Prior research has suggested that emotion and working memory domains are integrated, such that positive affect enhances verbal working memory, whereas negative affect enhances spatial working memory (Gray, 2004; Storbeck, 2012). Simon (1967) postulated that one feature of emotion and cognition integration would be reciprocal connectedness (i.e., emotion influences cognition and cognition influences emotion). We explored whether affective judgments and attention to affective qualities are biased by the activation of verbal and spatial working memory mind-sets. For all experiments, participants completed a 2-back verbal or spatial working memory task followed by an endorsement task (Experiments 1 & 2), word-pair selection task (Exp. 3), or attentional dot-probe task (Exp. 4). Participants who had an activated verbal, compared with spatial, working memory mind-set were more likely to endorse pictures (Exp. 1) and words (Exp. 2) as being more positive and to select the more positive word pair out of a set of word pairs that went 'together best' (Exp. 3). Additionally, people who completed the verbal working memory task took longer to disengage from positive stimuli, whereas those who completed the spatial working memory task took longer to disengage from negative stimuli (Exp. 4). Interestingly, across the 4 experiments, we observed higher levels of self-reported negative affect for people who completed the spatial working memory task, which was consistent with their endorsement and attentional bias toward negative stimuli. Therefore, emotion and working memory may have a reciprocal connectedness allowing for bidirectional influence.