While there is a growing body of psycholinguistic experimental research on mappings between language and vision on a word and sentence level, there are almost no studies on how speakers perceive, conceptualise and spontaneously describe a complex visual scene on higher levels of discourse. This book explores the relationship between language, eye movements and cognition, and brings together discourse analysis with cognitively oriented behavioral research. Based on the analysis of data drawn from spoken descriptive discourse, spontaneous conversation, and experimental investigations, this work offers a comprehensive picture of the dynamic natures of language, vision and mental imagery. Verbal and visual data, synchronised and correlated by means of a multimodal scoring method, are used as two windows to the mind to show how language and vision, in concert, can elucidate covert mental processes.