Language Relationships as a Graph
The first few lectures of the course made me realize that graphs are ubiquitous, appearing in just about every field of study. After taking historical linguistics last semester, I learned that all languages that have ever existed can be represented as one massive tree, branching off in many directions to form language families. I thought that perhaps there was another way to represent the relationships between languages; not in the form of a tree, but as a graph. After a bit of surfing the internet, I discovered something called lexical distance, which is essentially a measure of how similar words are between two languages.
Professor Konstantin Tishchenko created a graph depicting lexical distances between European languages in 1997, and a recreation of this graph by Teresa Elms is shown below. The nodes in the graph are languages, and the edges represent various degrees of lexical distances between languages.
Looking at this graph, we can apply many of the concepts we have learned in the course so far to analyze it. For example, we can see that there are many structural holes, or empty spaces, in the graph. These areas of low connection density represent the lack of similarities between languages of different families, whereas the areas of high connection density represent the language families, like Romance, Celtic, Germanic, and Slavic.
We can also interpret this graph as having strong and weak ties. Let’s say that ties representing lexical distances of <= 25 are strong, and all other ties are weak. In the image, the strong ties are solid lines while the weak ties are all of the different dotted lines. For the most part, languages do not violate the Strong Triadic Closure Property. Intuitively, I believe this makes sense; if two languages are very similar to the same language, then they are probably pretty similar to each other as well. However, there are some languages that violate this property. For example, Spanish has strong ties to Portuguese and Catalan, but Portuguese and Catalan do not have a tie. Perhaps this is because Catalonia and Portugal both neighbor Spain, but relative to each other are far away, so the languages changed in different ways. This is just speculation, but I find it interesting how we can use the graph theory we’ve learned to analyze a graph like this from a different perspective.
Sources:
Graph: https://elms.wordpress.com/2008/03/04/lexical-distance-among-languages-of-europe/
Article: https://alternativetransport.wordpress.com/2015/05/04/how-much-does-language-change-when-it-travels/
Article with fancy looking graph: https://bigthink.com/strange-maps/a-map-of-lexical-distances-between-europes-languages