The Global Language Network: How Influential Are Different Languages? : Networks Course blog for INFO 2040/CS 2850/Econ 2040/SOC 2090

The Global Language Network: How Influential Are Different Languages?

In a world with over 6,000 distinct languages (a statistic published by Ethnologue), it is important to be able to assess just how influential these languages are. Linguists have long utilized different measures to assess a language’s global influence, including but not limited to factors like income and wealth of groups that use that language. This methodology for gauging a language’s global influence, however, is outdated; as new ways to communicate have developed (through the internet and social media), wealth and income have become increasingly obsolete in capturing a language’s global influence. Additionally, wealth is only distinguishable to specific regions, making it difficult to use as a tool in measuring global language influence. To truly find out which languages have the biggest global influence rather than simply which are most-used, researchers at MIT’s Media Lab have mapped out different multilingual content (book translations, Wikipedia pages, and Twitter users) online to see how interlingual interactions occur. From these mappings, they were able to create “global language networks” (GLNs) that revealed and reinforced many different economic and political patterns. For example, linguists were able to observe that Arabic having fewer book translations corroborated its position in the GLN for Twitter and Wikipedia based on the idea that fewer translations correlated with less external information and knowledge being brought into Arabic-speaking countries. The figure below, created by the MIT Media Lab, depicts the global language network graph for book translations.

When interpreting the MIT Media Lab’s GLN models, we can see how languages that hold more influence or power than others can be classified as central “hub” languages, with other languages having edges connecting to them. Since our graph represents an expansive global language network, we see that a giant component (as described in Networks, Crowds, and Markets: Reasoning about a Highly Connected World) is formed stemming from the English language node in the center. Because of this, the position of a language node has a significant impact on how information is spread for that respective population. The more edges stemming from a node, the more connected that language is, indicating that more information is able to be shared with that node. Being able to numerically express the number of connections between languages (within each respective realm of book translations, Wikipedia page editions, and Tweets) allows us to assess how influential certain languages are, irrespective of whether or not they are one of the most-used languages. These GLN maps are also indicative of the influence of multilingual and bilingual individuals: bilingual speakers are able to connect “nodes” or people who are monolingual, crafting connections between individuals across different languages and language families. We can even go so far as to assess the influence of the degree of each node: the higher the language’s degree, the more likely it is that content will be created in or translated to that language, as the MIT researchers described. Evidently, global language networks are a more effective method of analyzing a language’s influence in the sense that they draw conclusions from the interconnectedness of languages, rather than using localized information from the population who speak that language. Mapping out GLNs allow us to draw conclusions and make choices about translation and language usage with respect to the degree of influence we are seeking.