We propose that speech comprehension involves the activation of token representations of the phonological forms of current lexical hypotheses, separately from the ongoing construction of a conceptual interpretation of the current utterance. In a series of cross-modal priming experiments, facilitation of lexical decision responses to visual target…