A glossary for MT–terrific! MT on a glossary—horrific!

In the last few months, I have been reading quite a bit about machine translation. And I also took the opportunity at the recent LocWorld in Seattle and the ATA conference in San Diego to attend sessions on MT.

In Seattle, TAUS presented several real-world examples of what can today be done with the Moses engine. It was refreshing to hear from experts on statistical MT that terminology matters, since that camp, at least at MS, had largely been ignorant to terminology management in the past. Here are a number of worthwhile tutorials on the TAUS site for those who’d like to stay abreast of developments.

At the ATA, the usual suspects, Laurie Gerber, Rubén de la Fuente, and Mike Dillinger, outdid each other once again in debunking the myth around MT. When fears did come up in the audience about MT and its effects, I had to think of a little story:

In the mid-90s, five of us German translators at J.D. Edwards were huddled in a conference room for some training. Something or someone was terribly delayed, and while chatting we all started catching up on the translation quota due that day. You know what that involved? It involved finding a string that came up 500 to 800 times. After translating it once, you could continue your chat and hit enter 500 to 800 times. See the screen print of the translation software to the right and you will realize that the software didn’t allow a human to translate like a human; we translated like machines, only worse because…oops, the 800 strings are through and you are on to the next source string. Some would call this negligent behavior and I am glad that today we have better translation software and that machines assist with or do the jobs that we are not good with.

Here is an example of where MT should not have been used. Those of you who know German, check out this text translated by MT and you will recognize very easily that, oops, it doesn’t make sense to have a glossary translated by a machine, least of all if there is no post-editing.

Tidbit from the ATA Conference

Yesterday, I ran into a fellow graduate from the Monterey Institute. K. has been a freelance translator for many years and shared some interesting insights that I thought others might like to hear.

My students in particular are wondering whether they should spend time managing their terminology. Most of them are planning careers as freelance translators and are unsure, at least at the beginning of our course.

K. and I started talking about work, and her comments were completely unsolicited. I was also glad to see that I didn’t trigger the uncertain feeling that the mere presence of a terminologist sometimes sets off in my freelance friends. Instead, she was frustrated when she mentioned that her “brain” had reached capacity, if you will, and she could no longer remember things she used to remember early on in working for a particular end-client. That term that had always been on the tip of her tongue just wouldn’t be available.

Furthermore, she mentioned that she had worked for one end-client for many years and diligently set up documents with glossaries. Her direct client, an agency, recently shifted her to a new end-client in the same industry. She said she was almost relieved because handling the many glossaries had become rather difficult.

Terminology technology is much more advanced and much more widely available today than it was when K. started her career. It no longer has to be difficult to set up a system that allows us to be fast. You also don’t have to manage huge volumes of data—after all freelance translators are likely to “only” drive one process with their terminology data, i.e. their translation process. But my advice is to follow terminology best practices, such as concept orientation, term autonomy, data elementarity.

Whatever you do, no need to look guilty.

ATA impressions

I have traveled quite a bit during the last four weeks and it is high time for an update. Let me start with a review of yet another great conference of the American Translators Association in Boston.

At last year’s ATA conference in Denver, I was still stunned because the Association still seemed to catch up with technology and the opportunity to embrace machine translation. This year, I saw something completely differently. Mike Dillinger gave a well attended, entertaining and educational seminar on machine translation. He certainly lived up to his promise of showing “what the translator’s role is in this new business model.”

It was so clear that editing for MT is a market segment on the rise, if not during Mike’s seminar, then during Laurie Gerber’s presentation on the specifics of editing machine translation output. She also shared tips on how to educate “over-optimistic clients”. You add to that Jost Zetzsche’s presentation on dealing with that flood of data, and the puzzle pieces start forming a picture of new skills and new jobs.

Jost’s presentation is very much in line with an article by Detlef Reineke and Christian Galinski in eDITion, the publication of the German Terminology Association, DTT, about the flood of terminology in our future (“Vor uns die Terminologieflut”). To stem the flood, it helps to think of “data,” as Jost did, rather than texts, documents or even segments. He also declared the glossary outdated and announced a bright future for terminology databases. To think about texts, documents, segments, concepts and terms as data is helpful in the sense that data along with solid corresponding metadata have a higher reuse value, if you will, than unmanaged translation memories or the final translation product. That has been terminologists’ message for a long time.

I also attended sessions on translation education, one by the University of Illinois at Urbana-Champaign and one by New York University. Since I will be working with the Translation Center of the University of Illinois on a small research project and am currently preparing the online terminology course that will be part of the M.S. at NYU starting this spring, it was nice to meet my colleagues in person.

ATA Preconference Seminar

The program for the 52nd ATA conference in Boston was just published. This year, Sue Ellen Wright and I will offer a preconference seminar.

Terminology Management for Translators
Barbara Inge Karsch and Sue Ellen Wright
(Wednesday, 9:00am-12:00pm; All Levels; Presented in: English)

This seminar will discuss best practices for translation-oriented terminology management, emphasizing pragmatic solutions for working translators designed to ensure long-term viability of terminological data. Topics will include fundamental principles, basic data fields for term entries, strategies for establishing target equivalents, and the avoidance of future problems and data loss. The benefits of following best practices include increased translation efficiency and accuracy, better source-language documents, reduced quality assurance costs, and an overall improvement in translation workflow and quality.

