pay little or no attention to memorizing characters (I would have been content with actively mastering 25 or so very high frequency characters and passively recognizing at most a hundred or so high frequency characters during the first year)
focus on pronunciation, vocabulary, grammar, particles, morphology, syntax, idioms, patterns, constructions, sentence structure, rhythm, prosody, and so forth — real language, not the script
> pay little or no attention to memorizing characters
I'm not usually one to speak bluntly, but, at least for Japanese, speaking from a lot of experience, this is flat out the wrong approach.
Older books like the romanzied "Japanese for Busy People" took this approach. I started with this book, it felt easy at the time, but it was immensely impractical - you don't build up enough of a base to learn from any other material. It's completely useless. After over a year, in a formal setting, I ultimately decided to just completely start over.
Currently Japanese courses mostly just focus on Hiragana and Katakana, since material for children usually has pronunciation guides over the kanji anyway. This is workable, but it just gives you no leverage where you could have some - you end up memorizing a thousand indiviaul words, some of which maybe kind of sound similar in parts, when you could instead learn a dozen charcters, and know that they share root words. Phoenetic approaches work for children who are in a situation where they can memorize from immersion with native speakers, and have additional clues like context and intonation to emphasize root words, to pick up new vocabulary.
When I started studying Chinese, we instead went with a "radicals-first" approach, and after just a short time of that, my Japanese vocabulary exploded - I could suddenly pick out all the seafood on a menu, and I started asking people mid-conversation "oh, does that use these characters?", and I could understand what it meant without them having to struggle to translate to English: Heart+Electric+Map - it's probably an electrocardiogram, especially if we're talking about hospitals. If you haven't learned the kanji, you don't even know if you can make that association from the sounds, much less how to ask that question.
That's how you get to fluency, by building up a foundation that lets you learn without active study. You can't get there by avoiding the basis of word composition in the language. Apologies if my frustration with characters-last approaches came through a little strong - they wasted a lot of my time.
Mair's point is that "building up a foundation" means learning the language not the script.
> characters-last approaches
I think it's more accurate to call what Mair's advocating a "language-first-characters-later-if-necessary" approach.
Here's what Mair said about Chinese classes:
It’s a tragedy that so many young Americans spend years stuffing their heads with hundreds of Chinese characters, gaining no usable proficiency, and then forgetting them all by the time they’re 25. If the Chinese would wake up and permit pinyin to function as part of a genuine digraphia, then I would say it might make sense for maybe 2 percent of the population to learn up to third-year level of Mandarin—strictly romanized, mind you. But there are exceedingly few teachers who are enlightened enough to teach it that way.
They're not reasonably separable. Memorizing romaji, or even strings of hiragana, removes the associations between root-words that are related, like those radicals that let you identify menu items, and removes the visual aid for identifying root words in context. Memorizing associations between arbitrary groups of sounds without the radicals is actually harder, it just sounds easier to those intimidated by the idea of learning characters.
Given, much to Mair's displeasure, that actual native media is written with kanji, they're immediately necessary, and not studying kanji effectively bars students from any learning outside the classroom until they have - which is completely counter to giving students either the context necessary for immersion or the reinforcement for traditional study to work.
Maybe it's merited for Chinese classes, with its larger character set, but Mair's approach just sacrifices students' ability to learn on their own in favor of lowering attrition rates of intro-level classes. It's actively harmful to students aiming for fluency to build up the characters as this insurmountable mountain in the distance before they can actually use their knowledge, rather than the useful compositional blocks and visual aids that they are.
(Notice the native narrator has no trouble reading the kana-only writing.)
It's true that much of current native media still uses kanji, but they aren't "immediately necessary". There are books/manga with furigana and software tools like Rikaichan/kun for online articles.
People should be able to learn whatever they want outside of the classroom, including kanji (being able to recognise commonly used characters certainly helps in understanding native media). It's just that grammar, vocabulary, etc (real language) are much more important than kanji for beginners.
Sure, Korean is phonetic, and so learning it phonetically makes sense, since you won't be forced to re-learn it again to be able to use your knowledge. Speaking from experience studying Korean, though, attaining a useable vocabulary through rote memorization is dramatically harder than if you're familiar with the 한자/漢字 words are derived from, for exactly that reason - even though hanja's not used directly so much in korean, we started studying it in the third semester of Korean class, and it was helpful, though I'd agree maybe a bit excessive there.
Games that use a strict hiragana character set are readable by natives because they're already fluent in speaking; they're unnecessarily difficult for non-native speakers, particularly since they tend to only have the bare minimum of separators between words. After living in 4 prefectures, working for a small Japanese company, and dating a native speaker for roughly a decade, I still find those games hard to read - it's just hard to identify the shapes of words, like reading English with no spaces; you have no choice but to guess where the boundaries are and sound everything out syllable by syllable, because everything's written differently from how you'd see it anywhere else.
If some people want to put off kanji or memorize phrases from a travel guide, that's their prerogative, but not teaching the compositional aspects of kanji, and leaving students to fend for themselves after selling the idea that it's all rote memorization (and seemingly to push this implausible ethnocentric dream of the language one day being exclusively romanized) is just a disservice to the students, as it was to me.
> strict hiragana character set are readable by natives because they're already fluent in speaking
This is an argument for spending more time trying to become fluent in speaking (by learning vocab/grammar/etc) than spending time learning how to draw a bunch of characters.
The use of spaces would be necessary if Japan were to switch to kana, and that's one of the 3 things Kanamoji-kai recommends as well:
Without spaces kana-only writing would be hard to read for most native speakers.
I honestly don't see any issue in not teaching kanji in Japanese classes, any more than not teaching cursive writing or Latin in English classes. They are just not essential. (Please do not start a lecture on how Latin helps you understand English.)
> ethnocentric dream of the languge one day being romanized
I'm not saying Japan should switch to romaji. I think in Japan's case kana would be much more realistic.
Also, I'm not advocating abolishing kanji either (people should be able to write kanji all day everyday if that's what they want to do).
I just want the government to (a) stop teaching kanji in public schools, and (b) mandate the use of kana (or some other phonetic writing system) in government documents.
I don't see how it's "ethnocentric". For most non-native speakers it's much easier to learn this way. Only non-native Japanese learners I've met who found it "easier to read" kana-kanji writing were people who already knew kanji e.g. Chinese people who were learning Japanese & were not fluent in speaking. The true of the matter is they were just familiar with the characters and they often didn't even know how to pronounce the words.
I have to agree with T-R. Japanese is much easier to read with Kanji. If you run into a kanji you don't know, that's too bad, but I run into words I don't know in English, too. You just look it up. English spelling is just as complicated, in my opinion.
Also, "stop teaching kanji in public schools" is pretty much the same as abolishing kanji.
You read my conversation with T-R and your takeaway is I think teaching kanji to beginners is a bad idea because "I run into a kanji I don't know"? Please.
> English spelling is just as complicated, in my opinion
That's a separate issue, but I agree with you to some extent. Here's a Feynman quote (which I agree with 100%):
If the professors of English will complain to me that the students who come to the universities, after all those years of study, still cannot spell "friend," I say to them that something's the matter with the way you spell friend.
> "stop teaching kanji in public schools" is pretty much the same as abolishing kanji
I don't think so. There are many things that aren't taught in public schools but taught in private classes (calligraphy/ikebana/yoga/etc).
It's good that people who care about these cultures enjoy them in their own time and other people are left alone and not being forced to learn something they think has a low ROI.
If kanji is such a good idea as you claim, it will survive and thrive without government coercion, don't you think?
This book is not for people who want to stay a beginner for a year, this is for people who want to learn the language at the same pace as a typical university student: you get a few weeks to learn hiragana and katakana, then the romaji is taken away. After a year, you'll already know anywhere between 1000 common kanji and the full jouyou set.
