When the “I” Looks at the “Me”: Autobiographical Memory, Visual Perspective, and the Self

Abstract

This article presents a theoretical model of the self processes involved in autobiographical memories and proposes competing hypotheses for the role of visual perspective in autobiographical memory retrieval. Autobiographical memories can be retrieved from either the 1^st person perspective, in which individuals see the event through their own eyes, or from the 3^rd person perspective, in which individuals see themselves and the event from the perspective of an external observer. A growing body of research suggests that the visual perspective from which a memory is retrieved has important implications for a person's thoughts, feelings, and goals, and is integrally related to a host of self-evaluative processes. We review the relevant research literature, present our theoretical model, and outline directions for future research.

When the “I” Looks at the “Me”: Autobiographical Memory, Visual Perspective, and the Self

Imagine Jane, a young college student, giving her first public presentation in front of her professor and classmates. As she begins to talk, she looks out at the audience and freezes. Unable to continue, she leaves the room and subsequently receives a failing grade for her presentation. The memory of this experience remains salient to her as she progresses through college and ultimately enters law school. Eventually, it becomes incorporated into Jane's self-concept and shapes important aspects of her identity and self-representations.

A growing body of research suggests that the visual perspective from which Jane retrieves the memory has important implications for her subsequent thoughts, feelings, and goals. Autobiographical memories can be recalled from either the 1^st person or 3^rd person perspective (also labeled the “field” or “observer” perspectives, respectively). When recalling memories from the 1^st person perspective, individuals see the event associated with the memory through their own eyes, as if they were reliving the experience. In contrast, when recalling memories from the 3^rd person perspective, individuals actually see themselves in the memory, as if they were an observer watching the remembered event (Nigro & Neisser, 1983). Research suggests that approximately one-third of autobiographical memories are spontaneously recalled from the 3^rd person perspective, and the rest from the 1^st person perspective (Nigro & Neisser, 1983; Robinson & Swanson, 1993).

Researchers from disparate areas of psychology have long recognized the significance of visual perspective. Freud (1899/1959), in his conceptualization of “screen memories,” argued that visual perspective plays a critical role in psychodynamic processes related to early childhood memories and that 3^rd person memories demonstrate the reconstructive nature of these memories. Cognitive psychologists have studied the shift from the 1^st person to the 3^rd person perspective as a function of the time interval between encoding of the event and retrieval of the memory (Berntsen & Rubin, 2006; Nigro & Neisser, 1983; Schacter, 1996). Social-personality psychologists have examined whether 3^rd person memories serve a distancing function when the individual recalls a memory of an event that is incongruent with the current self (Libby & Eibach, 2002; Libby, Eibach, & Gilovich, 2005). Clinical psychologists have discovered that traumatic events are often recalled from the 3^rd person perspective, presumably as a way to buffer the person from both psychological (McIsaac & Eich, 2004; Kenny & Bryant, 2006) and physical pain (McNamara, Benson, McGeeney, Brown, & Albert, 2005).

Together, these various lines of research suggest that visual perspective is a significant phenomenological characteristic that influences the way memories are used to regulate emotions and maintain a coherent identity over time. Despite this growing literature, however, there has been little integration of theory and research on visual perspective and memory.

Much of the research linking visual perspective and memory points to a central role for self-related processes and motives (e.g., Libby et al., 2005; Frank & Gilovich, 1989; Wilson & Ross, 2003). Yet current models of self and memory do not typically take into account visual perspective (e.g., Conway, 2005; Conway, Singer, & Tagini, 2004; McLean, Pasupathi, & Pals, 2007). Furthermore, although these models emphasize the importance of the self to memory retrieval, many do not identify and elaborate on the specific self-evaluative processes that influence the retrieval of autobiographical memories. The articulation and integration of these processes in memory retrieval will help organize the visual perspective literature and offer new directions for future research in this area. Thus, the current paper has dual aims: to articulate the self-evaluative processes that play a role in autobiographical memories and to discuss how visual perspective influences these processes in the context of autobiographical memory retrieval.

Accordingly, in this paper we will (1) review previous research on visual perspective, (2) introduce a theoretical model of the self processes involved in autobiographical memories and propose competing hypotheses about the role of visual perspective, and (3) identify directions for future research on visual perspective and the role of the self in autobiographical memories.

Previous Research on Visual Perspective

Visual perspective research has focused primarily on two broad issues: the determinants and the correlates and consequences of visual perspective. We review the relevant research on each of these issues below.

Determinants of Visual Perspective

In their seminal paper, Nigro and Neisser (1983) speculated that the visual perspective adopted at recall is a function of the characteristics of the original event, the time period since the event occurred, and the individual's motivational and emotional state at the time of retrieval.

Nigro and Neisser's (1983) findings support their speculation that features of the event at the time of encoding influence the perspective from which the event is subsequently recalled. Using a multidimensional scaling approach, they asked participants to recall a specific memory for each of eight situations (e.g., giving a public presentation), rate the similarity of the 56 possible pairs of memories, rate each experience for its emotionality and degree of focus on the self, and indicate the visual perspective from which they recalled each memory. Their findings suggest that highly emotional or stressful situations that involve a focus on the self at encoding (e.g., giving a public presentation) may be more likely to be recalled from the 3^rd person perspective, whereas situations low on both of these dimensions (e.g., watching television) are more likely to be recalled from the 1^st person perspective.

