Memories of Who We Are: A Preliminary Identification of Autobiographical Memory Functions in Recall of Authentic and Inauthentic Events

The story of who we are is central to our sense of authenticity and this story is constructed from our autobiographical memories. Yet we know surprisingly little about the functions that autobiographical memories of being authentic serve. This study provides a preliminary examination of the self, social and directive functions used in autobiographical memories of being authentic and inauthentic. Participants recalled times they felt they had been authentic or inauthentic at work. Analyses revealed that the self and directive functions were significantly more prevalent than the social function. In addition, authentic memories were most strongly associated with the self function while inauthentic memories were more likely to be used for the directive function. This may indicate that recall of an authentic experience serves to support one’s current self-identity, while recall of an inauthentic experience provides an opportunity to direct future behaviour towards a more authentic response. This study provides some of the first evidence for how autobiographical memories of being authentic or inauthentic may function in developing a coherent story of self that is needed for a sense of authenticity.

• We provide first evidence that memories of authentic and inauthentic moments in an everyday context (the workplace) are associated with distinct autobiographical functions. • Recalling a time we felt authentic serves to develop or support our self-identity, contributing to the story of who we are. • Recall of a time we felt inauthentic is used to direct our future behaviour: inauthenticity, although a negative experience, has a role in goal-directed behaviour.
Authenticity, or the sense of being true to oneself, is held to be an essential component of a fulfilled life by both philosophers and psychologists. A recent meta-analysis collating data from the past decades provided strong support for this view by demonstrating that authenticity is positively related to engagement in work and to well-being in general, and is therefore of distinct value for individuals, employing organisations and society at large (Sutton, 2020). Autobiographical memories have a significant role to play in our understanding of authenticity: they form the basis of our life story, the construction of 'who we are' that is key to any sense of authenticity (Harter, 2002). Although progress is being made on identifying the functions utilised in different types of memories, such as single or repeated events  or distant and recent memories (Lind et al., 2019), we know little about what functions may be served by memories of specific times one felt authentic or inauthentic. Evidence indicates that negative memories can be both adaptive (for example, in providing motivation for change) or maladaptive (for example, an associ ation with depression) (Burnell et al., 2020). Inauthenticity is commonly perceived as a negative experience but could memories of these times also fulfil adaptive functions in the ongoing composition of our life story?
In this study, we seek to identify the autobiographical memory functions used when people recall times they felt authentic or inauthentic at work. We use the work context as previous research has called for creative research approaches to explore how personal authenticity interacts with organisational constraints such as job requirements or power imbalances (Cha et al., 2019). Identifying the functions used in these autobiographical memories will help to elucidate the role that autobiographical memories may play in navigating the everyday challenges and benefits of authenticity at work. time: the true self is therefore captured by those personality traits that remain stable and consistent (Sheldon et al., 1997). However, research evidence indicates that people can feel authentic even when they report differing personality traits in different roles or relationships (Sutton, 2018). The second view of authenticity therefore holds that an authentic person is one who has a coherent or congruent understanding of the self, rather than an unchanging set of traits or behaviours. While this understanding may change or even be inaccurate, it serves important functions in goal-setting and aiding social interactions (Sheldon, 2013).
Authenticity involves both self-knowledge and self-expression, or as Harter (2002) notes, owning one's personal experiences and acting in accordance with the true self. The construction of an authentic self is therefore a continuous process of crafting the story of one's life, including sometimes inconsistent behaviours, in a way that forms a coherent self-concept (Boucher, 2011). This construction relies on the recall and integration of autobiographical memories and autobiographical memory is recognised by researchers in both authenticity and memory as being central to maintaining a sense of self-continuity (e.g., Baldwin et al., 2015;Bluck & Liao, 2013).
The first steps in understanding the effect of recalled memories on authenticity have already been taken. For example, recalling a memory of helping other people or achieving a goal is associated with a higher level of authenticity (Smallenbroek et al., 2017), while nostalgic memories both increase past-self authenticity and may act as a resource for people in contexts that do not allow for authentic self-expression (Baldwin et al., 2015). However, the functions served by autobiographical memories of experiences of being authentic or inauthentic have not yet been directly addressed.

