Sunday, February 20 2005

How I’m trying to learn Japanese

Between my long-standing interest in anime and my on-and-off interest in Japan, I eventually became motivated to acquire at least a working knowledge of the language. Unfortunately, I know from long experience that I don’t do well in a traditional classroom setting (last I checked, the buzzwords were “underachieving gifted”), and in any case my rather irregular work schedule would make it difficult to even try.

As a result, I started looking at the available software packages, to see if any of them had any real merit. After a few weeks, I settled on Rosetta Stone, occasionally available in stores, but mostly acquired either by mail-order, online purchase, or at one of their mall kiosks (new one just opened up at Valley Fair in San Jose, by the way). They have standalone, online, and classroom packages, a pretty good reputation, and a significantly higher price than most of the alternatives.

A lot of the other software uses romanized text instead of the real Japanese writing systems, which makes things easier for absolute beginners, but effectively cripples them in the long run. RS allows you to work with romanized text, but offers both kana and kanji as options, so that you can acquire reading skills that will actually be useful.

It does not, however, teach them. For learning the hiragana and katakana syllabaries, I used the commonly recommended Easy Kana Workbook, along with LanguageBug’s free Java flashcards for cellphones. I went through approximately half of the RS Level 1 course this way, with all dialog spelled out phonetically.

Meanwhile, I started working on kanji, using the appropriately-named A Guide to Writing Kanji and Kana. It’s slow going, due in large part to the hand cramping I developed from writing each character 100 times, but I’m making real progress, to the point where I can slowly, painfully read magazine and manga text with the help of The Kodansha Kanji Learner’s Dictionary (not to be confused with The Learner’s Kanji Dictionary…).

A while back, I finished my first complete pass through the RS Level 1 software, and it’s done a lot for my comprehension and reading skills. I’ve gotten less out of the voice-training feature, largely because the feedback is unclear; it’s extremely finicky about intonation and cadence, but lacks the ability to highlight your mistakes in a way that you can correct. If you don’t have a good ear for language, you could sit there for hours without making it through the first lesson. Even if you do (and, thanks to a youth spent in my father’s office in the Language Department at UD, I generally do), it may still take a while for you to figure out what kind of mistake you’re making.

For instance, I had an early tendency to fail on any phrase that included the word otokonoko, and only that one word. Why? My cadence was off. I was making some syllables shorter than others, and the waveform display didn’t show it. It finally dawned on me that I got better results with my eyes closed. Shutting out the picture and the waveform display allowed me to focus on more precisely mimicking the sample phrase.

Was it worth the cash? Yes. I get a lot more out of spoken Japanese in anime and when I shop at Kinokuniya Books and Mitsuwa Market. Spending an hour a day listening to native speakers pronounce realistic phrases that I know the meaning of has done a lot for my understanding, and seeing them written out in kanji and kana at the same time has given me some basic reading skills.

What can’t I do right now? Carry on a conversation. Write an original paragraph without a reference book handy. Software won’t supply those skills, but it can give me a decent framework to build them on top of. The Level 2 software (which, sadly, has a completely different set of native speakers, one of whom speaks in an extremely annoying staccato) will contribute further to that framework.

Is Rosetta Stone the best language software out there? Is their learning model pedagogically sound? I can’t give definitive answers to these questions. I do know that, within the limits of the supplied vocabulary, their model produces comprehension without translation. This is dramatically better than the approach used in my high-school German and French classes twenty-plus years ago, which left me with some vague recognition skills and the ability to sing Du kannst nicht immer siebzehn sein.