Clinical research provides additional evidence that, for highly emotional and stressful events, 3^rd person memories may begin at encoding of the experience. In their study of visual perspective and PTSD, McIsaac and Eich (2004) found that nearly half of the participants who recalled their trauma from the 3^rd person perspective reported that they experienced the trauma from the 3^rd person (the validity of these retrospective reports, however, is unknown). McIsaac and Eich suggest that the 3^rd person perspective may serve as one cognitive strategy to decrease the anxiety associated with trauma.

The second component hypothesized by Nigro and Neisser, the time interval since the event occurred, also influences whether the experience is recalled from the 1^st or the 3^rd person. Recent autobiographical memories tend to be recalled from the 1^st person perspective, whereas remote memories, particularly early childhood memories, are more likely to be recalled from the 3^rd person perspective (Berntsen & Rubin, 2006; Kihlstrom & Harackiewicz, 1982; Nigro & Neisser, 1983; Talarico, LaBar, & Rubin, 2004; Terry & Barwick, 1995-1996). Further, the proportion of 3^rd person memories retrieved increases across adulthood, continuing well into old age (Piolino et al., 2006). Flashbulb memories may be one exception to this shift: in contrast to everyday memories, which tend to shift from 1^st person to 3^rd person over time, flashbulb memories tend to remain in the perspective adopted immediately following the event (Talarico & Rubin, 2003).

Finally, Nigro and Neisser's third hypothesized component, the individual's motivational and emotional state at the time of retrieval, has received empirical support. Individuals experimentally induced to focus on their feelings about an experience are more likely to recall the event from the 1^st person, whereas individuals induced to focus on the concrete, objective circumstances of an experience are more likely to recall the event from the 3^rd person (Nigro & Neisser, 1983). Consistent with these findings, D'Argembeau, Comblain, and Van der Linden (2003) found that memories that individuals currently experience as positive or negative for personal reasons (i.e., events that would not necessarily be emotional for most people) are more likely to be recalled from the 1^st person perspective, whereas neutral memories are more likely to be recalled from the 3^rd person perspective. In contrast, consistent with the idea that a focus on the self at encoding is associated with 3^rd person memories, D'Argembeau and Van der Linden (2008) found that memories of highly self-conscious experiences, such as events that elicit pride or shame, tend to be retrieved from the 3^rd person rather than the 1^st person perspective. Although only a few studies have examined the influence of the individual's emotional state on visual perspective, the research to date suggests the provocative possibility that the emotionality at time of retrieval increases the likelihood that the memory will be recalled from the 1^st person, whereas emotionality at time of encoding increases the likelihood that the memory will be recalled from the 3^rd person.

Other factors at retrieval, in addition to emotion, may influence visual perspective. Libby (2003) asked participants to retrieve either a childhood experience that they remember doing or a childhood experience that had happened to them drawn from a standard set of childhood events. Participants in the remember doing condition reported more 1^st person memories, whereas participants in the happened to them condition reported more 3^rd person memories. These findings point to the potential malleability of visual perspective at retrieval.

Correlates and Consequences of Visual Perspective

In addition to the determinants of visual perspective, researchers have examined its correlates and consequences. For example, individuals make more dispositional attributions for their behavior in either naturally retrieved or manipulated 3^rd person memories than in 1^st person memories (Frank & Gilovich, 1989). Memories naturally retrieved from the 3^rd person are related to less reliving, fewer visual images, less sensory information, less importance, and less certainty that the event occurred as the individual remembers it (Berntsen & Rubin, 2006). These latter findings, however, were not replicated when visual perspective was manipulated by asking participants to retrieve memories from a specified perspective.

In the clinical domain, there has have been a particular interest in the phenomenological correlates of 1^st and 3^rd person memories of trauma and the therapeutic benefits of consciously switching from a 1^st to a 3^rd person perspective. For example, patients with PTSD who recalled their trauma from the 1^st person reported more emotion and anxiety associated with the memory than PTSD patients who recalled their trauma from the 3^rd person (McIsaac & Eich, 2004). In addition, the content of these 1^st person memories had more references to affective reactions, physical sensations, and psychological states, whereas the content of 3^rd person memories had more references to spatial relations, self-observations, and peripheral details. These findings suggest that 3^rd person memories serve as a distancing mechanism to dampen the impact of the trauma, although McIsaac and Eich point out there were no differences in the severity of symptoms between patients who recalled their trauma from the 1^st or the 3^rd person.

A somewhat different pattern of results emerges when considering memory and trauma in nonclinical populations. Berntsen, Willert, and Rubin (2003) found that students who met the criteria for PTSD tended to retrieve memories of their trauma from the 3^rd person perspective, whereas students with a similar trauma history but who did not meet the PTSD criteria tended to retrieve their trauma memories from the 1^st person perspective. For students who met the PTSD criteria, 3^rd person retrieval of their trauma memories was related to greater frequency of avoidance symptoms. Berntsen and colleagues (2003) suggest that this may serve a distancing function because “a reduced field [1^st person] perspective can be taken to indicate reduced emotional reliving” (p. 687).

