‘Am I saying it wrong?’ Progressivity-related troubles and instructional opportunities in child-robot L2 interaction

Abstract

Participants’ orientation to progressivity of action sequences is a fundamental feature of human social interaction, but less is known about how progressivity is maintained in human-robot interaction (HRI). We explore this by drawing on c. 14 hours of video recordings showing small groups of primary school children interacting with Nao, a programmable humanoid robot. Facilitated by a teacher, the children in our data are completing a short robot-assisted language learning lesson aimed at training English vocabulary and oral skills at a Swedish-speaking school in Finland. We investigate how the teacher and the pupils address emerging troubles in a word repetition sequence which the robot is programmed to carry out with one pupil at a time. Our analysis focuses on two kinds of troubles related to sequence closure: the robot’s so-called ‘third’ turns that either 1) do not ratify the pupil’s just-prior word repetition as ‘correct’ or 2) are (treated as) incongruent within the sequential context. We show how the human participants make sense of such conduct, recruit the teacher’s assistance to secure closure of the activity sequence, and orient to pronunciation instruction in situated ways. The results shed light on how children accommodate to, and are socialised into, human-robot interaction.