(Note: The paper was presented at the “SOAS African Literatures Conference – 55 Years After the First Makerere African Writers Conference” by Richard Oduor Oduku. The conference was organized as a memorial event and took place on 28 October 2017, at the University of London’s School of Oriental and African Studies (SOAS), with a keynote speech by Wole Soyinka)


A few weeks ago, Google launched Bluetooth earbuds called the Pixel Buds. The developers announced that the technology had the ability of translating over 40 languages, simultaneously, from one earphone to another – that I could speak in one language and 40 different people speaking different languages could comprehend what was saying, in their language. Another company, called Bragi, released their own version, Dash Pro earbuds, branded as “true wireless intelligent earphones.” Dash Pro also has the ability to translate over 40 languages, real-time, using iTranslate app on the iPhone. Microsoft’s Skype Translator is touted as having the ability to translate up to 50 languages, making conference calls between nations and languages easier. If there is a recurring characteristic in all these technologies, is that, they are grounded on the current developments in machine learning and artificial intelligence. All these applications seek to break down language barriers between peoples of the world.

But there are a few problems. An article in the Guardian written by Nigel Kendall intimated that these technologies “remove speech from the process of communication” by simultaneously translating spoken language into the ear of the listener. He goes on to say that while all this is good, they remind him of the Babel fish in Douglas Adams ‘The Hitchhikers Guide to the Galaxy’ – the aquatic creature that excretes simultaneous translations directly into its host’s ear canal. Such a creature, Adam believed, would translate both the vocabulary of languages and the cultural baggage that goes it.

This possibility was first shot down during the early days of computational linguistics, when Altavista, a web search engine that came before Google, was asked to translate the sentence “the spirit is willing, but the flesh is weak” from English to Russian. Its translation, if I I’m to use literal English equivalent, was, “the vodka is excellent but the meat is lousy.” Now we all know that was not the intended meaning.

The loss of meaning and misrepresentations is one of the biggest challenges of computational linguistics.  While such technologies are useful in the literal transfers of meaning in knowledge-oriented genres of writing such as encyclopedic entries, standard academic texts and newspapers, they are not particularly useful in literary translations, because the core linguistic operations in these texts are both literal and metaphoric, and employ complex forms of writing in embedding meaning in texts.

Literary translations demand, not just the promise and capabilities of translation technologies, but most importantly, the greater capabilities of meaning-making that can only be provided by a human translator. Meanings are context dependent. As Professor Ngugi wa Thiong’o reminded us, “languages are not inert objects used simply for communication, rather, every language is “both a carrier of communication and a carrier of culture”. Language is not only a reservoir of culture but also a vehicle through which it is transported either in text or speech.

The second problem with these translation apps is the invisibility of African languages. They are rarely included in the advertised promise of, “this application can translate speech into 50 languages simultaneously.” Therefore, utilizing technology in translation work to increase the representation of African languages in digital platforms, requires a different hybrid approach, employing both the expertise of literary translators and the capabilities of new digital platforms.


Since the late 1990s, stand-alone or networked computers, have become new playgrounds for thriving literatures. These literatures have a tendency to work against or challenge the dominating tendencies of publishing and scholarship. For instance, while the national print book market is publisher driven, and publishers leverage on market share when introducing new literatures in the market, digital literatures are mainly author-driven and owned. While print publishing has stringent boundaries, digital publishing presents opportunities for breaking established boundaries and mutating into uncharted waters. While traditional publishing is still controlled by traditional power actors like publishers, literary agents, and distributors, digital publishing is adaptive, fluid, and suspicious of gate-keepers. While traditional publishing format is limited to text, digital literatures can exist in multiple forms, as texts or audiovisual productions.

The Jalada Translations Issue was enabled by these capabilities, but grounded on the understanding that translation is a process through which language transports meanings from one culture to the other. We are cognizant of the politics around the hierarchies of languages, but were interested in the idea that translation can offer a space for languages to meet and converse, and share the cultural wealth that they carry with them as they walk the physical and virtual worlds. Instead of starting another raging debate on the reasons behind underproduction in African languages, we opted for what the Managing Editor, Moses Kilolo, called a “practical vision”, one that focused more on production, and less on a thousand reasons why it could not be done.

It was an experiment. Jalada began by building a network of translators, editors, and proofreaders, across the continent. It started with emails to writers Jalada has worked with in earlier anthologies, and became a web of digital identities across language and physical geographies. The web platform, as both the space for publication and enabler of cross-cultural communication and networking, offered a dynamism a static book format could not. As an ongoing exercise, publication online provided multiple pages that could be updated periodically, as translations in different languages were submitted, edited, proofed, and published. The translation was initially published as one story and 33 translations in different languages, making the short story the single most translated story in the history of African writing. Currently, the story has been translated into 65 languages, 44 of which are African languages.

All experiments are a response to specific problems in the society. Experiments may borrow from what is existing, may necessarily diverge and establish new ways of thinking. The Jalada Translations work, just like all digital literatures, represent creative ways of circumventing traditional limits on production and consumption of literature. It is important to note that, preferring an experimental approach to exploring and expanding the boundaries of translation work does not in any way denote a lack of grounding, but should be understood as a scientific method, in which continuous testing of possibilities help to break old paradigms and create new possibilities for capturing and celebrating our ways of being.

We are hope we have lit a fire in the hearts of many Africans, and other peoples of the world, having a desire to write and translate into their languages. The future lies in utilizing the diverse web and mobile platforms and applications to advance literary translations in ways that were initially unimaginable.

The experiment continues. Maybe one day, in the far future, humans will succeed in developing AI-powered translation systems capable of detecting the nuances of literary texts – both the complexities of language and cultural IDs – that make us uniquely human in this ‘Pale Blue Dot’, as the American Astronomer Carl Sagan called it, this earth, ‘this mote of dust suspended in a sunbeam.’

Sign up to receive the most diverting fiction, essays, analyses and news across Africa in your inbox, on Monday every week.

We don’t spam! Read our privacy policy for more info.