Wednesday, April 20, 2011

Exploring The Future of Translation in Rome

Rome is a special city in so many ways, but especially because you can see ancient, medieval and modern elements of the city, side by side at almost every turn. Rome seems to have a different sense of time. Yesterday, now and forever. At the beginning of April, Rome was also host to two conferences focused on  professional translation. I had the good fortune to be at the the LUSPIO Translation Automation Conference #LTAC held at Libera Universit√† degli Studi per l'Innovazione e le Organizzazioni.  The event was a very pleasant and positive surprise for me, and I think it is worth watching closely to see what the organizers do in future.
The conference was curated by Luigi Muzii (@ilbarbaro) with assistance from Anna Fellet and Valeria Cannavina who I fully expect will be future translation industry stars. Possibly even shaping the agenda and discovering best practices for the next generation of the professional translation industry across the globe.


“It’s not what you look at that matters, it’s what you see.”    Henry David Thoreau


The conference was special for many reasons:
  • The large number of fundamental questions about the effective use of technology that were raised and explored during the sessions
  • The balance of theory and practice, and careful observation of experiments presented by many speakers
  • The large presence of many young translators (and their teachers) whose dispassionate view of technology suggests that new approaches are possible and even likely
  • Presentations in both English and Italian
  • All the sessions were available in both Italian and English via simultaneous interpreters (eager interpreter students, who were quite effective)
  • The very thoughtful curation of the content that made the program stand out as particularly insightful amongst the many conferences I have attended
  • Many sessions drew you in and left you with new questions. Have you noticed that  people who are asking questions often are the most likely to discover new things? 
I have always felt that it is very valuable to have translators participate in industry events as they are instrumental in getting the work done and it is where the rubber meets the road as they say in English. In this event you had an auditorium full of people, with many translators, who share both a skepticism about language technology but also have an interest in where technology has been successfully used in practice. There were many who were sincerely exploring how/where/when technology makes sense and sharing their findings.
Unfortunately it has become increasingly difficult to gather the stream from Twitter, after the fact, to get the real-time feel of the event. (Let me know if you know of a good way to retrieve or show this). Some highlights of the questions explored and insightful observations included:
Anthony Pym: We do not know how to talk to each other and listen to each other, e.g. What does quality mean? To whom? Why? He also pointed out that research takes too long to get to commercial users and this needs to be sped up to be relevant for the real world and asked what academia should focus on in translation education to be more relevant?
Renato Beninatto talked about will be talking about the same things in future that we are talking about today: automation, process improvements, more languages, faster and increasingly more voice and video. He also said that there would be no project management in the cloud and collaboration infrastructure would become increasingly more important as Mygengo and dotSUB already show. He also said: “Instead of finding mistakes, translation reviewers should be rating readability or functionality” for the business purpose.  
Fernando Ferreira Alves raised the inevitable questions about the economic value of translation, made a case for translators being revenue drivers not cost centers, and suggested that large translator networks are emerging as a new production model. He also characterized  “Translation as Cinderella – indispensible but neglected” and admitted that it is hard for translators to make a living nowadays and made an urgent case for more certification to reshape the professional image. (This was very similar to what I heard the President of the Indian Translators Association say a few months ago in Delhi.)
Luigi Muzii (who presented in Italian) had an interesting and deep view of “controlled language” in his “Dov’e la lingua?” talk. He pointed out how most manuals are written under time pressures by reluctant authors and thus often not useful from the start. I liked how he used quotes to point out what the intent of “controlled language” is or should be e.g. “Make things as simple as possible but no simpler” Einstein or “The most valuable of all talents is that of never using two words when one will do” by Thomas Jefferson. He pointed out that tools do not replace knowledge and that translators should not be afraid or hostile to them. He also suggested that translators not get stuck with the view that translation should be literary. It was an insightful presentation that is worth looking up, scholarly yet practical. There were also several other presentations on controlled language that to me, all seemed to be more enlightened and profound than most discussions I have heard on the subject at other events. The term controlled language was used in it’s broadest possible sense. (I think that maybe all Italians have a little bit of Michelangelo in them).
Ana Gueberof talked about her tests in measuring the relative productivity of MT vs. TM based processes which was done to better understand how to pay translators for MT post-editing work. She made some interesting findings: 1) MT segments had the highest productivity but varied greatly by translator, 2) Some translators are better suited to MT work than others, 3) The slowest translators benefited the most from MT, 4) MT errors were more easily identified than TM errors and TM generated much higher levels of final errors than MT processes.
Alessia Lattanzi had an interesting presentation that explored what type of content is best suited to a controlled language process and what impact would this have on different kinds of MT systems. In her tests Google MT produced the best results and SDL the worst results by far. Interestingly, she also the monolingual translator quoting a study by Philipp Koehn which showed that target language competence and subject matter expertise plus MT can often outperform bilingual translators.
Valeria and Anna (they are a really good team and kind of a dynamic duo) provided great insight in a case study for Arrex, on how to manage the translation process in a highly leverageable and efficient way, that links content creation, customer support and international sales to the an informed and constantly improving translation process. This is a great example of how to do business translation right.  They later told me that they “strongly believe that companies wishing to change they way they write and translate their content, like ARREX did, will drive change (in the translation industry), not LSPs” and this case study certainly was a strong validation of that. 
Isabella Chiari also made an interesting presentation of corpus linguistics and made an interesting statement: “Relevance in corpora is a dangerous word, as quality is in translation”. She shared many sources of both monolingual and bilingual data.
Federico Garcea from Microsoft provided an overview of CTF (Collaborative Translations Framework) that brought MT, community TM and broad web-based collaboration together in a very cool collaboration environment that is being used by translation “amateurs” to undertake serious long-term translation projects. I could not help but be struck by how much nicer, elegant and free-flowing these tools looked compared to the clunky TMS/TM tools created by SDL et al.
Artur Raczynski of the EPO described the EPO experience with customized MT for European languages, saying that it was not really better than raw Google MT and thus their decision to just go forward with that. (However, he has not tried the Asia Online approach and should perhaps take a peek at the Lexis Nexis experience).
Jaap van de Meer rushed in from another conference in Pisa and gave his future of the industry talk that drew some strong interest. There were several more presentations on controlled language. When I questioned why such a strong focus on CL, I was told it was because authoring is a relatively easy place to bring improvement into the overall corporate globalization strategy, it created clean data for MT and helped to define winning versus losing authoring processes. It also appears to be a good way to get product, sales and translation people talking productively together. 

I was struck by the level of understanding and insight offered by many of the speakers and how it all came together as a cohesive picture of what is going on in translation and why these issues matter. It was also interesting to see the friction and tension between the old, abstract and theoretically dominant views and the new urges to make everything more relevant, practical and useful now. I found the curators ability to bring controlled language, MT, terminology management, enlightened authoring and corpus linguistics together into such a cohesive picture truly commendable. The only thing missing is community collaboration / crowdsourcing and I think they will get there soon.

2011-04-08 06.55.00
In my opinion, Anna & Valeria can offer great value to many companies looking to overhaul their translation processes and looking to update their translation automation strategies, and also to bodies like the CNGL and EU who could fund useful research through them that would help us all. I hope they get more involved in new case studies and that we hear more from them at AMTA, as I am sure they will continue to produce useful and insightful information for the industry. If you are in the market for these services, or just want a reality check, they are worth talking to, smart, focused and know how to look out into the distance as you can see above.

While in Rome, I had the good fortune to have a wonderful dinner in the Trastevere with David Orban, CEO of dotSUB and Renato (referenced in his blog)  where I even got to practice my Hindi with the waiter, and also visit a famous gelatteria with the lovely Sara Nicolini who is returning to Ireland shortly, discussing the challenges of working in Italy. I won’t even mention the other amazing meals with friends, wonder filled walks around the art of the amazing human creative explosion they call the Renaissance period, and the glory that was Ancient Rome that can be seen at virtually every turn. I hope I get to return, for Rome indeed is a special place.



  1. No comment on the dinner entre itself Kirti? Also are you dissing the entertainment industry that keeps the world spinning in our home down by the sea here in SoCal by making things look bigger than they really are?

  2. @Kumagoro

    As I said the dinner was wonderful.

    I am not dissing anybody here, just reporting a comment I overheard from one random tourist while walking past the Colosseum on that particular day.


  3. Great description of why the event stood out for you. I particularly like "the large presence of many young translators (and their teachers) whose dispassionate view of technology suggests that new approaches are possible and even likely" and "Many sessions drew you in and left you with new questions. Have you noticed that people who are asking questions often are the most likely to discover new things?"

    But best of all is how things came together as "**a cohesive picture** of what is going on in translation and why these issues matter".

    I'm jealous I wasn't there!

  4. Great Kirti! Excellent description
    Pity you did not mention me...