This study examined the evolution of visual perspective-taking skills in relation to the comprehension and production of first, second and third person pronouns. Twelve French-speaking and 12 English-speaking children were observed longitudinally from 1;6 until they had acquired all pronouns and succeeded on all tasks. Free-play sessions and three tasks were used to test pronominal competence. Four other tasks assessed Level-1 perspective-taking skills: two of these tasks required the capacity to consider two visual perspectives, and two others tested the capacity to coordinate three such perspectives. The results indicated that childrenʼs performance on perspective-taking tasks was correlated with full pronoun acquisition. Moreover, competence at coordinating two visual perspectives preceded the full mastery of first and second person pronouns, and competence at coordinating three perspectives preceded the full mastery of third person pronouns when a strict criterion was adopted. However, with less stringent criteria, the sequence from perspective taking to pronoun acquisition varied either slightly or considerably. These findings are discussed in the light of the ‘specificity hypothesis’ concerning the links between cognition and language, and also in the context of the recent body of research on the childʼs developing theory of mind.