Linguistic Change in Online Communities
My article (No Country for Old Members: User Lifecycle and Linguistic Change in Online Communities, Danescu-Niculescu-Mizil et al, 2013) talks about the connection between user behavior at the individual level and at the community level, how local changes in an individual’s language affect and respond to linguistic trends of a whole social network. The researchers looked at archives from two beer forums, BeerAdvocate and RateBeer, which were chosen for their long histories, active and longstanding users as well as a narrow linguistic domain which readily allows for linguistic innovation. Using statistical models, they identified a very consistent trend that almost all users follow in their “life-cycle” as part of a community: they learn the niche terminology of the community, but then stop embracing new terms introduced after they’ve settled into the community. From there, the language gap between the user and the community widens until the user leaves the community.
This trend raises significant questions about the nature by which linguistic change is propagated through the edges of a social network, since the evolution of an individual’s language clearly does not mirror the evolution of language in the community as a whole, and can even respond counter to it. Using the language of strong and weak connections, a possible explanation for this trend may be tied to the idea of triadic closure: previous psychological work has shown that close ties between a pair of people can lead to “priming”, or the copying of language between individuals. As such, language should be most similar among a group of close friends. Since longstanding friends are more likely to have developed strong ties than recent acquaintances, it is more likely that users prime off of the language of fellow older users rather than that of unfamiliar new users. If this picture were correct, it might be seen that change across the community arises in waves, with new language becoming more mainstream as old users fade out of the network. This “wave” picture to be supported by folk observation — the idea that older people as a group tend to resist change — as well as the sociolinguistic phenomenon of “adult stability”, which the beer forums seem to follow, which posits that the language used by mature members of a community closely follows that of the community at an earlier point in time. It might be interesting to explore the extent to which actual “cliques” develop in response to opposing linguistic trends, which might involve finding components in the network connected by local bridges, and then comparing the dominant terms used by each group.
http://www.cs.cornell.edu/~cristian/Linguistic_change_files/linguistic_change_lifecycle.pdf