By Wendy McBride, Inhouse Translator

JAT’s Tokyo Activities Committee’s April seminar was presented by Yuji Yamamoto, a leading light in Japanese translation circles, to around 50 E-J and J-E translators, about half of whom were non-JAT members.

The main topic of the seminar was Universal Terminology Exchange (UTX), a glossary format developed by Yamamoto-san and his team at the Asia-Pacific Association for Machine Translation to establish “a set of specifications for sharable dictionaries, which can be used across different MT systems.”

Offering a simple, open, high-density and versatile glossary format, UTX means to nix problems with and barriers to the creation, management and use of translation glossaries, including cost barriers, competing/incompatible formats, lack of consistent glossary style, incompatibility with other translation tools, and glossaries that are not computer friendly. Such issues mean that great volumes of useful information currently available cannot be freely utilized for translation purposes (by humans or machines). Yet, as Yamamoto-san stressed, a glossary of just 1.2% of the terms in a document (e.g. 50 terms for a 4,000 word document) can improve translation quality significantly.

The glossary format itself is a spreadsheet with columns for source term, target term, source term part of speech, target term status (provisional, approved, non-standard and forbidden) and comments. Term status information looks to be a good addition to any glossary. Non-standard (alternative but not recommended glosses) and forbidden (alternative glosses that are not to be used) information, in particular, could save a great deal of translator head scratching when several possible glosses exist.

In any UTX glossary, each term will have one definition, so like terms with different glosses need to be recorded in separate glossaries (by context). Status information is managed by a glossary administrator, but, given that each glossary will be context-specific, I could easily imagine a situation where the number of discrete glossaries becomes so great that actually finding a particular one becomes difficult. A lazy translator (like myself), might find it easier to keep their glossaries all together in their CAT tool.

Moreover, while the idea of sharing glossaries of technical terms among translators is appealing, and we JAT members have collectively proven ourselves extremely generous in sharing knowledge, I do wonder how far that generosity would extend if our hard-earned collective wisdom might actually hasten the day when a machine could translate the recent MT embarrassment that was the Visit Japan Tohoku portal “better than a human” could, as Yamamoto-san suggested!

The second, much shorter section of the seminar was on practical Japanese. Yamamoto-san’s suggestions for writing clearly in Japanese rang many bells with all the NSEs around me. Writing for understanding “in one read” is something we have been taught to strive for since picking up our first pencils. Interestingly, Yamamoto-san emphasized natural, easy-to-understand Japanese over technically correct Japanese. As he noted, there are plenty of examples of correct but difficult-to-understand Japanese.

If Yamamoto-san’s latest book, a stylebook for practical Japanese in the IT era, a copy of which was given away in a lucky draw at the end of the session, encourages greater conciseness and clarity and less ambiguity in written Japanese, readers and translators of Japanese will rejoice alike!

While I wasn’t totally sold on the idea of UTX as a means of turbo-boosting my productivity, Yamamoto-san gave me quite a bit to ponder in his somewhat challenging seminar.

There is also a report by Akiko Chiba on this seminar in Japanese.