Linguistic Feature Spread in Online Social Networks

Abstract

Sociolinguists studying computer-mediated communication often study the effects of categorical variables on online language use, but studies of the effects of community membership have lagged behind. Such studies of traditional, in-person communities have found that individuals tend to speak more like those in their immediate community, and less like those with whom they are distant, but comparable studies have not been carried out for online communities. To explore whether this trend may hold true online, I have conducted a study of the use of conversational, text-based communication on the social network site Twitter. I find that a variety of linguistic features characteristic of computer-mediated communication correlate with community membership in their usage, indicating that Twitter users do tend to “speak” more similarly to those in their immediate social circles.