I wrote this book based on that need. So step 1: learn hiragana, because after the introductory chapter you're not getting romanised Japanese anymore. However, because it's not a writing course, all kanji have furigana so that you can read every sentence. This book is not going to calmly and gently lead you through the Japanese grammar, it's meant as a proper "you need this book for a year at most, after that you should already know everything that we cover". Perhaps not to the degree that you can duplicate the book, or in a way that makes you remember all the technical terms, but all the constructions and how they use them: this is the pace you run through learning Japanese at first year university level.
Mair's argument is that it is a waste of effort to spend time on learning the characters, in the first stages of learning the language. They are more easily learnt systematically somewhat later on, when you know the corresponding words to which you can anchor them. Note that the people advocating for this are university language professors and teachers, who are claiming they have data backing up the claim.
As an adult student of the language, my experience is that I was glad to start without the crutch of romaji, for a simple practical reason - Japanese quickly becomes tedious to read Romanized. Kanji + kana are much more compact, and things like the conjugation of a verb and doubled letters which are common in the language are immediately obvious - because the writing system is designed to accommodate them.
Learning 2500 kanji from the beginning is a major cognitive task that shouldn't be a primary focus, nor is it in native Japanese language pedagogy. But learning the kana is simple, again I learned them as an adult and had them down cold in a few weeks of active classwork. My kanji retention is terrible, but learning materials almost always give the reading of kanji using furigana, and there are useful tools for annotating web pages, etc. with furi readings.
> Kanji + kana are much more compact, and things like the conjugation of a verb and doubled letters which are common in the language are immediately obvious - because the writing system is designed to accommodate them.
My grasp of Japanese is virtually non-existent, but I noticed this as well. Taking an example from the post, I find something like しっけ easier to parse than the romaji shikke. The latter form just feels clumsy for reasons I can't really articulate.
It also makes the pronunciation a little easier at times. If I read "desu", I read it with a fairly pronounced "u" sound. Reading です, on the other hand, I know without really thinking about it that the う sound is typically very short.
I think it's because most hiragana/katakana is in the form "consonant-vowel" meaning that almost all words are made up of collections of 2 letter sounds/syllables. This gives words a certain cadence which is easily repeated. When you convert that to romaji you begin to parse the words as though they were english and the discrete "consonant-vowel" sounds get broken down and become harder to see/read.
Really though, you can learn hiragana in a single sitting using mnemonic trickery and you can commit it to memory using a spaced repetition or flashcard system in a few days. Not learning hiragana and katakana seems much more of a hindrance than "wasting time" by learning it.
Right, my counter to that is that's great if you want to leisurely learn Japanese, but also not very relevant for the type of course this book supports. These typically start you on kanji pretty much in week 3, in conjunction with the word lists that support the kanji you're learning every week (and constantly use throughout that week and subsequent ones), specifically because of the argument raised: learning kanji without the words they're in, and even if that, without using those words, is ridiculous.
The argument above applies specifically to university courses, and was made by someone who is familiar with teaching such courses. So I don't think you can just discount it with handwaving, especially as there are claims about better success rates.
I'm not discounting anything, but different people want, or need, different courses.
My book is geared specifically towards people who are to enthusiastically committed to learning Japanese, and want more information about the language they're paying money for to learn, with some assumptions about what that means: they, like me, and like many of my fellow students at the time, but not like everyone, want to learn Japanese properly, without hand holding because some grammar or some orthography is deemed too hard. We'll see whether that's the case as we learn, not as an assumption baked into the course.
Of course that method doesn't work for everyone, but for me, and for many others, but again not for everyone, getting kanji early was exciting, motivating, and awesome. I would have been angry at a promise of a better approach without kanji for half a year as a paying student.
So that's the book you get here, too: I am not making assumptions on how slowly you want material, this is a grammar for supplementing your course work, or deep diving if that's what you want to do. And if the Japanese used in the book is too much, too soon: it's a free book, find one you feel fits you better. But we're better off with a spectrum of works: rather than every book following the best methodology, we need all those methodologies, and then people who go hard, fast, can pick the hard, fast books, and people who don't can pick the slower, "better success rate" approaches and do well with those instead.
