Multilingual Users of Twitter: Social Networks Across Language Borders or How a Story Could Travel the World

Social media is international: users from different cultures and language backgrounds are generating and sharing content. But language barriers emerge in the communication landscape online. In the quest for language diversity and universal access, the vision of a cosmopolitan Internet has stumbled over the language frontier.

This image by Eric Fischer illustrates language and national borders on Twitter.

Language boders in Twitter.

Language communities of Europe in Twitter. The tweets are geolocated over a map of Europe and the colors represent the different languages of the tweets. Work by Eric Fischer.

However, this video from the Twitter Blog shows how news spread across languages and countries. It represents the global flow of tweets after the earthquake in Japan in 2011.

How can social media stories travel across the world? Expatriates, minorities, diaspora communities, and language learners play an important role in forming transnational networks, creating social ties across borders. Ethan Zuckerman coined the term the “bridgebloggers” to define those bloggers who were trying to connect the local communities with a wider global audience. And I wondered if there are “bridge-twitterers”.

With the help of Prof. Jennifer Golbeck, I searched for multilingual users of Twitter to see if they are connecting different language groups in their social network. Using the social network analysis tool Gephi, I visualized the Twitter networks of 92 multilingual users. We also created network statistics to distinguish the types of networks using measurable properties. We found that social networks with two language groups can be classified in three types: separated language groups, integrated language groups, and peripheral language group type. The visualizations below are examples of  these three types of bilingual networks.

How to “read” the visualizations of Twitter social networks? Each visualization represents with nodes the followers and followings in Twitter of one multilingual user. The lines with arrows are edges representing the relationship “follower of” and “followed by”. The colors of nodes represent the language each person use in Twitter, or if they are multilingual. We determined the language of every person with automatic language identification of the last 30 tweets the person wrote.

Gatekeeper network

Separated language groups in the Twitter social network of a user writing in English and French. There is a group of French-writing people on the right side (green) loosely connected with an English-writing  group of people on the left (pink). Dark colors represent bilingual users and yellow nodes have no data. Visualization made by Irene Eleta with the Gephi tool.

Union type

Integrated language groups in the Twitter social network of a user writing in Greek and English. The Greek-writing group of people on the left (turquoise) is merging and mixing with the English-writing group on the right (pink), and there are many bilingual users (violet and dark green). Yellow nodes have no data. Visualization made by Irene Eleta with the Gephi tool.

Peripheral type

Peripheral language group in the Twitter social network of a user writing in English and Portuguese. The peripheral language group of people writing in Portuguese is on the right side (green) of the dominant English-writing group (pink). Dark colors are bilingual users and yellow nodes have no data. Visualization made by Irene Eleta with the Gephi tool.

Social media, mass media and people are interconnected in complex networks. For instance, we share a video from Youtube, a picture from Pinterest, a BBC news report with our followers in Twitter. We connect with people in many ways: sharing, sending messages, commenting, liking, following. In this wilderness of connections, some network types might be better than others for spreading news across languages and increasing our awareness beyond our comfort zone. We suspect that integrated language groups are so interconnected that stories have a better chance to jump from one group to the other. Ultimately, we want to foster the creation of paths across languages that didn’t exist before.

Instead of constraining multilingual users to one language choice, social media platforms should let them select multiple languages dynamically, and should facilitate translation with language tools embedded in the interface and a “Translate” button that links translations with the original tweet like the “Retweet” button does. Also, social media platforms could recommend resources and people across languages. If we make it easy for multilingual people to subtitle or translate their favorite content and share it with the audiences they want, they will be connecting the language islands of the Internet. And stories will travel the world.


You can find more details about this research in the journal Computers in Human Behavior. The preprint is available at www.ieleta.com. Authors: Irene Eleta and Jennifer Golbeck, iSchool and Human-Computer Interaction Lab, University of Maryland.

Deja un comentario

Tu dirección de correo electrónico no será publicada. Los campos necesarios están marcados *

Puedes usar las siguientes etiquetas y atributos HTML: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>