When they are writing material in English, I sometimes see native speakers of Japanese misspell English words that were derived from Japanese.

For example, I've seen "tunami" written instead of "tsunami", and "ninjya" instead of "ninja". I've also seen "geisya" instead of "geisha", and "Kyusyu" instead of "Kyushu".

My suspicion is that they know how these words are spelt and pronounced in Japanese, and that they convert it into the latin alphabet in a way that is different to how the word is spelt in English.

For example, with "tsunami", they've taken the first syllable, , and rather than converting it to "tsu", they converted it to "tu". Possibly because in certain ways of writing Japanese using the latin alphabet (romaji), the way that つ would be written is "tu".

Are there general rules about how English words that are derived from Japanese words are spelt? For example, are they generally spelt in a way similar to Hepburn romanization, except often without the macrons?

Even if the rules don't apply 100% of the time, rules that apply 90% of the time, with just a few exceptions may be preferable to some than memorizing all such words by rote.

An exception here is for wasei-eigo words such as "salaryman". When converting made up English into English, the English words that were used in the wasei-eigo are used ("salary" + "man", rather than "sararyman")

Any possible answers that a person who is not a linguist can easily understand?

Meta: I previously asked this on English Language Learners, but it was closed there (currently with one re-open vote), and in this comment on Japanese Language & Usage (not my fault!), it was suggested I ask on English Language & Usage.

  • 3
    After looking at your other questions, and rereading this one, it is still unclear to me what you want. The ostensible answer is 'There exist transliteration methods (e.g. Hepburn)' and that's enough. But You already know that so you must be asking something else. Are you asking why Japanese make mistakes in using some of these systems? That's because people aren't perfect. Are you asking for general rules that explain these mistakes? Well, that's a tall order requiring a list of mistakes which could be many. Or are you asking something else? What exactly are you asking?
    – Mitch
    Mar 27 '13 at 3:31
  • 3
    Hmm... that would seem to require someone proficient in Japanese transliteration. Also, to justify it, one would need to get a list of Japanese loan words and see how many used whichever transliteration method.
    – Mitch
    Mar 27 '13 at 3:56
  • 3
    Research is hard, let's go closevoting! Mar 27 '13 at 4:19
  • 7
    In Japan, two ways of phonemicizing Japanese characters / words into English Alphabet have stood side by side. One is ‘Hepburn method’ invented by James Curtis Hepburn (1815-1911), medical doctor and missionary who came to Japan in 1859 and influenced on westernization of Japan greatly. The other is the ‘Kunrei Romanization method’ proclaimed and promoted by the Ministry of Education in 1954. Major differences between two methods are Hepburn method (called Hebon style in Japan) transcribes Japanese sound し、ち、つ、じゃ、じゅ as ‘shi, chi, tsu, ja,’ while Kunrei method spell them as ‘si, ti, tu, jya.’
    – Yoichi Oishi
    Mar 29 '13 at 0:41
  • 5
    Cont. while Kunrei method spell them as ‘si, ti, tu, jya.’ After the government’s proclamation of Kunrei method, kunrei Romanization had been predominantly used. However, after the World war II, Hepburn method seems to have been used more often than Kunrei method, which became obsolete these days. The confusion of ‘tsunami’ and ‘tunami’ and ‘ninja’ and ‘ninjya’ as you point out all comes from the coexistence of two Romanization method in Japan.
    – Yoichi Oishi
    Mar 29 '13 at 0:43