A Translation App for the Olympics

As the 2012 Summer Olympics kick off and London prepares for the onslaught of foreign visitors, a new iPhone app is being released to make it easier for all of those people to communicate.

The app, called VoiceTra4U-M, works for both telephone conversations and face-to-face conversations in 13 different languages, including American English, Hindi, Indonesian, Japanese, Korean, Mandarin, Malay and Vietnamese. Text-only translation is available for another 10.

The app is the brainchild of the Universal Speech Translation Advanced Research Consortium, otherwise known as U-STAR. This research group was formed with the objective of “breaking language barriers around the world and implementing vocal communication between different languages.” It really is a perfect fit for the Olympics.

Of course, there are other translation apps for the iPhone, but VoiceTra4U-M has the advantage of openness, meaning that a country could set up its own servers and use the app to translate its local language, which might give it an advantage when it comes to attracting tourists.

Like all machine translation technology, VoiceTra4U-M is not perfect and does have some drawbacks. For one thing, the actual translating takes place on remote servers, which implies that you do need to have a data connection available to use it. And depending on your carrier, the fees for that data could add up quickly. Then, there’s the fact that the conversation will inevitably be full of awkward pauses as the data gets beamed to those servers, translated and sent back.

Finally, the U-STAR team chose to sacrifice breadth of translation ability in favor of accuracy (an understandable decision given the limitations of available technology). As the New Scientist notes,

“U-STAR has also initially focused on translating words and phrase related to tourism, making it 80 to 90 per cent accurate versus Google’s 40 to 60 per cent accuracy – though of course, this falls if you want to discuss a topic not covered by the app.”

That’s not to knock the potential usefulness of this app for travelers, though. Oftentimes as a tourist, all you need is to be able to accurately ask a simple question like “Where is the bathroom?” and then understand the answer.

Troubles with Gmail – Get the Send/Receive to Work

Written by Vic Marcus, Vice President of Business Development at Northwest Interpreters, Inc., a language services company based in Vancouver, Washington USA.

First of all, how does this article relate to the language services industry you may ask?  It does if you use Gmail as your e-mail service provider.  Whether you are a language service company owner or employee, freelance interpreter or translator, or just a Gmail user who looked around the web and couldn’t find a solution, this article may have the answers you are looking for.

Read more

A Lost Language, Set to Music

New Orleans has always been a melting pot for different cultures. A prime example is this interesting musical collaboration, featured on The Raw Story, between French jazz ensemble Mezcal Jazz Unit and the Native American blues/jazz-influenced Grayhawk Band. Grayhawk Band is headlined by Grayhawk Perkins, a historian for the Muskogee Nation, and the Mezcal Jazz Unit makes a point of seeking out indigenous artists from around the world to collaborate with.

In the article, Mezcal Jazz Unit bassist Emmanuel de Gouvello explained his group’s approach:

“We have to do something that is not usual world music, you know, just putting some drums or electronics on it. We have to respect the tradition, but do something new.”

The two bands collaborated on songwriting long-distance, then met up in New Orleans to practice before touring the state. The result was a one-of-a-kind sound. Perkins said,

“It was really intriguing for me to have him come in and say ‘Hey, I’d like to take that traditional [sound] and see what I can do…I can feel that French jazz style to it, which I don’t get here. I get more of that New Orleans jazz-funk style when I do my music.”

Called “3 Moon” after the Muskogee calendar, the band’s songs are based on Muskogee folklore and traditions. Even more intriguing: they are all performed in Mobilian, a pidgin language that different Native American groups used to communicate hundreds of years ago.

Nobody’s sure how long Mobilian was in use. Some scholars say that the tribes used it to communicate before Europeans arrived on the continent, and that influences from French, Spanish and English were incorporated later. Others say that it arose as a response to European settlement. Either way, it’s been extinct since the 1960’s. If you want to get a feel for it, I found a tutorial for you below. Hearing it set to music must be pretty cool. As Perkins described it,

“Here we are, doing almost exactly what our ancestors did 300 years ago. It’s pretty cool. It’s definitely a historic moment.”


