Abstract: A user views a video for learning a target language. The video is displayed with a caption that the user may interact with to select a phrase of interest to the user. A translation of the phrase is determined and a definition retrieved for the phrase. The translation is provided in the user's native language, and the definition may be provided in the target language. In addition, occurrences of the phrase may be identified in the remainder of the video, and a marker indicated in a timeline of the video to indicate when in the video the phrase appears, permitting the user to easily view additional contexts of the selected phrase. The user may also hide and display various interface elements to reduce reliance on the translation provided by the language learning system.