Explaining orthographic variation in a virtual community : linguistic, social, and contextual factors

The purpose of this project is to investigate factors that can be used to explain orthographic variation in City of Heroes (CoH), a virtual community based in an online role-playing game. While a number of models of variation exist for speech, to date, no statistical models of orthographic variation in virtual communities exist. By combining traditional variationist methods with computational text processing, this project documents socially meaningful alternations in the linguistic code regarding two types of sociolinguistic variables, namely spelling and use of abbreviations. For each of the two variable types, two dependent variables are posited, i.e. the alternation between: 1) –ing and –in in durative verbal aspect marking in forms such as coming and comin, 2) –s and –z markers of plurality in words such as cats and catz, 3) abbreviated and full forms for referential abbreviation in terms such as Atlas Park and AP, and 4) abbreviated and full forms for conative abbreviations in terms such as looking for team and lft. The study investigates the role that the following factors play in explaining orthographic variation in CoH: 1) message length, 2) standardness of the immediate linguistic environment, 3) cognitive load, 4) relative proximity in the virtual space, 5) degree of message publicness, 6) experience in the community, 7) avatar gender, and 8) social group affiliation. Through mixed-effects, multivariate models, the study demonstrates that each of the predictors has some role in explaining the orthographic variability observed in the textual record of the community. Moreover, interactions between some of the predictors prove to be significant contributors to the models, which highlight the importance of addressing interaction terms in models of language variation. The findings from the study suggest that the socio-contextual meaning of particular structures in the CoH community lead authors to make linguistic choices, which are realized as alternations in the linguistic code. Finally, implications for the study of language variation in general are discussed.