The purpose of the present study was to examine empirically the influence of past memories on current perceptions of relationship functioning. Michael White's therapeutic metaphor, which posits that individuals have the ability to overcome the constraining effects of socially-constructed meanings or narratives that negatively impinge on their lives, served as the primary conceptual lens. Research from the fields of cognitive psychology and neuro-lingusitic programming also provided crucial conceptual structure. In the current experimental study, three groups of university students participated in a memory recall exercise that attempted to access multi-sensory information about positive, negative or unspecified memories of experiences with a romantic significant other. Participants completed a dyadic measure that measured the level of perceived conflict with the significant other, and a mood scale based on bi-polar mood states. Mood was measured before and after the positive, negative, and unspecified treatment conditions. It was predicted that positive memory recall would result in lower scores of perceived conflict than the other two conditions; that memory type would affect intensity of affect; and that post-treatment mood would relate to the emotion associated with the type of memories recalled. The results indicated that type of memory recalled (positive, negative, unspecified) had a temporary effect on intensity of affect elicited, but did not affect ratings of perceived conflict. Type of memory recalled was not significantly related to post-treatment mood. Pre-treatment mood was the best predictor of post-treatment mood. The results are discussed in terms of future implications for the use of language to facilitate memory recall and affect mood state.