Making the mental switch from romaji to hiragana is vitally important.
This may seem like advice directed at beginners, and it is important for them. However, what I have to say is important for serious immersion Japanese learners at all levels, so I would ask even more advanced learners to stay with me and read this article.
Some people have suggested that it would be all right to learn Japanese, at least at first, using only romaji. Most serious Japanese learners disparage the use of romaji but often without explaining exactly why. I am going to start by explaining why it is so important to make the mind-switch from romaji to hiragana and actually think in kana.
When thinking about the meanings of words we should, and as we advance inevitably will, start thinking in kanji. However, when we think of the sounds of words we need to think in kana, because kana are precisely adapted to expressing the Japanese sounds in the way Japanese people perceive them.
Many more advanced learners, even after they have become fluent readers of Japanese, do not fully break the mental link between Japanese sounds and romaji. They have not fully switched from romaji to hiragana in their minds. We are going to talk about why this happens and what can be done about it. But first of all, let’s look at why it matters.
Why we need to mind-switch from romaji to hiragana
What is wrong with romaji in the first place?
Very simply the fact that it does not accurately represent the sounds of Japanese. If we continue to think that あ=a, し=shi, ふ=fu (or hu, depending which system of romaji you use), we will have a fundamental misconception about Japanese sounds.
Thinking in kana will not automatically teach good Japanese pronunciation, but thinking in romaji will make it much harder.
There are different systems of romaji transliteration and all of them have faults. The reason there are several is that it is a trade-off between one set of faults or another. The Hepburn system (which is currently the most usual in the West) is not the one commonly used in Japan. In some ways it is a good system and in others it perpetuates some very wrong ideas about the kana structure.
When you think about it, it is actually not possible for romaji to represent Japanese sounds accurately because romaji itself does not have fixed sound-values.
If a North American speaker pronounces the word ほとんど as if it were a word spelled “hotondo”, what comes out is something more like はたあんど. In traditional standard British English, the Japanese あ sound is closer to the short “u” than to most of the various sounds that “a” makes in English. And so on.
So this is the first reason for switching from romaji to hiragana. We need as early as possible to start associating Japanese sounds with Japanese characters and cutting out the intermediary of Roman characters, which necessarily misrepresent the sounds as ones we are familiar with in English.
The second reason is that romaji, and especially the usually-used Hepburn system, misrepresents how Japanese people think about kana and even gives a false idea of certain points of grammar.
For example, on the Wikipedia article on Japanese verb conjugation we are told that:
The eba provisional conditional form is characterized by the final -u becoming -eba for all verbs (with the semi-exception of -tsu verbs becoming -teba).
But つ-verbs becoming てば is not any kind of exception. It is 100% regular, and the apparent “semi-exception” only exists in Hepburn Romaji.
In a variety of ways, thinking in romaji causes us to see Japanese in ways that are a) different from how Japanese people see it, and b) an impediment to understanding how the language actually works.
In the Wikipedia example above, instead of thinking in terms of “the final u becoming -eba” (a purely romaji-based concept) we should be saying that the final う-row kana becomes the equivalent え-row kana plus ば. And then there are no exceptions, semi or otherwise, and we are seeing the language the way it actually is, instead of through romaji glasses.
Romaji to hiragana: still a problem for more advanced learners?
“But I know hiragana inside out. I read Japanese books all the time,” the more advanced learner may say.
Yes, I do too. But I have caught myself, and other more advanced learners, making little errors that indicate that we are still thinking somewhat in romaji. I have puzzled over why this may be. After all, we left romaji behind a long time ago.
Well, in a way we might not have. First we learned the sounds of Japanese as romaji equivalents. But perhaps more importantly, we type Japanese in romaji every day.
Now this may be less important if you hand-write Japanese a lot (perhaps in school). But many of us hardly hand-write at all in this electronic age, and we have argued that it may not be necessary.
However, if we mostly type Japanese on a keyboard via romaji, without even being aware of it, we are continually maintaining and strengthening the romaji to hiragana link.
Especially if we touch-type, this is largely subconscious. But we may actually have a much stronger mental connection to how the kana are made up from romaji than how they are made up in themselves.
When you try inputting kana directly (say putting your name in a DS game) you can find yourself wondering momentarily where the “cha” character is before remembering that you want ち plus ゃ. Of course you know ちゃ instantly when you see it written, but it isn’t the way you are used to inputting it.
This does have an effect on how you think about Japanese. I suspect the author of the Wikipedia article cited above was advanced in Japanese, but she was still thinking in romaji-linked patterns.
As full-immersion learners, we want to complete the romaji to hiragana mind-switch and break the mental romaji link.
Because of the strong subliminal influence of the keyboard in this process, I have been experimenting with typing directly in kana on my computer. This is possible with a Western keyboard (I will show you how to make it easy), and I have found it much easier than I imagined to re-learn touch-typing with a completely different keyboard layout.
Whether you want to do it or not is another matter. Most Japanese people type kana via romaji, but they are thinking in kana to begin with, so it doesn’t pose a problem for their Japanese! (It might adversely affect their relation to kana for English, but then they primarily use romaji as a minor part of Japanese).