I was specifically thinking of whether the voiceless velar fricative [x] as in Polish could be represented in Japanese, but [x] would be the same or very similar in every language which contains it, right?

Anyway, just wondering. Off the top of my head I can't think of a way, nor find one by perusing the IPA for Japanese article or online books.

  • What do you mean by "translate"? In the normal meaning of "translate", hito translates person and vice versa -- they mean the same thing. Japanese does not have [x], and the Polish velar fricative does not mean anything (sounds do not mean things, they are elements of words, which mean things).
    – user6726
    Commented Aug 30, 2016 at 5:08
  • What about ⟨x⟩ within a name? Commented Aug 30, 2016 at 5:12
  • AFAIK Japanese does not have [x] in names. Anyhow, what would it mean to translate x in a name?
    – user6726
    Commented Aug 30, 2016 at 5:25
  • Take, for example, the surname Bach. Commented Aug 30, 2016 at 5:25
  • See my answer for Bach and a bunch of other examples.
    – jogloran
    Commented Aug 30, 2016 at 5:32

1 Answer 1


I will assume that by "translate" you mean which syllables in words loaned by Japanese correspond to [x] in their source language.

The answer is that words containing [x] which come directly from languages with that phoneme are sometimes rendered ッハ hha. The most common examples would be Bach バッハ Bahha and Mach マッハ Mahha.

Other times, it is simply rendered ハ (cf Zakharov Zaharofu, Halacha ハラーハー), or particularly word-finally, フ (cf Lech Wałęsa Refu Vawensa). Note that while ending not in /x/ but /h/, the Arabic name for God is sometimes transcribed as アッラーフ Arrafu.

Words of Greek origin beginning in Greek with χ, however, tend to have that sound represented instead with syllables from the か行 (k- syllables), in line with English.

Another subtlety is that the German orthographic ch which before high vowels is a palatal fricative [ç] and not a velar [x] is almost universally transcribed as ッヒ hhi, as in アルベリッヒ Aruberihhi for "Alberich".

  • I don't think "transcribe" is correct either, since Japanese speakers do not transcribe words, except a handful who are linguists. Rendering a foreign language letter sequence vaguely believed to have some h-like quality is not even phoneme-adaptation, unless done by a Japanese speaker who is actually hearing a Polish speaker utter <ch>.
    – user6726
    Commented Aug 30, 2016 at 5:29
  • 1
    True, "transcribe" is not correct either, as the goal is not fidelity. However I believe this is what the OP wants to know -- how loan words containing [x] in their source language are rendered in the target language. I will edit the answer to skirt the terminology issue.
    – jogloran
    Commented Aug 30, 2016 at 5:32
  • Yes, that is what I would like to know. However, do inform me of why transcribe is incorrect here, in case there is a way of properly rewording my question. Commented Aug 30, 2016 at 5:34
  • The goal of transcription is to, ideally, map one-to-one from phonemes in a source language into some unambiguous representation in a target writing system. For instance, if I wanted to represent American Sign Language in a way that records it faithfully, I might use something like SignWriting, which ideally would allow me to map the source language into a written form and then back again. When a language loans words containing sounds it does not have, the mapping may not necessarily be faithful. Hope this helps.
    – jogloran
    Commented Aug 30, 2016 at 5:37
  • When the Japanese writing system really needs to (properly speaking) transcribe sounds that the standard language doesn't have (for instance when linguists are doing Ainu work or transcribing Japanese dialects), here's an example of what it tends to do.
    – jogloran
    Commented Aug 30, 2016 at 5:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.