To register, click on the image below.

What is terminology management?

Years ago at a party in Denver somebody asked me what I did. I said I was a translator. “That’s like straight out of a circus,” is what the response was. I wonder what that person would say today.

At the Conference of the American Translators Association in Denver there were many signs that translators as well as interpreters play a more central role in American business and government. In fact, the professions made it onto the shortlist of the 50 best careers of 2011, as compiled by US News and World Report.

Terminology management is a niche, and many of my translator colleagues have asked me over the years what terminology is. The scientific answer is “terminology is the study of terms and concepts within a subject field.” Each domain or subject area uses specialized terms (as opposed to words) that have very specific meaning to the people using them.

In terminology management, we coin, collect, research, document and distribute these terms, most often via terminology management system (TMS). Each term is documented in a terminological record that is not just used once and then discarded. Instead, it is set up so that it can be used over and over again similar to a paper dictionary. But because it is in electronic format, it can be connected to the authoring environment of a technical writer or the translation tool of a translator. And again similar to a paper dictionary, a terminological entry should follow standards. These standards and formats are different from a dictionary entry. In addition, the entry also is not a single, stand-alone record; it is connected to the related terms of the subject field which together form a system.

Terminologists extract terms from documents, either manually or with the help of term extraction tools. They evaluate which terms should go into the terminology database. They do more or less extensive research and document their findings in the database. When the collection, e.g. for a particular project is done, they make sure that users know and have access to the data.

And lest I get heat from my expert readers: terminology management can happen in monolingual environments. It is most often used in settings where source-language texts are translated into multiple target languages. But even then, terminology tasks happened long before the first translators touch a text.

Going back to the circus reference from way back when: translation has become more mainstream in the US since. Terminology management has a long way to go. And I have likened terminologists to jugglers before.

Machine translation and excellence

What does machine translation do in a blog on terminology management? And how on earth can MT and excellence appear in the same sentence? Terminology management is one of the cornerstones of MT. And excellence isn’t tied to a technology, it is tied to you!

After a seven-year hiatus, I finally joined my friends in the American Translators Association again this year. And last week, I was in my former US hometown of Denver at the annual conference of the ATA. I was surprised about a couple of things.

For one, the profession has matured incredibly since I last attended the conference in Phoenix 2003. The number of government representatives who attended and spoke coherently about the field bears witness to that. There was great news coverage. But most of all, I could sense a different attitude among attendees many of whom I have known since I first joined the ATA in 1996: There is pride in what we do and the courage to stand up for it!

The other surprise was that this apparently was the first time a real interaction between representatives of human AND machine translation took place (“Man vs. Machine,” a panel discussion of representatives from both camps moderated by Jost Zetzsche). That is stunning to me, since I spent the last six years in an environment where MT is routine.

As terminologist, I look at MT as an opportunity for cooperation. In fact, it was Microsoft Research, where machine translation research is located within Microsoft, who declined, when we first suggested collaboration years ago. It made sense to us to supply well-researched terms and appellations to support MT. Terminology from Term Studio has since been integrated into the MT process at Microsoft.

I suppose it is fear that still seems to have a hold on translators. It might be the fear of losing market share, of needing to change to more tools or automation, or of failing with clients. Let’s go through this one by one.

In my mind, there is one market for translation; we could even say one market for content production in target and/or source languages. This market consists of different segments, and we, as language professionals, have a choice on where we want to play. The segments are not strictly delimited, meaning a translator could move between them, but let’s focus on the following three.

• A Chris Durban, who represented human translators in the panel discussion and who is serving the high-end translation market (marketing, financial reports, etc.), chooses to stay away from automation. I venture to say that her work is better carried out without automation. The key is that she achieves excellence in what she does. And she asks to be paid for it, and paid well.
• Another translator might choose to focus on the high-volume market of manuals and handbooks in a particular industry. He will work with what Jost Zetzsche calls translation environment tools, short TEnTs. That will enable him to produce higher volumes than Chris, but with equal excellence.
• And then there are those, such as the translators at PAHO, the Pan American Health Organization, who post-edit machine translation output. Again, they have a different environment, but they do what they do successfully, because they strive to do it well.

At times, an individual might choose to stay in the segment they are in or to make a transition into another segment, which requires flexibility and diligence. If you thought you could get away with less than hard work, you might have chosen the wrong profession (or planet, for that matter). I believe in that, but I also believe in the fun and gratification that comes from delivering excellence.

The last point is working with clients. The need for client education is high. The ATA has contributed a lot in that area. If you are a member, just check out all the different resources available to us. True, it is tough to do client education when you are making a living being paid only for the word in the translation. It takes skill to find the right balance. Nonetheless, clients must be informed about what they ask for, especially those who say that “quality doesn’t matter,” because very likely they have no clue. Once you have done your duty and the client still insists on some form of “quick and dirty,” you can always say no to the job. I saw projects not succeed despite warnings and suggestions. But it is not the end of the world when someone insists on, say, machine translation without preparatory or complementary work and then fails with their own customers. You could consider it self-education. You just don’t want to be in the middle of it.

In my experience, if we aim for excellence, we will be financially successful and professionally gratified. Then, it doesn’t matter so much whether we chose “pure human translation”, decide on some form of translation automation environment, or focus completely on terminology management.