Translating Ancillary Justice.

I’m surprised to realize I haven’t mentioned Ann Leckie in any LH posts (though I have in comments); her novel Ancillary Justice is the best science fiction I’ve read in years, and it has an obvious linguistic hook, which I’ll let Alex Dally MacFarlane explain at the beginning of her very interesting and informative essay “Translating Gender: Ancillary Justice in Five Languages“:

In Ann Leckie’s novel Ancillary Justice (Orbit Books: 2013), the imperial Radch rules over much of human-inhabited space. Its culture – and its language – does not identify people on the basis of their gender: it is irrelevant to them. In the novel, written in English, Leckie represents this linguistic reality by using the female pronoun ‘she’ throughout, regardless of any information supplied about a Radchaai (and, often, a non-Radchaai) person’s perceived gender. This pronoun choice has two effects. Firstly, it successfully erases grammatical difference in the novel and makes moot the question of the characters’ genders. But secondly, it exists in a context of continuing discussions around the gendering of science fiction, the place of men and women and people of other genders within the genre, as characters in fiction and as professional/fans, and beyond the pages of the book it is profoundly political. It is a female pronoun.

When translating Ancillary Justice into other languages, the relationship between those two effects is vital to the work.

After reading a comment by the Hungarian translator, Csilla Kleinheincz, posted on Cheryl Morgan’s blog, we wanted to know more about this. We invited the translators of the novel into Bulgarian, German, Hebrew, Hungarian and Japanese to discuss the process, with particular interest in the translation of gender. What emerges is an insight into the work of translators and the rigidity and versatility of grammatical gender in the face of non-standard demands. Where necessary, translators turned to innovative and even inventive ways to write their languages.

It’s fascinating stuff; I’ll quote this bit because of its resonance with yesterday’s post:

Hebrew presented considerable challenges, as Emanuel Lottem explains in shared correspondence with Ann Leckie: “All verbs and adjectives vary according to gender. An exception is the pronoun ‘I’, which is the same for both females and males, but when a woman says ‘I sit’ it’s ‘ani yoshevet’, whereas a man would say ‘ani yoshev’. Moreover, all objects, concrete or abstract, are either feminine or masculine, so that in a sentence like ‘a large room (m.) with a large door (f.)’, the word ‘large’ would be ‘gadol’ in the former case, ‘gdola’ in the latter. This is not much of a problem when Breq tells her story, but it is when non-Radchaai are talking.”

This even presented a problem for the novel’s title: “The word for ‘justice’ – tzedeq – is unfortunately masculine, and I’d hate to use it in the title because of that.”

(I once posted three times in a row using Latin titles, entirely without premeditation; now I’ve similarly done two in a row on gender. Blogging is almost as full of synchronicity as life itself.)


  1. This probably is the best thread as any to ask a question that mildly interested me for some time. Using she as unmarked pronoun is a nice new feature of politically correct English. But does it work if the referent of the pronoun is doing something less than admirable? For example, “If the book’s author doesn’t know what beanball means, she must be really stupid” or “When politician talks about hard issues, she is often less than honest”, etc.

  2. Of course, just as it works when the referent of the pronoun is doing something admirable. The point is not to make you think women are wonderful creatures but to make you rethink the automatic assumption of maleness.

  3. In one of Delaney’s lesser-known stories (too hard to find out which at the moment), everyone is called she except when they happen to be sexually attractive to the viewpoint character, in which case he is used.

  4. Le Guin rewrote the first chapter of The Left Hand of Darkness (which in the original uses only masculine forms) several times: once with feminine pronouns and titles for the androgynous characters (Genly the Earthman remains he, of course), and again with the invented pronoun a (possessive form uns, I think). Gale and I, as well as Le Guin, agreed that the feminine version worked not at all: Argaven the Queen is just a totally different person from Argaven the King, even though a does and says exactly the same things, and even more so for Lady Estraven (shudder). Gale and I thought we’d need to hear the whole book with a to judge it fairly, but it seemed unproblematic, which would not have been true of sie or thon or any of the multitude of other less minimalist invented pronouns that have been tried. Le Guin’s unpublished screenplay uses a throughout.

    In the passage where Estraven goes into estrus (as female, Genly being male), she also gives us a few pages in the a-dialect, shifting to she as (or even just before) Genly becomes aware of it. She also pointed out that her neologism parent in the flesh was totally unnecessary; it should have been simply mother. However, habent sua fata libelli, especially after so long. There’s nothing to stop you, if you happen to be reading it, from substituting the a forms in your own head, or saying them out loud if you are not an Augustinian.

