« Audio cassettes | Main | Chinese drone »


John Dowdell

Some of the variables include:

-- Whether the audio requires use of a book, or whether the book is optional, or whether there is no book.

-- The blend in the recording between source and target languages. Some immersive tapes are all-target, phrasebooks are often 50/50 source/target, some courseware has lots of filler in the starting source language.

-- The time chunks of each unit, particularly if you're on CD or MP3 for looping. Pimsleur lessons are a half-hour, Penton VocabuLearn is usually 6-12 minutes per list... conversations between native speakers may last only a minute or two.

-- Whether the material builds and progresses, so that you have to follow it in a certain order. Also whether there is graduated review of past material, or whether new material is independent of what the program assumes you know.

This morning I understood one way that the Pimsleur Method is not appropriate. I hadn't had any coffee and was out walking, listening, but couldn't get my brain in gear. At one point I realized that it was hard to get a flow in target language when I had to keep coming back to English to understand what they wanted to hear next. This morning I would have preferred to just be in Cantonese, without having to switch my head back and forth translating -- I wanted to respond, not to translate.

One of my favorites right now is "Elementary Cantonese Conversation", two tapes and a booklet. They're published by "General Language Record Co." of Hong Kong, and are available in Chinatown/Clement bookstores for $27. I've seen two sets, maybe three on Cantonese, and two sets on elementary Mandarin.

It's a phrasebook. No filler. First comes the English phrase, then slow & clear Cantonese, then normal-speech Cantonese. The booklet has English, romanization, and characters. Phrases are collected by subject area ("At the hotel" etc). There is intermittent repetition of important/useful phrases.

I like that they start with the English, so I can focus on finding meaning first, figuring out how the sentence would be structred. The slow & clear Cantonese then helps me make sure I have the elements in the correct order, fills in any questionable vocabulary or tone issues. The final normal-speed reading lets me finally move on to the correct overall sound.

But all these vocabulary/phrasebook approaches chunk 'way too large -- you can repeat a segment, but not an item. And the device is profoundly stupid about learning what I'm learning.

Ideally we need a portable audio device with some elementary input, like a mobile phone. It needs to hold each vocabulary item in separate files for source and target language -- a thousand phrases, two thousand sound files. Then it needs to deliver these to me in graduated order, with smart repetition and reinforcement of terms, recognizing some type of input feedback when I'm confident of an item.

Audio flashcards, with a display algorithm optimized for actually learning and retaining the material -- it's a natural. Combine that with a mobile subscription service for graduated fresh immersive content, source-language grammar or usage tips, a whole programming model... I'm seeing a future where perpetual learning is always available to those who wish it....

The comments to this entry are closed.