Klin, Ralano, and Weingartner (2007) found transfer effects when a phrase, described as part of a note one character had left for another, was repeated across two passages. However, when the phrase was part of a note in story A and part of a conversation in story B, transfer effects were eliminated (Klin & Drumm, 2010). Klin and Drumm concluded that readers encoded the perceptual features of story characters' linguistic exchanges and that the mismatch (visual vs. auditory) eliminated transfer effects. The present experiments support this conclusion and also demonstrate that readers encode details of the social interaction that surrounds the characters' linguistic exchanges: Effects were reduced when the phrase in story A was part of a direct social interaction between the characters (e.g., phone conversation), whereas in story B, the interaction was indirect (e.g., voicemail). More generally, readers are exquisitely tuned to subtle aspects of characters' linguistic exchanges.