Reconstructing maps from text

• Main Authors: Avery, J.E (Author), Goldstone, R.L (Author), Jones, M.N (Author)
• Format: Article
• Language: English
• Published: Elsevier B.V. 2021
• Subjects: article; Artificial intelligence; Cognitive systems; Co-occurrence; Embodiment; Experimental control; human; human experiment; language; News corpora; semantic memory; Semantic memory; Semantic Model; Semantic representation; Semantics; Spatial cognition; Statistical sources; Vector-space models; Word co-occurrence
• Online Access: View Fulltext in Publisher

 Previous research has demonstrated that Distributional Semantic Models (DSMs) are capable of reconstructing maps from news corpora (Louwerse & Zwaan, 2009) and novels (Louwerse & Benesh, 2012). The capacity for reproducing maps is surprising since DSMs notoriously lack perceptual grounding. In this paper we investigate the statistical sources required in language to infer maps, and the resulting constraints placed on mechanisms of semantic representation. Study 1 brings word co-occurrence under experimental control to demonstrate that standard DSMs cannot reproduce maps when word co-occurrence is uniform. Specifically, standard DSMs require that direct co-occurrences between city names in a corpus mirror the proximity between the city locations in the map in order to successfully reconstruct the spatial map. Study 2 presents an instance-based DSM that is capable of reconstructing maps independent of the frequency of co-occurrence of city names. © 2021 The Authors