The purpose of this page is to provide my own opinions on several issues that come up very frequently on the forums. I am currently a graduate student at The Ohio State University, where I teach Japanese and study various things, including pedagogy (the study of how to teach).
I am certainly not saying that I'm right and other people are wrong -- the primary reason for these topics is to present the other side of issues that are often very one-sided on the forums, and to give people something to think about. The page will be organized as a series of questions, although it's not really a FAQ as such.
At the moment, the page is still in the process of being constructed, so some of the answers are less complete than they should be.
Why do you disagree with learning hiragana first?
To be clear, what I specifically disagree with is the idea that you *must* learn hiragana first, before doing anything else. My preferred method is to use a primarily audio-based approach, with romaji used only for reference, for at least the first few weeks. If you are primarily interested in the spoken language, you can use it for longer -- the program I teach in does not introduce katakana until 7 weeks into the program, and it's not until about week 15 that actual sentences written in hiragana appear.
My primary objections to the hiragana-first method are: 1. Learning hiragana is easier if the learner has something concrete (i.e. words, sentences) to link the hiragana to. 2. Sentences written in all hiragana are only marginally more authentic than sentences written in romaji. 3. It takes a long time to learn enough kana and kanji to be able to read any authentic Japanese, but basic conversation can be put to use immediately.
Romaji leads to bad pronunciation, though.
Using *any* writing system leads to bad pronunciation if it is not accompanied by audio practice. If you learn hiragana from a chart that gives romaji equivalents for each kana, that's no better than using romaji directly as far as pronunciation is concerned. On the other hand, if you know the sounds from audio practice, whether you use romaji or kana will make no difference in your pronunciation.
But doesn't using romaji hamper your ability to learn hiragana and kanji later?
Only if you let it become a crutch. It's not as if using romaji for a while will build some permanent barrier in your mind that prevents you from learning kana and kanji.
The earlier you learn hiragana, the more practice you get with it.
While this is technically true, it's not as relevant as people usually think. It's not as if "learning hiragana" takes a fixed amount of time that never changes no matter when you start learning it. Also, a delay of a few weeks will not make any difference in the long run of your Japanese studies.
Romaji isn't even Japanese, so why should it be used at all?
Romaji is Japanese -- the roman alphabet is officially recognized by the Japanese government as a part of the Japanese writing system, and there is an official romanization system (Kunrei-Shiki). This should put to rest any claims that romaji is "not Japanese" (or "English" as people sometimes claim).
Now, of course romaji is not what is usually used to represent Japanese, although it is used in some situations -- probably the most common modern use of romaji is on computer systems that (for whatever reason) do not support the display or entering of kana and kanji. Learning to deal with romanized Japanese, then, is not a completely useless skill.
But I really like kana and I want to learn it!
Go ahead. Personal motivation is a big factor in language learning, and if you are studying on your own, go ahead and focus on what you like. On the other hand, if you find hiragana frustrating and would rather focus on something else, don't feel like you *must* master hiragana because that's what some textbooks have you do.
Why should someone learn katakana before hiragana?
Why shouldn't we say that the subject of a sentence is always marked by が (ga)?
What's wrong with Heisig's books?
Why do you say that kanji are not necessary to represent Japanese? What about all the homonyms?
This issue is well covered in two books: 'The Chinese Language: Fact and Fantasy' by John DeFrancis, and 'Ideogram' by James Unger. These books explain this issue in detail, much better than I could do on this page.
To briefly address three points that often come up:
One common claim is that a Japanese friend said they couldn't read romaji/all-kana (or couldn't read it quickly), so that proves you need kanji. The problem with this is that people can read whatever they are used to the quickest. By applying this same logic to English, you can prove that the current spelling system is required to represent the language. (I've heard anecdotes that if native speakers work with romaji for a little while they don't have any problems with it.)
Along the same lines, learners often find that they have trouble with kana/romaji writing, and so they assume kanji must be necessary. In addition to the previous paragraph, learners often find kanji helpful because they do not have the linguistic knowledge necessary to determine the meaning of a word from context, and the kanji can help them figure out which word it's supposed to be. Native speakers do not have this problem. (This problem is particularly acute for learners trying to read/decode something far above their level of Japanese.)
Another claim is that there are so many homonyms in the language that you need kanji to disambiguate them. However, context will disambiguate almost every case. An important thing to remember is that if you claim kanji are necessary to read Japanese, then you are effectively saying that Japanese cannot be read aloud. Also, it's not enough to construct one contextless sentence containing a homonym -- the issue has to be more pervasive than that if kanji are really necessary.
I make a common challenge to people on this subject -- go to any major newspaper website (i.e. Asahi) and find five sentences that cannot be understood if read aloud to a native speaker (in the context of the whole article). If the homonym issue is so important, it should be very easy to do this.
(Finally, please note that "kanji are not necessary to represent Japanese" is not the same thing as "Japan should get rid of kanji" -- there are social and cultural issues that make this difficult.)