In this article, we consider influential theories and models of simultaneous interpreting through the prism of current theories and findings in psycholinguistics. We review evidence suggesting that meaning is routinely accessed during comprehension, and, on this basis, suggest that it is unlikely that interpreters produce translation equivalents without accessing meaning. We also reflect on data showing cross-linguistic activation in two-language settings, and suggest that such cross-linguistic activation regularly takes place in simultaneous interpreting, without interfering with the processing of meaning. Finally, we consider the role of prediction in simultaneous interpreting. Evidence suggests that prediction regularly takes place during simultaneous interpreting and may facilitate the interpreting process. On the basis of this theoretical analysis, we consider that simultaneous interpreting involves a synergy, rather than just an overlap, of comprehension and production. We outline the pedagogical implications of our conclusions, and explain why viewing simultaneous interpreting from this new perspective may open up directions for future research.