A Longitudinal Analysis of Vocabulary Changes in Social Media

Abstract

The vocabulary size of a language indicates the evolution of the language. The way people use their vocabulary in social media has changed, especially with the appearance of pictorial representations of ideas (e.g., emojis, memes). The adoption of emojis in the last few years motivated us to look into possible effects on vocabulary sizes in social media and maybe understand a little more about language evolution. In this paper, we do a longitudinal analysis of the vocabulary size used in social media for 14 different cities in the USA for a period of 2010–2015. We are especially interested in the relationship between vocabulary and education attainment. We computed the size of the vocabulary for each of the cities over time and compared that to the emoji usage for the same period. We found that emoji usage increases with time. Interestingly, the average size of the vocabulary behaves erratically with increases in the first two years, then reductions from 2012–2014, and then increases again in 2015. We investigated two factors that could be related to such pattern in vocabulary usage: (i) increase of reliance on emojis instead of words, which is negatively correlated with the growth of the vocabulary; (ii) increase the educational attainment, which shows a positive correlation with the increase of vocabulary for a specific time and place.