Narrative Identity: What Is It? What Does It Do? How Do You Measure It?

This special issue of Imagination, Cognition, and Personality is devoted to the topic of narrative methods for studying autobiographical memory in the fields of personality and cognitive psychology. But the wide assortment of excellent papers gathered here is as much about concepts and processes as it is about the niceties of research methodology. The editors expected as much when they emphasized, in their opening article, the extent to which methodological choices reflect theoretical stances and assumptions (Grysman & Mansfield, this volume). This is why researchers rarely, if ever, employ narrative methods to assess individual differences in personality traits or in intelligence. The theoretical assumptions we have about the nature of traits and intelligence suggest that these concepts are not likely to manifest themselves, at least not in reliable and readily detectable ways, in the stories people tell about their lives. Instead, researchers typically use self-report ratings or behavioral observations for traits, and objective tests of cognitive ability for intelligence. In these theoretical or methodological contexts, asking people to tell stories would be a waste of time.

But storytelling methods seem to fit well the phenomenon of autobiographical memory. If you want to know what people remember about particular events in their lives (what happened, who was there, what they were thinking and feeling), it makes eminent sense to ask them to tell the stories of the remembered event. In this way, the concept of autobiographical memory is operationalized through the research participants’ narrations of memories. The value of narrative methods would seem to rise even further in those cases where you want to know what the memories mean for the people who remember them, especially if you are interested in how people generate meanings spontaneously, in the very act of narrating (e.g., Alea, this volume).

Once a narrative method allows personal meanings to come to the fore, the concept of narrative identity (McAdams & McLean, 2013) steps into the theoretical arena, as it did in nearly all of the papers in this volume. As it turns out, narrative identity is the concept around which the different narrative methods described in these papers revolve. Or at least that is how I read them. This gives me an opportunity, then, to reflect on the papers as they relate to the concept of narrative identity, asking: What is narrative identity? What does it do? How do you measure it? I will address each of the three questions, in turn, from an historical perspective.

Form: What Is Narrative Identity?

Before the 1980s, psychologists did not typically associate the idea of “identity” with the idea of “narrative.” For me, the realization that these two concepts might be linked arose shortly after I taught a graduate seminar, in the summer of 1982, on “self and identity.” The reading list included William James, George Herbert Mead, Jane Loevinger, Heinz Kohut, Robert Kegan, and a large number of authors in developmental and social psychology. And, of course, Erik Erikson. The students and I spent a great deal of time puzzling over passages in which Erikson explicitly described the concept of identity. Identity is a “configuration,” Erikson (1959, p. 113) wrote, an “integration” of “childhood identifications,” “the vicissitudes of the libido,” “aptitudes developed out of endowment,” and “opportunities offered in social roles,” all working together, as it were, to confer upon a life a sense of “inner sameness and continuity” (Erikson, 1963, p. 261). An identity, moreover, should usher in adulthood, a new stage in life wherein a person now understands “life in continuous perspective, both in retrospect and prospect” (Erikson, 1958, p. 111). As such, the person who “has” identity is now “able to selectively reconstruct his past in such a way that, step for step, it seems to have planned him, or better, he seems to have planned it” (Erikson, 1958, p. 112).

Reading passages like these, I asked the students this: “If you could see an identity, what would it look like?” What form would it take? We entertained a number of possibilities. In an influential paper, Epstein (1973) had proposed that “the self-concept” is a “theory.” If you could see an identity, would it look like a theory? Complete with axioms, hypotheses, and so on? Probably not, we thought. People are not that systematic, when it comes to selfhood. Maybe it would look like a circus? Three rings of activity, loosely orchestrated by a ringmaster. Maybe it would look (and sometimes sound) like a symphony orchestra? Perhaps geological images were more suitable, suggesting literal and historical depth. As you dig deeper, you might unearth buried (and psychologically older) layers of identity.

We never hit upon an image that worked. But a few months later, I began playing around with the idea that identity, if you could see it, would look like a story. After all, a story potentially integrates different psychological elements, brings a certain kind of narrative order and logic to the chaos of experienced life, and, most importantly, seems to capture the temporal aspects of Erikson’s concept. When you ask people how they came to be who they are, and when you ask them to talk about the future, they typically tell stories. Imagine identity as an internalized and evolving life story, providing, in Erikson’s evocative words, both a “retrospective” and “prospective” sense of a life in time, such that “step for step, it [the story] seems to have planned him [the person as protagonist] or better, he [the person as narrator] seems to have planned [written] it [the story].”

In McAdams (1985), I articulated a model of identity as a life story, complete with setting, scenes, characters, plots, and themes. I imagined it as a big story, an integrative autobiographical project, a personal myth (McAdams, 1990) that situates a person in the world, integrates a life in time, and provides meaning and purpose. If you could literally see it (and read it), it would look like a bound novel. You would see or read the chapters, and you would likely focus in on particularly important, self-defining scenes (Singer & Salovey, 1993), episodes that jump out for their psychological import, like high points, low points, and turning points.

The concept of identity as a life story—what we now call narrative identity—has evolved considerably over the past three decades. Researchers and theorists today imagine narrative identity as something rather more evanescent than what was originally set forth in McAdams (1985), something more variegated and dynamic, more culturally contoured, and more situated in, and even constitutive of, ongoing interpersonal relationships and conversational performances (Freeman, 2011; Grysman & Mansfield, this volume; Hammack, 2008; McAdams, 2013; McLean, Pasupathi, & Pals, 2007; McLean & Syed, 2015). Whereas some contemporary scholars approach narrative identity as a big story that integrates many different chapters, scenes, and characters, others focus on small stories as they appear in circumscribed domains and contexts.

In the preceding papers, Alea (this volume) seems to have the big picture in mind as she codes specific scenes within the life story—high points, low points, and turning points—for the affective themes of redemption and contamination. Similar strategies appear in place for Banks and Salmon (this volume), who explore relations between high points, low points, and turning points, on the one hand, and symptoms of depression and anxiety on the other; Holm and Thomsen (this volume), who examine how self-event connections in narrative contribute to overall self-unity; and Liao, Bluck, and Westerhof (this volume), who consider how self-defining memories may promote overall positivity in personality. Studies like these often aggregate scores across scenes to arrive at broad, summary indices of a person’s overall style or propensity in narrating a life (e.g., McAdams & Guo, 2015).

Other approaches, however, examine more specialized stories, like narrative accounts of harm-doing (Pasupathi, Wainryb, Bourne, & Posada, this volume) or intercultural encounters (Kober, Weihofen, & Rennstich, this volume). The dynamics in these smaller stories are psychologically revealing for their own sake, even if they do not reflect on or foreshadow broader narrative tendencies across other life-narrative scenes. These kinds of studies dig deeper into a specific domain in narrative identity. They suggest that people tell different kinds of stories about different things, and they aim to show that these differences make a big difference.

In his Narrative Identity Structure Model (NISM), Dunlop (this volume) aims to discern a form for narrative identity that incorporates both the big and the little stories. The basic idea of NISM is simple, but the implications are substantial. Dunlop proposes that highly contextualized role-based stories rest below and contribute to a broader narrative about who I am (and how I came to be) more generally. My stories of myself as a father, then, sit next to my stories of myself as a professor, and both sets are subordinate to and feed into the higher level story of Me. NISM is hierarchical, like other models of the self that are popular in social psychology (e.g., McConnell, 2011) and certain trait models in personality psychology (e.g., McCrae & Costa, 2008). This means that smaller stories arrange themselves beneath and link upward to bigger stories, which in turn may link up with even bigger and more encompassing stories. NISM suggests a tightly organized form, with supersets subsuming sets subsuming subsets, all the way down the line.

In making conceptual room for big and small stories, NISM represents a step forward in theorizing about narrative identity. But as a form of friendly amendment, let me express mild skepticism about the tidiness of this model. As with Epstein’s (1973) conception of the self-concept as a theory, I wonder if narrative identity can be so neatly arranged. I agree that small, role-based stories must occupy key positions in the broader narrative scheme for identity. But rather than comprising a strict hierarchy, I propose that the form might more closely resemble an anthology of related stories, or what appears in literature as the single-authored short-story collection.

A good example of what I have in mind is Strout’s (2008) collection entitled Olive Kitteridge, for which she won a Pulitzer Prize. All of the stories in Strout’s collection are set in a small town in Maine. The central character in many, though not all, of the stories is a middle-aged woman named Olive Kitteridge, whom we follow from roughly her 50s through her 70s, with occasional flashbacks to earlier years. Many other characters play major roles as well, including Olive’s husband, her son, her son’s wives, her husband’s assistants at the drug store where he works, and a number of other people living in the small town. The individual stories in this collection are not like chapters; they do not add up to a novel. Each has its own plot line, but the plot lines intertwine in some cases, and many (though not all) reflect common themes. Among the most striking themes are intergenerational misunderstanding and ambivalence (in even the most generative settings), the unquenchable longing for quotidian social contact, and the remarkable power of sensual desire, even among the elderly. The tone is bittersweet but hopeful.

There is a big narrative in Olive Kitteridge, running through and amidst the smaller individual stories that comprise the volume. But the literary form is more organic and associative than hierarchical. The stories overlap in somewhat unpredictable ways, and a few of them seem to go off in their own peculiar directions. What holds the small stories together is the authorial voice of Strout, which itself is reflected in the overall tone of the book, the recurrent themes I enumerated above, and the fact that nearly all of the stories relate, in one way or another, to Olive Kitteridge as she develops over time. Maybe this is what narrative identity looks like.

Function: What Does Narrative Identity Do?

Psychology’s turn toward narrative in the 1980s was a logical extension of its gradual emancipation from the behaviorist grip. It may have been inevitable that once empirical psychologists defied the strictures of behaviorism to peer inside the black box of the human mind, as they began to do in the late 1950s and 1960s, they would eventually happen upon the idea of story. After all, human beings the world over love to tell and hear stories, as Bruner (1986) and Sarbin (1986) both observed. Human beings routinely adopt a narrative mode of thought and expression, Bruner wrote, when it comes to explaining why people do what they do. He distinguished the narrative mode from the paradigmatic mode of thought, which employs logic, evidence, and argument to explain instead how the (physicochemical) world works. Sarbin went so far as to anoint narrative as the new root metaphor for psychological science. Human beings are storytellers by nature, Sarbin argued. Human conduct seems to obey narrative rules. People think about their own lives, and the lives of others, in narrative terms, as stories unfolding over time (Polkinghorne, 1988).

Outside of psychology proper, social scientists and humanists of many different persuasions became enamored with narrative in the 1980s and 1990s. A central question running across many disciplines during this time concerned the function of narrative: What do stories do? First and foremost, they entertain us, some scholars argued (Brewer & Lichtenstein, 1982). Stories engage human emotions, and when they do not, they fail. What is the worst thing you can say about a story? That it is boring. From the parables of Jesus to Dickens, stories also provide instruction on virtue and morality, on how to live a good life (Coles, 1989). Throughout human evolution, even before language when people enacted narrative in gesture and dance, stories have functioned to simulate social experience (Mar & Oatley, 2008). When we read a good story or watch a good movie today, we observe social interactions up close. We witness the clash of human intentions and the timeless social conflicts and motivational dilemmas that characterize so well what human life has always been about. It is probably no exaggeration, then, to claim that stories teach us how to be human (McAdams, 2015).

Narrative identity is a special kind of story—a story about how I came to be the person I am becoming. With this special status comes the special function, a function that Erikson (1963) assigned to identity itself. It is the function of integration. Narrative identity brings things together, integrating elements of the self in both a synchronic and a diachronic sense (McAdams, 1985). Synchronically, narrative identity integrates different social roles (Dunlop, this volume), values (Pasupathi et al., this volume; Wang, Song, & Koh, this volume), attitudes, and performance demands in the variegated here-and-now of life. A person’s story, thus, explains how he or she continues to affirm a sense of “inner sameness and continuity” (Erikson, 1963, p. 251) across different situational and role contexts. The life story also integrates life in a diachronic sense, that is, over time, ideally showing how the self of yesterday has become the self of today, the very same self that hopes or expects to become a certain kind of (different but still similar) self in the future. Concerns about both synchronic and diachronic integration—self-unity in space and time—are salient in Holm and Thomsen’s (this volume) study of self-event connections, self-concept clarity, and dissociation.

Since the 1980s, psychologists have identified a number of other potential functions of narrative identity. As the most notable example, Bluck and Alea (2011) have enumerated (and developed a measure to assess) three primary functions of autobiographical memory in everyday life. People may call upon stories about their personal past to serve social, directive, or self functions. Telling autobiographical memories may promote social relationships; people enjoy sharing stories about their lives with each other. Autobiographical memories may also provide guidance (directives) for life. When confronting a difficult decision, for example, a person may call up memories of similar events in his or her life, consulting them for advice, mining them for insights that may prove helpful in the current situation. What Bluck and Alea put into the domain of functions serving the self includes promoting self-continuity (diachronic integration) for sure, but it also includes the ways in which memories may be called upon to boost morale or sustain positive self-regard. In this light, Liao et al. (this volume) found that positive meaning making in self-defining memories predicted enhanced self-esteem one year later.

In adopting a developmental framework for understanding narrative identity, Fivush, Booker, and Graci (this volume) bring together issues regarding both function and form. They point out that life story construction is constrained by the exigencies of the developmental period during which a narrator aims to make sense of the past. The same event, then, can mean very different things for the same person at two different points in time (Josselson, 2009). At an early age, for example, the narrator may lack certain skills in autobiographical reasoning that would otherwise enable him or her to discern a significant theme or insight from the event, or connect the event to similar others (Habermas & Bluck, 2000). When those skills come online later in development, the person may now understand that same event in very different terms. In this regard, McLean, Breen, and Fournier (2010) have shown that unlike older individuals and unlike females, early-adolescent boys who narrate negative experiences in highly elaborative ways do not enjoy higher levels of psychological well-being. Young adolescent boys may lack the autobiographical skills to process aversive life events in a psychologically productive manner.

Whereas developmental level may constrain meaning making in narrative identity, meaning making efforts may also catalyze development. Fivush et al. (this volume) describe the process of making narrative sense out of life as a mechanism for self-development. The performance of narrative identity may function, therefore, to refine meanings and thereby help the narrator attain a better understanding of self and reach a higher developmental plateau. Elaborating upon the distinction between narrative as window and narrative as process, introduced by Grysman and Mansfield (this volume), Fivush et al. (this volume) contend that narrating life experiences is indeed a window into the current developmental dynamics and parameters that prevail in a given life, but also a process that may promote development itself.

Method: How Do You Measure Narrative Identity?

The research program for narrative identity first laid out in McAdams (1985) drew heavily from the tradition of assessing individual differences in social motives through the Thematic Apperception Test (TAT; Murray, 1938). Beginning in the late 1940s, McClelland et al. developed standardized protocols for administering the TAT in research settings and for scoring TAT responses for variation on achievement (McClelland, Atkinson, Clark, & Lowell, 1953), affiliation (Atkinson, Heyns, & Veroff, 1954), and power (Winter, 1973) motivation. For each picture cue in the TAT, the participant is given 5 minutes to tell or write an imaginative story that describes what is happening in the picture now, what led up to this situation, what will happen in the future, and what the characters in the story are thinking and feeling. Then, the researchers conduct content analysis to detect the presence or absence in each story of a series of objectively defined themes associated with whatever motive is being studied.

In the first version of a life story interview, McAdams (1985) repurposed the standard TAT instructions to apply to the narratives respondents might provide of key autobiographical memories, such as life-story high points, low points, and turning points as well as other prompts. He developed content analysis systems for assessing variations on themes of agency and communion, among other things, following strategies validated in the TAT tradition. He also drew heavily on the concept of ego development (Loevinger, 1976) and the measurement logic employed for assessing the complexity of sentence-completion responses for individual differences in ego development. Whereas that TAT tradition captured the kind of content themes that might run through life narrative accounts, studies of ego development provided insight regarding structure, as in how structurally complex and coherent a narrative might be. McAdams reported a series of preliminary studies showing that (a) high scores on TAT power motivation were associated with agentic themes and images in life stories; (b) high scores on TAT intimacy motivation were associated with communal themes and images in life stories; and (c) high scores on ego development were associated with structurally more complex life stories, which tended to include a wider range of plot lines and what we would now label as more sophisticated forms of autobiographical reasoning.

Over the past three decades, empirical research on narrative identity has often adopted the convention of focusing on a handful of key scenes or events in a person’s life, as if they were individual TAT stories. Moreover, many researchers have followed McClelland’s lead, and that of McAdams (1985), in scoring open-ended narrative responses for objectively defined constructs, such as agency and communion themes, redemption and contamination, elaborative processing, meaning making, and coherence, among others (Adler, Lodi-Smith, Philippe, & Houle, 2016). The emphasis throughout has been on giving research participants freedom to craft their own narrative responses, reading the stories in full, and assigning scores or codes to the stories based on considerations of content and structure.

Holm and Thomsen (this volume) offer an alternative approach. They ask: Why not simply have respondents rate their own narrative accounts? Rather than read and code narrative responses, Holm and Thomsen (this volume) obtained self-ratings from respondents on what they describe as positive or negative self-change and self-stability. The self-report procedure saves time and energy for the researchers, and it may also have the advantage of tying the resultant scores themselves directly to the respondent’s own perspective on the particular memory being rated. After all, the memory belongs to the respondent, not an objective coder, so the respondent may be in a better position than an outside reader to understand what the remembered event was all about, and what it means for the respondent today.

Panattoni and McLean (this volume) approach the same issue from a different angle. They note that when respondents rate their own narrative accounts on a given dimension those ratings may not correlate highly with the scores obtained from objective coders on the very same dimension. These “puzzling mismatches” may effectively indicate lack of interrater reliability (if we think of the respondent as simply another rater), or they may suggest that self-ratings and objective codes are getting at two different things. Panattoni and McLean settle for the second interpretation, suggesting that the two different methods tap into two different features of narrative identity. Self-ratings may signal the respondent’s current conscious interpretation of the remembered event, whereas coding may tap into a mix of both consciously and unconsciously motivated aspects of narrative identity. Panattoni and McLean frame the distinction as reflecting two different conversational contexts.

The distinction between self-report and narrative coding is reminiscent of a parallel phenomenon in the literature on assessing social motives through the TAT. Since the 1960s, it has been known that self-report scores on scales designed to measure constructs like achievement motivation do not correlate highly with scores obtained from coding TAT stories on purportedly the very same constructs. Which of the two measures—self-report or the TAT—is the better measure of individual differences in achievement motivation? Like Panattoni and McLean (this volume), McClelland (1980) has long argued that the two methods tap into different things. Self-report scales measure conscious values or traits, he contended, whereas the TAT taps into less-than-conscious, or implicit, motivational trends in spontaneous narrative thought. In recent years, research on motive congruence has examined what happens for any given respondent when the two different scores are simpatico with each other (say, when a person scores high on both self-report and TAT achievement motivation) versus when they are not (Hagemeyer, Neberich, Asendorpf, & Neyer, 2013). The literature tends to suggest that congruence between implicit (narrative TAT-based) and explicit (self-report questionnaire) motives indicates an especially strong orientation in the given motivational direction and, as a result, higher levels of personality integration (Brunstein, 2008).

It is too soon to know if the methodological lessons learned in the TAT tradition are directly relevant to the issue of mismatches between self-ratings and third-person codes of life-narrative accounts. Following Panattoni and McLean (this volume), I agree that the two different methods tap into two different aspects of self-storytelling. But I would frame the distinction in a slightly different way. In the case of self-ratings, what the respondent is judging is his or her own memory as it is currently recalled and processed. In the case of third-person coding, by contrast, what is being assessed is the manifest narrative itself. The coder does not have access to all of the details, associations, and idiosyncratic meanings that the narrator has at his or her own disposal when remembering the event and telling it (or writing it down). Instead, the coder simply has the text. This would seem to give the narrator a certain advantage in being able to tap into the exigencies of the memory itself, including features that are not explicitly told or written down in the session. Therefore, if I (the narrator) rate how much emotional positivity prevailed in my recollection of the high school prom, I am rating what I remember not what I wrote down when asked to remember. I am not limited by the story I wrote or told. The coder is limited to the story. But perhaps this gives the coder a certain kind of advantage, too, though a different one. The third-person coder has no choice but to focus on the story itself. And might not the story itself—as written or told in the research moment—be a more representative expression of narrative identity itself? After all, narrative identity is not memory; it is the story told about memory.

In any case, the intriguing questions raised by Holm and Thomsen (this volume) and Panattoni and McLean (this volume), along with all of the other excellent papers compiled into this collection, show how far research and theory on narrative identity have traveled since the 1980s, and they signal important methodological and conceptual challenges that lie ahead. The papers also exemplify the creative synergies that occur when researchers from different traditions come together at the nexus of autobiographical memory, identity, and personality.