Found in Translation

Tools of the Trade
Translators are accustomed to using various keyboard systems to input
languages into their computers. Until recently, however, the writing systems of
many languages could not be combined within a single document. While we
had software that supported the individual writing systems of most major
languages, it wasn’t possible to combine Chinese and Thai, Arabic and
Hebrew, or Japanese and Russian in a single document or web page. Talk about
frustrating.
Then came Unicode. A system whose development started in the 1990s, it
can now encode far more than a hundred thousand characters from virtually all
living and many dead written languages. It also includes mathematical
symbols, musical annotation marks, emoticons, and many other writable
characters. It’s a translator ’s nirvana.
Translators’ use of technology doesn’t stop there. There’s a whole software
industry that addresses the needs of professional translators in most fields—
especially those working with technical, medical, legal, and business-focused
texts.
16
One of the most important technology tools for translators is called
“translation memory.” Think of it as a high-end consignment shop: It’s a
database that stores past translations for reuse rather than translating from
scratch. As any good “thrifter” knows, there’s a certain risk involved in the
thrill of the hunt—the quality of the inventory goes down if the translation
wasn’t good to start with or if the same sentence should be translated
differently in a different context. Hardly any translations are suited to all
occasions—like showing up for a picnic wearing a tuxedo. For this reason,
many translators prefer to shop their own closets first. If the database of past
translations consists mostly of their own work, they can be assured of the
quality, and they know how to mix and match them.
Another technology translators use is called a termbase. Termbases are also
databases, but they contain more detailed data, with lots of descriptive
information about each term such as definitions, grammatical information,
explanations of when to use or avoid using a specific term, and possible
synonyms. A termbase is less like a brick-and-mortar consignment shop and
more like eBay, with lots of product information at your fingertips.
Translation is rarely a one-to-one paradigm, in which a word in one language
perfectly matches a word in another language. Terms can be translated in
myriad ways, and a termbase helps us choose the correct translation for each
context. And just like translation memories, termbases can be shared among

---

many translators working in real time in virtual teams.
Finally, some translators also use machine translation, in which a software
program or online tool automatically translates the text according to its own
set of rules. Believe it or not, this can sometimes require more work than
translating from scratch, because we still need to do intensive editing to clean
things up afterward. Perhaps our analogy here might be a personal shopper. If
someone else buys them, you’d have no guarantee that the clothes would fit the
required size or style, so you might end up with a lot of costly alterations
before you could actually use them.
Yet for some projects, companies just want an approximate translation—they
don’t care if the clothes fit perfectly, so long as they can be worn. For these
projects, we human translators train the translation programs, teach them the
correct vocabulary, and tweak their understanding of grammar. Automated
translation does have its place, especially in certain industries. For example, in
the legal field, machine translation is often used to mine extensive amounts of
data, such as case law, to flag items that might be relevant for attorneys
working on a given case. In the manufacturing sector, machine translation is
sometimes used for the extensive support content and documentation that often
has a high degree of predictable structure and repeated terms and phrases. In
these instances, automated translation does not replace human translation. It
serves a different purpose, but it still requires professional translators.