Behind the Scenes of the English translation of The Labyrinth House Murders
Pushkin Press published my English translation of The Labyrinth House Murders (Meirokan no Satsujin) by Yukito AYATSUJI in October 2024. This is the third entry in the series, and I was absolutely thrilled when I first heard Pushkin wanted to continue the series after The Decagon House Murders and The Mill House Murders, and that they wanted me continue with the series. But, I have to admit: simple joy was not my initial, honest response. For my first thought when The Labyrinth House Murders was mentioned to me, was: "Hey, is that book actually going to work in translation?"
Because the (shin) honkaku mystery genre is very much about being modern fairly-plotted mystery stories, sometimes you face a problem: a trick that works in the original language, but wouldn't work in the target language as is. It could for example require a lot of footnotes to explain what the deal is about in the first place, but that would of course give the game away, as it draws your attention to something that is apparently important to solving the mystery, even though in the original book it might have been hidden very cleverly.
In the case of The Labyrinth House Murders, I had read the book about a decade before the offer and I remembered there’d be aspects of the book that would be a bit challenging in terms of translation. People who have read the original Japanese novel might remember one ‘major’ point especially that it is very closely tied to Japanese language. Reading the book again after the initial offer to see how doable the book was in translation, I was however reminded there was another aspect that would either require a lot of footnotes to even explain to the English-language reader, or it would require me, as the translator, to change a few things in order to keep the fair-play spirit of the book intact in the English translation. In the latter case, it would be of course be preferable if I could also preserve the underlying ideas of the tricks as imagined by the original author Ayatsuji. So yes, a part of the mystery has been altered by me from the original Japanese. Fortunately, I was able to consult Ayatsuji himself, who gave his blessings when I explained the why and how I wanted to resolve it.
I will go into more detail regarding the translation challenges of The Labyrinth House Murders in the part below, so I have to warn you for major spoilers for the book. I advise you to read the novel first before continuing.
SPOILERS BELOW ** SPOILERS BELOW ** SPOILERS BELOW ** SPOILERS BELOW
1 A Minor Change
Sometimes, a trick doesn’t work ‘as is’ when translated, but fortunately might only require relatively small changes from the original to work as intended. An example of this appears at the end of Chapter Ten, when Shimada points out the stories “The Minotaur’s Head”, “A Yearning for Poison”, “The Ghost in the Glass” and “Killing Wings” spell out “MYGK”, or “Miyagaki”, the signature of the culprit.
In the original Japanese, these stories had slightly different titles. Minotaurosu no Kubi (The Minotaur’s Head), Yami no Naka no Dokuga (The Poison Fang in the Darkness), Garasubari no Dengon (The Glass Message) and Kikei na Tsubasa (Malformed Wings). Because the way the Japanese language works, if you take the first hiragana (something like a syllable) of each of these titles, you end up with MI(-no-ta-u-ro-su), YA(mi-no-na-ka-no-do-ku-ga), GA(-ra-su-ba-ri-no-de-n-go-n) and KI(-ke-i-no-tsu-ba-sa), spelling out the name MI-YA-GA-KI. In English of course, it wouldn’t make sense to arbitrarily take the first *two* letters of each title, instead of simply the first (minus the article “the”), hence the titles here spelling MYGK (which is still recognizable as “Miyagaki”), while not straying too far from the original Japanese titles. This is an example of something that can be resolved quite easily in translation.
2 A Major Change
Whereas the problem above was relatively minor, the one I’ll discuss here required far more effort and it was the one thing I wanted to have Ayatsuji’s permission for. When Pushkin first asked me about The Labyrinth House Murders, I actually thought the issue I’ll mention as the third point (see below) would prove to be the most difficult problem to resolve. I wasn’t even sure whether it was doable, so I told them I wanted to re-read the book just to be sure, only to realize I had forgotten about this problem, and it was this that would turn out to be the most challenging one.
At the end of Chapter Nine and the beginning of Chapter Ten in the translation, we finally learn the meaning of “nuei”, the mysterious message left by Hayashi on his word processor. Utayama points out Hayashi had learnt to blind type, that he had probably misplaced his initial starting position when typing the message, and furthermore, Hayashi's keyboard had sticky keys. From there, Utayama tries to surmise what Hayashi might have tried to type instead of "nuei", and they arrive at the message “miro” (Hayashi's starting position having been one key to the left of what he had intended). Accounting for the sticky keys gives them the message “mirror”: the location from which the murderer appeared and fled. So Hayashi had intended to type out the word "mirror", which would become "nueeie" because of the wrong starting position of his fingers, which in turn, due to the sticky keys, became "nuei" as the "e" button wasn't always responsive.
