Theorising digital personhood: a dramaturgical approach

ABSTRACT This paper reports on findings from a study which investigates our digital identities. Through examination of the process of constructing biographical films derived from users’ social media (SM) data, we progress understanding of the digital self. Building on dramaturgical understandings of performance of self, we challenge the dominant views which argue that SM users operating as their ‘authentic selves’ can be empowered by having the potential for contemporaneous multiple selves. Through the introduction of the concepts of SM leakage and multiple temporal selves, we note the challenge of living with these digital self-representations which are sustained over time. We propose strategies for dealing with temporal shifts, as well as dispensing with the notion of the separation of these selves.


Introduction
longitudinal study of consumers' relationship to their music collections highlights interest in researching consumption and identity, and the practices involved in adhering to a specific identity. They found that identity is dynamic, with their participants trying out different identity practices as they grew up and their lives progressed. Some of these identities were cast aside as inauthentic. While Shankar et al.'s (2009) participants could leave behind their old identities and move into new roles and reference groups, the key challenge of the SM age identified in this paper is the difficulty of leaving these other past selves behind. This notion of old and new selves assumes a progressive nature of time, which is challenged by Turner's (1974) analysis of social dynamics. The digital age has been recognised by researchers as requiring new modes of enquiry, new concepts and new ways of understanding how our digital selves are formed, reformed and presented both online and offline (Dholakia & Reyes, 2013;Hansen, 2013). Therefore, we are concerned with the way Digital Personhood (DP; a term used to discuss recognition of a human being as having status as a person in the electronic realm) is being assembled, depicted and mobilised through various social networking sites (SNS) over time. New SM has led to an explosion in personal digital data that encompass both the expressions of self chosen by the individual, as well as reflections of self provided by others. Our DP identities are used by potential employers, romantic partners and by commercial organisations to engage with us and, therefore, understanding the complexity of these digital selves is important.
This desire for a more advanced and nuanced understanding of DP is particularly relevant in an era where our digital selves are being profiled, aggregated and reduced to code in order for organisations to target us as consumer subjects (Pridmore & Lyon, 2011;Zwick & Denegri Knott, 2009). Our study contributes to the field of digital identity by highlighting the nuanced nature of DP and the liminoid quality of our digital selves and, therefore, the shortcomings of an over reliance on aggregated data as a method of knowing consumers. The fetishisation of large data sets of aggregated information about consumers (McAfee, Brynjolfsson, Davenport, Patil, & Barton, 2012) can overemphasise the utility of such data in understanding individual consumers more holistically than in the past. If big data are seen as an opportunity to develop increasingly personalised marketing practices, there is a need to pause and consider the personal within this context (cf. Bollier & Firestone, 2010). Our paper highlights the challenges of DP and, in doing so, identifies the need to develop ways of understanding DP in the social media (SM) age, other than as simply aggregates of data.
This paper reports on an innovative study which considers our SM selves through the use of qualitative filmic methods. We worked with professional film-makers in order to produce film-like life documentaries, made up entirely of users' SM data. This study, in line with other narrative approaches to understanding identity (see Shankar et al., 2009), recognises the dynamism of narratives and questions notions of fixed identity. In so doing, we introduce the concepts of multiple temporal selves and SM leakage which need to be conceptualised and considered in order to develop SM literacy. We explore DP through the lens of performance, using Goffman's (1959) presentation of self as well as Turner's (1960Turner's ( , 1974 dramaturgical approach to challenge existing theories of DP. While Turner's dramaturgical conceptualisation helps to shape our analysis, by adopting Znaniecki's 'humanistic coefficient' as highlighted by Turner, we also draw on Turner to justify our methodological approach, through the inclusion of the subjective interpretation of DP from our research participants, ourselves as academic researchers and professional storytellers. Our findings further progress understanding of DP by applying Turner's (1960) conceptualisation of liminality to DP. This allows us to highlight the challenges involved in transitioning from different 'states' (Turner, 1960) as digital selves, as well as illustrating the importance of our digital traces in providing opportunities for reflection on such transitions. While much attention has been paid to understanding how SM allows for multiple and discrete selves, there has been less focus on accounting for liminality, identity shifts and SM management practices. Shields (2002) highlights the appropriateness of referring to virtual spaces as liminoid, as they can be seen as distinct from everyday life and consist of temporary participation. As Turner (1983) characterises liminoid as a playful and voluntary state betwixt and between, as opposed to liminal, which comes from clear social and cultural obligation, it would seem appropriate to speak of these spaces in relation to DP as liminoid.