Research from other domains also suggests that naturally occurring 3^rd person memories serve a distancing function, helping to minimize emotional and even physical pain. For example, college students who reported intrusive memories from the 3^rd person perspective reported greater detachment and numbness from the event in the memory than students who reported similar intrusive memories from the 1^st person (Williams & Moulds, 2007). In a study of patients suffering from chronic physical pain, participants who retrieved a memory of a significant pain experience from the 3^rd person perspective reported less severe current levels of pain than participants who retrieved a similar experience from the 1^st person (McNamara et al., 2005).

The 3^rd person perspective might be emotionally adaptive because it allows the individual to face the trauma, without having to psychologically re-experience the event (Wilson & Ross, 2003). The distance provided by the 3^rd person perspective may allow the individual to more objectively observe the situation and subsequently reframe, work through, and ultimately leave the traumatic experience behind them. It may be easier to engage in this process of re-construal when individuals can detach themselves from the emotional pain associated with the trauma.

Libby and her colleagues (Libby & Eibach, 2002; Libby et al., 2005) also postulate that 3^rd person memories serve a distancing function, but suggest that these memories have the additional function of maintaining continuity and coherence in the self. According to Libby, whether the 3^rd person serves a distancing or continuity function depends on the individual's motivational state at retrieval. When motivated to see change in the self, the 3^rd person perspective serves a distancing function, highlighting differences between the past and current selves, whereas when motivated to see similarity in the self, the 3^rd person perspective serves a continuity function, highlighting similarities between the past and current selves.

Libby et al. (2005) reported a series of studies that support their conceptualization of 3^rd person memories. When participants were motivated to focus on differences between their remembered and current selves, retrieval of memories from the 3^rd person perspective led to perceptions of greater self-change than did retrieving similar memories from the 1^st person. However, when participants were motivated to focus on similarities between their remembered and current selves, the 3^rd person perspective led to perceptions of greater self-continuity than did retrieving similar memories from the 1^st person. Thus, an individual's motives at the time of retrieval may influence the function of visual perspective. More generally, Libby's work highlights the importance of the self in memory retrieval and provides critical evidence for the interplay between self processes, visual perspective, and the retrieval of autobiographical memories. Her work is indicative of a more complex psychological process involved in the retrieval and subsequent functions of visual perspective than has been assumed in the past.

Theoretical Model of the Self Processes Involved in Autobiographical Memories

The emerging literature on visual perspective has provided new insights into the determinants, dynamics, and functions of visual perspective, but the field lacks an integrative theoretical framework. This may be due, in part, to the fact that current models of self and memory (e.g., Conway & Pleydell-Pearce, 2000) provide only a general framework for conceptualizing the role of the self in memory, making it difficult to incorporate processes related to visual perspective. That is, although these models emphasize the role of the self in autobiographical memory retrieval, they do not detail the specific self-processes involved. Therefore, as a starting point toward an integrative theoretical framework, we present a model that specifies the self processes involved in autobiographical memory retrieval and discuss how visual perspective might influence these processes. Our model is based on the assumption that the processes underlying the retrieval of autobiographical memories are driven by the same cognitive and motivational processes that govern all forms of self-evaluation.¹

To develop our model, we drew on ideas and findings from the literatures on autobiographical memory, self processes, and visual perspective. We first give a brief overview of current conceptualizations of the role of the self in autobiographical memory. We next describe each part of our model and then, after describing the entire model, discuss the role of visual perspective. Throughout this section, we will continue to use the example of Jane to make the relatively abstract model more concrete.

Autobiographical Memory and the Self

Although we carry around a vast storehouse of episodic memories, only some of these memories – autobiographical memories – are relevant to the self. According to Roediger and Marsh (2003), the “critical defining feature for autobiographical memory is the importance of the information to one's sense of self and one's life history” (p. 485). Individuals may have episodic memories of trivial events, but only memories that are relevant to a person's self are truly autobiographical. We focus on autobiographical memories, rather than episodic memories in general, because self-evaluative processes should be most apparent in memories that are important to the self.

Autobiographical memories are the end product of a reconstructive process. Since the pioneering work of Bartlett (1932/1967), memory reconstruction has been shown to be strongly influenced by a variety of factors, including schemas (Bartlett, 1932/1967), personality (Rubin, Schrauf, & Greenberg, 2003; Rubin & Siegler, 2004; Singer & Salovey, 1993), and the self (Conway, 2005; Conway & Pleydell-Pearce, 2000). This body of research implies that both the content and phenomenology of memories fluctuate over time and are influenced by the individual's current needs, goals, and motives. Consistent with this research, we assume memories are reconstructed, but we propose that they are reconstructed, in part, to serve self-evaluative motives.² This process should apply to all autobiographical memories, regardless of how or why they are retrieved (e.g., consciously or non-consciously, voluntarily or involuntarily). In addition, because memories are reconstructed at retrieval, it is likely that the product of this reconstruction influences subsequent retrieval of the memory. Thus, both the content and the meaning of a memory can change over time. Our model highlights the role of self-evaluative motives in this reconstructive process.

Conway and his colleagues developed a comprehensive model of autobiographical memory that emphasizes the self in memory retrieval (Conway, 2005; Conway et al., 2004; Conway & Pleydell-Pearce, 2000). According to Conway's model, memory is a powerful force that acts to construct and maintain a coherent self over time. As such, memory content can be enhanced or diminished, edited or distorted, amplified or suppressed to maintain such coherence. Autobiographical memories likewise ground the self in reality and help establish important goals. Although Conway (2005; Conway & Pleydell-Pearce, 2000) implicates the self in autobiographical memory and argues that memories help maintain self-coherence, he does not articulate the specific self processes involved in autobiographical memory retrieval.

