2021

Breaking Out of the Mold of Conventional Education:
An Attempt at Self-Determined Digital Language Learning

Inspired by the principles of Montessori education I conceptualized and developed a software application for self-determined language learning that uses computational methods to liberate users from the constrained practices of conventional education still found in many modern learning environments today. Instead of having to learn a language in a strictly linear way, the application allows learners to aquire a foreign language on their own terms: Using the app, users are constantly given choices to determine their personal learning path. To enable this, I designed algorithms that ensure no matter what choices are being made, the structure of the curriculum and the learning content of lessons constantly adapt accordingly. This creates a dynamic learning experience where learners are free to explore, yet new lessons always coherently connect to what has been learned before.

The work is an ongoing research project. At present, it is more of an experimental software platform for the development of computational methods for non-linear language learning. Usability and user experience questions are only barely adressed at this point and may play a larger role in the future.

Concept graphic illustrating the structural make up of the application. This state of the UI is supposed to reveal the technical workings of the application and will likely change with ongoing progress as usability aspects will play a greater role in the development of the application.

With ongoing progress the curriculum dynamically adapts to the individual choices of the learners. This results in an entirely different language course for every user.

Prototype created as a web application with a Node.js backend running in a Browser

Main Screen of the Prototype

In the center of the web app interface lies the course structure. Instead of having to passively follow a linear course structure, learners can take different learning directions and actively construct their personal language skill. This means that depending on the choices learners make, every course structure will look differently, reflective of the learners individual abilities. White lessons have been already finished, Gray lessons are yet to be taken. When a new lesson has been accomplished, an algorithm analyzes the learners newly expanded knowledge and uses it to find new lessons that fit best with the learners current skill and adds them to the course structure.

Watch the technical prototype in action


The main design work of this project has so far been the development of an algorithmically aided learning principle and was thus rather technical. As of right now the interface design is intentionally kept generic and technical and has been created with functionality in mind rather than usability. At the moment it has been designed in a way to provide insights into the workings of the application so it becomes clear what decisions it makes. Many elements (icons, lesson titles and descriptions) still act as placeholders and do not reflect real meaning. Only once the technical principle has matured to a solid state, usability aspects will start to play an important role in the design process and many aspects of the interface may change entirely.

Lesson Overview

When selecting a lesson, a pop-up window will display relevant information such as an introduction and explaination of the subject matter the learner will face in this lesson. Additionally, a sidepanel can be opened that reveal which sentences will be used in the lesson to teach its concepts. These sentences have been algorithmically selected and reflect the individual skills of the learner that have been previously acquired. If the learner would have decided to take a slightly different learning path before this lesson, these sentences may differ and reflect this path.

Lesson overview with placeolder icon, title, and description. A list of algorithmically selected sentences to be used as teaching material for providing context for the concepts taught in this lesson are shown on the right

Learning Session

Entering the learning session, learners are asked to continuously answer quiz questions and complete tasks to train themselves on the new subject matter. In the future, a wider range of question types should be used, such as cloze tests, audio recognition tests, and writing tasks.

Custom Lessons

because of its unique approach of dynamically assembling lessons with the help of algorithms, learners can easily create their own sets of custom lessons with topics that are of great interests to them. These lessons fuse with their previous learning process and seamlessly integrate into the existing dynamic curriculum

Due to its unique technical architecture courses for other languages can be added without a lot of effort: