Spatially coded semantic information about geographical terms

The cognitive representations supporting our conceptual knowledge of geographical terms such as country and city names have rarely been explored. It has been hypothesised that semantic information about geographical terms is spatially coded (Crutch & Warrington, 2003). The aim of the current study was to assess whether the findings from this previous single case study generalise to other patients and to different classes of geographical terms. Two globally aphasic stroke patients were administered a series of spoken word to written word matching tasks in which they had to repeatedly identify places from the same geographical area (e.g. Norway, Sweden, Denmark, Finland) or different geographical areas (e.g. Norway, India, Spain, Brazil). The patients replicated previous findings by showing that response accuracy was significantly affected by the real-world geographical proximity of the places being examined. Error rates were significantly higher when identifying geographically close as compared with geographically distant countries and cities. They were also significantly less accurate identifying the names of towns in Greater London drawn from the same geographical area (e.g. northwest, northeast, southwest, southeast) than from different geographical areas. Furthermore, it was found that geographical proximity effects occurred even between different types of geographical terms (e.g. England, London, Surrey, Thames). These results suggest that the comprehension of spoken and written geographical terms is directly influenced by their real-world location.