No unimodal vestibular cortex has been identified in the human brain. Rather, vestibular inputs are strongly integrated with signals from other sensory modalities, such as vision, touch and proprioception. This convergence could reflect an important mechanism for maintaining a perception of the body, including individual body parts, relative to the rest of the environment. Neuroimaging, electrophysiological and psychophysical studies showed evidence for multisensory interactions between vestibular and somatosensory signals. However, no convincing overall theoretical framework has been proposed for vestibular–somatosensory interactions, and it remains unclear whether such percepts are by-products of neural convergence, or a functional multimodal integration. Here we review the current literature on vestibular–multisensory interactions in order to develop a framework for understanding the functions of such multimodal interaction. We propose that the target of vestibular–somatosensory interactions is a form of self-representation.