In academic translation theory, there's a distinction between two broad approaches: domestication and foreignization. A domesticated translation makes the text feel native to the target language — idioms are replaced with equivalent idioms, cultural references are swapped for ones the reader will recognize, the reading experience is smooth. A foreignized translation retains traces of the source culture — it keeps the original rhythm, preserves terms that don't have direct equivalents, lets the reader feel that they're reading something from somewhere else.
Neither is inherently right. Both involve trade-offs. And the argument between them has been running in literary translation circles since at least the early nineteenth century, when Friedrich Schleiermacher wrote about whether the translator should bring the reader to the author or bring the author to the reader.
In scanlation, this abstract debate becomes concrete, specific, and surprisingly heated. It shows up in arguments about honorifics, about how to handle Japanese words with no English equivalent, about whether to explain a food reference in a translation note or just swap it for something the reader will recognize without needing to stop. Every one of these decisions is a choice between the two poles Schleiermacher described, and different translators and groups land in different places.
The honorifics question
This is probably the most visible version of the debate, because it's the one that most directly affects the dialogue — what characters call each other is everywhere in a manga, panel after panel, chapter after chapter.
Japanese has a well-developed system of honorific suffixes: -san for neutral politeness, -kun for male peers or subordinates, -chan for familiarity or cuteness, -senpai for someone senior in a school or workplace context, -sensei for teachers and doctors and masters of various kinds. These carry information about relationships, social position, and emotional distance that English doesn't encode at the name level.
A domesticated approach removes them. "Tanaka-san" becomes "Tanaka" or "Mr. Tanaka" depending on context. The relationship information gets conveyed through other means — through dialogue, through action, through narration — the way English would normally convey it. The reading experience is smooth. Nothing stops you.
A foreignized approach keeps them. "Tanaka-san said he'd be late." This preserves a layer of meaning that the domesticated version loses — you know, as a reader, that the character is being addressed with a certain kind of politeness. You also know you're reading something that came from somewhere with different social conventions, and that this is part of what you're experiencing.
Groups in the early IRC era tended toward foreignization. The reader base was self-selecting for people willing to learn some Japanese context, and the culture of the community was one where knowing the distinction between -kun and -chan was a mark of engagement with the material. As the audience for scanlated manga grew larger and less specialized, more groups shifted toward domestication — or at least toward a hybrid approach that kept the most culturally significant terms while smoothing out others.
The untranslatable term problem
Some Japanese words simply don't have English equivalents that carry the same weight. Nakama is the obvious example — it's often translated as "friend" or "comrade" or "crewmate" or "companion," all of which are slightly off. The word specifically means the people who are on your side, who you've chosen, who you'd protect. The closest English gets is "found family," which is a phrase rather than a word and has its own specific connotations that don't quite align.
There are a few options here. You can leave it in Japanese, italicize it, and include a translator note. You can choose the closest English equivalent and accept the loss. You can try to reconstruct the meaning through surrounding context — have characters demonstrate what the word means rather than stating it — and hope the reader picks up on it. None of these is obviously right.
What's interesting is that readers who've consumed a lot of scanlated manga often develop an understanding of certain Japanese terms through accumulated exposure. They don't need a translator note to explain nakama because they've seen it enough times in enough contexts. This creates a kind of reader who is partially operating in a Japanese cultural framework, which is unusual and arguably something that domesticated translation can't produce.
Cultural references and localization
A character in a manga makes a reference to a specific type of Japanese food, a regional dialect, a particular school tradition, or a pop culture touchstone that a Japanese reader in their twenties would immediately recognize and find funny. What does the translator do?
Full localization would replace it with something an English-speaking reader would recognize. The risk is that the replacement has different connotations, different cultural weight, different comedy timing. Something that's warm and nostalgic in the original might come out as generic or oddly specific in the localized version.
A translator note explains the original reference. The risk is interrupting the reading flow and making the reader aware that they're missing context they don't have. It can also stack — if you're adding notes every other page, the experience starts to feel like reading an annotated text rather than a story.
Leaving it untranslated assumes the reader either knows or is okay not knowing. This works better in some contexts than others — a historical manga about feudal Japan can probably have some untranslated period terms without confusing readers, because the expectation of unfamiliar context is built into the genre. A contemporary high school comedy where characters reference specific convenience store snacks is harder.
A position
For what it's worth, the view here is that "good" translation means having a considered position on these questions and applying it consistently — not being right in some absolute sense, but being coherent. A translation that mixes domesticated dialogue with foreignized honorifics and inconsistent handling of cultural references doesn't fail because it chose the wrong approach; it fails because it didn't choose at all.
The argument between domestication and foreignization isn't going to be resolved. It's been running for two hundred years and there are good reasons for that. What can be assessed is whether a translation knows where it stands and maintains that position through the work. That, at minimum, is what "good" might mean.