The traditional formulation of the relation of language and thought has been in terms of linguistic determinism-the view that a language's grammar constrains the way its speakers perceive the world and mentally represent what they perceive. Best exemplified by the "Whorfian Hypothesis," this view has received little empirical support in the psychological literature. C.Y. Chiu (of Hong Kong University) and I have proposed an alternative to the Whorfian Hypothesis-namely, that it is language use rather than language structure that affects cognition. We hypothesize that using language creates or activates mental representations, and that such mental representations can influence subsequent cognition. Ordinarily, we think about communicative messages in terms of the effects they have on their addressees. However, an implication of our hypothesis is that factors (like audience design) that influence the specific form of the message takes can also influence the way the speaker mentally represents the message's contents. Early results show that speakers' mental representations of graphic stimuli differ depending on whether they have described the stimuli to children or adults. For a discussion of this approach and some preliminary results, see Chiu, C.-y., Krauss, R. M., & Lau, I. Y.-M. (in press). Some cognitive consequences of communication. In S. R. Fussell & R. J. Kreuz (Ed.), Social and cognitive approaches to interpersonal communication. Hillsdale, NJ: Erlbaum..