JQ Magazine: JQ&A with Jim Breen of Monash University on WWWJDIC

I just spent several weeks in Finland and Sweden, where the level of English speaking is extremely good. No JET Programmes there. You can't even argue that they are related languages—Finnish is just as foreign as Japanese, and probably has fewer gairaigo (borrowed words). How do they do it? By teaching it properly in the first place, with trained and qualified teachers. (Courtesy of Jim Breen)

“I just spent several weeks in Finland and Sweden, where the level of English speaking is extremely good. No JET Programmes there. You can’t even argue that they are related languages—Finnish is just as foreign as Japanese, and probably has fewer gairaigo (borrowed words). How do they do it? By teaching it properly in the first place, with trained and qualified teachers.” (Courtesy of Jim Breen)

By Tim Martin (Fukui-ken, 2006-08) for JQ magazine. Tim is a neuroscience researcher and swing dancer based in New York City. He runs a blog called The Floating Lantern, where he writes about humanism and other things that matter. Lately he is trying to learn more about effective altruism and the science of applied rationality.

Jim Breen is the man behind a resource that probably every English speaker trying to learn Japanese has used: the massive WWWJDIC online dictionary. In the 1980s, Breen developed an interest in Japanese that led to him programming a Japanese dictionary for DOS as a hobby. While a professor of digital and data communications at Monash University in Melbourne, Australia, Breen continued working on the dictionary, until eventually it bloomed into an interface that connects and cross-references hundreds of thousands of entries for words, names, and kanji.

Now a recognized authority on lexicography and the Japanese language, Breen continues to work on his “hobby,” and is pursuing a Ph.D. in computational linguistics. In this exclusive interview, JQ spoke with Breen to find out how it all began, his thoughts on language teaching and the JET Programme, and how he thinks technology will affect our experiences with foreign language in the future.

How did you develop your interest in, as you say on your website, “things Japanese”? Is there a specific part of Japanese culture or media that got you hooked?

I guess my interest in Japan over other foreign countries began around 1977 when my eldest daughter, then six years old, began to study the violin using the Suzuki Method. My wife, a musician and music teacher, had heard about the Suzuki Method from a lecture and demonstration, had been extremely impressed, and had expressed an interest in our children studying in the method. From that point on Suzuki began to play an increasingly bigger part of our lives as our second and third children began to study within that method. Also, my wife began to explore teaching in the Suzuki Flute Method.

None of this interest was particularly focussed on Japan itself. In 1980, I took most of the year off work to complete my MBA (I was a budding junior executive in Telecom Australia in those days). A fellow student in our classes had studied Japanese, and impressed me by translating some of the titles of pieces in the Suzuki books. Also among the visiting lecturers was a former trade commissioner in Japan, who spoke eloquently about the importance of Japan and the need for people trained in Japanese. I recall going home that night and saying to my wife: “I think I’d like to study Japanese eventually.” She didn’t think much of the idea, and I concentrated on other things like finishing my MBA and completing a music performance diploma.

In mid-1981, my wife said one day that she thought she really should go to Japan to study teaching Suzuki flute with Toshio Takahashi at the Suzuki headquarters in Matsumoto, as there was no one in Australia teaching Suzuki flute. I liked the idea, and we agreed to go the following December and January, which is Australia’s summer period when schools are closed. Our kids, then aged ten, seven and three, could easily miss a couple of weeks school in December. I arranged two months’ Long Service Leave from Telecom (LSL is a employee right in Australia after you have worked more than 10 years with an employer), and after struggling for a short while with Teach Yourself Japanese (all Kunrei-shiki romaji), I also arranged to have weekly Japanese lessons with Brian Drover, who trained as a Japanese linguist in the Australian army during WWII.

So, in late November 1981, five Breens arrived at Narita, made our way into Tokyo, were popped on a Chuo-sen train to Matsumoto by people from the Suzuki organization, and later that day found ourselves being greeted by a welcoming party from the Nagano Girl Scouts(!). (My eldest daughter was a Brownie, so we had set the international tom-toms working.)

We spent two months in Matsumoto living in two tiny six-mat apartments rented to us by the mother of some local Girl Scouts. My wife had piano and flute pedagogy classes, my kids had lessons in violin and piano (the former with Shinichi Suzuki himself), and I did the housework, shopped, minded kids, tried to study Japanese, etc.

I guess I don’t have any particular parts of Japanese “culture” I concentrate on—my tastes are rather catholic in this regard.