Autobiographical Memory
Autobiographical (AB) memory consists of a coherent story of self developed from mem ories of events and knowledge of the self (Fivush & Graci, 2017). These AB memories are used in three different ways in everyday life: fulfilling self, directive and social functions (Bluck et al., 2005). While other functions have been identified, for example, emotion regulation or teaching others, they can be viewed as sub-categories of these three broad functions (Burnell et al., 2020).
The self function is concerned with self-knowledge, using memories of past experien ces to develop continuity and coherence in the self-concept. For example, a memory of the past may be used to determine the extent to which one has changed or developed over time (Bluck & Alea, 2011). This evaluation is seen as providing a link between who I am now and who I was then, creating a coherent developmental understanding of sometimes discrepant behaviours or attitudes.
The directive function incorporates the use of memories to problem-solve, provide guidance or plan for the future. This can include using the past to understand unexpec ted events or experiences, to aid with emotional regulation and for constructing a sense of meaning (Bluck et al., 2005). Finally, autobiographical memories can be used to build and maintain social bonds. This may be done by sharing those memories with others, using memories to learn more about the other person, and developing intimacy and closeness in relationships (Bluck & Alea, 2011).
In this study, we test the extent to which these three autobiographical memory functions may be used differentially in memories of authentic and inauthentic events. We expect that memories of being authentic will serve different functions to memories of being inauthentic.
First, as noted earlier, self-knowledge is an essential component of definitions and measures of authenticity, as without knowledge of our past behaviours, thoughts and feelings, we cannot build the story of who we are or evaluate our current authenticity (Kernis & Goldman, 2006;Wood et al., 2008). This knowledge of how the self has developed or remained stable is a key element of the self function and we expect, therefore, that recall of times when a person felt authentic will most frequently use the self function of autobiographical memory.
Second, authenticity is closely related to goal-driven behaviour, from philosophical definitions of authenticity as concerned with owning one's actions through to psycho logical conceptions of authenticity as self-determined behaviour (Harter, 2002). Research has shown that individuals' reflections on times they behaved inauthentically reflect a concern with problem-solving or directing their current and future behaviour (Sutton, 2018). Recall of times a person felt inauthentic is therefore likely to serve the directive function more than the self or social.
The potential relationship between the social function and memories of being (in)au thentic is less straightforward. Although the social function is often considered to be the most important function for autobiographical memories, it does not always appear to be the most frequently used in either self-report or expert coding measures (Bluck & Alea, 2011;Waters et al., 2014). In fact, the social function is more likely to be used in recall of extended and recurring events than in recall of single events . As we are investigating recall of single events in this study, we expect to find only a minor use of the social function. Further, there is something of a paradox when considering authenticity in a social context. While authentic self-expression is important for healthy relationships (Brunell et al., 2010), self-expression that is tempered by the social context can still feel authentic (Wang, 2016). Given these complexities, we do not expect to find that the social function is used any more or less in authentic or inauthentic event recall.
In summary, in this study we examine participants' recall of times they felt authentic or inauthentic for the prevalence of autobiographical memory functions. We expect that authentic events will be more likely to serve a self function, inauthentic events will be more likely to serve a directive function and neither type of recall will strongly serve the social function.

Method
To test our hypotheses, we reanalysed archival data that was collected as part of a larger research project which included qualitative accounts of authenticity at work (Sutton, 2018). Participants provided weekly, diary-like written narratives over the course of six weeks in response to prompt questions. We focus here on two of these narratives: the responses where participants were asked to recall a time they felt authentic or inauthentic at work. We adopt a quantitative coding scheme to evaluate the extent to which autobiographical memory functions are used in the recall of these events.