The characteristics of social media
There are various definitions of SM, but the common theme between all definitions is the focus on the creation and interactive exchange of user-generated content via Web 2.0-based technological applications. As Kaplan and Haenlein (2010) explain, SM is a collection of Internet-based applications built on an interactive platform, driven by webbased and mobile technology (known as Web 2.0), enabling the exchange and cocreation of user-created information in various forms between users. Chung and Austria (2010, p. 581) define SM as media that is published, created and shared by individuals on the internet, such as blogs, images, videos and more as well as online tools and platforms that allow internet users to collaborate on content, share insights and experiences and connect for business or pleasure.
These definitions characterise the ideological and technological foundation of SM and highlight different ways in which people share and exchange media content, created by themselves and which are available publicly (Kapoor, Jayasimha, & Sadh, 2013). Chung and Austria's (2010) definition notes the collaborative nature of SM, where content is shared, repurposed, commented on and reacted to, indicating the dispersed nature of SM identity formation and interpretation. Kaplan and Haenlein (2010) organise SM activities into six distinct categories, based on existing media theories of social presence and media richness. These are (1) selfpresentation, (2) self-disclosure, which include blogs, microblogs (Twitter, Weibo) and collaborative projects (Wikipedia), (3) SNS (Facebook), (4) connect sharing communities (YouTube), (5) virtual social worlds (Second Life) and (6) virtual game worlds (World of Warcraft). The characteristics common to most SNS include the construction of user profiles, the specification of relationships with other users and access to information as a result of that relationship (Pike, Bateman, & Butler, 2009). boyd and Ellison (2008, p. 211) define SNS as web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system.
Here, boyd and Ellison (2008) identify what might be seen as an ideal SNS, where it is possible to have a bounded system and maintain privacy within a chosen network. However, what we have seen as SNS have developed, is the difficulty of maintaining such boundaries and privacy, as the desire for data from corporations and governments clashes with the desire for control over our data from consumer citizens (Liang, Zhang, Shen, & Lin, 2014). SNS allow simultaneous one-to-one, one-to-many and many-to-many communication via text, photographs, instant messenger, direct messages (like email) and videos (boyd & Ellison, 2008). Millions of users interact on these platforms each day and currently, Facebook holds over one billion active users worldwide. This is a dramatic increase in active Facebook users from just 10 years ago, when there were approximately six million users (Facebook Newsroom, 2015), despite claims that Facebook is losing popularity (Oremus, 2016). A primary function of SM, and, in particular, SNS, 'is to consume and distribute personal content about the self' (Ellison, Vitak, Steinfield, Gray, & Lampe, 2011:19). Indeed, the literature discusses SNS as a way of connecting with known and unknown others (Sheldon, 2008;Special & Li-Barber, 2012), therefore enforcing differences in sharing behaviour when users are aware of their audience (Vasalou, Joinson, & Courvoisier, 2010), and variations concerning the presentation of the self and identity development (Marder, Joinson, Shankar, and Houghton, 2016;Schau & Gilly, 2003;Zhao, Grasmuck, & Martin, 2008). Arguably, the challenge of segregating multiple personas is becoming increasingly difficult in a digital age (Belk, 2013(Belk, , 2014. Marwick and boyd (2011a), in their study on the motivations behind 'tweeting', focus on the increasing difficulty of maintaining multiple digital identities to suit different audiences.
While self-presentation in the presence of multiple audiences has been associated with positive effects (e.g. Leonardi, 2014), other studies highlight the negative implications, since users find it difficult to meet the expectations of multiple audiences simultaneously (e.g. Lang & Barton, 2015). Marwick and boyd (2011a, p. 11) claim that this creates 'a lowest-common denominator effect, as individuals only post things they believe their broadest group of acquaintances will find non-offensive'. Consequently, previous research has illustrated that people constrain the self they present online, due to increased audience surveillance on SNS. Certainly, SNS typically converge different relationship types into one group of 'friends', allowing users to connect to and view information about others, hold conversations, share photos and videos, and build and maintain social relationships (boyd & Ellison, 2008;Ellison, Steinfield, & Lampe, 2007).
