Abstract
The semantic theory of emotion words recently proposed by Johnson-Laird and Oatley (1989) was empirically investigated in three studies. In all three studies, I assessed, for different samples of German nonbasic emotion words: (a) subjects' judgements of the conditional probability of experiencing basic emotions, given the experience of nonbasic ones; and (b) their beliefs about whether it is possible to experience nonbasic emotions without also experiencing basic emotions. In Study 1, I examined the proposed semantic relations between 48 nonbasic and their defining basic emotion words, as well as 14 of the proposed semantic relations among nonbasic emotion words. In Study 2, these tests were repeated using object-focused test sentences. In Study 3, the semantically based relations among 12 emotions were compared to all of the nonsemantic relations existing among these emotions, and the theory was additionally tested by examining self-ascriptions of emotion words in concrete situations (hypothetical scenarios). I found that (1) the semantic theory of emotion words proposed by Johnson-Laird and Oatley (1989) was consistently unsupported for the disgust-derivatives, and that in a substantial number of cases a second nonbasic emotion was nearly as prominent as the modal one (Study 1); (2) the conditional probability and possibility relations between allegedly semantically connected emotion pairs were frequently no stronger or even weaker than those between semantically unconnected ones (Studies 1, 3); (3) in terms of absolute judgements, the data fell considerably short of the theoretically predicted results (all studies), particularly when (4) object-focused emotion words were used (Study 2); and (5) no more favourable results were obtained when subjects' self-ascriptions of emotion words in concrete situations were examined (Study 3). These findings call in question Johnson-Laird and Oatley's semantic theory of emotion words and potentially also their theory of emotions.