Katakana: The usage of ー

I have studied katakana for around 3-4 months :sob: , but I still do not know when to use ー in a word, for example tomato is to-ma-to and coffee is co-ffee but when in katakana they are different in length、トマト vs コーヒー, is there any rule to remember which we help me determine when to use ー in a katakana word.


There isn’t really a rule that can be applied. The differences are just quirks of the importing process. You just have to remember “this word is spelled with a ー” or “this one is spelled without a ー”. The same goes for other katakana loanword quirks. The fact that different words were imported at different times, or sometimes through languages other than English (despite similar words existing in English) contributes to the inconsistency. For instance, コーヒー was imported from Dutch, not English.

You basically just have to treat them as any other word in Japanese and learn them for what they are, and not base it on what you know in other languages.

It’s not as bad as English spelling though.


It is probably best when you learn loanwords just like normal words. Japanese is special with how they integrate words into their own culture and like said in the post before, there is no inherent rule to it that you can just learn.
Just treat コーヒー the same as 愛情 while learning, as loanwords are used a lot and often times (especially in the beginning) when you just hear one, you will not necessarily realize its a loanwords anyway.


Yeah the “trick” is to familiarize yourself with the Japanese pronunciation, and then transcribe that. It’s very tricky at first but it does get easier with practice.

It’s even trickier as a non-native English speaker because most katakana loanwords are from English and you have to imagine how a Japanese person would imagine an American person pronouncing the word and then transcribe that.

For instance a very common use for ー is effectively to transcribe English ‘r’ as in アーチ (arch), アナウンサー (announcer), マーク (mark) etc… It makes sense but it took my French brain a while to get used to it because I would intuitively have guessed “アルチュ”, "アノウンセル”, “マルク” or something like that.

I really struggled heavily with katakana at first, but fortunately it’s easy to get a lot of exposure to it. One thing that helped me early on was this Anki deck: https://ankiweb.net/shared/info/2015522924


That’s not transcribing American English, it’s transcribing a southern UK non-rhotic English. When I say “arch” or “mark” I don’t sound the ‘r’ at all, so the katakana transcription makes perfect sense to me :slight_smile:

The Dictionary of Intermediate Japanese Grammar has an appendix giving “rules” for katakana transcription, which might be of interest, but there are enough exceptions and words that turn out to have been borrowed from French or Dutch or whatever that in practice I agree with the other advice here and suggest treating the spellings as words to be learned in the same way as native Japanese origin words. This has also the advantage of reinforcing the correct pronunciation of them – トマト isn’t pronounced like English “tomato” even if it’s spelled the same in romaji, and コーヒー and “coffee” are even further apart…


I actually hesitated to make that claim because that did sound a bit bri’ish to me (as in, I feel like I sound more british when I pronounce “mark” as “ma’k”), but then I thought that Japanese people must have had a lot more exposure to American English since WW2? That’s why I assumed that it would be their reference for pronouncing the words, but maybe it’s more complicated that that.


My guess is that the conventions for English loanwords were set in the late nineteenth and early twentieth century, when British English was still the prestige dialect that was generally taught to second language students.


I found a book Loanwords in Japanese, which says:

As noted by Quackenbush (1974: 61), when many dictionary traditions were being formulated in the late 19th century, English language scholars were heavily influenced by foreign advisors (oyatoi gaikokujin: cf. §2.4.4) from Britain. It is probably for this reason that adaptation patterns for, amongst others, the vowel
found in the word lot (89–90), as well as the English rhotic vowels (111–115), mirror English English and not American English pronunciation. These dictionary traditions based on English English still hold, despite American English being taught exclusively in secondary and overwhelmingly in tertiary education.


I just found this on the Tofugu site this morning.


Fascinating, thank you for looking that up!

1 Like