Our study illustrates that the challenges related to SM use are tied to temporality and change, balancing multiple roles and the public/private intersections that occur through SM use. While studies such as those by Vican and Ulusoy (2012) and Denegri-Knott and Molesworth (2013) consider the creation of alternative or multiple selves as well as consumer desire in digital virtual worlds, there is a need to understand the less instrumental uses of SNS, which collectively form our digital selves. Although some individuals choose to not accurately represent their authentic self via their digital profiles (Stone, 1996;Turkle, 1995), often accuracy in our digital representations is important, such as communicating with friends and family or developing a professional profile. Furthermore, our current digital identities are increasingly being profiled and targeted by companies through the aggregation of our data (see Xu, Frankwick, & Ramirez, 2016). Zwick and Denegri-Knott's (2009) article on biopower and database marketing aims to understand marketing in the age of increasingly integrated and networked customer databases, where we are, in fact, viewed as aggregated data. Following Pridmore and Lyon (2011), who describe consumers as being constituted of collections of their data, it is necessary to question the impact of perceptions of us as collections of data and the narratives that are emerging as a consequence of this. Thus, understanding the complexity of our DP is important, because we are profiled from our online aggregated data.
The drama of the self and online identities Goffman (1959) offers a dramaturgical approach to explain how an individual presents an idealised rather than authentic version of his/her self. Individuals thus engage in a theatrical performance created for a specific audience at a specific time, leading to front stage and back stage personas. One core assumption of the dramaturgical approach is that activity takes place in specific bounded settings. Regarding the front stage persona, Goffman suggests that individuals are like actors and are conscious of being observed by an audience, so will perform to those watching by adhering to certain rules and social conventions. Individuals portray a positive and desirable character to the outside world, attempting to guide their impression on others by altering their own setting, appearance and manner. The back stage persona is a hidden, private area where individuals can be themselves and drop their societal roles and identities. This continued presence allows individuals to adapt their behaviour accordingly, a process he termed impression management. Self-presentation is considered a process of social exchange, whereby an ideal self is presented to and interpreted by others, then either accepted or, if rejected, adjusted (Goffman, 1959).
In addition to the impressions we give and give off, Goffman (1959) considers further analogies, for example, the mask as a means for deception in face-to-face environments. A mask can be held in place from within an individual, bringing forth certain features while simultaneously marginalising others. The individual is not changing their identity, but rather, as Goffman argues, both the mask worn and the hidden individual behind it are facets of the same person. Indeed, individuals contain multiple contemporaneous selves, whereby people act differently in different social settings.
While Goffman's interest lies in the drama of everyday life, seeing performance as central to all interactions, Turner is concerned with social dramas, which occur at times of conflict. Turner (1985) draws on Schechner's analysis of dramaturgical approach to differentiate between Goffman's interest in everyday performance and his own in conscious acting, where conflict provokes the need for asserting a desired role. This interest in conflict resulted in Turner's development of theories of liminality, the state of transition of being 'betwixt and between' (Turner, 1960) one state and another. Turner (1960Turner ( , 1974 identifies a number of stages to go through in moving from one state to the next. Turner (1960, 46) defines state as 'any type of stable or recurrent condition that is culturally recognized'. Turner was interested in the movement between these states, which sparked his interest in the concept of liminality. Goffman's presentation of self has become popular as an explanation for online participation (Bullingham & Vasconcelos, 2013;Hogan, 2010). It has been proposed above that online environments provide their users with the potential to perform and present different multiple identities. Bullingham and Vasconcelos (2013) argue that the distance between performer and audience that physical detachment provides makes it easy to conceal aspects of the offline self and embellish the online. Goffman (1971) might consider this to be a reflection of the splitting character of the self during interaction. Table 1 contains some of the many articles that use Goffman for this purpose: A common thread running through these articles is that individuals employ impression management and selectively disclose personal details in multiple online spaces in order to present a fantasised 'ideal' self (Higgins, 1987;Leary & Kowalski, 1990). To date, Goffman's self-presentation theories within online environments have only been applied on a contemporaneous level and, therefore, do not consider the importance of a longitudinal view. This need to consider a longitudinal perspective leads us to alternative or complementary dramaturgical approaches, such as Turner's (1974) consideration of social time. For Turner, social events, or rituals, signal the passage of time; or an individual's progress through a series of events indicates the passing of time. Turner's more dynamic dramaturgical theories, with a focus on temporality, offer us further insight into the nature of DP, the performance of digital identity and the need for the development of practices to deal with the challenges of the digital self.
