Using affective knowledge to generate and validate a set of emotion-related, action words

摘要: Emotion concepts are built through situated experience. Abstract word meaning is grounded in this affective knowledge, giving words the potential to evoke emotional feelings and reactions (e.g., Vigliocco et al., 2009). In present work we explore whether differ extent which they ‘specific’ knowledge. Using a categorical approach, an ‘context’ created, it possible assess proportionally activate knowledge relevant different states ‘sadness’, ‘anger’, Stevenson, Mikels & James, 2007a). We argue that method may be particularly effective when assessing of action Schacht Sommer, study 1 use constrained feature generation task derive set participants associated with six, basic (see full list Appendix S1). Generation frequencies were taken indicate likelihood would state had been paired. 2 rating was used strength association between six most frequently generated, or ‘typical’, corresponding emotion labels. Participants presented series sentences, (typical atypical) labels paired e.g., “If you feeling ‘sad’ how likely act following way?” … ‘cry.’ Findings suggest typical associations robust. always gave higher ratings vs. atypical label pairings, even (a) direction manipulated (the verb appeared first sentence), (b) behaviours performed by rater themselves, others. Our findings emotion-related vary for states. When measuring grounding, then appropriate conjunction unimodal measures, ‘magnitude’ Newcombe 2012). Towards aim provide words, accompanied frequency data, show strongly each evokes