NMT Language Spotlight: Papiamentu – ABC Islands

Most of us have had some extra time on our hands with the restrictions placed upon us from COVID 19. Some of us are engaging in new or forgotten hobbies, some exercising more (or less), and some of us devising projects to progress businesses or satisfy new interests.

In the Dutch Carribean island of Curaçao, Manuel Maduro has taken on a unique challenge – to build a Neural Machine Translation (NMT) engine for Papiamentu (PAP) – a Creole language derived from Portugese, Spanish, Dutch, English and French, as well as Arawakan and African languages. Today, PAP is spoken by roughly 300,000 people on the ABC Islands and by a diaspora of approximately 100,000 in the Netherlands. 

The word Papiamento comes from  ‘papia’, ‘pap(e)o’ or ‘pap(e)ar’ – ‘to chat’, ‘to talk’ and while in Curaçao and Bonaire it is known as Papiamentu, Arubans say Papiamento. For the islanders, language has been an important part of the island’s history and cultural development having experienced many influences over the years, from Spanish and Dutch colonisation to a large population of slave settlers from Africa.

Today, Curaçao has three official languages – PAP, Dutch and English, with Spanish also widely spoken. Because there is a need to have documents in many of these languages, there has long been a requirement for translation services. 

Curaçao

“Over the years there has been continuous demand for translation services,” says Manuel. “There is still only a small pool of translators in Curaçao, however the quality of the translation is increasing as Papiamentu education is improving. PAP became a compulsory language in schools from 1986.”

While translation quality has improved to meet demands there is also rising expectations for faster turnaround times and a relentless push to keep costs down which means technology is key to support translation efforts.

“We have to work faster and cheaper,” says Maduro. “That’s what got me thinking. In the past I was more old fashioned when it comes to translation technology, but I understand its benefits – maybe not for translating poems or songs, but for a lot of use cases it’s the way forward. It allows for smart efficiencies and scalability and allows us to be able to deliver on customer demands.”

Manuel is new to Machine Translation and decided to select a user friendly option to ensure a smooth transition to this type of translation technology.

“When I decided I wanted to start this project, I looked into a number of options, and while there were some viable ones, I found that the set up in KantanMT platform was the most intuitive and user friendly and it would likely lead to the most amount of success given the time I have to commit to the project.  I got in touch with Tony O’Dowd, Founder of KantanMT and he was very welcoming of the project.”

Because PAP is a new language, there are few resources to support translators. The first record of the language was in 1704 and the first written works date back 150 years. In the 1970’s there was a move by the government to establish the language which meant official words and spelling were established and a dictionary was released. However, despite the fact that PAP has gained a firmer foothold than other Creole languages, there is still a ways to go. 

“For PAP, there aren’t many tools available – it boils down to a few dictionaries and a word processor. It is a small language in the process of being codified, but with two different official spellings (Aruba and Curaçao decided to go their separate ways) and a basic set of grammar rules. It’s an ongoing process.”

Manuel is currently preparing to build his NMT engine and is excited about the prospect.

“I’m gathering all my translation memories at the moment. Over the past 20 years I have collected quite a large volume but as you can imagine, there is a little prep work to do.” says Maduro. “It’s an exciting project and I’m really interested to see the output. It will be a bit of a learning curve for me but I’m happy to share the results and update when the engine is ready for use.”

The team at KantanMT looks forward to seeing Manuel’s progress and is delighted that the platform was chosen for such an interesting and important project. We will update early in the New Year with results from the project. 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s