Earlier this week, we got into the speed with which translators translate and how the three T’s (Technology, Traffic management, and Throwing-enough-translators-at-the-job until it’s done) affect delivery and deliverable. Most readers manage around 200 words per minute, and some can read much faster, especially if they scan for just the good parts. Written translation has great shelf life and wears very well. When properly managed, written translation is a gift that keeps on giving, communicating a message to whomever, whenever, wherever, no matter how long ago the translation was completed.
But while translation is asynchronous, meaning readers can read it anytime, not so with interpretation. The spoken word is synchronistic, meaning it’s driven by the clock. Once you say something, meaning expires faster than fresh-caught fish, and if the response lags, then the thread is broken…end of transmission.
So if translation is a painting, interpretation is a performance, and once the curtain rises, the show must go on.
Most likely you speak at the rate of 150 words per minute, perhaps a little slower. But for TV, “wall-to-wall” (as in some guy talking without a break), the figure is 150 words per minute. Conversations slow the speed rate, but on TV and radio, they tend to move a little faster, since broadcasters know that fast talkers command more attention than slow pokes.
Stick an interpreter into that conversation and things begin to slow down. Even a simultaneous interpreter is not simultaneous, since a lag of a few seconds is unavoidable. The interpreter has to listen, sentence by sentence, phrase by phrase. In some languages, meaning isn’t clear until the last syllable of the last word, so the interpreter is waiting — or finishing up the previous sentence while listening to the next (the level of concentration these guys bring to the conference table is truly amazing, with the same game face you see on top of the hill at the Olympic Super G). But despite all that brain power at work, the lag can be quite distracting for everyone involved, which is why sound isolation between the different languages is so important for a successful simultaneous event. Simultaneous interpreters also have to squeeze their translation into the time allotted, which is why, if you are working with a simultaneous interpreter, you-should-talk-slow-ly. Would it be as slow as 100 words per minute? I’m not sure.
Now, when it gets to consecutive interpretation, things slow down much more dramatically. Because it’s consecutive, every statement is repeated, and because it’s translated, the translation is always longer than the original. We tell our clients that it will take about 2.5x longer to cover the same topic when using a telephone interpreter.
So send in the clowns…er, I mean the machines, right? No dice. Fast as it is, machine translation is too error-prone to work with voice recognition systems, which are pretty error-prone themselves. So unless input is tightly controlled, automatic voice translation lies far in the future, never mind what Google says. If it can be done, it will take a lot of processing power, which takes a lot of time too.
Those lags that people encounter in interpretation are real conversation stoppers and make the process of communicating with people speaking another language a draining experience. My current theory is that it’s a major reason why people avoid using interpreters — because they find the delays tedious and frustrating.
Texting, subtitling, and dubbing are also part of the translator’s art, and they impose their own time constraints and opportunities, but I’ll save that conversation (monolog?) for another time.