The great myth of our times is that technology is communication.
– Libby Larsen
One of the recent – more noticed by the mass media – incidents with computer translation (or rather, mistranslation) happened when a news agency decided to use machine translation for the comments given by Prime Minister Justin Trudeau on the White House lawn in March this year.
He happened to speak in English and in French. Closed captioning in English showed, among other pearls of computer wisdom “We also announced its Nazi innings” and “us old guys and Houston days eight days it’s going”.
Of course, computers make our lives infinitely easier, and it is true for simultaneous interpreters as well. Notebooks and iPads help in the booth immensely to read original speeches and charts, verify terminology, consult online glossaries etc.
What happens in a simultaneous interpreter though that cannot happen in a computer? Why cannot computers handle simultaneous interpretation?
The simple answer is: because they are not alive. Human communication evolved over millennia and topics we discuss and how we discuss them becomes more and more complicated. It is not just what is said, but how and when.
Yes, you may say, but isn’t translation just a process of taking words in the original language and using the words in the language of the translation (called the target language)?
Yes and no. Yes, because this is exactly how computer algorithms and databases work: you take a word and draw an equivalent of the word from a database. Voila! No, because as soon as you try to do it, you encounter ambiguity in language and so many things you cannot really fit into a rigid computer algorithm.
What is a “foot”? A human foot or a unit of measurement? Or, maybe, a foot of a mountain? When you heard “foot” have you also thought of “root attachment” in aviation i.e the lower part of the blade of an aircraft propeller? Or, maybe, foot of a perpendicular?
What about that story about two brother politicians whose last name was Foot? One was much more conservative than the other and once someone said about them: the right Foot does not know what the left Foot is doing. It is not about feet any longer, but an allusion to the New Testament used with a family name.
A word does not exist in isolation, it is only an abstract shell or concept: you need to know context!
Google Translate made a step in the right direction because they had virtually unlimited access to an immense amount of bilingual texts. It allowed the computer logic to get at least a glimpse of the context by analyzing topics of already translated texts. It made their translations hmmmm a tad more accurate compared to what used to be before.
However, devil is in details and no computer today would be able to “understand” the biblical illusion simply because computers are not self-aware.
There are other things, however, beyond understanding context. A lot of decisions in simultaneous interpretation are made by the interpreter subconsciously (a “gut feeling”) or based on previous experience.
There are certain paradigms and ways to handle difficulties conference interpretation students are taught and some of them are not for the faint-hearted: you have to make split second decisions, take certain risks, predict and anticipate, and learn to recover from situations when you sometimes have put your foot (here we go, again) in your mouth!
Therefore, simultaneous interpretation is a huge Juggling Act.
Granted, computers handle bits and pieces of information quickly, but in most cases it is like hunting bacteria by using a hammer.
The list of additional difficulties can take many pages. Let’s consider just 3 more:
- Accents, for example, or the entire problem of speech recognition. Speech recognition software can perform better if properly “trained” by a particular user: not the privilege simultaneous interpretation software may enjoy.
- Regional variations of the same language. In Chilean Spanish “guagua” is a “baby”. In Cuba, the Dominican Republic and Puerto Rico a “guagua” is a “bus”. Beware of computers bearing gifts: “You’ve just given birth to a healthy baby bus, ma’am!” Not very good news of a new mother!
- And, finally, professional language interpretation is not substituting words and not even conveying meaning. It is facilitating communication.
This is a much more difficult task than just translation: as a professional interpreter you must make sure that parties are not hindered by cross cultural differences. In other words, in some cases the intention of the speaker becomes more important than the words the speaker says.
Sweet words said with a threatening intonation are not as sweet anymore: another thing computers cannot take into consideration.
Or an example from Russian culture: in Russian it is considered perfectly normal and neutral to say: You are wrong, while an English native speaker may interpret this literal translation as confrontational or a very strong disagreement. What is there for an interpreter to do? “You are not quite right, I think” seems to be the option that really facilitates cross cultural communication.
In the final analysis, what is extremely surprising is not even the fact that people en masse do not understand how computers work with language. Few people really analyze language, they just use it.
What is surprising is a childish belief that someday (shall we say 10 years from now) computers will produce perfect simultaneous interpretation. These “predictions” were done 10 years ago, they are done today.
It does damage on at least two fronts: it discourages learning foreign languages and oversimplifies the role professional interpreters really play.
What is also surprising is how someone may prefer a computer translation to a human interpreter in the first place for any application beyond child play.
Maybe someday false memories taken from a professional simultaneous interpreter may be implanted into a computer that will therefore have access to years of experience and hundreds of conferences? Perhaps.
But until then – do hire a professional simultaneous interpreter instead of trusting a bunch of silicon chips!