The Impact of Semantic Neighborhood Density on Semantic Access

Theories of word recognition agree that multiple word candidates are activated and compete for recognition when a word is read or heard. Orthographic and phonological density metrics have provided significant, but partial, accounts of this activation and competition process. We extend this approach to examine the impact of semantic neighborhood density on visual semantic access. We compare measures of semantic neighborhood density derived from three distinct semantic representations: elicited features, associations, and text-based co-occurrence measures and examine whether embellishments such as frequency-weighting and distinctiveness-weighting improve measures of semantic neighborhood density. According to some metrics, responses are faster and more accurate in dense semantic neighborhoods, but other metrics show exactly the opposite pattern. We propose that distant neighbors facilitate semantic access but near neighbors inhibit semantic access.