Abstract When subjects are read a list of stimulus words and are asked to g enerate an associate to each word, recall for associates given to emotionally toned stimulu s words is poorer than recall for associates given to stimulus words of a neutral affective tone. The present study reports the results of 4 experiments designed to investigate this finding. Expe riment 1 replicated the effect and showed that when subjects associate to an emotionally toned stimu lus word, that associate is likely to be a word with a congruent emotional tone. Experiment 2 demonstrated that reducing the number of emotionally toned stimulus words in the word list re sulted in an improvement in recall for associates generated to these words. In the third exp eriment recall for associates generated to emotionally toned stimulus words was poorer if the s timulus words were similar in meaning (eg. worry, anxiety, panic, fear) than if they were diss imilar (eg. anxiety, death, passion, greed). Experiment 4 found that stimulus word properti es other than emotionality could be manipulated to demonstrate a differential recall. The resu lts throw doubt on an explanation involving impaired consolidation of the associate due to the a rousing properties of the emotional stimulus words. Rather, they are best explained in terms of th e inability of the retrieval system to discriminate amongst competing sets of responses which are s emantically and emotionally congruent.Resume Lors de la lecture d'une liste de termes a des sujets, en leur deman dant d'associer a chacun un autre mot, I'evocation de mots associes a des termes de caractere emotionnel est moins vive que celle de mots associes a des termes neut res. Quatre experiences ont evalue ces conclusions. L'experience 1 a demontre que lorsque les sujets associent un mot a un terme de caractere emotionnel, ce mot a tendance a renfermer la meme valeur emotionnelle. Les resultats de l'experience 2 indiquent qu'une diminution du nombre de termes de caractere emotionnel de la liste rehausse l'evocation de mots associes. Au cours de la troisieme experience, I'evocation de mots associes aux termes de caractere emotionnel etait moins vive lorsque ces termes avaient un sens semblable (p. ex., inquietude, angoisse) que lorsqu'ils n'etaient pas semblables (p. ex., angoisse, mort). Lesresultats de l'experienc e 4 ont indique qu'il etait possible d'obtenir differents resultats d'evocation en manip ulant d'autres caracteristiques des termes, a part leur caractere emotionnel. La disc ussion des resultats porte sur l'incapacite du systeme de recuperation de discerner ent re des groupes de reponses en concurrence les uns par rapport aux autres.In an experiment examining recall for self generated associates to stimulus word s, Jones, O'Gorman, and Byrne (1987, Experiment 2) presented subjects with 30 stimulus wor ds: 10 pleasant, 10 unpleasant, and 10 of a neutral affective tone. Subjects were aske d to generate an associate to each word and, following this, were again presented with each stimu lus word and asked to recall the associate previously paired with it. The results showed tha t the recall of associates given to affectively toned stimulus words (both pleasant and unpleasa nt) was poorer than the recall of associates given to affectively neutral stimulus words. Simi lar findings have been reported by other workers (Levinger & Clark, 1961; Parkin, Lewinsohn & Folk ard, 1982; Rossmann, 1984; Sinton, 1981).Levinger and Clark (1961) argued that some inhibitory process involving the emotional words might account for this finding but they did not elaborate on how this migh t occur. Other authors, (Eysenck & Wilson, 1973; Parkin et al., 1982), have suggested that the effect is due to the arousing properties of the emotional words interfering with the consolida tion of the associate. A number of studies have investigated this hypothesis starting with the idea that arousal impairs immediate recall but facilitates long term recall (see Eysenck, 1976, for a review). …