Yet, it is well documented that memory retrieval processes are influenced by self-motives, self-representations, and other self processes (e.g., Fiske & Taylor, 1991). For example, the literature on the self-reference effect indicates that material related to the self tends to be remembered better than non-self-related material (Rogers, Kupier, & Kirker, 1977). Further, individuals recruit memories consistent with their self attitudes and, when their attitudes change, adjust their memories accordingly (Ross, 1989). When motivated to believe they possess a desirable attribute, individuals selectively retrieve memories that support this belief (Sanitioso et al., 1990). Self-esteem is related to a positive memory bias: High self-esteem individuals remember events as more positive than when they actually occurred, whereas low self-esteem individuals remember events as more negative (Christensen, Wood, & Feldman Barrett, 2003). To maintain or increase self-esteem, the self in a distant memory may be disparaged to inflate how much the individual has improved over the years (Wilson & Ross, 2001).

Thus, self processes, particularly motives for self-congruence and self-esteem, play an active role in the retrieval and reconstruction of personal memories. Our model identifies and integrates specific self processes previously implicated in the retrieval of autobiographical memories and we use these processes to formulate specific hypotheses about the role of visual perspective in memory retrieval. In the Self-Memory System of Conway and Pleydell-Pearce, these processes are likely part of the verification criteria for generative memory retrieval. That is, as a memory is being retrieved, the autobiographical knowledge activated is continually evaluated by control processes. These criteria help monitor the content and experience of the memory to ensure that it is consistent with the individual's current self-goals.

Appraisal Processes

Figure 1 shows our model of self, memory, and visual perspective. A fundamental step in the retrieval of an autobiographical memory is to determine (i.e., appraise) whether it is relevant to the self, and therefore truly autobiographical in nature. To make this appraisal, which may occur either implicitly or explicitly, individuals compare the memory to their network of self-representations, including representations of their actual, ideal, and possible selves. For example, if Jane decides to pursue a career in law, recalling her failed class presentation may activate representations of her possible self as a trial lawyer. Retrieval of this memory will be very relevant to her possible self as a trial lawyer and to her ideal self as an adept public speaker. If, however, Jane decides to become a librarian, this memory becomes less relevant to Jane's possible self and therefore less salient over time.

Figure 1.

Model of self-processes involved in the retrieval of autobiographical memories from the 1^st person and 3^rd person visual perspective. PA = Positive Affect; NA = Negative Affect.

According to our model, memories with self-relevant content are then appraised for their congruence with and threat to the current self. Congruence appraisals refer to whether the self in the memory is consistent with the current self, and threat appraisals refer to whether the self in the memory enhances or diminishes self-esteem. For example, if Jane has a representation of her self as a good public speaker, then retrieval of her disastrous first talk would be appraised as incongruent with and a threat to her self-concept. However, if Jane has a representation of herself as a poor public speaker, then the memory would be congruent with her self-concept, but nonetheless threatening to her self-esteem. In this case, Jane would be caught in what Swann and colleagues (Swann, Griffin, Predmore, & Gaines, 1987) referred to as a cognitive-affective crossfire, where her memory is cognitively consistent with the self but affectively inconsistent with her need to maintain and enhance self-esteem. In the cognitive-affective crossfire literature, when these two motives conflict, cognitive consistency tends to “win” over self-esteem enhancement. For example, low self-esteem individuals prefer feedback congruent with their negative self-views to positive feedback that feels good, but is inconsistent with their selves (Swann et al., 1987). When the two motives are congruent, such as when high self-esteem individuals receive positive feedback or low self-esteem individuals receive negative feedback, they work together to amplify or dampen the feedback. As discussed below, both conflict and congruency likely impact visual perspective.

Activation of Self-evaluative Motives

To regulate the psychological consequences of congruence and threat appraisals, self-evaluative motives are activated, including the motive to maintain a consistent and coherent identity (when incongruence occurs) and the motive to feel good about the self (when threat occurs; see Figure 1). The motive to maintain a consistent and coherent identity is referred to as self-verification (Swann, Rentfrow, & Guinn, 2003). That is, individuals seek information that is consistent with their self-views; this coherence gives meaning to the self, organizes and predicts behavior, and guides social interactions. Consistent self-views promote a coherent social environment that further serves to stabilize self-views.

The motive to feel good about the self and avoid feeling bad is referred to as self-enhancement (Swann et al., 2003). That is, individuals seek to maintain and enhance their self-esteem through self-serving attributions, biased processing of self-relevant information, downward social comparisons, and other cognitive processes (e.g., Taylor & Brown, 1988). According to our model, visual perspective may be one additional mechanism through which individuals maintain consistent self-views and enhance their self-esteem. Indeed, Wilson and Ross (2003) suggest that visual perspective may influence how people remember more so than what they remember. Below we describe how these self-evaluative processes work in the context of our model.

