yndigo

How to achieve a good translation

Although finding the right translator for the job is arguably the most important task an agency performs, most translation agencies do more than just outsource translations. They edit and proofread them, too. And while there’s no such thing as perfection in the translation business, we, like many other firms, try to get as close as we can.

But there’s no agreement on how to achieve a good translation. Even the words editing and proofreading don’t have the same meaning throughout the industry. In journalism and publishing, the meanings are established: editing is a review of the substance and structure of an article or book. Proofreading is a review of the grammatical and typographical elements.

In my experience, editing in the translation industry refers to a review, based on a side-by-side comparison, of how faithful a translation is to the source, especially in terms of accuracy but also style. It requires a second translator who is also fluent in the language pair and conversant in the subject matter. Proofreading on the other hand is a line-by-line check to ensure the translation is free of typographical, grammatical and numerical errors.

I’ve known of agencies, however, that define proofreading as a cursory accuracy check of a translation against its source text, and editing as a review of the translation alone for its readibility and grammar.

While I do not believe the latter process will produce the most accurate translation, I understand agencies’ temptation to adopt it. It’s quicker and less costly and it’s likely to generate a prettier final product, one that will have the appearance of a good translation especially to a client that does not read the source language.

Obviously budget and deadlines have an impact on the quality of translations — and communicating this with the client is essential — but we must begin with a clear idea of what it takes to produce a really good translation. Here is my opinion:

1. Review the document. Most agencies have the know-how to distinguish between languages they don’t speak and to determine the subject matter of a document they don’t understand. But before outsourcing a translation, the entire document should be reviewed both for the level of difficulty and any surprises. Intros can be easy, but several pages in all of a sudden things get tricky.
2. Choose a good translator. This is so important. In an era of growing databases both online and proprietary–filled with translators from across the globe–we can start to think of translators as fungible. Yet knowing the true strengths of your translators is the single most important step to getting a translation right.
3. Choose a good editor. Although I would argue the translator should be the strongest member of the Translator-Editor-Proofreader (TEP) team, the editor must be careful and astute enough to rout out problems. And, if time permits, the editor’s mark-up should be sent back to the translator for final approval.
4. Choose a good proofreader. Seeing a pattern? A proofreader’s job is essential because — and it still surprises me — despite a careful review of a translation, more mistakes inevitably crop up. This is because the proofreader’s focus is different. Their eyes are less on the message and more on the formal qualities of the text, which can be missed by even the most thorough editor.
5. Welcome feedback. Although our clients are not translators, they often work day in and day out on topics that we encounter only for the short amount of time we are translating. All feedback should be considered and often improves our translation and teaches us something in the process.

The Seach Engine Giant enters the World of Human Translation

Google can barely blink without making headlines, and its recent launch of Google Translation Center is no exception. The Center aims to match translation customers with professional and volunteer translators in 40 languages.

Why would Google, a company not know for its human-powered solutions, want to take on this role? Some have suggested — because the results will be stored on Google’s servers — it is an effort to feed their statistical machine translation tool, Google Translate, with corpora of words in language pairs in which they are currently poorly stocked. If this is true, Google is essentially building the Translation Center’s future obsolescence into its much grander plan of seamless international seach capabilities.

Who will be translating? Well, in its appeal to get translators involved in the new service, Google asks, “Passionate about bringing content into your language? Implication: some translators will work simply based on their desire to make information accessible to other speakers of their language. Other translators will work for pay.

Whether fueled by passion or a paycheck, translators will log on to the service and respond to requests from people who have uploaded a document for translation. All negotiation will then take place between the customer and the translator. And Google will ensure neither quality in one direction or payment in the other. Google will provide its translation tools however, which will include databases of Translation Memory.

As with machine translation in general, dire predictions of upheaval in the translation industry have accompanied the launch. Some say the first victims will be translation agencies, whose most appreciated (and most scorned) role is that of middleman (although some of us would argue our services go far beyond this). Then, as gaps are closed in machine translation quality by a company with Google’s power and scope, human translators will soon follow.

