DETECTING UNNATURAL LANGUAGE IN INDONESIAN LEARNING APPS

Abstract

Language learning apps are increasingly popular for acquiring new languages, including Indonesian. However, there are concerns about the unnaturalness of sentences presented by these applications, often due to direct translations from English. This study examines unnatural language patterns in popular Indonesian learning apps, such as Duolingo, Pimsleur and Glossika, by analyzing sentences and comparing them with natural Indonesian expressions. This research uses a qualitative analysis to categorize the patterns related to grammar, vocabulary, and contextual appropriateness. The findings aim to highlight common unnatural patterns in sentences and provide recommendations for improving linguistic accuracy in digital language learning tools.