Skip to main content
Innovation Design Engineering (MA/MSC)

Francesca Nava

Fablo: Bedtime stories and minority language preservation

Languages are dying at an unprecedented rate, there are thousands of endangered languages all over the world. Even though this is a pressing issue there aren’t a lot of minority language resources and tools. This is due to the lack of data available for minority languages. Here is where Fablo comes in. Fablo gamifies the experience of learning and archiving a minority language. Fablo is a collaborative storytelling tool, that through the universal act of telling stories, helps communities and families pass the knowledge of a language from a generation to the next, allows researchers and linguists to gather a rich amount of data for minority languages, and makes the collection of the data-set necessary to develop further technology possible.

I am a Digital Service Designer with an art background: before studying Innovation Design Engineering at the Royal College of Art and Imperial College I completed a BFA at Oxford University.

As an artist I was interested in observing people and their behavior and making them interact with my artworks, now, as a Designer, I am passionate about Human Centered and Experience Design. My main objective is to find simple and elegant software solutions to important societal issues.

Fablo: How does it works

There are 6,500 languages globally, but by 2100 that numbered will be halved and, according to some more pessimistic predictions, even reduced to just 10% of the existing living tongues. It is unprecedented for languages to be dying at such a large scale, and it is an entirely new modern phenomenon.

This is an urgent problem:

Losing all minority languages would lead us to have a small sample of languages; researchers and linguists would no longer be able to analyze language and extrapolate its secrets.

Most importantly, communities affected usually want to retain their languages but are pressured, for lack of resources, technological and economic reasons not to do so. Still, their mental and physical well-being is often profoundly affected by language loss.

I individuated two main problems in language preservation:

  • Difficulty in archiving there are not enough researchers to archive all the endangered languages in the world, and a large data-set fundamental to develop technologies such as spell checks, language learning apps, and translation software that are now missing for minority languages.
  • The Generation gap is also one of the leading causes of language loss, and it is when one generation speaks the language increasingly less than the previous one.

I chose bedtime stories because they are a universal aspect of most cultures and because the evening is when parents have the most time with their children. I asked some minority-language-speaking families to map their day to discover when they have more time to speak with their children and when they are more likely to speak the minority language. Most of them reported having more time in the evening. Bed-time stories are already often used in these families as a vehicle of language transmission, but not in an efficient and consistent way

How do people tell stories — Exp 1. I analyzed how people tell stories, particularly in a minority language. They struggle with two main things: remembering words they are unfamiliar with and coming up with new ideas/introducing new elements of the story. Once they are suggested an idea when struggling, they would continue with it.
Collaborative story telling: Human and Machine — Exp 2. I created a chatbot that would help the user tell stories to see if collaborative storytelling could be automated. The chatbot wasn’t good at coming up with coherent stories, but it excelled at introducing specific elements of the story, precisely what people needed.
Game system and learning method — Exp 3. I created a card game entirely based on telling stories in a minority language through the help of card suggestions. I tested it on 4 families (8 children) to see if they would enjoy the game-play and how much they would learn. The children enjoyed the game a lot: enough to play it even when the test was over. They improved their language proficiency by 52% in a week.
Archiving through story telling — Exp 4. I developed an archiving prototype to test if families could archive a language through telling stories. I tried it on 3 families; they each contributed around 10 words each in the span of a story. I decided to improve those numbers by increasing the time available to a week and expanding the user test to an entire language community. I posted the link to my prototype to various websites and forums dealing with the Furlan language and achieved a data-set completion of around 80%.

Fablo has the double aim to archive and preserve the language through the bedtime stories told by the grandparent (and share the data with linguists and researchers) and to teach the language to the parents and the child. There are 2 phases: 

  1. The grandparent archives the language by telling stories to the grandchild.
  2. Fablo then suggests words to the parents and child while telling a story.

Fablo is structured like a game. The game is composed of different digital cards distributed to the players while telling the story. The cards may contain verbs, nouns, and adjectives. The player has to incorporate them in the story they are telling. To use a card, they have to press the record button. In this way, the computer learns the sound of the word, but also its usage. The more words are archived, the more the suggestions become accurate and precise. The child and parent can later use it to keep telling each other stories in the minority language and reach fluency.