Techniques for dynamically selecting example passages to output on an electronic device in response to a user selecting a word or other portion of a content item are described herein. In one example, a user selects a word from a rendered electronic book. In response, the device outputs an entry for the word from a reference work (e.g., a definition from a dictionary, etc.). In addition, the device outputs passages that include the word from other electronic books stored on the device. These passages are more likely to be recognized and appreciated by the user than generic passages, since the dynamically-selected passages are found in electronic books that the user has previously acquired and potentially previously read.