Abstract:

Learner-centered learning design has become widespread as well in the field of Japanese language education to meet the diversifying needs of learners. Collaborative learning has become the center of the classroom, while self-paced learning, such as grammatical practices, is often positioned as out-of-class learning.

However, most of the materials available to self-learners for grammar retention are pattern drills. To learn something means to acquire both the ability to have knowledge retention and knowledge transfer. Prior research suggested that transfer can be sub-categorized into near and far transfers. Pattern drills are practices that promote near transfer. For example, the practice of substituting the given words like ka-u (to buy), tabe-ru (to eat) into the sentence pattern koto ga deki-ru (can do -ing) is an exercise to promote near transfer. Far transfer occurs when learners retrieve some sentence patterns from their knowledge to another situation where they want to say something. (e.g., Where can I buy a ticket? (at the station), I cannot eat shrimp. I am allergic to shrimp. (at a restaurant))

Therefore, we designed and developed a self-learning management system that allows learners to learn new sentence patterns and incorporate retrieval practice in a different context. This system stores tasks made by experts in Japanese language education and presents them to the learner. Those tasks are to say what is required in a given situation using sentence patterns that they should already know or that are appropriate to start learning for the learner. Each learning object of sentence pattern included in the task has a URL link to learning resources, making it easy for learners to learn new sentence patterns or recall them if they have forgotten something during retrieval practice. The system marks the learning object as learned once the learner completes the task. In the task-selecting procedure, the system first calculates their similarity to other learners using a recommendation engine, then sorts them in order of likelihood to be preferred.

Next, the system examines the sentence patterns in each listed task to see if any of the sentence patterns in each are:
1) already learned by that learner or
2) a new learning item that the learner satisfies the learning prerequisites.

When the top three tasks that meet the above criteria are selected, they appear to the learner. The learner can choose any of the three they wish from them.

A series of evaluations showed the effectiveness of this system. As a result of an operational acceptance test, the system performed as expected. The evaluation test by specialists in Japanese language education showed that they successfully registered their own designed tasks to the system. They also gave positive feedback on the system. The evaluation test by Japanese language learners proved that they could accomplish tasks with this system. Furthermore, their feedback contained "ease of imagining situations" and "feelings of fun."

Keywords:

LMS, language learning, retrieval practice, recommendation, learner-centered, task-based.