Japanese has some sets of characters which look very similar or even identical. Obviously, context is usually more than enough to distinguish which character is intended, but I'm wondering if there are subtle differences which can be used to disambiguate. It's easy enough to compare characters in computer fonts, so my question is really more about handwriting.

  • ー (katakana), 一 (kanji) and — (em dash)

    The first two can be distinguished from the last one in Minchō-type fonts, and usually it is possible to distinguish between all three. In vertical writing the katakana and the em dash are oriented vertically so is easily distinguished from the kanji ichi.

    What about in handwriting or Gothic-type fonts? I suspect they are indistinguishable then, but counterexamples are welcome.

  • ロ (katakana) and 口 (kanji)

    In print, katakana ro is usually printed a little bit smaller than kanji kuchi. In some of the fonts I have the final stroke of katakana ro protrudes to the right, while the penultimate stroke of kanji kuchi protrudes downwards.

  • カ (katakana) and 力 (kanji)

    Like the above, katakana ka is usually printed a little bit smaller than kanji chikara; but I've also noticed that the little hook is subtly different for katakana ka, and that in some fonts katakana ka has a bit of a rightward slant or curvature.

  • ニ (katakana) and 二 (kanji)

  • エ (katakana) and 工 (kanji)
  • タ (katakana) and 夕 (kanji)
  • ト (katakana) and 卜 (kanji)

    Other than size, it seems like the only way to distinguish these pairs in Minchō-type fonts is to rely on the tendency for katakana to have more brush-like strokes than the kanji.

  • ハ (katakana) and 八 (kanji)

    These are usually easy to distinguish in print, even in brush-type fonts, since the last stroke of katakana ha is of a different type from the last stroke of kanji hachi. What about in handwriting?

  • え (hiragana) and 之 (kanji)

    These are easily distinguished in print and good handwriting, but I have trouble making them distinct in my handwriting.

  • へ (hiragana) and ヘ (katakana)

    Is there any difference between these two? They are identical in Hiragino Mincho, for instance...

  • 1
    Em-dash is not a Japanese character. – user458 Sep 3 '11 at 19:34
  • 4
    @liori and are very clearly different to a native speaker, and confusing them is actually a good criterion for identifying a non-native writer. There is a slang イソターネット, which is making fun of a noob who does not know what the internet is. – user458 Sep 3 '11 at 21:51
  • 3
    @sawa Can I take it you mean that ソ and ン can look similar to native speakers too, and it is inability to infer the difference from context that identifies non-native speakers? (Thus leading to a situation where even native speakers can err when the context is an unfamiliar-to-then word like インターネット... if ソ and ン are always visually distinguishable to native speakers, irrespective of context, it's hard to see how the イソターネット gag could arise.) – Matt Sep 3 '11 at 23:09
  • 3
    You are asking about subtle differences in shapes. But because subtle differences are subtle, they differ from handwriting to handwriting, and you cannot rely on them when reading handwritten text. Traditionally, katakana ロ and kanji 口 have different shape in the lower-right corner (katakana ロ has a longer vertical bar and kanji has 口 a longer horizontal bar), but not everyone follows this “rule” even when one writes carefully. – Tsuyoshi Ito Sep 3 '11 at 23:26
  • 1
    Probably I misremembered the difference of the shapes between ロ and 口. In the MS Gothic font, both have the vertical bar longer, and in the MS Mincho font, the shapes are opposite of what I wrote in my previous comment. – Tsuyoshi Ito Sep 4 '11 at 14:11

I noticed that most of the pairs in your list are between kana (mostly katakana) and kanji, with the only exception of へ. In my opinion, in most situations you can infer whether it's the kana or kanji symbol from the surrounding text. I think katakana symbols rarely sits alone on its own because we would find them in a bunch of at least 2 characters in a sequence. So if we find 私の力 and 入り口 then we would know that they are the kanji ちから and くち because otherwise the katakana would be alone by themselves. On the other hand, if we find バカ and クロ then we would know they are the katakana 'ka' and 'ro' for the same reason that the katakana バ and ク cannot be alone by themselves. This is, however, just a theory of mine so I don't have anything to back them up.


In the case of a kanji and kana that are similar as in your examples (ロ、カ、ニ、エ、タ、ト、ハ), the kanji of the pair is slightly bigger, almost as if it's been zoomed in on a little bit. Depending on your font, you can even see this when placed side-by-side. However, I personally think and are different enough that you should easily be able to distinguish them.

As far as the hiragana and katakana , I was always taught that in writing them, the hiragana should be slightly rounded at the top whereas katakana is an exact point. Again, that's what I was taught and don't know if this is a hard rule or not. But as you point out, many fonts show them to be exactly the same, and even in practice I don't always round off the hiragana.


re: カ (katakana) and 力 (kanji)

Well, I can normally distinguish them quite well. Now, that is. When you draw an imaginary basic bottom line to the kana per line, the chikara kanji normally looks a bit as if it is too far below, by fractions of an inch; whilst the katakana ka does not cross that line by any means. Hmm, do I have some back up? Yes I do. I had to translate a Japanese sentence using Google Translate and both characters were in that text: the chikara kanji and the katakana ka. My observations described above were plain to see once I had my browser zoom cranked up quite a bit. So maybe they will help someone out there after all... :)


if you want good handwriting, you need to learn to write with a brush. (and when you buy one, don't make the mistake i made and go for the biggest one. choose one that's as long, small-circumfrence, and thin as possible. 10 dollars, or five pounds, is a good price to pay. there's no need to pay for ink or paper at this stage. black watercolor on white printer paper works well. i store my colors in little jam jars, the kind they have at hotels.)

you can also write with something brushlike, such as a fudepen or a ballpoint pen, which can approximate brushstrokes because it responds to pressure well. if you choose to go with a fudepen, i recommend this one.

as a side note, i have seen some excellent ballpoint pen calligraphy. in china, it became somewhat of a fad after the revolution. and comes under the banner of 硬筆書法.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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