Pomax, I met you in an IRC channel years ago (2005 I believe), used a very early rendition of your book to teach myself Japanese (it was just a Japanese grammar dictionary at the time, written in DocBook), and promptly ended up become one of the many teenage anime fansubbers of the era. I made so many good memories fansubbing anime and made so many friends who I still cherish today, that I thank you for the amazing introduction to this language that I still relish to this day.
I'm sorry but I think the book is getting priorities wrong when the learner's goal is to learn practical Japanese.
Also I think some example sentences in the book would sound awkward to native speakers. (e.g. "今日自転車で帰り中で転んでしまいました", "もっと安いのありましょうか", "お箸の使い方に慣れてない人に難しいです" (this may be a typo), "きれいじゃないで、明るい部屋です", etc.)
I think this girl gives better advice on how to learn Japanese:
Then it should please you to know that is not the goal of the book. This is a grammar (a type of educational book), not a "Learn Japanese" course book, and is to supplement courses (institutional or personal) in Japanese, to act as reference and deep-dive material =)
As for odd sounding sentences: nice! let me know about them on the github repo. Link's in the side nav as the github icon, but the direct link is https://github.com/Pomax/nrGrammar/issues
Even if research has shown advantages to waiting to teach the writing system, this approach has become extremely unpopular, to say the least. There are only one or two places still using the textbook "Japanese: The Spoken Language," and all other modern textbooks teach the writing system from the beginning.
Even textbooks that used to use romanized Japanese for fairly logical pragmatic reasons, like "Japanese for Busy People," which is aimed specifically at busy adults who are trying to learn the spoken language in the least amount of time/effort, have been rewritten to eliminate romanized Japanese! Many people studying Japanese go so far as to consider the presence of romanized Japanese, even in combination with normal written Japanese, a black mark against a book, a viewpoint which I think is a bit extreme.
Additionally, even if there is a benefit to early spoken fluency, it's going to really annoy students when after two years they suddenly have to start learning the writing system from the very beginning. With this approach, students wouldn't be able to use any modern instructional or reference materials for 2-3 years. Basically, they wouldn't be able to study on their own at all.
Also, from the standpoint of practical language instruction, since many students drop out of Japanese classes the second they're asked to learn hiragana (much less katakana and kanji), it is much more efficient for all parties involved to have this occur earlier in the process.
Many people studying Japanese as a foreign language actually also 1) find it very discouraging to discover that they can't read words they know because they haven't been taught the characters, and 2) find it very difficult to parse romanized Japanese since it is harder to tell where words end and distinguish particles/verb endings from nouns and verb stems.
> The problem was that all of my language teachers insisted that I memorize hundreds of characters right from the very start.
For every person who is mad that that they were made to learn kanji from the beginning, there is another who thinks their Japanese ability was permanently impaired by not being taught kanji early enough (I have met plenty of the latter). Regardless of the method, there are people who are going to object to it.
You will never remember, or "passively recognize" any characters, (especially low-frequency ones) unless you specifically focus on remembering them. Also, at least 75% of words meanings become trivial once you learn what their corresponding characters mean. Or you can spend all your time getting all those similarly-sounding kanji compounds mixed up in your head because you can only write them in hiragana.
He's mostly talking about Kanji from what I can tell. Learning Hiragana is probably a big different. It takes about a weekend and helps with pronunciation. Everybody is different, but I found it useful to at least know about...
I agree, I'd recommend learners to learn kana because you can learn it in a few weeks (as opposed to many years for kanji, and even after that many years, people - including natives - don't remember them very well).
For example Prof Victor Mair is against teaching Chinese characters to beginners:
http://languagelog.ldc.upenn.edu/nll/?p=10554
http://languagelog.ldc.upenn.edu/nll/?p=11580
pay little or no attention to memorizing characters (I would have been content with actively mastering 25 or so very high frequency characters and passively recognizing at most a hundred or so high frequency characters during the first year)
focus on pronunciation, vocabulary, grammar, particles, morphology, syntax, idioms, patterns, constructions, sentence structure, rhythm, prosody, and so forth — real language, not the script