This PhD project focuses on adaptivity within complex language tasks (for second and foreign language acquisition) by means of simulated dialogues with chatbots. By complex tasks we mean language learning activities in which language is used in a functional way and in which language skills, knowledge and attitudes are acquired in an integrated way. Complex language tasks go beyond the closed exercise, and come today within reach of those who want to practice language skills in a personalized way without having to rely on native speakers via chatbot technology.

However, Linguineo has noticed that users of its online language offerings centred around chatbot technology mainly drop out when the content is too difficult or convenient. In order to match the challenge within tasks to the level of the learner, there is a need for adaptivity. Although adaptivity can be achieved to a large extent within closed exercises, it is an unsolved problem within complex language tasks. Moreover, in a blended learning context, the question is what is the ideal mix between support offered through technology (adaptivity) and support offered by the language teacher.

The project therefore aims to (1) conceptualize and validate a data-driven framework for realizing adaptivity within complex language tasks, (2) evaluate the effects of adaptivity in online learning environments for complex language tasks, and (3) examine how findings from (1) and (2) can lead to better support for the teacher. More specifically, the project will (1) make language acquisition in complex language tasks measurable in real-time, (2) make the learning environment adaptive between and within tasks, and (3) align the support provided by the technology and the instructor.