The authors develop a new approach to the brain mechanisms of speech and perception  based on the hypothesis that speech and perception are forms of one common structure — «language structure». In contrast with speech, which often appears the most developed and «genuine» form of language, perception seems only indirectly linked to the latter.  For this reason, the article focuses on perception as a particular form of language, and it analyses recent theoretical and empirical research. Perception can be described as a «visual language» with its own alphabet (characters), morphology (parts of the «visual speech» — visual syllables, words and combinatory rules) and with its own visual syntax (punctuation marks).