Recently a company I do R&D consulting work for asked me to serve as a translator on a conference call with a potential client in Japan. Although I consider myself a fluent Japanese speaker—I literally have no problem carrying on a conversation if roused at 3:00 a.m. (as actually happens sometimes)—this task was somewhat intimidating for three reasons:
(1) Intepreting from Language A to Language B is a specialized skill, for which knowledge of both Language A and Language B is necessary but not sufficient. The ultimate in interpretation is the simultaneous interpreter, who manages to translate the first part of a sentence while listening for how the second part will turn out. But even at a simpler, taking-turns level, interpretation is trickier than you might think if you haven't tried it.
(2) The subject matter under discussion—digital imaging technology—is something that I don't usually discuss in Japanese, so there would be a lot of unfamiliar vocabulary.
(3) (This is a particular issue with Japanese:) the stylistic register of the conversation, being a business meeting, was different from what I am most used to, which are informal, friendly conversations.
I had about three days to prepare. In that time, there was little I could do to address points (1) and (3), but I could hope to do something about point (2)—vocabulary. I did my best to bone up using available tools (all of which happen to be available for free).
I was lucky enough to receive a 30-minute video showing a conversation similar to that I would be required to interpret. I listened to the video and jotted down (in Notepad) words and phrases it seemed would be useful to know off the top of my head. I ended up with a list of roughly 200 terms. I then processed my notes as follows:
1: I copied the list of terms into a single column of an Excel spreadsheet. I could have typed them into Excel to begin with, but this took only a few seconds.
2: I uploaded the Excel document into Google Docs and converted it to a Google spreadsheet.
3: Using the GoogleTranslate() function, I created a second column containing the machine translations of the terms in the first column.
4: Mrs. Gorodish—a native Japanese speaker with however no expertise in digital imaging—did a sanity check of the translations. Most of them did not seem to need fixing. I went back over the list and together we decided on translations for the most easily misunderstood terms (think, for example, how many possible meanings the word "form" has).
5: I downloaded the spreadsheet back into Excel format and cut-and-pasted the contents into a tab-separated text file.
6: Finally, I imported the text file into my Anki deck of Japanese vocabulary.
The foregoing prep time took perhaps an hour, with the bulk of time spent on step 4. It then became a matter of studying the terms over the next three days, taking perhaps an hour a day. Many of the terms were actually easy-to-remember adaptations of English words, but I intentionally "failed" them in Anki in order to make sure they would be presented the next day.
The meeting itself came off well, and all parties declared their satisfaction with the translation. Of course, how would they know? ...since I was the only bilingual present I could have committed the most outrageous translation errors and no one would ever know (including myself).