Self-verification

Autobiographical memories are one potential source of information about the self that can serve to verify self-beliefs (Conway et al., 2004; Pasupathi & Rich, 2005; Ross, 1989). A self-relevant memory will be appraised as either congruent or incongruent with the current self, resulting in feelings of authenticity or inauthenticity, respectively (see Figure 1). When self-verification processes are activated, the self is motivated to either connect or distance the self from that memory, depending on whether the self in the memory is congruent or incongruent with the current self (Libby et al., 2005).

The 3^rd person perspective is related to the distinction between the self as perceiver (“the “I”) and the self as object of perception (the “Me”) and extends early theorizing on how the self maintains continuity over time. Cooley (1902), for example, theorized that we judge ourselves through the eyes of significant others, and invoked the metaphor of a “looking glass” through which we see our reflection via the perceptions of close others. Mead (1934), extending Cooley's ideas, conceived of a “generalized other,” abstracted from society's rules and norms, watching over our behavior. When appraising whether a memory of a past experience is congruent with the current self, recalling a memory from the 3^rd person perspective allows the individual to take the perspective of another to judge his/her behavior based on the point of view of another.

Self-enhancement

The retrieval of an autobiographical memory likely engages self-enhancement, as well as self-verification, processes (see Figure 1). When the memory is negative and appraised as a threat to the self, the activation of the self-enhancement motive will serve to distort the memory (e.g., the individual may selectively retrieve positive aspects of the experience but fail to retrieve negative aspects), reduce its salience, and/or distance the current self from the remembered self, dampening the negative emotions associated with the memory. When the memory is positive and appraised as enhancing the self, activation of the self-enhancement motive will serve to make the memory more salient and decrease the perceived distance between the current and remembered self, amplifying the positive emotions associated with the memory.

Several lines of research highlight the use of memory to enhance the self. As mentioned above, Christensen and colleagues (2003) found that high self-esteem individuals rated positive memories as more positive over time. Wilson and Ross (2001) demonstrate that individuals use memories of their past selves to make themselves feel better in the present. Although they did not directly address self-enhancement, Walker and his colleagues found that negative memories are more susceptible than positive memories to the fading affect bias (Walker, Skowronski, Gibbons, Vogl, & Thompson, 2003; Walker, Vogl, & Thompson, 1997). In a series of studies, they demonstrate that negative affect associated with a memory decreases faster than positive affect and that this is independent of memory age, original intensity of the event, or extremity of the event. Both processes—dampening and amplifying of memory-related emotions—may be facilitated by visual perspective, as discussed below.

Self-Regulatory Processes and Visual Perspective

How might visual perspective promote self-verification and self-enhancement regulatory goals in the context of autobiographical memories? Two broad views can be derived from the research literature. The first view, which we label the Dispassionate Observer, hypothesizes that the 3^rd person perspective promotes an objective, “dispassionate” perspective on the self, which distances the past self from the current self and thereby reduces feelings of inauthenticity when an incongruence appraisal has been made and reduces feelings of negative emotion when a threat appraisal has been made. The second view, which we label the Salient Self, hypothesizes that the 3^rd person perspective increases self-focused attention, both visually and emotionally; that is, the self moves into the foreground and, as gestalt psychologists would say, engulfs the visual field. This heightened self-focused attention promotes congruence between the past and current self and thereby magnifies feelings of authenticity when a congruence appraisal has been made and magnifies positive emotion when the memory is self-esteem enhancing.

Below we discuss these processes in greater detail, and show how these two views generate competing hypotheses about the consequences of the 3^rd person perspective.³ Because autobiographical memory is the end product of a complex process, it is likely that many factors moderate this process. After presenting the two views, we discuss several moderators that have been identified in the visual perspective literature.

Dispassionate Observer

The Dispassionate Observer view suggests that the 3^rd person perspective serves a distancing function for the individual. By watching his/her actions from a detached 3^rd person perspective, the individual is able to more objectively evaluate his/her behavior and psychologically disengage from the situation in the memory. Ultimately, this would lead to a dampening of the emotional response associated with the memory and a distancing of the current self from the self in the memory. When a memory is appraised as incongruent and the self-verification motive is activated, the Dispassionate Observer view predicts that the memory will be retrieved from the 3^rd person perspective as a way of distancing the current self from the past self. Similarly, when a memory is appraised as threatening and the self-enhancement motive is activated, the Dispassionate Observer view predicts that the memory will be retrieved from the 3^rd person as a way of dampening the emotional experience associated with the memory. From this view, memories that make the individual feel either inauthentic or bad about him/herself are more likely to be retrieved from the 3^rd person. In contrast, memories that make the individual feel authentic or good about him/herself are more likely to be retrieved from the 1^st person.

Using our stage-frightened student as an example may help clarify the Dispassionate Observer view. By the time Jane reaches law school, she may perceive herself as an adept public speaker. In this case, 3^rd person recall of her disastrous first speech would help distance herself from that failure and highlight for her how much she has changed. In addition, 3^rd person recall of this experience would dampen the negative emotion related to this failure, helping Jane to feel better about herself and buffer her against the self-esteem diminishing implications of the experience.

Several studies support the idea that 3^rd person memories can reduce the emotional impact of an experience, relative to 1^st person memories. In one experiment, participants completed a series of manual tasks, such as molding clay into an object of their choice (McIsaac & Eich, 2002), and were subsequently asked to recall and write about the tasks they just completed from either the 1^st or the 3^rd person perspective. Memories described from the 1^st person perspective were more affective (based on content analyses and self-reports) and contained more references to psychological states and physical sensations than 3^rd person memories. In contrast, 3^rd person memories contained more statements about how the participant looked, what they did, and the spatial layout of the objects in the room. Robinson and Swanson (1993) asked participants to recall eight experiences from their preferred perspective and then, two weeks later, from a specified visual perspective (either 1^st or 3^rd). Participants who shifted from the 1^st person to the 3^rd person showed a decrease in their emotional reactions to the memory (although the complementary effect, an increase after shifting from the 3^rd to 1^st person, did not occur). Likewise, Berntsen and Rubin (2006) found that changing from a 1^st person to a 3^rd person perspective reduced all of the qualities related to the reliving of the memory, such as intensity and visual clarity (although, similar to Robinson and Swanson, the complementary increase did not occur when memories shifted from a 3^rd person to a 1^st person perspective).

According to the Dispassionate Observer view, 3^rd person retrieval should dominate when memories are negative; however, when the memory is positive but incongruent with the current self, producing what Swann et al. referred to as a cognitive-affective crossfire, it may be retrieved from the 3^rd person perspective as a way of distancing a memory incongruent with the current self. For example, clinically-depressed individuals, when asked to retrieve both positive and negative memories from different periods throughout their lives, retrieve more positive memories from the 3^rd person perspective than non-depressed controls (Lemogne et al., 2006; but see Kuyken & Howell, 2006). This pattern holds even after the depression has remitted (Bergouignan et al., 2008). Positive memories are likely to be incongruent with the depressed individual's self and thus activate the self-verification motive and produce feelings of inauthenticity. According to our model, to reduce inauthenticity, the 3^rd person perspective may be employed to distance the current depressed self from a past self enjoying a positive experience. Although this mechanism likely promotes authenticity, retrieving positive memories from the 3^rd person perspective may perpetuate the depressive state by down-regulating positive emotions. Together, these studies support the idea that the 3^rd person perspective is more dispassionate and reduces emotional engagement.

Salient Self

The Salient Self view suggests that the 3^rd person perspective serves to magnify the emotions associated with a memory and strengthen the connection individuals feel between their current and past selves. When a memory is appraised as congruent and the self-verification motive is activated, the Salient Self view predicts that the memory will be retrieved from the 3^rd person as a way of strengthening the connection between the current and past self. When a memory is appraised as enhancing self-esteem and the self-enhancement motive is activated, the Salient Self view predicts that the memory will be retrieved from the 3^rd person as a way of amplifying the positive emotional experience and increasing positive feelings about the self. From the Salient Self view, memories that make the individual feel either authentic or good about him/herself are more likely to be retrieved from the 3^rd person. In contrast, memories that make the individual feel inauthentic or bad about him/herself are more likely to be retrieved from the 1^st person. Note that the Salient Self and Dispassionate Observer views generate opposite predictions in this context.

Once again, Jane's memory of her poor performance will help clarify the Salient Self view. If Jane still perceives herself to be a poor public speaker, recalling this memory from the 3^rd person perspective may actually highlight her failure in this domain, heighten her anxiety over public speaking, and increase the negative emotions associated with the experience due to the heightened self-focused attention of seeing herself in the memory.

Findings from the literature on self-focused attention lend support to this view. Self-focused attention, such as looking into a mirror or watching a recording of yourself, generally magnifies an individual's emotional experience and makes perceptions of the remembered self more congruent with the current self (Carver & Scheier, 1981; Robins & John, 1997). Conceptually similar to the third person, the self-focused attention findings suggest that the 3^rd person perspective should intensify a person's emotional response and increase feelings of authenticity. Consistent with this possibility, 3^rd person memories are rated as more vivid and emotional than 1^st person memories (Terry & Barwick, 1998-1999).

Evidence from the clinical domain also supports the Salient Self view for memories that are congruent with the self. For example, patients suffering from social phobia report memories of socially anxious events from the 3^rd person and memories of non-social events from the 1^st person, whereas non-anxious controls report both types of memories from the 1^st person (Wells, Clark, & Ahmad, 1998). In addition, social phobics tend to rate themselves more negatively than controls when retrieving memories of highly anxious social situations (Coles, Turk, Heimberg, & Fresco, 2001). For these patients, memories of socially anxious situations are highly congruent with their self-conceptions as socially anxious individuals, and thus the 3^rd person perspective may amplify the emotion associated with these memories. Although magnifying negative memories conflicts with the self-enhancement motive, as noted above, cognitive consistency tends to win over affective enhancement when the two motives are in conflict (Swann et al., 1987).

Moderators

Autobiographical memory retrieval is a complex process, and just as multiple factors influence other aspects of memory retrieval, many moderators likely influence visual perspective. Below we discuss several moderators that have been discussed in the visual perspective literature. These moderators point to conditions under which the hypotheses made by the Dispassionate Observer view will dominate and conditions under which hypotheses made by the Salient Self view will dominate.

First, as reviewed in the background literature above, an individual's motivational focus at the time of retrieval moderates the effect of the 3^rd person perspective (Libby et al., 2005). When focused on the similarities between a current and a past self, the 3^rd person perspective serves to increase feelings of connection with the remembered past self (Salient Self). In contrast, when focused on the differences between a current and a past self, the 3^rd person perspective serves to decrease feelings of connection with the remembered past self (Dispassionate Observer). These findings support both the Dispassionate Observer and Salient Self views and provide evidence for the circumstances under which each view will dominate.

Recent work has extended the impact of visual perspective to imagined future desirable behaviors (Libby, Shaeffer, Eibach, & Slemmer, 2007; Vasquez & Buehler, 2007). Libby et al. asked participants, the night before the 2004 presidential election, to visualize themselves voting from either the 1^st person or the 3^rd person perspective. They found that visualizing voting from the 3^rd person led to more actual voting than visualizing voting from the 1^st person. Mediator analyses showed that participants in the 3^rd person condition were more likely to vote because this perspective encouraged a stronger pro-voting mindset. Libby et al. suggest that 3^rd person images increase dispositional attributions, making individuals more likely to perform the desired behavior, thereby increasing self-congruence. Vasquez and Buehler (2007) extended this reasoning into the academic domain and showed that students who visualized success from the 3^rd person perspective increased in achievement motivation. These findings dovetail nicely with our Salient Self hypothesis. Although the exact processes underlying imagination may differ from those underlying autobiographical memory, both may engage similar self processes (Frank & Gilovich, 1989; Libby et al., 2007).

Second, Kross, Ayduk, and Mischel (2005) found that emotional focus (focusing on which emotions occurred vs. why they occurred) moderate the effects of a “distanced-self perspective,” which they conceptualize as a 3^rd person perspective, on participants’ emotional experiences. In their distanced-self condition, Kross et al. asked participants to “take a few steps back and move away from your experience...watch the conflict unfold as if it were happening all over again to the distant you...” (p. 711). Thus, similar to 3^rd person perspective memories, participants visualized themselves when thinking about the experience in the memory. When retrieving an anger memory from this perspective, participants instructed to focus on which emotions were experienced tended to show greater implicit and explicit anger and higher overall negative affect (Salient Self), relative to participants instructed to focus on why the emotions were experienced (Dispassionate Observer). Kross and Ayduk (2008) subsequently replicated these findings using depression memories instead of anger memories. Again, these findings support both the Salient Self and Dispassionate Observer views and identify emotional focus as an important moderator.

Third, research from the clinical literature also suggests circumstances under which each view will dominate. Patients who retrieved memories of their trauma from the 3^rd person perspective immediately following the trauma had more avoidance symptoms one year post trauma, whereas participants who had shifted from 1^st person to 3^rd person over this period of time did not have any avoidance symptoms (Kenny & Bryant, 2006). This shift from the 1^st to the 3^rd person perspective may serve to distance the trauma from the self (Dispassionate Observer), whereas continual retrieval from the 3^rd person may amplify the emotional experience of the trauma memory (Salient Self).

Finally, most previous studies of visual perspective have focused on memories of negative emotional experiences. Although memories of positive emotional experiences tend to be retrieved from the 1^st person (e.g., Berntsen & Rubin, 2006), there are many factors that might moderate this tendency. We suggest three. First, everyday positive experiences do not typically require regulation, and therefore may be less likely to be retrieved from the 3^rd person. However, memories of unusual, or intensely positive experiences may be more susceptible to the effects of self-regulatory processes and may be more likely to be retrieved from a 3^rd person perspective. Second, the context of the event might matter, specifically whether the event occurred in public or private. It may be easier to retrieve a memory from the 3^rd person perspective when a clear 3^rd person perspective actually exists; that is, when the event occurred in a public context. Third, stable individual differences variables may moderate this process. For example, narcissistic individuals might be more likely to retrieve memories, particularly positive memories, from the 3^rd person, as narcissists enjoy looking at themselves in mirrors and evaluate themselves more positively when watching themselves on videotape (Robins & John, 1997). Seeing yourself in a memory is conceptually similar to seeing your self in a mirror or on videotape and thus all three experiences may activate similar self-evaluative processes. Consistent with this view, narcissists report more positive affective content in memories retrieved from the 3^rd than the 1^st person perspective (Sutin & Robins, 2005).

Future Directions

In this final section we present several avenues for future research on visual perspective, including: (a) testing components of the theoretical model, (b) exploring the consequences of pitting the self-enhancement motive against the self-verification motive, and (c) examining the similarities and differences between 3^rd person memories and false memories.

Systematically Test Components of the Theoretical Model

The theoretical model presented in this paper has multiple components that could be manipulated or assessed to test its validity. Because considerable evidence already supports the self-evaluative processes described by the model, we specifically outline ways to test the two competing views of visual perspective.

First, to test predictions made by the Dispassionate Observer and Salient Self views, it would be informative to manipulate threat and congruence appraisals. For example, to manipulate threat, participants could be asked to retrieve a memory of a positive, self-enhancing experience or a negative, self-diminishing experience. Similarly, to manipulate congruence, participants’ self-conceptions could be assessed and then they could be asked to retrieve a memory that is congruent with an aspect of their self (i.e., a self-verifying memory) and a memory that is contrary to how they see themselves (i.e., a non-self-verifying memory). As noted in Figure 1, the Dispassionate Observer view predicts that for negative and incongruent memories, retrieval from the 3^rd person will dampen the emotional experience and distance the current self from the self in the memory, whereas the Salient Self view predicts that for positive and congruent memories, retrieval from the 3^rd person will amplify the emotional experience and strengthen the connection with the current self.

