shelley’s Frankenstein, and Lewis Carroll’s Alice’s Adventures in Wonderland. The kinds and volume of materials differ by work, but the framework is the same. If we handled several works at once in one text, we could not properly show any single set of materials. We therefore chose to concentrate on one book.

There are three reasons we chose Pride and Prejudice. First, Korean readers know it well. Second, it is rich in social context: class, marriage customs, and period conventions naturally call forth different kinds of materials. Third, it uses the demanding narrative technique of free indirect discourse, where the narrator’s voice slips into the inner life of a character and speaks alongside it. This is one of the points most likely to misalign in translation, making the novel a fitting specimen of translation difficulty.

Free indirect discourse can be shown in a single short line. One passage in the novel reads:

Was she to be congratulated, or was she not? (Should she be congratulated, or should she not?)

Who speaks this sentence? It can seem as though the narrator is asking the reader; it can also seem like a character’s private question to herself. In fact, it is both. The narrator has slipped into the character’s mind and carried her hesitation in the narrator’s own voice. In English, tense and the handling of pronouns (she/her) naturally signal that subtle entry. To retain it in Korean, however, the translator must identify that point exactly. Render it as narration and it becomes a cool commentary; render it as the character’s speech and it becomes an explicit monologue. One line divides into two different kinds of Korean.

There are dozens of passages like this in Pride and Prejudice. That is why it makes a useful specimen of translation difficulty.

A Fourteen-Part Course

This text continues for fourteen weekly installments. Installments 1–3 cover the place and framework of ; 4–7 examine the inside of the Pride and Prejudice work folder; 8–10 take up one kind of material at a time; 11–13 cover character notes and folder appendices; and installment 14 addresses the integrity and final accounting of the materials.

If you join us for all fourteen installments, you will see through to the end how a book establishes its household among us before it enters translation.

Let us introduce three terms that will recur often from the next installment onward. A profile is a two-page document that reads a work and summarizes its character. A segment table is a set of eighteen tables made by dividing the work into eighteen units. An event master is a single chronological index of every event in the work. These three form the foundation of the data work.

To Begin

There is one final thing I would like to say.

This text is not a boast. It is not an argument that our tools are good or our procedures superior. It is closer to a report on how we sit before a book when we carry it to readers in another language.

An apple is not an apple. A pear is not as delicious as a Korean pear. To acknowledge this anew each time, to look again at every word, and to place its weight into materials: that is .

In the next installment, we will trace, one step at a time, the six stages this work follows in practice.