How did you study Japanese yourself? What did you use to help you, and did you wish a dictionary like yours existed at the time?

Back in Australia, I decided I wanted to go on studying Japanese; in fact, I didn’t want to go back to Japan until I could speak Japanese somewhat more, and read Japanese. I tried a few approaches—short courses, self-study books. The latter were pretty primitive 30-plus years ago. I thought I should do some formal study, but making the time was hard. As it happened, I was heading towards a period of career movement, and it wasn’t until 1986 when I had become an academic that I had the time flexibility to start studying properly, and I did three years of Japanese at Swinburne Institute (now University) of Technology in Melbourne. Swinburne’s course was innovative in that it concentrated on modern practical Japanese, and was taught entirely without use of romaji. It started from scratch and went a very fast pace—week one was hiragana, Week two was katakana, and from week three we were into kanji. I enjoyed it immensely, although it was a huge time demand.

In 2002, I went back to Swinburne and repeated the third year of the course as a “refresher.” The course structure and content had changed a lot, but I found it very useful.

The role and availability of dictionaries has changed massively. Back in the ’80s, we had one kanji dictionary (Nelson) and only a couple of word dictionaries. It was all on paper, and finding a word based on the kanji was a laborious task. The world has changed so much in this respect.

How has having increased fluency with a foreign language and culture affected you?

It’s had a massive impact on my life. My wife and I have been to Japan about 12 or 13 times, and we have travelled all over the place. I think my aggregate time is about 15 months, including a seven-month stint as a visiting professor. (Our other language/culture is France, where we go often, too.)

I enjoy Japanese history, architecture (I particularly like Meiji-era buildings), theatre (Noh, Kabuki, Bunraku). I’m not a fan of manga nor of anime, although I do like all the Miyazaki films. I really like
good Japanese cinema: Ozu, Kurosawa, Imamura, Kore-eda, et al., and writers like Tanizaki, Mishima, Murakami, etc. I particularly like to get out in the wilds in Japan (as far as it’s possible), and have done a lot of walking in the Alps, Shikoku, etc.

It’s not all just a big love affair with Japan—there are lot of aspects of Japan I think are pretty awful. I’ll avoid getting too far into the political system, or approaches to environmental issues. As for relations with neighbouring countries, and addressing the nation’s recent history…

What made you decide to create an online Japanese dictionary? How has WWWJDIC evolved over time?

As someone who had spent much of my life around computers, I had hankered to come to grips with handling Japanese text on computers. I had been told in Japan that it was too hard for Western computers to “do” Japanese because of the need for fonts, etc., so it was a refreshing surprise in late 1989 to read a message on the sci.lang.japan Usenet newsgroup, one of the first groups I subscribed to, that Mark Edwards at the University of Wisconsin was writing a free Japanese word processor that could run on ordinary PCs. Soon after that came the announcement of a kanji terminal emulator program (KD) for PCs. I downloaded Mark’s program (MOKE 1.0) and saw that it was indeed possible to see and enter kana and kanji on an ordinary 8086. From then on, I was hooked.

MOKE came with a rudimentary Japanese-English dictionary file, which was expanded somewhat (1,900 entries) in the commercial 2.0 release. I had long been interested in the idea of a computerized dictionary—indeed I had helped people at Swinburne publish a student dictionary—so using clues from the KD code and Ken Lunde’s pioneering “japan.inf” information document on Japanese coding (later expanded into his first O’Reilly book), I wrote a C program that searched the MOKE dictionary file and displayed selected entries. Of course the file was too small, so I added several thousand new entries, and in early 1991 released the software (JDIC for DOS) and the expanded file as freeware.

The rest, as they say, is history. The dictionary file, EDICT, has grown and grown, and in 1999 developed into a much more complicated and larger lexicographic database (JMdict), now supported by an online maintenance system. It is used by a myriad of servers, apps, etc.

A particular early concept and goal I had was to integrate kanji and word dictionaries, which until then were always separate. I was able to track down enough data to build up a kanji dictionary file, and I added that to JDIC about 1992. It meant that someone could look up kanji using radical/stroke count or reading, then jump straight to words containing that kanji. Also you could go from a word to the entries for the constituent kanji. This was quite pioneering—as far as I know, JDIC had this capability before any product or package in Japan did, and was very liberating to students as it reduced a lot of the drudgery.

