Two prominent theories of alignment (priming and grounding) are tested in human–human text-only computer interactions. In two experiments, dyads of strangers and dyads of friends conducted conversations using Instant Messenger. These conversations were either neutral in nature or interlocutors were told to disagree on a particular topic. Conversations were assessed for paralinguistic, linguistic, semantic, affective, and typographical alignment. Results show distinct differences in alignment patterns dependent on conversational dynamics. Grounding theory is supported and discussion includes examining how nonverbal cues are translated into text-only conversation.