The original Japanese message was not “nuei” but “wwh”. The change in message is not just a simple change because we went from Japanese to English. The concepts behind how the message is solved also involves the assumption Hayashi was desperately trying to type out a message pointing out where the murderer came from, and that his message had been ‘transformed’ without him realizing as he was dying as he typed the message. The problem is that in the original Japanese version, the message was “transformed” from its intention because Hayashi was used to working on a type of word processor with a special keyboard layout. This layout does not exist at all in English, and only exists in Japan to accommodate the Japanese language. In order to keep this part of the mystery solvable, the editor and I agreed this would have to be changed in English and I also discussed this with Ayatsuji himself.
There are two major methods to input the Japanese language on a keyboard. One is using the alphabet to type in the Japanese, and then you can convert the alphabet input into Japanese. That is what I use, so I type, in alphabet, “meirokannosatsujin” (The Labyrinth House Murders) and can have it automatically converted to the corresponding Japanese “迷路館の殺人”. Most Japanese keyboards follow the QWERTY layout for alphabet input. However, there’s usually also a key to alternate between this alphabet input method (known as romaji input) and direct kana input, kana being basically something like a syllable in the Japanese language, of which there are 48. When set to kana input mode, the keys on the (same) keyboard correspond directly to a kana of the Japanese language (or in combination with another key, their voiced/unvoiced versions). The button labeled “Q” in romaji mode for example will become the key to input the kana “た” (ta). So if you want to type in “meirokan” in kana mode, you'd first push the key on the keyboard that now also functions as the kana “me” (which is usually the question mark key on a normal QWERTY layout), then the one for “i”, then “ro” etc.
The issue here is that this is not the only layout used for direct kana input. The one I explained above is known as the JIS layout. In the original Japanese text of The Labyrinth House Murders, we learn the word processors provided at the Labyrinth House follow the JIS Kana layout (where the Q=た/ta etc.), whereas Hayashi at home had been using the so-called “thumb shift” layout, which was created by word processor maker Oasis. The “thumb shift” layout is a unique layout that features two special “shift” buttons, and each of them is operated by a thumb (so left thumb and right thumb), hence it being known as the "thumb shift" layout. This means all keys on the keyboard can have three modes: no shift, with left shift pushed down, or the right shift. This allows for a keyboard to cover all 48 kana of the Japanese language, with a limited number of keys. This makes it easier for a user to write the Japanese language, as they don’t have to move their fingers as much while typing, and it was therefore an input method much loved by people who have to write a lot, like novelists. The “key” thing here is to remember that 1) the layouts of the kana are different between the JIS and “thumb shift” layouts (so different keys correspond to a different kana depending on layout) and that 2) the “thumb shift” layout utilizes two shift buttons which don’t exist on a word processor with a JIS layout keyboard.
So that brings us to the original Japanese dying message. As mentioned above, the message left by Hayashi in Japanese was not “nuei” but “wwh”. While the characters speculate a few times about its meaning, it’s only at the end of Chapter Nine, Keiko remembers Hayashi had been complaining about having to use a different kind of keyboard at the start of the tale, and with the help of Utayama’s knowledge of the keyboard input systems, they first realize that Hayashi inadvertently had set the keyboard on alphabet/romaji mode, and that he had tried to type on the JIS layout keyboard of the word processor, as if it were a “thumb shift” keyboard layout, because he was used to working on an Oasis word processor.
This means that even though Hayashi typed the alphabet characters “wwh”, he had intended to type out Japanese kana, and with the intention of typing via the “thumb shift” layout, and not JIS. Because of how the “thumb shift” layout works, each key can represent one of three kana, depending whether a, and which shift button is held during typing. The alphabet key “W” could represent “ka”, “e” or “ga” in Japanese, while the “H” could represent “ha”, “ba” or “mi”. This means there are 3 * 3 * 3 = 27 different possible combinations of kana that “wwh” could represent, like “ka-ka-ha”, “e-ka-ha” etc. Utayama types out all of these combinations, as they can't tell just from "wwh" whether Hayashi had intended to push any of the shift buttons or not. Most of these tries result in nonsensical combinations, until Shimada focuses on the combination “ka-ga-mi”, which means “mirror” in Japanese, pointing the survivors to the mirror in the room.
