SingLing: Learning Languages by Singing Code-Switched Lyrics

Music is a powerful medium for memory and emotional engagement and thus provides an effective context for language learning.However, existing digital tools for language acquisition often rely on vocabulary drills or straightforward lessons that fail to sustain interests and motivation.This gap between engaging content and pedagogically effective language exposure limits the potential for intuitive, immersive, real-world learning experiences.Leveraging the widespread popularity of multilingual music and the cognitive benefits of contextual learning, we present SingLing, a web-based interactive tool that supports non-native language acquisition through code-switched lyrics.Drawing on the linguistic phenomenon of code-switching, SingLing generates a bilingual version of the lyrics in which selected words or phrases are translated and rhythmically embedded into the original song.These modified lyrics are rendered as synchronized subtitles, preserving the flow and structure of the music.With a large language model and a substitution engine, SingLing ensures that inserted translations maintain grammatical accuracy, semantic coherence, and rhythmic alignment with the original lyrics.Language learners reported higher vocabulary recognition and engagement compared to traditional methods.Participants highlighted the memorability and enjoyment of learning vocabulary through familiar musical contexts.Our results suggest that code-switched music-based interfaces like SingLing can support incidental vocabulary acquisition and cultural engagement in ways that are both intuitive and motivating.This work introduces a novel AI-enhanced multimodal system that leverages popular media to personalize language learning.By bridging linguistic theory and unique code-switched interaction, we advance the design of intelligent educational interfaces that are both culturally aware and technically innovative.