Key points are not available for this paper at this time.
Draft version 1. This paper has not been peer-reviewed. Word associates are commonly collected and employed in cognitive and clinical research, yetthe precise reasons why a particular word is activated as an associate for a given cue remain unclear.We examined the source of responses in a word association task using linguistic and sensorimotorrelationships between words as measures of relatedness, in order to assess whether responses are morelikely to be driven by the cue or by preceding associates. We asked participants to produce up to 20associates for each cue and analysed the strength of relationships and latency for each response.Results showed that word association responses had strong effects of local response chaining, where agiven associate was more likely to be related to the preceding associate than to the cue itself. Ingrowth curve analysis of the timecourse of producing responses, we found little influence of the cueon response times. Instead, the strongest source of facilitation was the sensorimotor and linguisticrelationship of each response with its immediately-preceding associate. These findings suggest thatlinguistic and sensorimotor information underpins word association mechanisms, whereby localchaining from the most recent response is the primary driver of which new associates are activatedand produced. Results support the linguistic-sensorimotor basis of semantic memory as well astheories that people search memory using local rather than global constraints, with implications forusing existing word association norms to investigate semantic memory structure and for the study ofsemantic relatedness in language production and comprehension.
Dymarska et al. (Fri,) studied this question.