A question for customers and translators alike: does the Google Translation Center sound like a step forward in the translation industry? Will you try it? Why or why not?

For more information, see here, here, here, and here.

I just wanted to pass along a photo from a Flickr member of a restaurant sign in China as an example of why humans should still be involved in the translation process:

This gem has already made its way around the internet, often noting how blind reliance on an internet translation tool led to the biggest translation goof we’ve seen in a long time.

Let’s look at the upside though. It’s unique. And it may just create the kind of buzz they were looking for.

Until you actually need one, it’s hard to imagine how many translators and translations agencies are out there. Especially the first time, how can you be sure you’re getting a good translation?

1. Get a recommendation: Ask a colleague or friend, even better if that colleague is a friend! Chances are good if they received quality and service at a fair price, you will too. Don’t forget to ask how recently or how frequently they’ve used the translator or agency.
2. Ask for samples: If you have the luxury of time, ask a translator or a translation agency for what they consider their best work in the subject matter you’re interested in. This step is really essential if you have an ongoing project.
3. Ask questions: Translators want to do a good job so most will not take on a project that is outside their expertise. You should ask questions, however, to learn how extensive their background is in the field, how they learned it, how they maintain their language skills, how they conduct research, what other types of jobs they have worked on, etc. If you’re speaking with someone at a translation agency, try to talk with the project manager (not just the salesperson) so you can ask about their quality assurance measures.
4. Be a partner: Seasoned translators have seen thousands of documents of all different sorts, but they haven’t seen yours. And you know your subject matter. Whenever possible, help the translator give you a high quality product by providing reference documents, glossaries of preferred terminology, style guides, background information, etc., and be open to questions about the project. It will lead to a much better translation.
5. Read ATA’s Translation: Getting it Right: Get ready, I’ve saved the best advice for last. You can find the ATA’s publication on shopping for translation services here. It’s chock-full of useful information in plain language.
If you liked this post, please share it: These icons link to social bookmarking sites where readers can share and discover new web pages.
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 

When a word has no good equivalent

In my last post, I talked about words that were so closely tied to a certain country’s sporting event, they are inevitably used internationally in their original language. But how about words that don’t translate at all because the situation they grew out of just doesn’t exist elsewhere?

Many websites have compiled long lists of these, such as the Mirror.co.uk. Transubstantiation, a blog I follow, covers the topic here and here. Read more of his posts and you’ll see how deeply he gets into the notion of the impossibility of translation in general.

Some “untranslatable” words are less translatable than others. Take the Yamana word dona which means “to take lice from a person’s head and squash them between one’s teeth.” American English has no word for that, thank goodness. And there would be no way to translate “dona” (though I can’t imagine ever needing to professionally) without a translator’s note, which can make the translation either messy or unacceptable.

There are some cases, however, when a more practical “untranslatable” must be used. Keeping the word in the original language is prettier of course, but that’s not really translation. Preserving the original also falsely assumes the reader’s familiarity with that language and/or culture. Somewhere in between these two options, perhaps, creative use of periphrases and neologisms may help to bridge the two languages, even if some of the cultural flavor is lost.

One term that appears on many lists of untranslatables is the French esprit d’escalier, which means “to possess a mind that thinks of comebacks too late, as in when you’re descending the staircase (escalier) on your way out of a party.” The difficulty here is not that the idea does not exist in English (why not ?) but that the French coined a term based on a setting where this term frequently came into play. So we can translate the idea — “delayed wit,” pehaps — but the cultural flavor will be lost. How do you say, “Monday morning quarterbacking” in French?

When one language dominates the terminology

“Kiss your Julys goodbye,” someone told me a few years ago when I started watching professional cycling. He was right. The Tour de France is upon us and now, along with WordPress, my e-mail, and a translation I’m reviewing, I have two browser tabs dedicated to today’s 4th stage, a 30 km individual time trial in Cholet, France.

