Modelling parsing constraints with high-dimensional context space

Deriving representations of meaning has been a long-standing problem in cognitive psychology and psycholinguistics. The lack of a model for representing semantic and grammatical knowledge has been a handicap in attempting to model the effects of semantic constraints in human syntactic processing. A computational model of high-dimensional context space, the Hyperspace Analogue to Language (HAL), is presented with a series of simulations modelling a variety of human empirical results. HAL learns its representations from the unsupervised processing of 300 million words of conversational text. We propose that HAL's high-dimensional context space can be used to (1) provide a basic categorisation of semantic and grammatical concepts, (2) model certain aspects of morphological ambiguity in verbs, and (3) provide an account of semantic context effects in syntactic processing. We propose that the distributed and contextually derived representations that HAL acquires provide a basis for the subconceptual knowledge that can be used in accounting for a diverse set of cognitive phenomena.