On Anthropomorphism in Technology-Enhanced Language Learning: Does Modality Matter in Agent- Based Multimedia Instruction on L2 Idioms?

The present study aimed to satisfy a twofold purpose: On the one hand, it sought to verify the postulation that agent-based instruction could offer a compromise approach to teaching L2 idioms where form and meaning would be equally emphasized during instruction. Given that anthropomorphism has not been much under scrutiny, this research, on the other hand, sought to ascertain whether learning and retention of English idioms would be differentially impacted when two different modalities of virtual tutors —anthropomorphic and nonanthropomorphic— were present in the tutorial. To this aim, the participants of the study received instruction on 128 English idioms from human teachers, a multimedia application featuring a humanoid virtual teacher, or a piece of multimedia courseware with a nonanthropomorphic virtual tutor. Analysis of the post-intervention measures of L2 idiom knowledge revealed that agent-based instruction had proved more effective in improving both learning and retention of the target idioms among the participants. A further finding was that despite the greater motivational benefits of the humanoid virtual tutor, it had not privileged the participants, performance-wise.