Participants
Participants (N = 29) were full-time employees (70% female) with a mean tenure of 4 years in their current jobs (SD = 4.6), mean age of 36.8 years (SD = 9.7) They had previously taken part in a research project aimed at differentiating between feelings of authenticity and role personality (that is, the extent to which an individual expresses dif ferent personality traits in different roles). From the original sample of 191 respondents, a stratified sample was invited to take part in this second stage: 12 participants were selected from each of four different 'strata' representing high or low scores on the two measures of authenticity and personality differentiation. Although not directly relevant to the aims of this study, this stratified sample ensured that data was collected from participants representing both approaches to understanding authenticity reviewed in the introduction, namely feelings of congruence and extent of differentiation between role personalities. From these 48 participants, 26 provided recall narratives for both events, two provided authentic event recall only and one provided inauthentic event recall only, giving a total of 57 narratives and an overall response rate of approximately 50%.

Procedure
Once per week over the course of six weeks, participants were emailed a unique link to a short online questionnaire containing open-ended questions on the following topics related to authenticity at work: personality differentiation, experience of inauthenticity, experience of authenticity, efforts towards self-integration at work, authenticity in work relationships, and reflection on the research participation. Participants were asked to give written responses to these questions, with prompt questions to encourage elabora tion. Reminder emails were sent after three days and those who completed all six weeks were given a small voucher as thanks for their participation.
In Week 2, participants were asked to recall an experience of feeling inauthentic at work: We all take on different roles in our normal lives, whether at work, with friends or family. Sometimes this can mean "playing a part" or acting in a way that suits the situation rather than a way that feels like the "real you". Can you think of a time you did this at work?
In Week 3, participants were asked to recall an experience of feeling authentic at work: In the last questionnaire, we asked you to write about a time when you behaved in a way that seemed to suit your work role but was not necessarily your "natural" or preferred way of doing things. This time, we would like you to think of a time when you felt able to really "be yourself" at work and weren't just taking on a role.
Each of these questions were followed by questions encouraging participants to elaborate on the memory. These prompts were structured using the critical incident technique (CIT) (Flanagan, 1954) which has developed into a valuable technique for gathering retrospective reports of psychological experiences (Butterfield et al., 2005). Participants were asked to describe what happened, using the following questions as prompts: • What did you do? How was that different to how you would normally be? • Why did you act this way?

Measures
Recall narratives were scored by the authors using Waters et al. 's (2014) autobiographical memory function coding scheme. This scheme assesses the expression of each of the self, directive and social functions in a recall narrative, on a 4-point scale from 0 (no content expressing this function) to 3 (extensive elaboration of this function). Each narrative there fore receives a score of 0-3 on each autobiographical memory function. All narratives in this dataset received at least one nonzero score.
In the original study reporting on this coding scheme, disagreement between raters was resolved by adopting the coding developer's scores . In our study, each researcher coded a sample of narratives independently, then met to discuss and agree on any discrepancies in coding before independently coding the remainder. Final agreement (Pearson's correlations) between the coders was good: r = .66, .70 and .84 (p < .001) for the self, directive and social functions respectively. The mean of the two raters' scores was therefore utilised in the following analyses.

Results
Descriptive statistics and correlations between variables are reported in Table 1. Al though the mean word count for the inauthentic event recall was significantly higher than for the authentic event, t(25) = -2.51, p = .02, word count was not significantly correlated with any of the function scores. Function scores were therefore not related to the length of the narrative. Function scores were also largely uncorrelated with each other, the only exception being the self and directive functions in the authentic event recall, which showed a moderate negative relationship (r = -.39, p = .04). To test the hypotheses, we used a 2x3 repeated measures ANOVA (authentic and inau thentic event x self, social and directive function). This revealed a significant main effect with a large effect size for function F(2,50) = 17.12, p < .001, η 2 = .23 and a significant interaction between event and function, with a medium to large effect size, F(2,50) = 12.23, p < .001, η 2 = .12, but no significant effect for event F(1,25) = 1.29, p = .27. Post-hoc comparisons using Bonferroni corrections showed that the self and the directive functions were both used significantly more than the social function (mean difference of .59, p < .001 and .61, p < .001, respectively).
The interaction effect between event recall and function is illustrated in Figure 1. Post-hoc comparisons using Bonferroni corrections showed several significant differen ces at the p < .001 level. As predicted, authentic event recall was significantly most likely to serve a self function, followed by directive then by social function. Similarly, inauthentic event recall was significantly more likely to serve a directive function than either self or social. However, the increased use of the self function in authentic than inauthentic event recall did not reach significance.

