On the language of space: Neurocognitive studies in blind and sighted individuals

Abstract

In 1688 Molyneux asked the question whether a blind man who knew the difference between a sphere and a cube from touch, with his sight restored would be able to distinguish between them based on vision. One interpretation of Molyneux’s question is whether spatial representations are modality-specific. In a series of experiments the nature of spatial mental representations and the link with spatial language and imagery was studied. First the current literature on this topic was reviewed. Spatial information can be obtained through modality-specific input (such as vision, touch and audition, etc.), but also through amodal verbal input. A model is proposed in which these inputs feed into a supramodal representation of spatial information. A supramodal representation contains the purely spatial information, but maintains links with input modalities. This model explains how different sources of input can yield behaviorally similar results. In line with this, results from early blind people, without any visual memories, indicate that they perform quite well on various spatial and navigation tasks, suggesting that vision is not the only modality that can convey spatial information. Following the model described above a series of empirical studies were conducted. The first part of this thesis deals with simple spatial sentences, such as “the ball is above the shoe”, and reference frame processing. Reference frames determine the point of view for communication and are very important for successful communication. Potentially, reference frames allow for communication between different sensory modalities, as illustrated by Molyneux’s question. This part discusses how these simple spatial sentences are represented using, verbal, visual and haptic comparisons. Spatial language hardly ever occurs in isolation but should also be considered in a larger context such as talking about space and getting from one place to the other. Therefore, the second part of this thesis examined the processing of complex spatial descriptions. The reference frames that clarify how objects are located support different perspectives and possibly different representations, both behaviorally and neurally. This part describes neuroimaging studies that investigate whether the representations built up from these complex descriptions are indeed spatial, that is whether they represent metric information about the locations described. Finally, the third part of this thesis focuses on how compelling such spatial information is when understanding language. Can spatial information also be accessed through language that is not necessarily spatial at first? This question links spatial representations to the influential idea of grounded cognition that language is understood by linking it to perceptual simulations. Three chapters directly tested early blind participants. This special participant group allowed to determine the role of visual experience in representing spatial information. Together the results from this thesis support the idea the spatial information can be obtained from multiple different input modalities and is represented supramodally. This supramodal representation can be addressed using spatial language. Moreover, vision is not a prerequisite to build up these supramodal representations, as demonstrated by the highly similar results for blind and sighted participants.