Translation Fail

Google Translate’s Egyptian #Fail

Hosni Mubarak: Egypt’s once and future president? Not in this universe, but Google Translate sure seemed to think so as Egyptians made their way to the polls two weeks ago. According to Voice of America, the sentence ““I will respect Egypt’s future president” turned into “I will respect President Hosni Mubarak” in Arabic once Google’s automated translation algorithm had its way with it.

Whoops! What happened? Google spokesman Jason Freidenfelds’ explained that the problem stemmed from the length of Mubarak’s tenure as president, which lasted from 1981 to 2011. Google Translate “learns” by matching translated material from one language to another, and seeing how different words and phrases correlate with each other. Since Mubarak had been president so long, every time the Google Translate’s algorithm saw the words “Egypt’s president,” Mubarak’s name was referenced.

Freidenfelds’ comments should reassure any conspiracy theorists who might have thought that Google somehow supported the dictator. Unfortunately, they also point to the limitations of using Google Translate (or any machine translation program, really) to translate important communications. This is especially true for languages like Arabic, which don’t have as much material available for the algorithm to work with and have significant structural differences from English.

As Freidenfelds explained to the VOA:

“For English and Spanish, there are tons of translations out there, and there have been for years. So you’ll get a high quality translation with the right phrases and words…In French or Spanish, you might flip the adjectives and nouns. That’s not that hard for machine translation to pick up on. But if the change pushes a word all the way to the end of the sentence, that’s a lot harder for the system to pick up.”

Google’s since fixed this particular translation glitch, but the overall issues with machine translation remain. Can you imagine how embarrassing it would be to have an error of this magnitude in your business communications? An experienced, reliable translation service is a much safer bet!


Alien Language in "Prometheus"

Have you seen Prometheus yet? If not, please consider this your spoiler alert and stop reading now if you don’t want this post to spoil your fun.

The latest Ridley Scott flick follows a team of humans and an android as they travel through space in search of the alien race who created humanity. Some are motivated by a desire for knowledge, though the elderly businessman who financed the expedition is more interested in immortality. Of course, everything goes to hell in a handbasket once they find representatives of the alien race, known as the Engineers.

But how can they expect to communicate with the extraterrestrials to begin with? What language would these aliens speak? In the film, the answer is apparently something close to Proto-Indo-European, otherwise known as PIE.

Proto-Indo-European is the common ancestor of all the languages in the Indo-European family. There is no written record of PIE, but linguists have reconstructed what they think it might sound like based on comparing all of the languages in the family.

In the film, the android David is shown studying PIE on the ship while the humans sleep in suspended animation, with the help of a holographic projection of a linguistics professor, a real-life language consultant named Dr. Anil Biltoo. Later, David briefly speaks with one of the Engineers. Here’s how Dr. Biltoo translated his speech for the Bioscopist:

“The line that David speaks to the Engineer (which is from a longer sequence that didn’t make the final edit) is as follows: /ida hmanəm aɪ kja namṛtuh zdɛ:taha/…/ghʷɪvah-pjorn-ɪttham sas da:tṛ kredah/
A serviceable translation into English is: ‘This man is here because he does not want to die. He believes you can give him more life’.”

Why Proto-Indo-European? After all, it’s not the common language of all humanity, just of the Indo-European language family. Commenters at Language Log suggest (I think accurately) that PIE was chosen because it’s the oldest reconstructed language that would still sound slightly familiar to the film’s target audience. Even so, the version of PIE spoken in the film was apparently simplified to make it easier for the actors to pronounce.

Perhaps the best commentary on the use of PIE in the movie came from this blogger. I can’t top this:

“That they [the Engineers] would go to the mythic lengths of murdering fathers with the skulls of sons, impregnating women with monsters, and driving companions into a cannibalistic frenzy for the inferred sin of using the debased newspeak of Proto-Indo-European, arguably makes them the ultimate linguistic prescriptivists.

Wow. Those aliens must really not like PIE!

Google Joins the Language Preservation Fight

Google’s commitment to its “don’t be evil” motto has been in question for some time. However, there’s no question that they do sometimes use their powers for good, and this week’s announcement of the Endangered Languages project is a perfect example.

