Active perspective-taking (PT) has been argued to improve performance on social coordination tasks (Galinsky, Wang, & Ku, 2005). In most such tasks individuals are free to use any natural form of communication, making it difficult to differentiate between the effects of the empathic and cognitive aspects of PT. The Experimental Semiotics research paradigm (Galantucci & Garrod, 2011) provides social coordination tasks in which freedom of communication is reduced in a way that minimizes emotional communicative exchange. Pairs of participants played a cooperative guessing video-game in which they could communicate only via limited graphical exchanges. The roles of Guesser and Signaler either alternated, providing opportunities for PT, or remained static, limiting the opportunities for PT. PT was found to enhance performance, suggesting that it facilitates social coordination even when it operates primarily at a cognitive level. The methods of Experimental Semiotics provide new opportunities for isolating cognitive processes in human interactions.