The emotional organization of autobiographical memory was examined by determining whether emotional cues would influence autobiographical retrieval in younger and older adults. Unfamiliar musical cues that represented orthogonal combinations of positive and negative valence and high and low arousal were used. Whereas cue valence influenced the valence of the retrieved memories, cue arousal did not affect arousal ratings. However, high-arousal cues were associated with reduced response latencies. A significant bias to report positive memories was observed, especially for the older adults, but neither the distribution of memories across the life span nor response latencies varied across memories differing in valence or arousal. These data indicate that emotional information can serve as effective cues for autobiographical memories and that autobiographical memories are organized in terms of emotional valence but not emotional arousal. Thus, current theories of autobiographical memory must be expanded to include emotional valence as a primary dimension of organization.
The autobiographical memory literature leaves an elementary question unanswered: Where does emotion fit into the organization of autobiographical memory? Two extreme positions could be endorsed. First, one could claim that emotion is one of the primary dimensions along which autobiographical memory is organized. Alternatively, one could argue that, although emotion colors much of our autobiographical experience, it does not serve as an organizing principle in memory. Unfortunately, the literature does not point unambiguously to either of these positions.
Some theorists have argued that any reasonably comprehensive theory of autobiographical memory must explain how emotion is incorporated into the memory (Rubin, 1998). Rubin supported this claim by noting that Proust's emotional reaction to the famed "petite madeleine" preceded both his verbal recognition of the memory and the image associated with the memory. Empirical evidence also demonstrates that mood has a profound effect on memory in general (Eich & Forgas, 2003; Ellis & Moore, 1999) and autobiographical memory in particular (Christianson, 1992; Christianson & Safer, 1996). Diarists typically report that emotional events are easier to recall than nonemotional events (W. F. Brewer, 1988; Wagenaar, 1986; but see Linton, 1986, for a different result). Conway (1989) found that emotion words were much more likely to spontaneously cue autobiographical memories than were concepts like "birthday presents" or "furniture." Finally, Pillemer, Goldsmith, Panter, and White (1988) found that one of the best predictors of the clarity of an autobiographical memory was its emotional intensity at the time of the event. These data suggest that emotion is one of the primary dimensions of organization for autobiographical memory.
However, the literature is also replete with data that suggest that emotionality is a relatively weak predictor of autobiographical memory retrieval. Rubin and Schulkind (1997c) found weak correlations between emotional ratings of word cues and the age and response latencies of the memories retrieved in response to the cues. Robinson (1976) found that emotional words elicited more recent memories and produced longer response latencies than did either objects or activity words (see also Conway & Bekerian, 1987). Finally, although Linton (1986) placed mood tone at the top of her hierarchical organization of autobiographical memory, she claimed that an emotional cue is rarely sufficient to elicit a specific memory. Furthermore, and in direct contrast with data reported by Pillemer et al. (1988), Linton (1982) found small, weak correlations between emotionality ratings taken when the event occurred and subsequent memory. These data suggest that emotional states are relatively low-level event information, analogous to the color of the shoes that one may have been wearing when one first met one's future spouse. …