The nature of inner language has long been under the scrutiny of humanities, through the practice of introspection. The use of experimental methods in cognitive neurosciences provides complementary insights. This chapter focuses on wilful expanded inner language, bearing in mind that other forms coexist. It first considers the abstract vs. concrete (or embodied) dimensions of inner language. In a second section, it argues that inner language should be considered as an action-perception phenomenon. In a third section, it proposes a revision of the « predictive control » account, fitting with our sensory-motor view. Inner language is considered as deriving from multisensory goals, generating multimodal acts (inner phonation, articulation, sign) with multisensory percepts (in the mind's ear, tact and eye). In the final section, it presents a landscape of the cerebral substrates of wilful inner verbalization, including multisensory and motor cortices as well as cognitive control networks.