The initial JDIC program was followed by a rather crude text-glossing program (JREADER), which in turn grew into the “xjdic” program for Unix X11 Windows, from which came MacJDic, GJiten, etc. In 1998, I finally got around to massaging the xjdic code into being the start of the “WWWJDIC” Web-based dictionary. WWWJDIC is now my only supported/maintained public-use software, and I tinker with it continually. It’s rather old-fashioned—definitely WWW 1.0—and would certainly do with a major rewrite. It now runs on six mirror sites around the globe, and updates the dictionary files daily. Despite its clunkiness, it gets a lot of use. As well as [combining] the main JMdict/EDICT and KANJIDIC dictionaries, it also uses a 740,000-entry name dictionary and a heap of subject-specific glossaries I have scraped together over the years.

What’s the word or phrase in WWWJDIC that you found most difficult to translate into English?

There are so many horrors that none really stand out now. Japanese has a rich lexicon and a lot of subtleties. Some of the entries in the major dictionaries go on through dozens of senses. Right now the JMdict editors are arguing over some of the more arcane uses of words like ところ and も. I have a particular interest in finding and translating new words or finding new meanings for old words. One I was working on today was 裾切り, which means cutting off the cuffs of trousers, shorts, etc., but now also has a new meaning in bureaucratic language of exempting some small operator from regulations, etc.

What is your favorite word or phrase in Japanese? How about your favorite kanji compound?

えええと! I can’t really say I have favourites there. I guess I’ve been around Japanese too long. I rather like phrases like なるほど in context.

 In what way do you think the new learning tools that are available, such as online dictionaries and Anki software, have made Japanese more accessible to foreigners?

Used properly, they can improve productivity by removing some of the drudgery. They also have a vast, and largely unexploited, potential. I have used online vocabulary and kanji testing a bit, but I still get mileage from traditional kanji flashcards, too. I have to say, though, that I think significant application of IT to language teaching, including Japanese, has yet to happen. Most of the systems I have seen simply automate traditional processes and don’t exploit what really can be done. I think it requires language teachers (and I am NOT one) and technologists to get together and think outside the box. The trouble is that language teachers don’t know the technology, and IT specialists usually don’t know anything about teaching languages.

Dictionaries are slowly getting away from simply being electronic copies of paper dictionaries, but it’s a slow process. We need real innovation in looking up words, associating dictionary glosses with text, etc. In fact, the whole concept of a “dictionary” needs to change. I think eventually word/phrase translation will become a standard part of display software, but there is a long way to go.

In one way, there is a bit of a trap in having dictionary information instantly available in that it can take away the pressure to memorize vocabulary and kanji, which is really an essential part of actually becoming fluent. I fall into this trap all the time as most of the Japanese text I see is onscreen and it’s so easy to associate it with its meaning. It’s only when I’m forced to read signs, newspapers, etc. when I’m in Japan that I actually have to make an effort to dig things out of my memory.

Have you noticed any other changes over time in how Japanese is studied or taught?

Not as much as I’d like. My own experience is rather limited, but apart from a few things such as dictionaries, nothing much had changed between the mid-’80s and the early 2000s. From what I hear, it’s not much different now.

For over a quarter century, the JET Programme has recruited native English speakers from around the world to strengthen English instruction in Japanese primary and secondary schools. Last November, the Tokyo Metropolitan Government decided to send their native Japanese English instructors abroad to enhance their teaching abilities. What do you think of these initiatives? How important is it to learn a foreign language from a teacher with native or near-native ability?

To answer your last question first, I think it’s very important. As Edwin Reischauer once wrote, the problem with English teaching in Japan (was) that there were 70,000 English teachers, few of whom could actually speak English. I don’t want to insult JET people, but I have often felt the JET Programme smacked of tokenism, and while it was of some positive benefit, it went nowhere towards addressing some very fundamental flaws in English education in Japan. Similarly, sending native Japanese teachers of English abroad will achieve little unless there is also a thorough revision of the goals, syllabus, etc. of the whole programme. It really needs to get away from being oriented to university entrance exams, and switched to achieving real working language skills. I simply can’t see that happening any time soon—the current rather hopeless system is so strongly entrenched, and a large proportion of the current teachers can’t work with anything else.

Have you followed the JET Programme much over the years, and are there any experiences with it that have informed your thoughts?

I’ve not paid a lot of attention to it. A colleague at Monash had done it and talked a bit about it—I gather she didn’t enjoy it much. A few people I’ve met in Japan first came there as JETs. Most of the anecdotes I’ve heard have been on the cool side.

Aside from issues with an English curriculum centered on university entrance exams, how do you feel about the effectiveness of bringing in foreign English speakers to teach?

