t a recent conference in Montreal I felt vindicated to
hear that I am not alone in my notion of the ideal use of terminology
tools among translators. Lynne Bowkers, who teaches translation
technology at the University of Ottawa, gave a talk on the discrepancy
between the terminology components that many translation environment
tools (aka CAT tools) offer and their actual and/or ideal use by
But let me start from the beginning. Although most of us already own
tools that support terminological work and maintenance, we typically
don't use this important part of the translation process and fail to
harvest its many benefits. While we may use translation environment
tools (TE tools) like Trados, SDLX, DéjàVu, or one of the
many others out there, we are quick to use the translation memory
component but we ignore the terminology database component. There are
several reasons for this:
We are quick to use the translation
memory component but we ignore the terminology database component.
- The often-used term "translation memory program"
seems to suggest that the emphasis is on the translation memory.
- There is a more immediate gain through perfect and
fuzzy matches on a sentence-by-sentence translation memory basis than
there is with terminology databases.
- Translation memories can be built up relatively
quickly by aligning existing translated file pairs and/or automatically
as you translate new texts.
- The construction of terminology databases is a
comparatively tedious process: terms have to be individually
highlighted in the translation or even entered into the terminology
management application, and additional information has to be entered.
These would all be really good arguments if terminology
databases were not so terribly useful! No matter how immediate the gain
through translation memories, and no matter how tedious it may be to
enter terms into the terminology database, it pays out. I promise!
I think of terminology databases as proactive, living
dictionaries that are completely geared toward your (or your client's)
preferred terminology. The idea of terminology databases is that you
can view any corresponding data to any source segment when you are
ready to translate it. You can then enter the displayed terms with the
help of keyboard shortcuts or, as in the case of some tools including
Déjà Vu, Similis, Multitrans, or MemoQ, even
automatically "assemble" target segments with the relevant terms.
Since this semi- or fully-automated use of terms is part
of the normal and desired workflow for the translator, the typical
concepts that many terminologists believe in do not necessarily apply.
describes the difference between the terminology work of traditional
terminologists and translators as that of systematic terminology
("which deals with all the terms in a specific subject field or domain
of activity") and ad hoc terminology ("which deals with a single term
or a limited number of terms"):
[A]d hoc terminology is prevalent in the translation
profession, where a translation for a specific term (or group of terms)
is required quickly to solve a particular translation problem.
Exactly! Not only is it unreasonable to cover all the
terms in a specific subject field, but realistically speaking it is
also unreasonable to enter all the descriptive data that a "true"
terminologist would enter to describe the term, including all the
relations that it has with its synonyms, antonyms, etc. This is what I
want to have in my terminology database for a term ("term" always
includes "phrases" as well): one (!) source term, one (!) target term,
the subject area, the client, and possibly a definition. In very rare
cases I may add some grammatical information (e.g., whether the gender
for the German translation of "URL" is feminine or neutral), and I
would expect information like date of entry or name of user to be
entered automatically. You may have noted that I exclamation-marked
(how's that for a cool new verb?) the requirement that there should be
only one source and target term. This is hugely important if you a)
want to search terms automatically ("URL, Uniform Resource Locator"
would neither be an automatic match for "URL" nor for "Uniform Resource
Locator") and b) want to use that (semi-) automated method of entering
target terms into your translation (and don't want to manually delete
the first or second part of "URL, Uniform Resource Locator"). The same
is true for grammatical information. An entry like "URL (noun)" would
be unhelpful as either source or target. If you need grammatical
information, enter it into a separate field.
Why the extra trouble of entering subject and client
information? Well, if your project is well set up, your tool should do
this automatically. And it is important information about the origin of
the term which allows your tool to prefer a certain term over another
with an identical source but a different target. Not all tools offer
that feature, but you'll love it in those that do.
And here is another important difference between
classical/systematic terminology and the translator-preferred "ad hoc
terminology." Most tools offer a fuzzy terminology search, so that
"Uniform Resource Locators" would be a match for "Uniform
Resource Locator." But what about matches between the singular and
plural of "mouse"? Only tools with extensive pre-configured knowledge
about the source language would recognize "mice" as a match for
"mouse." As a result, you are well advised to enter as many grammatical
forms into your database as make sense for your particular term and
source language. And even phrases like "click on the button," "click
OK," or "the dialog opens" may be helpful entries in a translator's
terminology database (but not a terminologist's repository).
So, let's look at the terminology component of many of
the more common TE tools. There's a really interesting difference
between the more traditional ones such as Trados and Star Transit and
newer ones like MemoQ, Similis, or Lingotek. While Trados and Star
Transit offer the complete range of functionality that terminologists
require, the terminology components of many of the newer tools are
simpler and more straightforward. Some companies, especially larger
corporate translation buyers, use the terminology component of tools
like Trados and Star Transit to their full extent, but the tools'
apparent complexity tends to frighten off the freelance translator.
What is not known to many translators is that in general the
terminology component is only complex when used in a complex way. The
ad hoc way described above is really quite simple, and has actually
become even simpler in the latest releases of most tools.
And if all this wasn't persuasive enough, many tools now
also offer a quality control feature: once you're done with your
translation, it automatically checks whether all the terms in your
project match those in your terminology database(s). Tools that offer
this include Star Transit, across, DéjàVu, Trados
(TagEditor), SDLX, MemoQ, and various others.
And last but not least, if for some reason you do not
use a TE tool, there are also stand-alone tools such as Lingo that
offer only terminology management. While a tool like this is not
integrated as well into the overall workflow, it at least offers quick
access to the data that guarantees the success of your project.
However, I'll still repeat my plea and hope that the
chorus grows louder as more voices join in: Translation workers of the
world, unite! Use your terminology tools within your TE tools!