The Endangered Languages Project is a searchable online repository for information about endangered languages that should foster collaboration among people interested in preserving them. As Google explained on its blog, the project

“gives those interested in preserving languages a place to store and access research, share advice and build collaborations. People can share their knowledge and research directly through the site and help keep the content up-to-date.”

Previously, archives of information on lost and endangered languages were scattered among various universities and other institutions. With The Endangered Languages project, these organizations can make all of this information available online, so others can access it without needing to travel.

The project is the brainchild of Google’s Jason Rissman, who noticed that scholars had already begun to use YouTube to store recordings of endangered languages and decided to get involved. Rissman told Time Magazine that when it comes to language preservation,

“There have been a lot of silent efforts. There have been a lot of exciting projects happening at the regional and community level, but this is the first time anyone is bringing it all together.”

Now that the site is live, Google is handing the reins over to language preservation experts like the First People’s Cultural Council and the Institute for Language Information and Technology. Anybody can contribute information about a given language, though presumably an expert will moderate content to make sure its accurate. Still, this should be a wonderful way for groups of language speakers to take the lead in recording their languages for future generations, assuming Internet access is available.

You also don’t have to be a scholar to enjoy browsing the site, which has information about 3,054 endangered languages (though some languages have more documentation than others.) Just a language nerd. Go check it out!

Solving a Linguistic Mystery

Spoken by the Burusho of Pakistan, the Burushaski language has puzzled scholars for decades. Like Basque, it is considered to be a language isolate, not related to any other language in the world.

Now, that may have changed. Linguistics professor Ilija Casule, of Australia’s Macquarie University, claims to have determined that Burushaski is, in fact, an Indo-European language. What’s more, his claim looks solid, as it has now been backed by other respected linguists and is due to be published in the Journal of Indo-European studies.

In an article on Sci-News.com, Professor Casule explained the problem he faced:

“People knew of its existence but its Indo-European affiliation was overlooked and it was not analyzed correctly. It is considered a language isolate – not related to any other language in the world in much the same way that the Basque language is classified as a language isolate.”

By carefully and comprehensively examining every aspect of the language, including its grammar, phonology, lexicon and semantics, he has been able to prove that it most likely descends from Phrygian, the same language spoken once upon a time by King Midas. Originally from Macedonia, the Phrygians emigrated to Anatolia in ancient times, then continued to move east.

It’s tempting to see this discovery as a confirmation of the Burusho people’s ancient legends, which claim they are descended from soldiers in Alexander the Great’s army. However, by the time Alexander was born in 356 BC, the Phrygians would have been gone for centuries. Plus, according to Wikipedia, a 2006 study found only a small amount of Greek DNA among Burusho males. So, that’s still up in the air.

In any event, it’s always interesting to see a long-standing mystery finally get solved. As Sci-News.com notes, Prof Casule’s methodology could be used to solve similar mysteries, as well:

Prof Casule’s work is groundbreaking, not only because it has implications for all the Indo-European language groups, but also provides a new model for figuring out the origins of isolate languages – where they reside in the linguistic family tree and how they developed and blended with other languages to form a new language.

An Interesting Way to Learn Mandarin

There’s no doubt about it…learning Mandarin is challenging.  But one website thinks it may have stumbled upon a winning formula, at least if you’re a teenage boy.  At SexyMandarin.com, online Mandarin lessons are delivered by lingerie-clad “tutors” striking a variety of suggestive poses.

For example, one lesson, entitled “What Time is It?”, gives the viewer a brief rundown of how to ask and answer that question in Mandarin. Demonstrating proper pronunciation are two barely clothed female “models” who have apparently snuck away from their husbands for some “girl time.”

The idea of someone actually using these videos to learn Mandarin may seem about as far-fetched as the idea of actually reading Playboy for the articles, but their creator, Kaoru Kikuchi, said the scantily-clad females make the language more approachable. As he explained to the Telegraph,

“If you go the textbook way with all these Chinese characters it just makes you intimidated. If you start with the colloquial way … or sexy clips it is a different story.”

