How to learn Chinese and Japanese

Language Log 2014-02-17

Some years ago (in 2008, as a matter of fact), I wrote a post entitled "How to learn to read Chinese". The current post is intended as a followup and supplement to that post.

Earlier this week, I received from Tom Bartlett, who has had a tremendous amount of experience teaching Chinese at some of the best universities in the world, the following paragraphs:

Y. R. Chao, in the Introduction to Mandarin Primer, recommends that students learn all 24 lessons of the book in GR only, to establish a strong basis in the spoken language, then to restudy the book a second time, only then starting to read in characters. At Princeton we did each lesson that way: one week entirely in GR then a second week in characters.

[VHM: GR = Gwoyeu Romatzyh / National Romanization (a type of tonal spelling devised by Y. R. Chao)]

Is Eleanor Jorden’s Beginning Japanese still much used? It took students through 40 lessons entirely in romaji, without a single kana or kanji in the entire two volumes. I did about 30+ lessons before starting to learn kana, and found it very effective for absorbing the spoken structures quickly.

I replied to Tom:

You see how smart and perceptive truly brilliant linguists like Y. R. Chao and Eleanor Jorden were! I completely agree with their approach. Ditto for John DeFrancis (JDF).

If I were the czar or god of Chinese and Japanese language pedagogy, I would not teach students a single Chinese character until they were relatively fluent — about two years. I've always said that we should learn languages the way babies do; they learn to speak long before they learn to write.

Eleanor Jorden (praise be her name!) understood that principle well, and she taught people like Ron Walton (at Penn [deceased now]) and Galal Walker (at Ohio State) who followed her precepts. Others who understand the importance of mastering the spoken language first are Bob Sanders (Auckland, NZ) and Cornelius Kubler (Williams College) — Neil was also trained under Eleanor Jorden (there are others who stress spoken language too; I cannot name all of them).

My dear wife, Li-ching, went as far as she could toward deemphasizing the characters during the early stages of learning Mandarin within the confining curricula of Harvard, Bryn Mawr, and Swarthmore.

Jerry Packard published a relevant study in 1990:

Jerome L. Packard, "Effects of Time Lag in the Introduction of Characters into the Chinese Language Curriculum," The Modern Language Journal, 74.2 (1990), 167-175. (Here is a pdf.)

Jerry found that the time lag of delayed character introduction improved students' ability to discriminate Chinese sounds, and improved their fluency.

Gene Buckley comments:

Thanks for pointing out that article; it’s very interesting that a rather modest time lag has this effect, but I guess it only takes a few weeks to persuade them that Chinese is more than characters! His explanation seems right to me.

I was fortunate that when I wanted to learn a bit of Japanese before traveling there in 2007, a colleague had Jorden’s books on his shelf. It was nice to see someone really pay attention to the fact that it’s a pitch accent language!

Long-term experiments in China have demonstrated the validity of the same principle (see the link at the very beginning of this post, also under "Z.T." here).

Jim Unger notes:

Jorden was even clearer, in her Written Japanese, which she co-authored with Hamako Ito Chaplin, on how it was to be used in conjunction with Beginning Japanese. The basic principal is: never demand that students read/write things they do not already know as language/speech. The writing system is introduced after a suitable foundation in speech is laid; gradually, as progress in both speech and writing continue, the gap is closed so that, in the end, students can start using authentic written materials on a self-study basis to improve their overall linguistic/cultural competence. The main thing is Jorden's rule of not trying to teach students to read/write that which they cannot say/understand already.

BJ and WJ were later superseded by Japanese: The Spoken Language and Japanese: The Written Language, co-authored with Mari Noda, which is very much still being used by strong programs around the country.

The closest thing to these materials for Chinese are DeFrancis's textbooks.

Stephan Stiller observes:

This follows the idea of primacy of speech for all languages (with the exceptions of Literary Sinitic (LS) and sign languages, and the exceptional status of LS has as I understand been partially questioned).

Elaborating on that idea, if speech is (on a cognitive level) primary, the characters are "attached" to already existing words (in a "lousy phonetic script" (JDF) manner) with some weak semantic clues. That characters are "attached" to words that are already stored in phonetic form (i.e.: as phoneme or syllable/"mora" sequences) seems certainly the only way one could possibly master Japanese kanji, as to predict their use from semantic considerations seems impossible (unlike for Chinese, where we can in many cases state "rules" like "的 is pronounced dì when it means 'target'").

Delaying characters will undoubtedly improve speech performance because characters are distracting and take up a lot of time and will be stressful to students without a good visual memory. But for that same reason one can argue that characters should be introduced earlier than (say) 2 years in, because they take longer to absorb.

Therefore I would think that the study one would ultimately want to perform is to monitor over a course of 4-5 years the acquisition of fluency and character knowledge, comparing 2-3 overall strategies:

simultaneous introduction
gradual explicit introduction
a total delay of (say) 2 years after which characters are meant to be learned through extensive reading

I'm thinking that one could split option 3 into 3a and 3b, where 3b would be as above and 3a would be to try to let the student learn passively from early on, which would mean "through extensive reading of simple texts where the introduction of characters is gradual" – akin to what the Sanders-Yao textbooks do.

Given my remarks in the following paragraphs regarding Japanese, I know that many (but how many?) Japanese textbooks limit kanji exposure in their vocabulary lists. I think characters there were delayed beyond what the scope of the jōyō kanji list would require, but an expert will know for sure. In general, it will be instructive to learn from Japanese language educators, who I'm sure have something to say.

Jorden's "Japanese: The Spoken Language" textbook series also caught my attention because it teaches pitch accent.

Almost all textbooks and other instructional materials ignore pitch accent, and many (though not all) dictionaries don't indicate it either. I never understood why.

Pitch accent varies regionally, but so do all aspects of Japanese (and in fact all aspects of any language), and Tokyo serves (as I understand) as the standard for all areas of Japanese, not only pitch accent. There's a reason there's the NHK pitch accent manual.

It's also not that pitch accent is better learned through immersion. It's like saying "we don't indicate stress in our English vocabulary lists because the student will learn it through immersion" or "we don't make distinctions re vowel length (or: the tense-lax vowel pairs) in our phonetic transcriptions for German because the student will learn them through immersion". Or if pitch accent is magically better suited for immersive learning, I really want to see an argument for that.

I think it was Haruhiko Kindaichi who wrote something similar (to try to learn pitch accent explicitly in order to not sound like a foreigner) in his book, but I don't have the reference at hand.

Might have to do with kana not indicating it and instructors not being familiar with it (for that very reason) and consequently language resources not showing it.

Bob Ramsey observes:

Like many in my generation, I, too, used Beginning Japanese for the first two years of instruction, and yes, we also delayed the study of writing.

Very few places these days use BJ, and most places that still follow Jorden use her later textbook series, "Japanese the Spoken Language", in three volumes that she co-authored with Mari Noda, now at Ohio State. As you probably know, that textbook series is used over the course of three years of pretty intensive instruction, and it's supposed to cover the "basic" structure of Japanese before more advanced students move along and branch into specialized or particularized Japanese instruction.

But some years after the completion of JSL, Jorden and Noda put out, in various field editions, a companion series, "Japanese the Written Language", which begins introducing written prose based upon vocabulary and patterns already covered in JSL. Instruction using that companion series (which is still not entirely complete) usually begins, at least at most places, a few weeks after starting JSL with katakana, followed a couple of weeks after that by hiragana and, much later, by the introduction of a few basic kanji embedded into basic sentence patterns. By the end of the first year, students usually finish Volume 1 of JWL as well as JSL, by which time they will have covered about 100 commonly used kanji, not learned in lists but rather embedded into natural Japanese written prose.

Let me add that one reason I like JSL far more than any other textbook is Jorden's beautiful, very pragmatic—actually brilliant—approach to teaching the structure of the language. What she says about the nature and use of, for example, that old bugaboo of language students and linguists alike, the particle wa, is elegant and more insightful than all those books about it more "theoretical" linguists have written over the years. The only thing that comes close is Martin's incomparable Reference Grammar of Japanese.

From Richard Warmington:

When I started learning Japanese 30 years ago, the course in which I was enrolled used a textbook which did indicate pitch accent. That textbook was Osamu and Nobuko Mizutani's An Introduction to Modern Japanese. However, I don't think we were given much guidance on pitch accent in classes, and I think most of us students ignored it for the most part. The course also used one of Jorden's books, but it was her "Reading Japanese," in which hundreds of kanji are introduced, rather than a textbook which only used romaji.

The following year, those first-year textbooks were replaced with materials produced in-house, and there was no longer any reference to pitch accent at all, I'm afraid. Hiragana were taught from day one, and students were expected to master that script within four weeks. Romaji were never used after that point.

I suppose one reason why many Japanese dictionaries don't indicate pitch accent is that they may be primarily intended for use by Japanese people, who may be expected to know the pronunciation of the words. Maybe it's assumed that they are just looking for definitions?

As for romanized Chinese: for better or worse, I tried to exploit my knowledge of kanji as much as I could when I started teaching myself Mandarin (years after learning Japanese). I was somewhat influenced by Chao's Mandarin Primer, though, and became quite comfortable reading GR — see my comments here (the last two or three posts on the page).

From Perry Link:

Bravo for this statement. You can add my name to the list of supporters. Native-speaking teachers who like to begin with characters should reflect that they themselves, as native speakers, in fact learned in the order Chao and Jorden prescribed: oral fluency first, writing next.

[VHM: Perry is referring to my reply to Tom Bartlett quoted above.]

N.B.: What is said about "Chinese" above naturally refers primarily to Mandarin (see the discussion here), but can also be applied to any of the many Sinitic languages and topolects.

I've been involved with Chinese language teaching long enough (well over forty years; I have taught 1st, 2nd, 3rd, and 4th year Mandarin [I once taught 3rd and 4th year Mandarin simultaneously], and have been teaching Literary Sinitic / Classical Chinese for thirty years) to realize that some students are attracted by the characters and that is one of the main reasons they opt to take Chinese language courses (this is especially true of so-called heritage learners). They think that the characters are beautiful or exotic, and sometimes their parents want them to learn the characters to maintain the culture. On the other hand, my experience, and the experience of my colleagues named above, abundantly attest to the wisdom and efficacy of learning the language first and the script later.

The advent of electronic devices for the computer assisted writing of characters has also brought about a sea change in the way they are learned and maintained. David Moser will write a guest post on this subject before long.

BOTTOM LINES

If you delay introducing the characters, students' mastery of pronunciation, grammar, vocabulary, syntax, and so forth, are all faster and more secure. Surprisingly, when later on they do start to study the characters (ideally in combination with large amounts of reading interesting texts with phonetic annotation), students acquire mastery of written Chinese much more quickly and painlessly than if writing is introduced at the same time as the spoken language.

[Thanks to Ben Zimmer, Bill Hannas, Brendan O'Kane, and Matt Anderson]