1 In this paper, I intend to examine the gaps between the authority of experience and the authority of expertise, the fissures between the theory of narrative research and the challenges of actual practices. These bring us to the heart of some of the enigmas in this work. To work narratively, to quote Clifford Geertz, is "to exchange a set of well-charted difficulties for a set of largely uncharted ones" (1983, p.6). Or, in Ruth Behar’s words, "Nothing is stranger than this business of humans observing other humans in order to write about them" (1996, p. 5).

2 Through narrative, we come in contact with our participants as people engaged in the process of interpreting themselves. What constrains us is the very thing that intrigues us: what we are dealing with are core meaning-making systems of real people and with issues that pertain to us, as humans, as well. The truths inherent in personal narrative issue from real positions in the world—the passions, desires, ideas, and conceptual systems that underlie life as lived. We work with what is said and what is not said, within the context in which life is lived and the context of the interview in which words are spoken to represent that life. We then must decode, reorganize, recontextualize, or abstract that life in the interest of reaching a new interpretation of the raw data of experience before us.

3 But as narrative researchers, we retain a responsibility to protect those who inform us, even as we return to our colleagues to relate our own narrative of what we believe we have learned. The limits that we must impose on ourselves here reflect the fact that the data that we are dealing with is core, central, important. These are not aggregated peripheral variables we are studying, but the axes on which people’s lives turn.

The Story

4 The story behind my rather unusual title begins several years ago when I was asked, along with four other qualitative researchers, to do a reading of an interview from a narrative research point of view with the intention of comparing our approaches and interpretations of the same interview and to present this at an annual meeting of the American Psychological Association. The research question was framed in terms of understanding something about processes of resilience and social support, and the interviewee, whom we called Teresa, was a 26-year-old graduate student in psychology. The question posed to the participant was to narrate a situation when something very unfortunate happened to you. Teresa told about how, at the age of 19, she had been studying to be an opera singer when she developed thyroid cancer. The surgery saved her life but destroyed her capacity to sing. Her story concerned her tragic loss of the only identity she had had or imagined for herself. "My voice was gone, so I was gone, and I’d never been anything but my voice," she said. But her story progressed to choosing another path and eventually becoming a graduate student in psychology who was taking a class in research methods, which is how she came to be interviewed. Five of us then read and interpreted the resulting narrative, with the aim being to compare our approaches. The other readers were Kathy Charmaz, working from a grounded theory approach; Linda McMullen, doing a discourse analysis reading; Rosemary Anderson, representing intuitive inquiry; and Fred Wertz, representing phenomenological inquiry. Fred had been Teresa’s professor and department chair and it was in his class that the interview took place.

5 After we did our analyses, we presented two symposia at the American Psychological Association Annual Meeting, the first offering our analyses and the second highlighting the similarities and differences we discovered among us. In these panels, we located some of the edges of our modes of inquiry and learned different things about resilience from our readings of the Teresa text. We decided to turn this project into a book (Wertz et al., 2011). Then someone raised an intriguing question: wouldn’t it be interesting to ask the real Teresa to read our analyses and comment—and to become a co-author of the book? I had my doubts about doing this, as others did, but our curiosity about this next level of response held sway. "Teresa" was quite willing to read what we wrote and to write a chapter herself.

6 Although none of us were very clear on just what we were asking of the real Teresa, I must admit that I was quite eager to read what she had to say about the analyses that we had worked so hard on. Our purpose had been to demonstrate our own approaches and also to find the intersections and clashes with the other approaches. But what did Teresa think about our analyses? Teresa, who was, after all, a graduate student in psychology studying qualitative research, was very articulate about her responses. Her chapter was exceptionally perceptive and eloquent. I read her chapter carefully, noting especially the sections—and there were many— where she felt that one or the other of us got her wrong, misunderstood or distorted her meanings. And in the midst of reading her chapter, what came to my mind forcefully as a response was "Bet you think this song is about you." I realized, powerfully, that the book was about the interview in which she was a participant—not about her. And it is this intersection that I think raises some fascinating and profound questions about the whole enterprise of narrative research in the social sciences.

7 I returned to a question that has occupied me for many years: how do we manage the problem of who in some way "owns" the material that we collect—and then interpret—in interview-based research (Josselson, 1996b)?

8 This story is undoubtedly unsurprising to those of you engaged in interpretive research. For those who conceive of their research as research with their participants, the issues are different. Here I want to consider the intersection between the authority of experience which belongs to the participant and the authority of expertise which belongs to the interpreter’s disciplinary approach. Even when we are trying hardest, though, to straightforwardly re-present the words of our participants, what we write will rarely feel entirely accurate to them. What I want to try to do is to theorize why this is so.

9 Phillip Steedman (1991), arguing a poststructuralist point of view, writes that if Puccini were to come back to life and protest current interpretations of Tosca, Steedman would reply to him, "Tormented shade, you plainly know a lot about Tosca but what we have is libretto and score; your unique role was in creating the work but now you, like us, are one of its interpreters" (p. 59).

10 The story of our five analyses of a single text gets even more complicated. After writing her chapter, the woman whom we called Teresa asked that her chapter be published under her real name—after all, there isn’t such a person as Teresa, and a very real person had crafted a response chapter. Her creative product, her chapter, she felt, belonged to her under her real name. This threw our group into quite a muddle. We consulted lawyers and ethicists who advised us that there was nothing either illegal or unethical about allowing this person, whose real name is Emily McSpadden, to do this. We cautioned Emily that she was, in her interview, referring to others who might object to being identified. But she checked with them, gave them transcripts of the interview and obtained their permission for her to publish under her own name. We warned her that in the future she might regret her decision—say, if she became a psychology professor and wouldn’t want such personal information available to her students and colleagues. Or the detailing of her medical history might later compromise her in other ways. She replied that she was aware of these possibilities and that she was an adult who could make decisions. So, somewhat hesitantly, we agreed that she could have her real name on her chapter and as co-author of the book.

11 But, I forcefully objected, I would not support changing the "Teresa" name throughout the book to "Emily." My analysis had been of the Teresa text, not of Emily. Here was an opportunity to clearly mark this separation. Emily now became one of the interpreters of the Teresa text through her interpretation of our analyses of it. Emily and Teresa, in my mind, were not identical. In addition, we have to recognize that Emily was also reading a text and the analyses of that text years after it was obtained. She herself acknowledged that her understanding of her experience had changed over the years. In literary terms, she became a guest in the text (Barthes, 1989).

12 Still, we as a group are left with a problem. From a postmodern point of view, which is foundational for a hermeneutic approach, we recognize that there are many truths, but how could we avoid giving the impression to our readers that Emily’s response to our analysis of Teresa is somehow a litmus test of the accuracy of our readings? How could we prevent suggesting that in some way she has the "truth" about the meaning of the interview? And how could we avoid our readers having the impression that somehow, in reading her response to our analyses, they were getting the right answer at the back of the book? It was even hard for me not to read her chapter this way. On what ground do we claim our own interpretive authority?

13 Our aim as narrative researchers is to seek general rather than biographical knowledge. Emily understood this. Yet, she was understandably more captured by what we said about Teresa than what we said about our methods. Her very personal response was in terms of in which segments of our interpretations she felt best understood, where she most found her experience of herself reflected and mirrored.

14 Now, we might ask, in what ways are Emily’s responses to our interpretations privileged? Although the researchers, on their part, viewed their analyses as concerning general psychological knowledge that might not match and might even contradict the participant’s self- understanding, we could not deny that the participant’s life is embodied in her data and that the analyses are based on her life as she lived it.

The Problem of Whose Narrative It Is

15 I am aware that some narrative research has as its aim re-presenting others’ experience (Josselson, 2004). This is often true of descriptive studies of marginalized groups where the aim is social awareness or action. In these cases, one can well return to the participants with a completed write-up to verify that one has accurately represented their experiences. In such cases, the issues of interpretive authority are different because there is little interpretation at levels beyond those of conscious experience, although the experience may be recontextualized.

16 I also want to say that I am aware of and forcefully advocate disguise so that anonymity and confidentiality can be preserved—so this is not an issue of any participant being recognized by others in their lives or publicly undressed. What I am interested in is either experiencing or imagining the problem of a participant recognizing his or her own narrative in our work. I think we can, in most cases, manage this adequately so that no one is harmed, but I believe that the conversation with participants, real or imagined, about what we say about them, has much to teach us about what we are doing.

Multilayeredness

17 Narrative research involves obtaining some phenomenological account of experience obtained from the person or persons under investigation, and our epistemological praxis relies on hermeneutics, a disciplined form of moving from text to meaning.

18 What we are analyzing are texts, not lives. What, then, is the difference between a life and a text? Here we face similar problems to literary analysts who struggle with the intersection of work and text. Roland Barthes writes that while we may respect the author’s intentions in a literary work, the text, in his words, "practices the infinite deferral of the signified" (1989, p. 169).

19 I am taken with Barthes’ explanation of the plurality of the text. He traces the meaning of the word "text" to its etymological origin in the Latin word "textus," which means "thing woven." The ancient metaphor, according to the etymological dictionary, is that thought is a thread, and the raconteur is a spinner of yarns. The storyteller is a weaver.

20 Some of the dilemmas of whose narrative it is in narrative research ensue from the problem of the multilayeredness of the text itself and the multivocality of human experience. There are always interwoven layers of meaning in any interview text. Beyond conscious intentionality there are other aspects that may be of interest to scholarship that lie beside or beneath intention.

21 To further complicate matters, the self is always multiple and multivocal. People are composed of discourses. Which genres they choose to frame their narration may reflect the taken-for-granted fabric of experience which may itself be of interest to scholars—or the framing may be a product of the relational setting of the interview.

22 First, I’d like to consider the implications of multilayeredness. Many of the meanings woven into the text are beyond the conscious intention of the person who creates the text with us in the interview situation. They join us as interlocutors in the conversation about their lives, but not as analysts of the resulting text. How could a participant possibly know, for example, that we will be paying close attention to the gaps in their speech, their use of "I don’t know," the sequence of topics they address or the ways in which words or images are repeated? As narrative analysts, we may take all these phenomena of textual discourse very seriously—and interpret them freely—while the participant, reading our analysis, is left only to somewhat inarticulately blubber: "but I wasn’t aware of that; I didn’t mean anything by that."

23 Every aspect of narrative work is interpretive, as everything implies meanings. When we use narrative interview texts in our research, we are co-constructing meanings at different levels of analysis. The research process itself involves socially constructing a world with the researchers included in the body of their own research. We, as researchers, "coproduce" the worlds of our research. We don’t simply "find" these worlds (see Gergen, 2009). Truth is primarily a matter of perspective, as the philosophers of hermeneutic science have argued.

24 I have elsewhere adapted from Paul Ricoeur a distinction between a hermeneutics of restoration and a hermeneutics of demystification (Josselson, 2004). A hermeneutics of restoration aims to be faithful to the text and restore its explicit and implicit meanings. The purpose is to absorb as much as possible the message in its given form and to re-present, explore or understand the subjective world of the participants or the social and historical world they feel themselves to be living in. The interview thus provides a window on psychological and social realities of the participant. By contrast, a hermeneutics of demystification regards the text as disguised. Signs are read according to some procedure of meaning-making, some preexisting code book. Within psychology, for example, psychoanalysis may offer a way of reading defenses or unconscious conflict expressed in symbols. Or we might do feminist or Foucauldian or Marxist readings. Analysis from a position of the hermeneutics of demystification may try to identify what is unsaid or unsayable through, for example, attention to negation, evasion, revision, denial, hesitation, and silence as indicative of important memories or psychological experiences that cannot be expressed. Especially in cases of trauma or abuse, significant biographical realities may be expunged from the narration but nevertheless cast an interpretable shadow on the text. From the position of a hermeneutics of demystification, attention is directed to the omissions, disjunctions, inconsistencies, and contradictions in an account. It is what is latent, hidden in an account that is of interest rather than the manifest narrative of the teller.

25 Both of these—the hermeneutics of restoration and the hermeneutics of demystification—are interpretive stances, not properties of a text itself. The closer we are to a hermeneutics of restoration, the more likely the participant is to feel well understood. As we move to position ourselves within a hermeneutics of demystification, the voice and logic of the interpreter take center stage and participants are less likely to find a match with their own experience. Think of Freud’s interpretation of the Dora case. He may have gotten Dora all wrong, but he did show us how to think psychoanalytically. As interpreters, we appropriate an interview text and subject it to our interpretive gaze. But this is seldom a process we explain to our participants.

26 To be sure, the person storying his or her life is already interpreting experience in constructing the account. The researchers’ task is to offer a telling at some different level of discourse.

27 Even the most faithful stance within a hermeneutics of restoration reorders meanings. If we are working from an interpretive rather than a purely descriptive approach, we are not speaking for our participants. Rather, we are speaking about the texts we have obtained from them. At some point in her response chapter, Emily realizes this. She writes: "After reading the different analyses of my data, I find myself asking an interesting question: What was being studied? Resiliency? Me as a person who exhibited resiliency? Or both?"

28 The primary goal of interpretation is not the passive repetition of what the speaker told us. When we listen to another's story our intention is to bring our own interpretation to his or her material. We take interpretive authority and we need to make this explicit. Even if we ask our participants to corroborate our interpretation, it is still our interpretive framework that structures understanding.

29 As with any work, each observer interprets from his or her own meaning-making horizon. It is our own positioning as creators of the research questions, as co-creators of the relational context of the interview, and as readers of the text that claim our ownership of the narrative we create. The meanings we derive from a text were not always already there in the participant. But I think that we sometimes get confused about this. It is hard to escape our embeddedness in a modernist, realist worldview—a worldview we can cognitively disown but still often emotionally rely on to ground ourselves. I say this out of repeated experiences of reading published work that declares itself as positioned in a relativist epistemology and then quickly becomes realist.

30 Co-construction of a text does not imply that we have similar aims as our participants or that we are somehow working together to produce the research results. If we have done our work well, we are likely, in some ways, to offer a dissonant counterpoint to their self-understanding.

31 One inherent and indelible contrast between us and our participants is that while we interrogate their specificity in quest of what may be generalizable, their interest remains lodged in understanding their own lives and, perhaps, how their lives compare to other lives. We write as scholars for our peers about some aspect of our participants’ lives that advances our understanding of some conceptual or social phenomena. But our participants are not, in most cases, invested in our scholarly aims.

32 When we write about people who have offered us finely tuned, carefully balanced stories of themselves, both the context and proportion are lost. The vital conversation of the interview morphs into the flattened discourse of scholarship. What we write about them may be an aspect of themselves which, while conscious at the time they revealed it to us, may be repressed and no longer available or real to them by the time it sees print. As a result, the written narrative is no longer emotionally "true." Participants may not be conscious of social forces which operate in and through them, social forces which become apparent only when narratives are juxtaposed. Such focus may feel odd or disorienting to the person whose discourse it is.

33 Labov and Waletzky (1967) depict all narratives as having clauses which either orient the reader to the story, tell about the events, or evaluate the story—that is, instruct the reader as to how the story is to be understood. We can get agreement with our participants about the events themselves. These are the "facts" of the exposition. But it is the evaluation of events that is of primary interest to the narrative researcher because this represents the ways in which the narrator constructs meanings. Such meanings, though, are neither singular nor static. As we read a narrative, we can detect the evaluation the participant offers—the cues that tell us how he or she thinks the story should be read, what it means to the participant. As interpreters, though, we will also be reading for different cues to different forms of meaning-making.

34 The best example from our experience with the Teresa texts came from Linda McMullen’s discourse analysis of the text. Discourse analysis is not about the person, but rather about a particular pattern of speech. It can easily be misunderstood because the exclusive focus on language patterns as social performance is quite different from the way we usually view language in everyday life. Linda McMullen’s analysis focused on the discursive performance of what she called "enhancing oneself and diminishing others" (Wertz et al., 2011, p. 211) as an instrumental social practice that might be related to resiliency. She theorized this talk as occasioned by a culturally-relative discursive pattern of "doing resilience" (p. 210). While Linda was studying talk as a cultural pattern, Emily took these findings in reference to real events in her life and found this analysis rather insulting. She wrote: "While I didn’t mention the involvement of people in certain elements of my experience of my illness and resulting hardships, I didn’t see how it might imply I was really ignoring a presence and support that was actually there … Had I really been such a self-aggrandizing braggart (or "self-enhancer") in giving my interview?" (Wertz et al., 2011, p. 362). Linda McMullen replied to Emily that she meant to suggest no such thing. She was, indeed, not making any personal comments about Teresa at all, rather regarding her as a conduit of social practice. It was of interest to me that I framed and discussed a similar issue in my thematic account, but Emily did not take offense to the way I phrased and contextualized it. Thus, it is not only the interpretation we make but how our writing about what we discern is phrased and how this phrasing is read by the participant. As always, there are layers of interpretation.

35 Emily found some of our accounts and re-presentation of Teresa’s experience "overly dramatic" and seemed unsure whether the drama had come from her or the interpreter. One account she found to be "idealized and aggrandizing" (Wertz et al., 2011, p. 344). She worried that my construing her text as a story was suggesting that her experience was fiction rather than fact. Only in the phenomenological analysis did she never feel misrepresented. But we as a group were left to wonder if this is because phenomenological analysis is the closest to experience or because this analysis was done by the one of us who knew her and was in a personal relationship with her.

36 When we write, we can usually only speak about one meaning at a time and showing the dialogic interrelationship of meanings as well as their intersubjective construction challenges even the most artful writer. Even as I tell you this one fairly bounded story, there are many layers of meanings and many that are omitted.

37 Thus, one explanation for the disjunction between our readings of a text and the person whose life it is based on has to do with the inherent multilayeredness of texts and experience.

Multivocality

38 Another way of looking at the tensions inherent in this disjunction concerns the view of the self as either singular or multiple. A traditional view of the self may investigate agency, morality, power relationships, or authenticity and regard people as having unified and static selves. The postmodern view of the self decenters the self, sees the self as a linguistic construction composed of shifting positions. If we posit a relationally-constituted self, then the self appears as fluid and multiple, and is recreated in different contexts (see Gergen, 2009). This view assumes the self to be polyphonic and dialogic, a set of characters or voices in conversation with one another and with others. Thus, the notion of "giving voice" to our participants just muddles our thinking. To which selves are we giving voice?

39 Here I rely on Bakhtin’s view of the self, which is relativistic. The self can exist only in relationship to some other, whether that other be another person, other parts of the self, or the individual's society or culture. In this view, the individual is always in process. Bakhtin says that if our image of being human includes the dynamics of inconsistency and tension, people no longer coincide with themselves and cannot be entirely captured by the plots that contain them (Bakhtin, 1981).

40 This, I think, is a key to another disparity between lives as lived by our participants and our work on their texts. Their protest can, in part be seen as their awareness of not being fully represented in the plots of their lives as they are constructed in the texts that we analyze.

41 In Bakhtin's view, reality is always too contradictory and multiple to be fit into a straightforward genre. Human beings are always in process, existing on multiple planes of present experience, poised in complex relation to the past and to the future.

42 The best we can hope for is to represent only some of the selves, or partial selves, of our participants. This is where the relational context of the interview becomes exceptionally relevant. When we assert that the interview is co-constructed, we are recognizing that we are co-creating a dialogue between aspects of ourselves and aspects of our participants. Which voices emerge to narrate is determined by our impact on the participant and his or her assessment of us in terms of who the participant thinks the audience is. Our participants are performing with us particular constructions of themselves in response to whom we seem to them to be and what we have asked them to tell us about. In the story of the Teresa texts, then, we can only understand Teresa as the parts of self elicited by telling her story to a fellow student who was doing his first interview, filled with preconceptions, within the context of a class. To whom the story is told shapes the telling by its calling out certain aspects of self. This is why I remained adamant that we continue to call the person whose texts we analyzed "Teresa." I wanted to mark forcefully that we weren’t analyzing Emily, who is a great deal more, in the sense of having many more selves, than Teresa. And I don’t know Emily.

43 I had a memorable experience of the multiplicity of selves recently after a 4-hour interview with one of the participants in my 35-year longitudinal study of women’s identity (Josselson, 1996a). I have re-interviewed these women every ten years. The most recent interview took place in my participant’s home in Maine and was an intense, open, and rich dialogue. This particular participant was one who told me that each interview over the years has led her to learn something new about herself and she was much looking forward to one more experience of what has been for her a quasi-therapeutic encounter. This has to do with her experience rather than anything I have been doing, but it has led her to be very self-revealing. In the interview, I felt I had had a view of many of her selves, most of her important experiences. I felt very close to her. The next day, after I left Maine, I had a logistical matter to settle and I called her to discuss it. Her husband answered the phone and told me that she had left right after the interview to attend her mother who was critically ill in a hospital in New Hampshire. I was dumbfounded by this news—and still cannot fathom why she had not told me this during the interview. What does this say about the selves that she kept in the shadow? What about my sense that she had been so fully disclosing and open with me? What explanation could there be for why she kept this hidden when she must have been in great distress all the time we were talking? I was powerfully reminded that we, as narrative researchers, seek understanding without certainties.

44 There is always more to the story with our participants. They may not even be aware, months or years after an interview, of what they told us and what they did not. To them, reading our analyses, an experience of what we left out may lead to a sense that we got it wrong. Any thread, encoded in a statement, exists in a complex context of many selves and we can never embed it fully in linear language. We give linguistic labels, often single words, to experiences that reverberate differently to the various characters contained in the narrative. In my analyses of the Teresa text, I regarded Teresa as narrating her rational, goal-directed self while embodying her frantic, terrified self in others. This was a property of the text, not of her experience. Emily’s protest that she had a lot of feelings at the time is, then, understandable, but it does not negate my interpretation.

45 Stories end. Selves evolve. Bakhtin stresses that human beings are defined by their "unfinalizedness." We retain always the capacity to surprise ourselves and others. Context, to Bakhtin, encompasses "infinite dialogue in which there is neither a first nor a last word" (1986, pp.167-68). As narrative researchers, we enter into this ongoing dialogue, both with our participants and with our colleagues.

Narcissistic Tensions

46 Most of you are familiar with the wonderful Carly Simon song from which the title of this paper is taken. The title of the song is "You’re So Vain": "You’re so vain, I’ll bet you think this song is about you," she sings (Simon, 1973). So what’s vanity got to do with all this?

47 Some years ago, in an effort to understand more about how what we write about our participants affects them, I undertook a project to re-interview the participants I had written about who I knew had read the books that resulted from the studies in which they had participated. Ever since, I have been attending to the responses of my participants to what I write.

48 Acutely aware of the power of the written word and also aware of the ways in which I was making use of my participants’ lives to further my own academic career, I approached these interviews about the effects of my writing about them with more trepidation than I had ever had in any interview (Josselson, 1996a). Although I had talked easily with these people about the most private details of their lives, I felt intensely anxious and uncomfortable talking with them about how it was for them to find their lives in print—and in words which I had written. My impression was that they were also anxious and uncomfortable with this topic. It was my vanity as well as theirs that was at stake in these conversations. There are narcissistic tensions aroused in all of us by this kind of work.

49 It was difficult to engage most participants in this discussion. They told me they hadn’t read what I wrote or didn’t remember what I said or that what I said was "fine." Elliot Mishler similarly reports that his participants "didn’t care" (Clandinin & Murphy, 2007, p. 649) about what he wrote about them. At some level, I think this is true as people have gone on with their lives and, while the publication of my book may have been a major event in my life, it was an often unnoticed blip in theirs. At the same time, I also think that our writing has effects on participants that may seem to them too internal, irrational, or idiosyncratic to talk about without a great deal of encouragement to do so. I tried to learn from what could not be said—or was said haltingly or incompletely. There seemed to be something discomfiting in these interviews. I felt I was being asked implicitly not to talk about this, much like pretending it didn’t happen. I was left with the distress of not being able to name and understand my own discomfort, which felt like some mixture of shame, guilt, and dread.

50 Participating in this process of sharing one's life to be written about by someone else stirs up a welter of narcissistic tensions in both the participant and the researcher. I turned then to theories of narcissism to better understand what was going on. Heinz Kohut, the premier psychoanalytic theorist of narcissism, has what I found to be a useful explanation. When we work empathically in the interview situation to understand our participants, we may be evoking what he calls a "mirror transference" and feel to them like we carry core aspects of themselves. Such an experience with another person, in Kohut’s formulation, is deeply rewarding but unconsciously evokes vulnerable aspects of the self, which he calls the grandiose self (Kohut, 1977). We have, indeed, aggrandized our participants by regarding them as important enough to write about – but the grandiose self is always tinged with shame. Some of my participants felt embarrassed about how pleased they were with what I wrote if it seemed positive to them. Others felt wounded by what I left out; experiences that seemed central to them were omitted from my written account— didn’t I get how meaningful these experiences were to them? For others, what was most wounding was being a minor character among a group of interviewees—as though I had indicated that they weren’t as important to analyze in detail as other people. And most wounding of all was being left out altogether.

51 One memorable interview for me was of a man, someone I know beyond his role as an interviewee, who recognized that the portrait of himself he had painted in his interview was of the false self he had been living. "Look," he said, "there is a lot I didn’t tell you—and you would have a very different impression of me if I had."

52 These issues are close to the core of personality organization: "How important am I?" "What selves am I actually living and presenting?" "What do you really think of me?" They are extremely difficult to discuss—but these are the personality dynamics that we rub against when we do narrative work with living people.

53 And vanity exists on our side as well. When Emily sent to our group the draft of her chapter, I skimmed through until I came to the section where she writes about my analysis. I know that at least two of the others did the same. What was on my mind first and foremost was: "What did you think about what I wrote?" Another version of "What do you think about me?"

54 We are, none of us, ever free of our need to be recognized for what we feel ourselves to be; to be, in other words, mirrored—or ever free of our wish to be valued. These are, in Kohut’s view, aspects of healthy narcissism, but these wishes are always complex and tinged with painful feelings.

55 I tried to analyze my own narcissistic dynamics that were causing me so much anxiety talking to my former participants about what I had written about them. I could acknowledge that I wanted to be affirmed and valued by my participants for what I had said about them, however irrational that wish might have been. But I wanted to understand better my own discomfort, the mixture of dread, guilt, and shame which goes with writing about others and then encountering them afterwards.

56 The dread is easiest to trace. There is always the dread that I will have harmed someone, that I will be confronted with "How could you say that about me?" I will discuss later the one time this happened to me.

57 The guilt is more complicated. My guilt, I think, comes from my knowing that I have taken myself out of relationship with my participants (with whom, during the interview, I was in intimate relationship) in order to be in relationship with my readers. I have, in a sense, been talking about them behind their backs and doing so publicly. Where in the interview I had been responsive to them, now I am using their lives in the service of something else, for my own purposes, to show something to others. I am guilty about being an intruder and then, to some extent, a betrayer. This, too, is a part of me and is narcissistically difficult to manage. I realize I need to bear this guilt rather than to build intellectual rationalizations to quell it.

58 My shame is the most painful of my responses. I suspect this shame is about my exhibitionism, shame that I am using these people's lives to exhibit myself, my analytical prowess, my cleverness. I am using them to advance my own career, as extensions of my own narcissism and fear to be caught, seen in this process. Narrative research is narcissistically unsettling both for us and our participants.

59 Oscar Wilde wrote, in The Picture of Dorian Gray: "The nineteenth century dislike of realism is the rage of Caliban seeing his own face in a glass" (1891/1988, p. 3). We confront a similar conundrum when we place our analyses of texts into the gaze of the person who authored the text. When we write about others, we entangle ourselves in others' intricately woven narcissistic tapestries as well as our own. And I see no way out of this.

Implications and Practicalities

60 What, then, are the implications of these considerations for our practices? Our challenge as narrative researchers in relation to our participants is both to respect their subjectivity and to claim our interpretive authority, which always involves objectifying them in some ways. My aim is to try to overcome the dichotomy rather than to take sides.

61 Narrative research involves exploring some conceptual question through interviewing people (or obtaining other narrative texts) and then analyzing these texts with reference to the research question. Our "five ways" group, in my view, was investigating the Teresa texts for a better understanding of processes of resilience following difficult experiences and in order to contrast our approaches. The focus was clearly on the processes, not on Teresa. I think that if we get this clear at the outset, some of these dilemmas about whose narrative is represented in the final report become less confusing.

62 When agreeing to participate in narrative research, a person voluntarily submits to become a subject of others’ interpretation, an object of the gaze of the other. It is this that, I think, we do not make explicit enough in our contract with interviewees about their participation. We recognize that our participants have their subjectivity, which includes their meanings, self-understandings, and self-knowledge. Their reading of our writing is from an assumption that we are writing biography. But our task is not biographical. Rather, it is to mine some understandings about general human processes from a meticulous and intensive reading of a particular person or several people’s storied experience.

63 We write, as scholars, for our peers about our participants. How can we keep the distinction between a focus on the person and a focus on a phenomenon or process separate when they are intertwined? This is both a distinction in terms of our understanding of what we are doing and an ethical question. We have to hold this doubleness at all levels. And I think that deluding ourselves into thinking that we are humanists or feminists and fully honoring our participants does not get us out of the dilemma—unless we publish nothing but our interview text, and even that has its problems. We have an ethical obligation to respect the relationship with our participant and also a scholarly obligation to be in conversation with our peers.

64 The ethical problems here ensue from the fact that, in order to obtain rich and meaningful material in the interview, we enter a relationship of trust, respect, and empathy with our participant—an "I-Thou" relationship in Buber’s terms. We then take ourselves out of this relationship to communicate about some conceptual matter with our peers, making use of the interview material in an "I-It" manner.

65 The concept of "informed consent" is a bit oxymoronic, given that participants can, at the outset, have only the vaguest idea of what they might be consenting to. In the chapter I wrote on ethics for Jean Clandinin’s Handbook of Narrative Research, I suggested that we need to request informed consent both at the beginning and at the end of the interview (Josselson, 2007). This is because participants often don’t know at the outset what they will tell us. If we interview well, we may often evoke disclosures that the participants were not prepared to reveal when they began to speak. Therefore, we need to ask them again if they consent to our using what they have revealed—after they have revealed it.

66 I now think that we have an ethical obligation to go even further.

67 As I said before, I have pretty systematically over the years asked my participants whom I have written about to react to what I have written. No one has ever said they were harmed, although many were affected in different ways. Most people have read what I have written with mild curiosity and gotten on with their lives. My studies, of course, are very important to me and relatively unimportant to my participants, most of whom enjoyed the opportunity to tell me about themselves but cared very little what I did with their stories. Gail Agronick and Ravenna Helson (1996) did an empirical, questionnaire-based follow-up study of the effects of being written about which demonstrates that different people react differently to this experience. Indeed, it seems that those most intent on their own personal growth are the ones who take our observations most seriously.

68 I want to tell you here about one of the participants in my 35-year longitudinal study of women’s identity development who had the most profound effect on me in regard to these issues —the one I mentioned before who caused me dread in feeling I had harmed her. This is a very self-aware woman—she has been that way over the 35 years that I have known her. From her I learned a lot about a particular form of identity formation that Erikson called identity diffusion. Debbie, the pseudonym I gave her, was a main character in both of the books I wrote about these women. When I invited her for an interview this last round, when she was 56, she thought about refusing to participate. But she chose to come and confront me instead. What she most objected to was being placed in the identity diffusion category, a placement that I had made according to well-standardized criteria when she was 21 years old. And, although she recognized that I represented a lot of her experience accurately, she was troubled that she couldn’t see her whole self in what she read as my portrait of her. Here is what she told me when last I saw her: "I was so angry. I was mad and sad. I was angry because I felt like there was not enough time for me to be heard, it was so unfair, and I was sad because I wanted to be heard. The experience was and has been over time that it’s not the whole story—it’s not all of me and it’s actually quite uncomfortable. And this time I actually thought about not coming because it’s difficult for me. I guess what I have to work on is the issue of being seen, which is so important to me. I want to tell you I haven’t felt heard and this is very difficult and I see you want me to tell you what I actually think in my head and that seems vulnerable and dangerous." I then tried to explain to Debbie what I thought I was doing; I tried to tell her that I was using what I learned from the interviews as exemplars of particular forms of identity formation. I said to her, "Someone who reads the book can learn something about their own life by reading an exemplar and it doesn’t really matter that something about a particular person is left out. Still," I told her, "I recognize that reading what I wrote must be like going into a fun house and looking in one of those distorting mirrors – there are some things that match and some things absolutely don’t." I invited her at this time to write with me about the dilemmas of writing other people’s lives, but she hasn’t had time to follow up on this. In preparation for this talk, I asked her about her further thoughts, and she emailed me the following:It is interesting that after we met the last time and I was able to speak to you about my feelings and you responded in kind, the intensity of my feelings about being a participant diminished greatly. Our genuine conversation changed my experience of the relationship from one of I-It to I-Thou and my feelings of being an object and not a subject, of being viewed through the narrowing range of a microscope, of being pinned down like a butterfly on an observation board were ameliorated as we sat and spoke face to face: two subjects, two voices, two women.

69 So this helps me think further about the ethical dilemmas here: I now phrase the problem to myself as how we stay both in relationship in the "IThou" sense and also warn our participants that we will be taking another stance in regard to them and making use of their narrated experiences for other purposes. We have to ask our participants to hold in their minds the same doubleness we strive to manage—the ways in which our writing both will be and will not be about them. Ethics may involve telling our participants at the end of the interview, in simple and direct language, that we have come to understand a great deal about them and have also learned a great deal from them about the topic that we are researching. We need to tell them that what we write will be about the topic and not about them so that what we use from the interview may not be exactly as they have told it to us and how we understand it at this moment. In other words, we have to be prepared to discuss with them something about our actual research process. We need to tell them that we will review the text and use material that we think will illustrate processes that may be true of others as well. Our accounts will be "based on" their interviews. We may use some of their material, depending on our purposes, or we may not be able to fit any of it into the final report. We caution them to expect the distorting mirror and remind them that this will not mean that our understanding of them that we are experiencing in the present moment is lost or less meaningful to us. We stress that we have learned from them and will use this learning in our work. We express our gratitude to them for teaching us about their experience.

70 I think also that reflexivity becomes ethically necessary in the written account, not only in relation to our participants but also in relation to our readers. We need to say who we are as interpreters who bring our own subjectivity to the topic or people we are writing about. Interpretive authority cannot be implicit, anonymous, or veiled. We have to come out from behind the curtain and say who we are who are claiming our authority. This is not an argument for self-indulgent autobiography threaded into our texts. Rather, I am advocating self-disclosure that reveals what we are bringing as interpreters of texts.

71 When Emily responded to our analyses of the Teresa texts, in which we made apparent who we are as interpreters, she commented on her appreciation of this. She valued both how we responded to her emotionally—all of us mentioned being moved by her experiences—and she was also responsive to our disclosures of what experiences we had had with cancer. I think that there was something about our willingness to expose something of ourselves that mitigated the "objectification" of our analyses. Such revelation also invites our readers to form an image of who it is who is doing the interpreting, who is guiding their understanding. Who we are constitutes part of our credibility and should be open to critical scrutiny.

72 In this talk, I have been exploring the dialectic between connection and otherness, between the relationality of the interview and the disconnection of the interpretive process, between the illusion of objectivity and the equally pernicious illusion that we can fully represent subjectivity. I think we have to find a way to, in effect, invite our participants to recognize with us these dialectics.

73 In the story of the five analyses of the Teresa interview that I have been threading through this presentation, we invited Emily to become a "guest in the text" she had created. In addition, we invited her to be an interpreter of the texts we created. Here were her final words on the matter:I see now that so much can be ascertained from one experience, that several different readings of a story, with an aim to understand what has been lived in it, can often bring about crucial insights about the human condition. I don’t know, even now, if the words I spoke and wrote, as they appear for the purposes of this project, can ever adequately convey my experience on the whole. I can at least attest to the fact that they faithfully represent at least some aspect of it, the one I was able to communicate in this way. If this has proven enough to shed some light on the research and its questions, I can ask for no better than that. (Wertz et al., p. 351)

74 Ruthellen Josselson, PhD, ABPP, is a Professor of Clinical Psychology at The Fielding Graduate University and was formerly a Professor at The Hebrew University of Jerusalem and a Visiting Professor at Harvard University. She is the author of Playing Pygmalion: How People Create One Another (Jason Aronson, 2007); Revising Herself: The Story of Women's Identity from College to Midlife (Oxford, 1998); and The Space Between Us: Exploring the Dimensions of Human Relationships (Sage, 1995). She has co-edited eleven volumes of the annual, The Narrative Study of Lives, and is a founder of the Society for Qualitative Inquiry. She received both the Henry A. Murray Award and the Theodore R. Sarbin Award from the American Psychological Association and a Fulbright Fellowship.