Modeling Adolescents’ Online Writing Practices: the Sociolectometry of Non-Standard Writing on Social Media

Abstract

The paper discusses four generalized linear mixed models fitted to capture distinct patterns of non-standard writing practices in Flemish adolescents’ social media messages. Apart from a general model that predicts the count of all “deviations” from the Dutch formal writing standard, additional models were fitted for specific types of non-standard features. These types relate to the so-called chatspeak “maxims” of orality, brevity and expressive compensation. While the general non-standardness model reveals interesting correlations between the teenagers’ online writing style and their socio-demographic profile, the more specific models allow for a better and more nuanced sociolinguistic understanding: for different types of non-standard writing practices, they reveal distinct dynamics between the social predictors gender, age and educational track. Strikingly different gender patterns are found for the oral features, representing traditional non-standard writing, compared to the expressive features, representing new kinds of non-standard writing, bound to digital media. Furthermore, gender does not appear to be a predicting factor for the brevity-related features, except for the most theory-oriented educational track. Consequently, we argue that non-standard writing on social media platforms should not be operationalized as one comprehensive cluster of deviations from the formal writing standard, but rather as different subsets of non-standard features that, by serving different purposes, appeal to a different extent to different groups of youngsters and consequently display distinct sociolinguistic patterns. In other words, although Flemish adolescents may have access to the same pool of non-standard markers, they do not share one and the same social “digilect”.