    The story “Winter’s King”, which she originally wrote in the masculine for the excellent reason that Gethenians were not yet ambisexual, was revised after the novel to use feminine pronouns but masculine titles. I have never read the older version, but Le Guin claims it’s quite different, much more Oedipal than the standard version. The Gethenians in “The Shobies’ Story” are mostly referred to in the plural, I think.

  5. «In one of Delaney’s lesser-known stories»

    Stars in My Pocket Like Grains of Sand. (Also that’s “Delany”, no second “e”.)

    I once altered an ebook of The Left Hand of Darkness to use feminine pronouns, so I could try reading it that way, and I thought it worked fine. But I only changed the pronouns, as in the revised “Winter’s King”, so Estraven remained a lord and Argaven a king. I had difficulty deciding which pronouns to use when Gethenians referred to Ai; I think I went with “she” on the grounds that they’d be using the same pronoun for him as for each other, but I was never completely happy with the decision. After all, that’s the kind of thing you’d normally change when translating from a language without gendered pronouns.

  6. The gender of the beloved changing within the poem happens in Arabic and Persian poetry. The first time I came across it, I thought it was a printer’s error. (Now that I look for an example, I get all sorts of inappropriate results, but hopefully someone knows what I am referring to.) But it is a different attitude towards consistency within a poem.

  7. Google Translate’s attempt to translate a phrase from the book into Mongolian:

    “She was probably male.” – > “Тэр нь магадгүй нь эрэгтэй байна.” (That person is probably male)

    That’s reasonably correct, Mongolian only has gender-neutral personal pronouns, so there is no way to convey strangeness of this sentence in the original.

    Of course, one can always translate as “This woman was probably male”, but it would be so plainly absurd that readers are likely throw up the book in disgust.

    Anyway, Mongolians are very practical people and simply don’t get SF, so the point is moot. Nobody will read it there.

  8. marie-lucie says:

    In one of Ruth Rendell’s detective novels a woman is found murdered and in trying to find a motive which might lead to the murderer the police discovers that she had kept a number of love letters dating from before she was married. The identity of the writer, whose unusual signed name suggests a man, remains a mystery until the fortuitous discovery that the writer was another woman. After reading the book I wondered how it would be possible to translate it into French in such a way as not to reveal the writer’s sex from the discovery of the letters, since almost every adjective used by the writer to refer to herself would be in the usually distinctive feminine form, as in Je suis heureuse ‘I am happy (fem)’ not Je suis heureux (masc). (I did not try to find a solution myself).

  9. Delany

    I know that perfectly well, but get it wrong more than half the time. Take it as a perpetual qere. 🙂

  10. “After reading the book I wondered how it would be possible to translate it into French in such a way as not to reveal the writer’s sex from the discovery of the letters, since almost every adjective used by the writer to refer to herself would be in the usually distinctive feminine form, as in Je suis heureuse ‘I am happy (fem)’ not Je suis heureux (masc).”

    I don’t speak French (unfortunately) but it seems to me that if it’s only a question of adjectives the problem could be sidestepped by using nouns in the translation:

    I am happy — I have never known such happiness
    I was so disappointed — disappointment descended on me
    I became angry — I was overcome by anger

    But no doubt this would have all kinds of unfortunate stylistic implications. In English, at any rate, this sounds really stilted, borderline Victorian.

  11. And the Victorians went to great lengths to keep their (robust) sexuality under wraps. See, it all fits.

  12. Alon Lischinsky says:

    I’ve been thinking for a while about the problem posed by translating the delightful ‘She was probably male’ passage to Spanish, simply because the pro-drop nature of the language would make it very stilted to keep the feminine pronoun there. Unfortunately, the stylistic effect would be much diminished without it, although there are other opportunities for gender marking in the paragraph.

  13. There’s another way to go with this/ Rather than simplifying the paradigm, which can make things seemed more than a little forced, you can do the opposite – make the pronominal paradigm so complex and arbitrary that you can refer to anyone however you decide. I saw this paper on the pronouns in Northern Iroquoian languages:,d.eWE

    …and I thought, hmmmm….a writer could have all kinds of fun with this.

  14. Interesting stuff. The paper is “A history of Iroquoian gender marking,” by Michael Cysouw, and the abstract begins “The North Iroquoian languages have a three-way gender division in the third-person prefixes.” Here’s a bit from the body of the paper that explains the gender thing:

    However, there are notable differences between the languages for the marking of feminine referents. Consequently, these prefixes are labeled in this paper by their joint meaning as identified here (i.e. “indefinite,” “masculine,” and “non-human”). These labels do not imply a complete characterization of their functions. For example, the “non-human” prefix is used in many languages for part of the feminine human marking, but not in all languages, and not always to the same degree. The name “non-human” only indicates that all languages agree on the fact that this prefix is used for at least non-human marking.