And yes, this was impossible to implement in the English translation in a fair manner. It would require far too much explanation about the Japanese language and input systems (as you can see here), especially if you’re just reading the book to have a bloody adventure with a fair-play mindset! So the challenge here was to create a dying message that 1) pointed towards the “mirror” in a way that an English reader would understand, 2) that relied partially on the notion of specifically Hayashi having trouble to type on his word processor because of acquired habits and 3) it needed to be a message that allowed for some meaningful speculation. The last point was also important: in the original Japanese, we have the characters come up with a few mystery-related ideas regarding “wwh”, like the suggestion it might be upside down, meaning it was “hmm”, indicating Hayakawa Mystery Magazine (an actual existing magazine) or how it could or not stand for the initials of any of the suspects. So if the message “wwh” was changed, these discussions would need to be altered too, but still follow the same discussion beats. But point 2 was also important as there is proper foreshadowing/clewing regarding Hayashi having a problem with using his word processor early on in the book (him complaining to Kiyomura and Ino), and I wanted the English message to also be clewed properly. And of course, most importantly, the rewritten clue had to fit in with the rest of the book, and fit naturally in the narrative.
After playing with ideas like Hayashi being used to an AZERTY layout, or for example other types of message that could also indicate a mirror/looking glass in one manner or another, I eventually settled on the idea of Hayashi still ‘touch-typing’ the wrong message because of learned habits, but with a keyboard with sticky keys (a matter he could complain about early on in the book, and yet have Ino be unable to help out). The new dying message of “nuei” allowed for some speculation regarding the “Nue”, a Japanese mythological being that like the Minotaur is a combination of animals and also allowed for proper mystery-related references/guesses, that wouldn’t feel too out of place in the context of the book. The idea of an "unfinished" dying message is a common trope of course, so it was quite natural to move from "nuei" to a possible "nue is..." I informed Ayatsuji about needing the change the trick regarding the keyboard and explaining what I had mind and how the clewing would be altered, and I was more than happy when he accepted these changes, despite the trick being a rather drastic change from the original (which is of course far more original and more complex than the one I came up with). Which is also a reason I wanted to explain the original trick here.
Minor note: this is also why the English translation actually has one less diagram than the Japanese version, as the original “diagram 3” related to the explanation of the “thumb shift” keyboard, which was of course not relevant anymore in the translation. I used the original "diagram 3" above.
3 Small Changes To Prevent A Major Change
When I was first asked about The Labyrinth House Murders, I remembered the book’s last big reveal: the fact the culprit was actually a woman, while the book makes you think/assume on your own she was a man. Of course, this is fairly easily done in Japanese, as you don’t need to use gendered pronouns as much as in English. The problem here is of course you can’t get away with this in a translation simply by using non-gendered pronouns like “they”, because the trick relied on forcing that specific character = male assumption on the reader, without actually using as many words.
So I first asked for some time to read the book again, and I have to admit I was surprised when I realized it could be done in English. I could easily imagine it done in a shorter story, but I was glad to see it was possible to do in a full-length novel too. Again, in Japanese this switcheroo is easier to pull off because you often refer to a person by name, rather than by (gendered) pronoun and in general, you can drop pronouns much more easily in the Japanese language. But as I re-read the book with a translation in mind, I definitely could see enough possibilities to do the same. This would require some minor rewrites of sentences at times, like shifting the subject of a sentence elsewhere (for example going from “X used item Y” to making “Y” the object of a sentence), referring to the culprit not by name but by profession to create some variety in the wording, grouping the culprit with other people in certain sentences/scenes so I could refer to a collective “they”: little tricks/changes like that that had to be implemented here and there depending on the sentence and scene, but it could be done. If for example the book had featured extended scenes with only the culprit and she’d be constantly described by the narration, then this would have been far, far more difficult to pull off.
Point 2 mentioned above required some creativity on my part, but I’d say that the mystery-related writing regarding point 3, mostly required concentration. Having read the book already, I of course knew that the culprit was a woman, and I had to be very careful not to accidentally type out a gendered pronoun whenever she was in the scene or referred to, as it is very easy to just let ‘her’ slip by. And at the same time, it was important to not be too blatant about avoiding her pronouns in the writing too much, as that in turn would also attract attention to her specifically. I was fortunately helped by the fact that in a way, the author wants the reader to notice the trickery going on, so some degree of unnaturalness can be forgiven, but still… But again, implementing this trick in the translation required a bit of creative wording to convey the original Japanese. I’d say this challenge was more about me having to be very careful and keeping that in mind the whole time, rather than requiring me to be creative at a certain mystery-writing level like the previous challenge mentioned. Though I did really like having to be “careful” all the time, as it really got me involved with the mystery-plotting throughout the whole novel: even in scenes without much going on, I could have been wrestling with a sentence that wasn’t important on its own, but happened to involve the culprit, meaning I needed to find the correct word choices for that sentence. It’s also here where the knowledge editors will go over my work and think along too, is a great relief, because it’s so easy to overlook one little word you might type by accident. Of course, I had to spoil this part of the mystery upfront to the editors, because this is one thing that could easily be ‘corrected’ by them in a first read of the first version of the manuscript.
And that wraps up my musings on the translation of The Labyrinth House Murders. I hope this was informative and given you more insight about what the original Japanese text contained.