Meanwhile, producer Mick Gleissner admitted the project is ” kind of ridiculous but it’s also fun.”

Ridiculous, yes. Fun, perhaps. But will it actually help you learn Mandarin? Wu Yue, a teacher from Beijing’s Mandarin Connection school, told the Telegraph she doubted it would:

“It is very entertaining, and might be good for marketing and promotion, but [it is] not good for serious language learning,” she said. “Students would get easily distracted during a class featuring sexual content.”

And in the end, marketing and promotion seem to be exactly what these videos are all about. At the end of the videos,  you get the tagline “When you’re serious about learning Mandarin, head over to NewConceptMandarin.com,” a language school that offers DVDs, online classes and immersion courses, with nary a word about sexy Mandarin “tutors.”

Happy Father's Day!

Today is Father’s Day in at least 72 different countries around the world, everywhere from the United States and the UK to Zimbabwe. But how did Father’s Day get started, and why do so many countries celebrate it on the third Sunday of June?

Father’s Day was the brainchild of Sonora Smart Dodd, a twenty-eight year old woman who helped her father raise her younger siblings after her mother died in childbirth. Only two years earlier, Mother’s Day had been established as a holiday in the US, and she felt that fathers deserved similar recognition.

The first Father’s Day was a local celebration held on June 19, 1910 by the Spokane Ministerial Alliance in Spokane, Washington, where the family had settled.  Father’s Day didn’t really begin to take off until the 1930’s, though, when Dodd returned from school and began promoting it again. She wasn’t alone in her fight, and while some of her allies (like the tie/menswear industry) had less-than-altruistic motives, they had the money and national clout needed to gain acceptance for the holiday.

The third Sunday in June was made an official US holiday in 1966, by President Lyndon Johnson. It’s since spread around the world, and even countries who aren’t celebrating it today generally have a day set aside on the calendar for it.

Perhaps the most unusual way to celebrate Father’s Day comes from Germany (though they celebrate it on Ascension Day instead of the third Sunday in June.) Here, it’s tradition for men of all ages to band together and go out hiking, pulling a wagon filled with beer, wine and food behind them. These days, however, many German fathers eschew this tradition in favor of more family oriented celebrations. After all, when your family is trying to celebrate your virtues as a parent, it’s simply bad form to get smashed.

Happy Father’s Day to all the Dads out there! Are you celebrating? How?

Photo Credit: Attribution Some rights reserved by LadyDragonflyCC

"Baby" Robot Learns Human Language

Researchers at the University of Hertfordshire have built a robot with a  creepy resemblance to a human baby. Why would they do such a thing, if not to haunt your nightmares tonight? (Seriously, watch the video if you don’t believe me.) To learn more about how humans learn language, of course!

According to Wired, the robot baby, called DeeChee, is learning English almost the same way a  human toddler would: by listening to human volunteers talk about colors and shapes. Here’s how the research team, led by computer scientist Caroline Lyon, described the project:

“Our work focuses on early stages analogous to some characteristics of a human child of about 6 to 14 months, the transition from babbling to first word forms.”

So, why did they need to build a baby? As creepy as the little fellow is, his resemblance to a human baby is what makes the human volunteers want to teach him to talk. He taps into our maternal or paternal instincts, which is important because, much like a human baby, it takes more than some baby Einstein videos to teach language to a robot. According to Lyon,

“Learning needs interaction with a human, and robot embodiment evokes appropriate reactions in a human teacher, which disembodied software does not.”

As Scientific American reports, DeeChee is programmed to recognize phonemes (sounds that are the building blocks of language) and then string them together, responding to praise from its “teachers” whenever it manages to repeat a word and use it appropriately. In the study, Lyon writes that “It is known that infants are sensitive to the frequency of sounds in speech, and these experiments show how this sensitivity can be modelled and contribute to the learning of word forms by a robot.”

Right now, DeeChee is still learning colors and shapes. Wait til he figures out how to say “I’m sorry, Dave…I can’t do that.”

Blog Posts

Available Pages


Archives by Month: