ABSTRACT
This study examines a large corpus of online gaming chat and avatar names to explore gender differences in virtual world (VW) language use. In particular, we examine the relevance of socio-linguistic observations of gender in face-to-face conversation to the contemporary space of VW chat interactions in online gaming environments. In addition, we study the relationship between a player’s gender and naming decisions for online avatars in terms of linguistic observations based on sound symbolism. Analysis shows that many of the existing socio-linguistic claims about gender and speech also hold true in the VW for many of the categories posited (e.g. swearing, hedging, empathy). For avatar naming conventions, results showed that these rules could predict gender with a high average accuracy (>0.7) for both males and females. Applying the same rules to avatar names from individuals whose real world (RW) and VW gender were different still enabled detection of RW gender at a similar high rate of accuracy, despite mismatched gender. We conclude that the predictions of socio-linguists about gender-linked behaviors and decisions in RW conversational interactions largely transfer across subcultures to VW environments such as online gaming chat and avatar naming conventions.
