Title: Debiasing semantic analysis: the case of the English preposition to
Abstract: The paper addresses the semantic question how the meaning of prepositions can accurately be distinguished from their interpretation in actual context.Tyler and Evans (2004) argue that 'motion' and 'path' are to be excluded from the core meaning of prepositions because these features are derived from collocating elements and from our general encyclopaedic knowledge of the world.The case is made that the same reasoning applies to the general notion of 'space'.Taking the English preposition to as a case study, I raise three arguments against the assumption of a core spatial meaning.First, a corpus based analyses of the uses of to in present day English indicates that a spatial reading of to is only pertinent when the preposition is combined with other linguistic elements associated with a spatial scene of events.Second, the assumption of a core spatial meaning is not supported by diachronic data, as there is no evidence that the oldest recorded uses of to were ever restricted to a spatial configuration.Third, the cognitive underpinnings of the core meaning of to are problematic because of two reasons.In the absence of non-linguistic evidence, any semantic analysis based on these image schemas is inevitably circular.Moreover, the hypothesis that linguistic forms are merely tagged onto existing pre-linguistic image schemas is inconsistent with findings on infant cognition and typological data on spatial categories.A semasiological analysis of the to is additionally proposed, which examines the 'instrumental' meaning of to in comparison with that of towards, at, until/till and into.The meaning of to is defined as 'establisher of relationship between X and reference point Y'.