The concept sounds good, but:

(a) From what I’ve heard and read, the ALTs actually get very little opportunity to teach, and in many cases, they are not that well-equipped to do it;

(b) I feel the whole programme is largely one of window-dressing. I think there is an awareness that English-language (well, all foreign-language) education in Japan is not up to scratch, and a perceived need to do something about it. Tackling it in a serious way is too hard, whereas bringing in a couple of thousand gaijin a year and spreading them around is relatively cheap and good publicity.

I just spent several weeks in Finland and Sweden, where the level of English speaking is extremely good. No JET Programmes there. You can’t even argue that they are related languages—Finnish is just as foreign as Japanese, and probably has fewer gairaigo (borrowed words). How do they do it? By teaching it properly in the first place, with trained and qualified teachers.

Two recent articles in the Japan Times caught my eye. The first is by one of their regulars, Teru Clavel, which rips into English teaching in Japan, and confirms many of my comments. It makes depressing reading. The other was a collection of follow-up letters, which basically agreed, and one pointed out [of] Teru’s article, “Here we go again…such articles have also become an annual tradition.”

For me, one of the sad things in Teru’s article was the feedback about the irrelevance of English skills to many of the parents. It’s so true of Japan—for many people, their real engagement with gaikoku is of little interest. And so much of this comes down to their education system.

How do you think the JET Programme could be streamlined or improved?

If Japan is going to seek external teachers to take part in English language education:

-It should recruit properly trained EFL teachers
-It should give them real teaching positions

Given the curriculum and constraints around employment, I don’t expect that to happen in a hurry.

Do you think there is benefit to the other aim of the programme—grassroots internationalization?

I’m sure there is some, but I feel it’s pretty marginal and not well focused. I suspect it’s more about the political justification of the programme than actually wanting real internationalization.

Anyway, them’s my thoughts on all this.

If you had developed an interest in some other culture, such as Thai, for example, is it possible that there would now be a Jim Breen Thai/English online dictionary?

Quite possibly, if it had happened around the same time. If my engagement with Japan had been a decade or two later or earlier, it would probably all be different. The dictionary started because there was the potential to do it, and nothing else was around. Now I can buy a really good electronic dictionary for ichiman-en or so, and I doubt I’d be driven to do it. Also, I had a bit more energy at 38 than I have at 66.

We’ve seen Google do some impressive work with their machine translation (MT) of foreign languages, but translations between English and Japanese are still often mangled. What do you think the future looks like in terms of using machines to do good English/Japanese translations?

I’ve been a keen observer of machine translation over the years, preferably from a safe distance. People outside the computational linguistics field are probably unaware that the development of much of the field and much of modern linguistics theory was driven by the early work in MT in the ’50s and ’60s, because that work highlighted just how little people really knew about the mechanics of languages. I’ve given talks on MT to translators (possibly a life-shortening exercise), and it’s interesting to experience both the loathing and the fear. MT is here to stay, and it’s getting better, albeit a lot more slowly than was hoped in the early days. Initially there were the transfer systems (Google used the commercial SYSTRAN package for years), then the statistical systems became the flavour (Google switched to its in-house SMT system for Japanese about six years ago), and now hybrids seem to rule.

I think eventually MT will reach a stage where it will be of an acceptable quality for most purposes. It may well be decades off, but I think it will get there. It’s already getting a massive amount of real-world use, and for many purposes is achieving a lot. The EU makes extensive use of a combination of controlled language and MT in generating documents for all the different languages. Recently, a friend of mine installed a Japanese package on a colleague’s iPad. All the documentation was in Japanese, and he was able to do the installation by pasting the instructions into Google Translate. The results were good enough, and that’s what many people are happy with, whether we like it or not.

Do you have any interesting plans for the future, regarding WWWJDIC or any other projects?

These days I’m not much of a tool-builder. WWWJDIC is the last of a long line—there are bits of code in there I wrote in 1991. It badly needs a total rewrite, but I don’t have the time and energy.

My focus is really on:

(a) The dictionaries themselves, especially JMdict/EDICT. I want to see it settled as a stable ongoing project covering Japanese with lots of other languages than English. I hope it will keep on growing after I am no longer involved.

(b) My own research work. I’m partway through a doctorate in computational linguistics, with my topic focussed on the detection and extraction of neologisms in Japanese texts.

Visit WWWJDIC online here. Smartphone users can download the app free for Android and iOS.

For more JQ magazine interviews, click here.