Along with the competition, I love the history and lore of this 105-year old event. Take Eugène Christophe’s 1913 ride over the Pyrenees, a time when Tour rules forbade giving assistance to riders. Having broken his front fork, Christophe carried his bike down a mountain to a blacksmith and repaired the bike himself, only to get hit with a time penalty because a boy had been seen helping him pump the bellows!

Cycling is one of those things, like cuisine and couture, that grew up in Europe and centered in France. And now, many cycling terms are French: tour, peloton, étape, prime, soigneur, directeur sportif, domestique, etc.

Now some of these terms have English equivalents, for example peloton = main field; étape = stage. Yet the French terms dominate even in races outside France. And other terms like tour, soigneur and domestique have never been translated to anyone’s satisfaction.

We cannot call the Tour de France the “Ride of France.” And “Spin Around France” would be taking the grueling three-week event a bit lightly. A soigneur has some of the duties of an athletic trainer but often takes care of food, clothes and even massages. Domestique — literally “servant” — is the term for cyclists on a team that help their leader win the race by pacing him, giving him a wheel, carrying water bottles, etc.

Thus it would be inappropriate to translate many of these words. Because I know something about the sport, I would know which words to render into English and which to leave alone. Yet I certainly can’t say the same for cuisine, and when it comes to couture I’d be totally lost.

Fisher Tool Co., Inc. v. Gillet Outillage

The U.S. 9th Circuit Court of Appeals issued an opinion the other day in a case that involved, among other things, an apparent mistranslation. Mécanisme à cliquet was translated into English as “ratchet mechanism,” a mistranslation the opinion claims “certainly isn’t clear on its face, since French-to-English dictionaries support it.” This is a case where cliquet is a broader term than ratchet; the French has several possible meanings, and the chosen translation did not fit the context.

It would be interesting to know whether the person translating the patent reviewed the figures, in which case he would have seen a “pawl” or “catch” — the intended meaning of the the French expression here — rather than ratchet, which, according to Wikipedia, is “a device that allows linear or rotary motion in only one direction, while preventing motion in the opposite direction,” i.e., a pawl plus other components.

The alleged mistranslation was part of Fisher’s suit claiming malicious prosecution (for a prior infringement case brought by Gillet), in which it argued that Gillet was aware of the the mistranslation and that the mistranslation invalidated the patent pursuant to 35 U.S.C. § 375(b), because it broadened the claim.

In the end, the translator was “off the hook” so to speak, as the judge found Gillet’s original case tenable and that it had probable cause to sue.

Translation pricing

When you talk price with someone who’s never bought translations before, there are frequently two reactions: laughter and then shock.

Laughter at the idea that translators charge by the word or that a penny higher or lower makes a big difference. For example, I was interviewed by SmartMoney.com about working with freelancers and the interviewer, Diana, chuckled a few times when I told her that talk among project managers will often sound like, “he’s only 11 cents… yeah but he’s not so great… she’s 13… really?.. is she worth it?.. well, she’s decent, but always busy; what’s your budget?” etc.

Shock at the total price. “It’s only a few pages!” Well, actually it’s 35 single-spaced pages from Korean to English. Although prices in the translation industry have generally been stagnant for years, I understand sticker shock, as I discussed in a previous post about misunderstood translation clients.

Which word?

In translation, we talk about source words and target words. Because source words are more easily countable these days, the trend in the industry is to price by the source word. In my segment — legal translation — per-target-word pricing still prevails but the industry-wide per-source word practice is slowly encroaching as more and more translators become accustomed to pricing that way.

Why not by the hour?

Corinne McKay added some great insights to this debate in a post a few months back, her conclusion being that good translators make out better overall when pricing by the word and that clients are better able to price their translations up front.

