J. Behuv. Thu. & Psy chr ur . Vol. 9, pp. 121.124 . Per gam”” Pr c,* zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Ltd., ,978. Pnnted Ill Great Brltaln Mx)S-7908/78/0643-0121 $02.00/O PROCESS STUDIES IN LANGUAGE CONDITIONING-II: THE ROLE OF SEMANTIC RELEVANCE IN CONDITIONING NEGATIVE EMOTIONAL RESPONSES IAN M. EVANS and ARNOLD R. WEISS Department of Psychology, University of Hawaii, Honolulu Summary-Continuing an earlier inquiry into the nature of language conditioning, undergraduate students were assigned to one of six groups in a 3 x 2 factorial design in which type of UCS (neutral, negative related, negative unrelated) and imagery were manipulated and the effects on attitude towards chocolate and jelly beans before and after conditioning were examined. Conditioning involved pairing the word “chocolate “with theappropriategroupof UCS words. The two groups of negative words were equally negative but were either related to the concept of chocolate (e.g. pimples) or unrelated (e.g. fever). The results revealed no effect due to imagery ability, but a significant change in attitude towards chocolate in the group conditioned with related negative words. The results indicate the importance of semantic relevance in the conditioning of negative emotional responses on the basis of UCS words that evoke these responses. In a previous study of language conditioning (Weiss and Evans, 1978) it was shown that an anxiety response could be more effectively reduced when the verbal representation of the phobic object was paired systematically with words connoting calmness than when paired with words of simply general positive affective value as is the usual practice (Hekmat and Vanian, 1971). The demonstrated importance of the exact nature of the UCS words is in ac- cordance with Staats’ (1968) model of language conditioning which involves transferring to the CS word the emotional response evoked by the ucs. Despite circumstantial evidence of this kind, the precise process underlying language con- ditioning remains unclear as the UCS words might be learned directly as paired associates. An interesting experiment would be to compare words chosen not for appropriateness of emotional quality, as in the previous study, but words chosen for appropriateness of thematic quality. It was hypothesized that UCS words, equal in their emotional properties, would be differentially effective in the language con- ditioning paradigm according to their thematic relevance to the CS word. In order to keep the study somewhat analogous to clinical treat- ment, the CS word chosen was chocolate, on the ground that aversion therapy for obesity or other consummatory excesses is designed primarily to diminish the positive responses to the abused substances. It was further argued that if the subjects were acquiring verbal associations with strong thematic links, the altered value of the CS word might come about through associated negative sensory images evoked by the word pairs. Thus subjects with good imagery should show greater change following exposure to thematically related negative words than subjects with poor imagery. METHOD zyxwvutsrqponmlkjihgfedc Subjects The subjects were 58 University of Hawaii undergraduate Requests for reprints should be sent to Ian M. Evans, Department of Psychology, University of Hawaii, 2430 Campus Road, Honolulu, Hawaii 96822. 121