The issues related to the polymodality of perception and various forms of intermediality are widely discussed in linguistics whereas the aspect of their interrelation is seldom considered. This research is aimed at tracing the participation of various modes of perception in the process of interaction with the world, the reflection of this process in our consciousness and memory in the form of various images up to the manifestation of those images in the intermedial relations in fiction text. The study draws on cognitive semiotics, text linguistics, polymodality and intermediality. The theoretical analysis sugessts that the polymodality of perception results in the formation of synesthetic mental images represented in the phenomenon of intermediality. The empirical material of the research includes text fragments from the works by Virginia Woolf, James Joyce, Joyce Cary and Adam Thorpe containing samples of linguа-artistic and linguа-musical techniques of writing as forms of intermediality. The intermedial analysis of the empirical material shows how the polymodality of perception determines the specificity of the authors’ creative thinking represented in the choice of language means that activate musical and visual images in the readers’ minds. As a result, such fiction texts are constituted as related to a specific art whose forms and devices are used in it. The analysis has also revealed that in the process of verbal representation of musical or artistic images language exploits all forms of its semiotic signs: symbols (for naming colours, shades of light, sounds of music), indices (allusions to works of art or their authors) and icons (the use of forms and devices characteristic of other arts). The study confirms the productivity of exploring intermedial interaction in the cognitive and semiotic perspectives to demonstrate the relations between perception, conceptualization and the embodiment of the results in fiction text.