1

一人 & 二人 are trivial to type with the (Windows) IME keyboard, but I can't get it to do 三人 through to 十人 just from typing the romaji (sanri etc.). It comes up with 三里 by default, with no option to choose the correct Kanji. Surely these are common words that should be straightforward to type? I can do it by typing sannninn but when used before います the word in hiragana is さんり so that feels pretty awkward, especially with that many n's!

Update: The "san-ri" (さんり) reading appears to be just a bug in Google Translate. Annoying because it's a lot less keys to type than さんにん!

5
  • 1
    Hmm? You mean, 「三人います」 is read as 「さんりいます」?
    – Chocolate Mod
    Dec 17, 2018 at 22:53
  • It does look like this is actually just a bug with Google Translate, and at least one other automated Japanese transliteration/speech synthesis tool I was using. I was pretty surprised when I saw it but given my level of Japanese was assuming it was correct!
    – Dylan Nicholson
    Dec 18, 2018 at 3:36
  • 1
    Machine translation is often actively harmful to language learning because of issues like these and countless others. It seems to most beginners like it's a great idea, but it's not and I highly recommend you avoid it.
    – user1478
    Dec 18, 2018 at 13:08
  • Understood, though of course humans can err too - and in this case it was machine translation (via the IME engine) that led me to realise that there was an issue! I've found generally if the IME is not finding the expected kanji for something you've typed it's a reasonably good chance you've typed something wrong.
    – Dylan Nicholson
    Dec 18, 2018 at 19:37
  • 1
    Input methods failing to convert is actually a useful signal, as you’ve noticed. You can get false positives, but it’s useful to double-check when you can’t convert something. It doesn’t fall under the umbrella of machine translation, however. Although Google Translate works as an input method, I was referring to the actual translation from one language to another. That’s the dangerous part that can slow down the learning process.
    – user1478
    Dec 18, 2018 at 20:00

1 Answer 1

Reset to default
4

人 is exceptional in that as a counter, it is read り for 1 and 2, but にん for 3 and higher. So さんにん is the only correct reading and your IME cannot be expected to convert さんり to 三人...

5
  • Hmm interesting, then google translate and various online Japanese teaching tools seem to be wrong...
    – Dylan Nicholson
    Dec 18, 2018 at 2:57
  • 2
    @DylanNicholson That is possible indeed... I cannot recommend using Google Translate for learning Japanese — I don't know about the "various online Japanese teaching tools" you use, but you can take a look at our (community-maintained) list of resources for learning Japanese
    – Earthliŋ Mod
    Dec 18, 2018 at 3:38
  • @Andrew T. I was able to reproduce it by translating "three people" which correctly gives "3人" but with transliteration "3-Ri"...
    – Earthliŋ Mod
    Dec 18, 2018 at 7:39
  • Yeah, I just noticed. Apparently, the bug in Google Translate is that an Arabic number followed by 人 is read as ri (3人 = 3-Ri), while number in kanji is read properly... (三人 = san'nin)...
    – Andrew T.
    Dec 18, 2018 at 7:40
  • Not always though, only in certain sentences, esp. of the form "xが三人います" (or いる).
    – Dylan Nicholson
    Dec 18, 2018 at 19:41

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .