Interpreters will be replaced by pre-recorded videos

According to INTERSECT, a newsletter of Cross Cultural Communications,

The Trump administration has announced that interpreters for initial deportation hearings will be replaced by pre-recorded videos in several languages.
 
[...] The videos will inform those who face deportation of their rights. How anyone knows whether the videos will be understood, or how those who watch them will ask questions, is unknown. Previously, interpreters were available for questions when judges informed those to be deported of their rights.
 
In addition, vast numbers of those who will watch the videos speak indigenous languages unlikely to be recorded in the videos. In an article in The San Francisco Chronicle,  one judge reported this change as a “disaster in the making.”
 
Language lobbyists suggest that U.S. residents write a letter to Congress to protest the change.

One more step in the Republican quest to transform the US into a country where the rights of the people count for nothing.

Instaspekers: even if it were legitimate, too dangerous to be trusted?

Over the past few weeks I received many emails from Instaspeaker, a new company which will soon launch a translation app for mobile devices. They are looking for experienced translators to provide interpreting and translation services on the fly through their app. According to their messages

Instaspeakers is a live on-demand translating app. Think Uber, but for translators/interpreters. Using our app customers can initiate a video call with a translator/interpreter [and get] video remote interpreting (VRI) [or] upload a picture or document [to get] an audio file [with the sight translation of the document].

On the face of it, this app is just one of the many that aim at providing linguistic services in real time for their users. I find these messages deceptive, disparaging and dangerous.

• Deceptive because they say "Earn on your terms," which, to me, implies setting your own conditions and rates — but the rates are instead set by Instaspeakers. Deceptive also because when you go to their website, they paint a very rosy picture of the translators' and interpreters' earning potential:

Elite translators have between 0-9 years of experience, and are billed at $1.50/minute. Elite translators can earn up to $73,000/year
Premier translators have between 10+ years of experience, and are billed at $2.50/minute. Elite translators can earn up to $134,000/year

...only to say in the footnotes that

Earning estimates are for explanatory purposes only, and the actual earning potential of each translator will be determined by the actual time each translator works and the rate for which their services are billed out. Earnings estimates are based on 40 hour work weeks over the course of 12 months.

So, to earn $ 73,000 dollars in 40 weeks, how much would an "Elite" translator have to work? If we multiply 40 weeks times 40 hours/week, times 60 minutes/hour, we get 96,000 minutes. If billed at $1.50/minute, the total would be $144,000 - but since Instaspeakers' earning estimate is 73,000, that means that only about half of the $1.50/minute would be paid to the translator or interpreter.
But nobody would be able to constantly translate 40 hours per week for them: even if it were possible, 40 hours of actual production work for them per week would mean not having any other customer, and takes in no account the time one would always need for administrative tasks, idle time, and so on.

• Disparaging because it treats translation and interpreting as a hobby "Instaspeakers allows you to earn extra cash in your spare time".
  
• Dangerous because they ask to their candidates (who, after the first 400 applicants, will be required to pay $15 for their own background checks) a wealth of personal information: address, social security number and bank account information. They say they need the bank information to pay you, and the address and social security to run the background check.
  

I have no reason to believe that this wannabe "Uber for translators / interpreters" are not a legitimate service, but...

But if I am wrong, providing them with all that personal information would mean providing someone we can not really check with all the information necessary to steal our personal identities, and, even if they are legitimate, providing a combination of name, address, social security number and bank information would mean that in case of a data breach (and we have seen how frequent such events are), the threat actors would have all of our personal ID, not only a mere email address or credit card number.

How to ensure poor translation results

 I've just received the following email from translated.net:

Ciao! Abbiamo appena ricevuto una richiesta di traduzione di 40.000 parole, con subject legal. Il documento (word caricato su Matecat) sarà diviso in link da 1.500 parole circa. I link saranno inviati domani mattina (ore 10.30 \ 11.00) e dovranno essere pronti entro le 17.00 del giorno stesso (14\06\2019).  Facci sapere se sei disponibile e se si per quanti link da 1.500 \ 2000 parole circa

For those who don't speak the execrable Italian-English mixture used in the above message, the following is a rough translation:

Hi! We have just received a request for the translation of 40,000 words, subject legal. The document (word loaded in Matecat) will be divided in sections of about 1,500 words each. The sections will be sent [to translators] tomorrow morning (by 10:30 - 11.00 AM) and must be ready by 5:00 PM on the same day (June 14, 2019). Let us know if you are available and, if so, how many 1,500-2,000 word sections you can take.

This means translating 1,500-2,000 words in a bare six hours or so (possible, though probably a bit tight, considering the subject), but it also means that the whole translation will be done by a group of between 20 and 27 translators. You'll notice there is no mention of terminology coordination (impossible anyway, given the time constrains), nor of editing or proofreading (and no sign of the rates offered).

Even assuming all translators who accept this "offer" are all good professionals (and, frankly, I doubt any good translator would participate in such a project), the resulting translation is likely to prove disastrous because of the inconsistencies that will inevitably crop in, given the likely number of translators and the little time available.

This is the recipe for a failed translation.

An article from the Economist on the art of translation

The Economist, has recently published Daniel Hahn on the art of translation, an article which deals, among other things, with the intersection between writing and translation:

“Translating becomes a sort of replicating of an idealised version of that writing experience, so that I can write a novel as if it’s from scratch, but I know it’s going to work because it’s been road-tested in another language. I’m constantly writing these great novels, which is an amazing privilege.”

New edition of Mats Linder invaluable SDL Trados Studio Manual updated for SDL 2019 SR1

Mats Linder's SDL Trados Studio - The Manual has now been updated with a new edition - the second for SDL Trados Studio 2019, to cover changes introduced with service release 1 (SR1) of the program.

The manual is now grown to 598 pages, and includes 30 pages on machine translation -- useful for many translators, considering the ever-increasing presence of MT in many projects.

The price is USD 52 or 49 Euro for new users (or half of that for those who bought previous versions of the manual).

I highly recommend it, since it complements (and mostly supersedes) SDL's own badly written documentation: Mats writes in a way that any user of SDL Trados studio will find useful.

Interesting article on post-editing machine translation

Recently, Isabella Massardo published on interesting article on post-editing machine translation:

5 Effective Strategies for Post-Editing MT

Because of the misleading fluency of NMT systems, we now have to get the meaning of the source text first and then compare the text with the MT raw output to make sure that the translation is correct and adequate.

Good to know that there are still people who realize that advocating for "Monolingual Post-Editing", as some zealots increasingly do is ever more dangerous, especially now that machine translation is becoming even more "fluent".

Edit:
The link was not correct but it should work now.

Term Extraction Using Concordance Tools

Yesterday, at the 9th annual conference of the Colorado Translators Association, I gave a new presentation: "Term Extraction Using Concordance Tools". I think it was well received: lots of colleagues attended, and several had various questions to ask at the end.
I've posted the presentation (in embedded form) in the "Other Presentations" tab of this blog, from which you can also download a copy. At the end off the presentation there are links to all the programs mentioned.

'It's a silent conversation': authors and translators on their unique relationship

From the Guardian, an interesting article on authors and their translators:

From Man Booker International winner Olga Tokarczuk to partners Ma Jian and Flora Drew … leading authors and translators discuss the highs and lows of cross-cultural collaboration.