Another way to explore the processes presented in the model would be to examine memories that are more closely linked to self processes, including self-defining memories and memories for experiences of self-conscious emotions such as shame and pride. Self-defining memories are characterized by affective intensity, vividness, repetition, and linkages to other memories, and, compared with general autobiographical memories, are more likely to be important to the individual, tap themes of self-discovery and self-understanding, and focus on unresolved conflicts or enduring concerns (Singer & Salovey, 1993). Thus, the self-regulatory processes outlined in our model are likely to be more pronounced for these types of memories. Memories of self-conscious emotional experiences may be similar to self-defining memories in this respect. Memories of shame and pride experiences are generally more self-relevant than memories of other emotional experiences because self-evaluative processes are integrally involved in the experience of shame and pride (Tracy & Robins, 2004). Although differences in visual perspective have been examined for emotional memories, this work has focused on memories for experiences involving non-self-conscious emotions such as happiness and sadness. If self processes play a central role in the retrieval of autobiographical memories, as our model suggests, then the dynamics and functions of visual perspective might be quite different for both self-defining and self-conscious emotion memories.

To more fully test the predictions made by the two competing views, visual perspective could be manipulated along with congruence and threat appraisals. By manipulating all three variables in a single experiment, the divergent predictions of the Dispassionate Observer and Salient Self positions could be fully contrasted. That is, participants could be instructed to recall a positive or negative memory that is congruent or incongruent with the self from the 1^st or 3^rd person. The Dispassionate Observer view predicts that retrieval of a self-threatening memory from the 3^rd person perspective leads to a dampening of the emotional experience relative to the 1^st person perspective, whereas the Salient Self view predicts that the 3^rd person perspective leads to an amplification of the emotional experience relative to the 1^st person perspective. With regard to self-congruency and visual perspective, the Dispassionate Observer view predicts that the 3^rd person perspective will lead to a distancing of the past self from the current self relative to the 1^st person perspective, whereas the Salient Self view predicts that the 3^rd person perspective will strengthen the connection between the past and current self relative to the 1^st person perspective.

Cognitive-Affective Crossfire in Autobiographical Memories

Although the self-verification and self-enhancement motives often work in concert to promote self-coherence, they can also work in opposition to each other (Swann et al., 1987). When the self-verification and self-enhancement motives conflict, the individual needs to process information that is either consistent with the self but feels bad or inconsistent with the self but feels good. In Swann et al.'s research, individuals caught in the cognitive-affective crossfire typically seek out and attend to self-verifying information over self-enhancing information—in other words, they choose cognitive consistency over affective enhancement. For example, when given the choice between positive and negative feedback, low self-esteem individuals will seek out and attend to the negative information.

Asking low self-esteem individuals to retrieve a memory of a success would provide an ideal opportunity to test the hypotheses made by the competing views outlined above. If a low self-esteem participant retrieves such a memory from the 3^rd person perspective, the Dispassionate Observer view predicts a dampening of the positive emotions and a distancing of that self from the current self, whereas the Salient Self view predicts an amplification of the positive emotions associated with the experience and a strengthening of the connection between the current and past selves. The opposite predictions can be made for memories retrieved from the 1^st person. In contrast, a high self-esteem individual would not be caught in the cognitive-affective crossfire and the motives would thus work together to amplify the memory. In this case, the Salient Self view predicts that the memory will be retrieved from the 3^rd person whereas the Dispassionate Observer view predicts that the memory will be retrieved from the 1^st person.

Examine Similarities and Differences Between 3^rd Person Memories and False Memories

In some sense, most 3^rd person memories are necessarily false memories because, unless the individual dissociates during the experience, it is impossible to experience an event from any perspective other than the 1^st person. Despite this link, there is little research on the similarities and differences between 3^rd person and false memories. Some studies suggest that false memories of early childhood are more likely to be recalled from the 3^rd than the 1^st person perspective, but true memories of early childhood also tend to be retrieved from the 3^rd person, although not to the same extent as false memories (Heaps & Nash, 2001). In addition, 3^rd person memories share many of the same characteristics as false memories. Compared to true memories, false memories are typically rated as less rich in recollective experience and visual imagery, less emotionally intense, and less important (Heaps & Nash, 2001). Likewise, naturally occurring 3^rd person memories are often less vivid, emotional, and important (Berntsen & Rubin, 2006; Libby & Eibach, 2002; McIssac & Eich, 2004).

In addition, Libby (2003) found that after participants visualized a false childhood event from either perspective, retrieval of the false memory from the same perspective as they had visualized it led to imagination inflation. That is, participants increased their ratings of certainty that the false event had occurred. By focusing on one aspect of the phenomenological experience of the memory (visual perspective, in this case), source-monitoring criteria may change; errors in source monitoring are likely to occur when a specific, salient criterion matches the salient phenomenological characteristic of the memory. Future research could use a similar manipulation, but add measures of other phenomenological characteristics, such as vividness and emotional intensity, to examine whether the salience of visual perspective influences the phenomenology of true and false memories in different ways.

In summary, the research literature on visual perspective and autobiographical memories is still in its infancy, but it is already beginning to reveal some provocative insights into the nature of self and memory. There are many exciting avenues for future research and we hope that the model presented in this article will help promote and facilitate research on the role of visual perspective in autobiographical memories.