I came across this encouraging (though thoroughly belated) news a few days ago, on twitter and elsewhere, from a friend familiar with my work, advocacy, and interest in facilitating the use of African languages, particularly Yoruba, in today’s world, particularly in Information Technology. Google is opening up its famous Google Translate machine to include a number of populous African languages.
This is encouraging for a number of reasons:
First, I have spent the last two years petitioning Twitter to include Yoruba as one of the languages in which the platform is being currently translated, without much luck. It has however led to an annual Tweet Yoruba Day – a day set aside in March of every year to document trends in the use of the language on electronic platforms, advocate/encourage continuous use, and celebrate the rich depths that the language brings to the world. Having Google take this step without a major public petitioning is heartening.
Secondly, the Google machine is a worldwide platform with reaches into the farthest corners of the earth. Having Yoruba join the league of other world languages, famous and non-famous, in which thoughts and opinions can be transmitted through translation is something to be proud of. I am proud of it. I am also glad that I am here to witness, and contribute to its development. (More on this later).
And third, the sample translations given on the page created by Google for freelance translators towards this project shows that – though very far from perfect, Google has put a lot of efforts into the initial work. That is admirable. Long before the project was announced, word-to-word translations for a number of Yoruba words were already sourced and documented. They will not suffice as far as the final translation engine is concerned, but this is an encouraging start. As the current state of the machine shows, the syntax is far from perfect. I’ll rate all the translations “poor”.
The next step now is to build a larger corpus that includes more than just word-to-word associations, but phrases, proverbs, aphorisms, colloquialisms, songs, and a number of other culturally relevant communicative utterances that make Yoruba a uniquely rich African language. Then, run them through different tense and aspect variations present in the language, and have translators/linguists tweak it until it is as close to perfect as possible. This will take a lot of time, a lot of effort, plenty texts from different levels of complexity in Yoruba speech (from poems to novels), and a number of dedicated people. However, this start is an important step.
I applaud, and will keep my eyes – and fingers – on it.