Measuring the linguistic similarity of discourse from open-world role-playing games to the real world through an additive multidimensional analysis

Abstract Digital games can provide rich sources of second language (L2) input; however, the extent to which gaming discourse is similar to real-world discourse has been a topic of debate in the computer-assisted language learning community. To quantitatively measure the extent to which gaming discourse shares linguistic similarity with real-world discourse, this study reports the findings of an additive multidimensional (MD) analysis comparing registers in open-world role-playing games to real-world registers using Biber’s (1988) Dimension 1: ‘Involved versus Informational Production.’ Results indicate that gaming discourse provides extensive language exposure that shares much linguistic similarity across a wide range of real-world contexts. Importantly, however, these similarities only become salient when the situational characteristics of gaming discourse are considered and parsed appropriately into register categories.