Alignment in communication, the adaptation to one’s communication partner, is a pervasive phenomenon in everyday life, which occurs at various linguistic levels (Pickering & Garrod, 2004). Numerous studies have shown reliable lexical and syntactic alignment in a dialog context (e.g. Branigan et al., 2000; 2011). Many of these studies have focused on alignment at one linguistic level, in a rather short time frame, and in a controlled experimental setting. This study focuses on how lexical and syntactic alignment evolves over time, that is, over a longer stretch of discourse.

Pairs of participants played a game together in which they matched each other’s array of objects in a grid. During each round of the game, one participant (the instructor) described the position of five geometric shapes in different colors to his or her interlocutor (the slider). Instructors could see the slider’s array and thus monitor successful placement of the shapes. Once all shapes were placed correctly, participants switched roles until a total of eight rounds were completed. All pairs of participants moved the same shapes to the same locations in the same order.

To test whether participants showed alignment in a setting where neither syntax nor lexical items were experimentally manipulated, we created a baseline by comparing the utterances of instructors and sliders who did not actually play together. The results showed reliable lexical and syntactic alignment compared to the baseline. In addition, we found that lexical alignment, but not syntactic alignment, increased over the course of the game. Furthermore, lexical alignment, but not syntactic alignment, contributed to task success: the higher lexical alignment, the faster participants completed the task. We discuss the results in relation to current models of alignment.