  15. marie-lucie says:

    Alon: translating the delightful ‘She was probably male’ passage to Spanish

    I think that in French I would write: (Elle/Cette femme,) c’était sans doute un homme. Adding the extra-sentential noun-phrase would make the sentence clear, without becoming the grammatical subject. The indefinite pronoun ce as subject is very useful as it never indicates gender and is therefore compatible with a wide variety of phrases, unlike English “it” which cannot be used for a person.

    Spanish does not have a full counterpart (at least I don’t think so), but one could still add a noun-phrase, as in “La seNora debia ser hombre” (correct me if needed).

  16. The challenge of this translation is not to hide the gender of the person, but rather to indicate that it is female and probably male at the same time!

  17. Alon Lischinsky says:


    Spanish does not have a full counterpart (at least I don’t think so), but one could still add a noun-phrase, as in “La seNora debia ser hombre” (correct me if needed).

    You could, but it would not sound idiomatic. The male patron has been discourse-topical for almost the entire paragraph, and using an NP to refer to her as a sentential subject would be odd at this point (at least in the standard register; it would be possible to use the pejorative la tipa ~‘cette garce’, but this would not reflect the way the narrator speaks in the original).

    I think the easiest way to solve this would be to employ a construction in which the patron appears as direct object (something like la juzgué probablemente macho), but I haven’t been able to come up with an elegant one.

  18. The Japanese translator Hideko Akao has this to say in the linked essay:

    “It is in colloquial speech that I had to make a particular effort in the Japanese translation. Most common English personal pronouns aren’t gender-specific but third-person singular is. On the other hand, Japanese has a great many first- and second-person pronouns and they may often differ depending on natural gender. In addition, Japanese has many different tones between male and female in conversational sentence, in particular in sentence-final expressions: if a character speaks in an ‘unfeminine’ manner, Japanese readers often tend to assume it is ‘him’ automatically.”

    In Korean, the basic third-person singular pronoun is 그 geu, which is not gender-specific. The pronoun 그녀 geunyeo “she” combining 그 geu “this” and 녀 nyeo “woman (女)” is a 20th-century coinage used for translating gendered pronouns in English and other languages, and is not used colloquially, only in translated texts. It still feels too stilted and artificial to use even in writing (though it does occur sometimes in writing that is not translated from other languages but composed directly in Korean), and I don’t see the point of trying to introduce gendered basic pronouns when Korean has done fine without them. In other languages some people are even going the other direction, trying to introduce gender-neutral pronouns.

    그녀 geunyeo is thought to have been inspired by the similar Japanese creation of 彼女 kanojo as a feminine pronoun for use in translation, but I get the feeling that the usage is different between the two languages based on the excerpt above. Is 彼女 kanojo more accepted in Japanese writing than 그녀 geunyeo in Korean writing? In Korean, 그 geu is still gender-neutral as it was traditionally; in Japanese, is 彼 kare considered gender-neutral or is it exclusively masculine now?

    In Korean, 그는 남자였던 것 같다 “he/she was probably male” would be perfectly mundane because 그 geu is gender-neutral. Writing 그녀는 남자였던 것 같다 “she was probably male” would be really weird because 그녀 geunyeo is an artificial pronoun used specifically to draw attention to her gender, so it is more like saying “that woman was probably male”. One would probably understand it as “she probably used to be male”.

    Korean conversational sentences also have many tones depending on sentence-final expressions, but it is much more difficult to mark them as male or female. I can’t really think of an utterance in Korean that would be considered ‘unfeminine’ due to tone alone such that readers would assume it is ‘him’ automatically (even if I consider the old-style speech used in historical dramas). I don’t speak Japanese so I am not in a position to compare, but based on the above I get the impression that Japanese is more gendered than Korean in some ways at least.

    The most obvious way that Korean is gendered is in the use of different familial terms according to whether one is male or female. For males, older brother is 형 hyeong and older sister is 누나 nuna; for females, older brother is 오빠 oppa and older sister is 언니 eonni. Because these terms of address or pronouns are often used affectionately for people who are not biologically related, they are quite common in colloquial Korean. I don’t know how common their occurrence would be in this particular story, though.

  19. Man, this book is even more difficult to translate than I would have guessed!

  20. Eimear Ní Mhéalóid says:

    The translation issue that would give me trouble, if I were to try to translate the book into Irish, is the word “ancillary” itself, which in the book is more often a noun than an adjective. There’s no single word which would have the same effect (nouns and adjectives have different forms).


  1. […] [Edit: 2015-11-09 10:48] Language hat quotes a post on translating Ancillary Justice, an interesting given how gender varies in […]

Speak Your Mind