Discussion
We have shown initial evidence that memories of authentic and inauthentic moments in an everyday context (the workplace) are associated with distinct autobiographical functions. The coherence view of authenticity allows for a process of integrating contra dictory or inconsistent behaviour into a coherent, ongoing understanding of the self (Boucher, 2011) and our evaluation of the functions that autobiographical memories fulfil at work provides initial insight into how this may occur. Recall of authentic events primarily serves to define or enhance identity while recall of a time the individual felt they behaved inauthentically serves the primary function of directing future behaviour. This shows similarity to the posttraumatic growth literature and other memory research that indicates people may value negative memories and are able to use them in adaptive ways (Burnell et al., 2020).
Authenticity is a prized and sought-after state (Sedikides et al., 2019), which raises the question of why people might engage in inauthentic behaviour and further, what function memories of these times might serve. Previous research has indicated that people may engage in strategic inauthenticity in the workplace, that is, behave in ways they recognise as inauthentic in order to serve a larger purpose or achieve a specific goal (Sutton, 2018). This study presents further evidence for the role of inauthenticity in goal directed behaviour. In this case, recall of a time when the person felt inauthentic served to direct and change future behaviour. So while individuals may engage in inauthentic behaviour in order to achieve goals, they also use their memories of these situations to guide and direct future behaviour. Authenticity is valued in and of itself, but this study reveals that memories of being inauthentic serve a valuable function too, providing further support for suggestions that AB memories are more adaptive than maladaptive (Burnell et al., 2020).
The distinctly lower use of the social function in both of these types of memories is notable. Dominant models of authenticity in the literature are drawn from primarily individualistic cultures, where being true to oneself is often couched in terms of resisting other people's influence (e.g., Wood et al., 2008). Emerging models of authenticity drawn from more collectivist cultures, however, recognise that suppressing one's own impulses and desires out of respect for the group can feel just as authentic (Wang, 2016). A replication of this study in a more collectivist sample may well demonstrate a greater social function in these memories. However, it should also be noted that recall of single events is less likely to use the social function than recall of extended or repeated events  and further research may need to take this into account when designing the type of recall to avoid ceiling effects.

Limitations
This study used a stratified sample representing a range of individual differences in terms of authenticity and personality differentiation, but the relatively small size means that findings await confirmation in larger and more diverse samples. Additionally, it would be beneficial to triangulate the coding of narratives by including participants' ratings of their own memory functions as previous studies have shown some distinctions in self-and other-report of memory functions for different event types . Related to this, it should also be noted that the researchers, although coding the narratives blinded for the type of recall, were not blind to the aims of the study and it remains possible that narrative scoring was biased by this.
Because this study drew on archival data from a non-experimental design, the order of the questions was not counterbalanced, meaning that order effects cannot be ruled out and further work using counterbalancing or separate experimental groups is needed to address this limitation.

Future Research
Authenticity is known to be associated with well-being, as are the self, social and directive functions of autobiographical memory (Sutton, 2020;Waters, 2014). In addition, previous studies have shown that authenticity can be increased by the recall of a nostal gic event and that this may occur through reflection on the authentic self (Baldwin et al., 2015). Several authors have suggested that this reflection is a fruitful potential target for interventions to improve well-being, for example, by re-interpreting negative experiences as motivating or re-integrating them into an overall sense of purpose and meaning (e.g., Pillemer, 2001). Studies like these indicate that it is possible to improve psychological outcomes through targeted memory recall interventions. Future research could identify whether interventions involving recall of authentic or inauthentic events might serve to improve well-being and the extent to which this is due to the differing functions of autobiographical memory. By clarifying the functions served by memories of being authentic or inauthentic, we hope to have provided the basis for developing future interventions that may improve authenticity and well-being.

Conclusion
In summary, this study provides first evidence for differential autobiographical memory functions in recall of authentic vs. inauthentic events. Recalling a time one felt authentic serves primarily to support current identity, while recall of a time one felt inauthentic provides an opportunity to direct future behaviour towards a more authentic response.

Funding:
The authors have no funding to report.