Bernie Bierman added a comment to Corinne’s post from his own article about how translator compensation has changed through the years. According to Bernie, “[t]he per-word unit has for at least one hundred years been the basis for determining a translator’s fee,” but Computer Aided Translation tools have completely changed the playing field and today’s translators are “like livestock marching in willing resignation and without protest to the slaughtering pens.” Strong words.

Will translators continue to make sense per word?

Technology is marching on and the translation industry is embracing it. If the efficient new translation model proposed by some in which texts are machine translated and then post-edited by a professional translator, who knows, maybe we’ll see translation priced by the hour, which is normally how editing is done. And if so, will those editors be paid more than today’s translators, or less?

Wictionary.org defines translatorese as: “(pejorative) Stilted or unidiomatic language produced by translation.” That doesn’t sound good.

However some contend there are certain types of documents and certain audiences where erring on the literal side is a safer bet. Not literature of course (ironically we can’t take literature literally as one of my grad school professors used to say). And not advertising or marketing. So which translations should be anything less than perfectly readable?

Well, none really. Translations need to be read, and struggling through clunky prose in any context can keep us from fully understanding the meaning of the original. But there are limits when it comes to transforming the original for the sake of style.

A translator’s creativity should at times be held to a minimum. Accuracy is paramount. And rendering a translation that is 100% accurate and reads as if it were written in the target language to begin with, while possible perhaps, would take far too long to produce, i.e, longer than most translators are afforded. Especially in certain language pairs.

And there are potential dangers if style is our ultimate goal. Readibility is easy. Just edit the translation until it reads well. If we’re not careful though, we can stray not only from the original structure but from its meaning, too. Thus in the time allotted, translators often have to strike a balance… and the balance often tips toward the literal.

Secondly, a client using translations from multiple translators will have headaches if one or more translators rework the text to such an extent that (1) referring back to the original will be difficult, and (2) piecing it together with other translations, impossible. This is especially true if an interpreter is using the translations or if references are made to other documents, which may already be translated.

Is there such thing as neutral Spanish?

This question has been asked time and time again in translation circles. To many Spanish speakers the very idea is laughable. When people from different Spanish speaking countries strike up a conversation, they recognize, often understand, and sometimes chuckle at, the words and expressions the other uses. But these regional terms rarely impede the conversation, and, although there are sometimes universal word choices that more easily bridge national divides, there is no manufactured Spanish language shared by all speakers.

For practical purposes, however, there is. And it consists not so much in creating words shared by all as it does eliminating words unique to just one locale. For translation customers, this is good news. If you want to market your products to various Spanish-speaking groups, you can save money by translating your advertising just once. Agencies and translators, although they sometimes discourage the practice — especially in the case of advertising, which many would claim should be “hyper-local” to be effective — increasingly honor this request and have become more adept at creating a neutralized language.

According to Guillermo Cabanellas de las Cuevas, in his article, Neutral Spanish: Is it Necessary? Does it Exist?, finding words without local flavor is not always possible. “Suppose that our hypothetical client wants to market peaches in the Spanish-speaking world. What word should he use on his cans? Melocotón (used in Spain and other countries) or durazno (used in Argentina and other countries, and even in certain parts of Spain)? There is no ‘neutrality’ for this conundrum. We cannot mix melocotón and durazno and get a ‘melozno’ or a ‘duracotón’.”

On the other hand, Prof. Isabel García Izquierdo, in her article highlights the fact that some see neutral Spanish “not as an artifice devoid of any identity. Instead, they see it as a way of bringing the 400 million speakers of Spanish closer together without losing their identity in the process.” The purpose of a Universal Spanish in this case, therefore, would not be market your products cheaper but to enrich understanding in the Spanish speaking world.

So why not Universal French? Or Universal English?

In all my years in translation, I’ve never received a request for Universal French. We translate for Canada or Switzerland or Belgium, or perhaps into “standard French” as the French and their Académie might have us call it. But the industry doesn’t seem to have coined the term “neutral French” as it has with Spanish.

And when we translate into English, we ask “British or American?” So what’s different between Spanish and other languages with regional differences?