Consequently, one of the greatest challenges facing SM and online platforms is how to manage one's multiple online identities simultaneously (as conceptualised by Goffman, 1959), encompassing both the expressions of self chosen by the individual, as well as reflections of self provided by others. By drawing on Turner's (1960) temporality sensitive dramaturgical theories, we also refer to the need to consider current and prior selves. We recognise that we may operate in relatively fixed states in relation to some of our online networks, playing the role of family member, colleague, school friend and so on. We may also simultaneously inhabit liminal states while transitioning from one state (professional context, sexual orientation or other state) to another. Whilst theory seeks to account for the nature of alternative or multiple digital selves, there is limited empirical evidence of how one self manages and/or provides coping strategies for this complex process. Moreover, previous studies have only Table 1. SM studies drawing on Goffman.

Authors
Goffman applied to online spaces Donath (1998) Utilises Goffman as an initiating point for signalling theory Schroeder (2002) Heavily employs Goffman's dramaturgy in his analysis of virtual worlds Schau and Gilly (2003) Builds on Goffman to examine personal Websites as a noticeable form of multiple, and, in particular, desirable self-presentation Zhao (2005) Shows that the online experience of teenagers revealed that telecopresent others in the online world do constitute a distinctive 'looking glass' which generates a digital self that is different from the self-constructed offline Ellison, Heino and Gibbs (2006) Investigates Goffman's self-presentation strategies among online dating participants (how participants manage their online presentation of self in accomplishing the goal of finding a romantic partner) boyd (2007) Uses Goffman to examine SNS activity as networked identity performance Robinson (2007) Argues the effectiveness of Goffman's approach over postmodern perspectives found in Turkle (1995) Lewis, Kaufman and Christakis (2008) Reflects on Goffman's front stage/back stage distinction to examine the notion of student privacy in SNS and, in particular, the factors that are predictive of a student having a private and a public profile Tufekci (2008) O ffers research on Facebook self-presentation around Goffman alongside Dunbar's social brain hypothesis Zhao et al. (2008) Expands upon Goffman's self-presentation theories to investigate identity construction on Facebook Quan-Haase and Collins (2008) Uses impression management as per Goffman to discuss the art of creating status messages in order to signal availability Menchik and Tian (2008) Utilises Goffman and symbolic interactionism to interpret 'face-saving' on e-mail mailing lists Mendelson and Papacharissi (2010) Demonstrates that pictures and photos on SNS conform to traditional notions of impression management Hogan (2010) Argues that self-presentation in SM environments can be split into performances (i.e. chatting), which take place in synchronous 'situations' (as per Goffman), as well as artifacts (i.e. status updates, photos etc.), which take place in asynchronous 'exhibitions' Bullingham and Vasconcelos (2013) Presents an exemplification and discussion of the contemporaneity of Goffman's work and of its applicability to the analysis of identity and presentation of self in the blogging and Second Life contexts Marder, Joinson et al. (2016) Examines theories of self-presentation within the context of audience surveillance in SNS, stimulating a self-comparison process that results in impression management Marder, Slade, Houghton and Archer-Brown (2016) Contributes to impression management theory, as per Goffman, within SNS by providing an understanding of the effect of visible affiliation on page 'liking' behaviour in the context of political parties SNS: Social networking sites.
considered our online identities contemporaneously and thus failed to acknowledge temporal issues and the different iterations of digital self-hood over time. Indeed, SM is constantly evolving and changing, with new platforms being introduced and operational rules changing in existing platforms. Therefore, current online self-presentations need to additionally consider the temporal aspects, since what we were like 5 or 10 years ago might not align with what we are like in the present day. Changes such as starting university, moving house, getting married, starting a family or coping with ill-health or bereavement of loved ones and so on warrant close analysis within a digital context. Most life transitional studies within a marketing and consumer-based context (e.g. Rindfleisch, Burroughs, & Denton, 1997;Schouten, 1991) have been rather celebratory about consumption and are set within an offline context. O'Loughlin et al. (2016) examine life transitions with more negative connotations, considering how European men cope with decreased social and economic circumstances due to the post 2008 recession. In doing so, they note the importance of the concept of liminality as people transition from one identity to another. Following Turner (1974), they identify the three core phases of liminality, separation, marginal and reaggrigation, as the liminars move from one stage to another.

Methodology
It is important to develop a detailed understanding of the emergent processes underpinning self-presentation and response to others within digital environments (Schau & Gilly, 2003). However, established data collection methods, such as participant recall, projective techniques and observational methods, fail to fully incorporate several important characteristics of SM. These include the longitudinal adjustment to social norms, self-censorship, the construction of the self as content and, above all, the unexpectedness and uncontrollability of SM interaction. In response to these challenges, our paper presents an innovative and self-reflexive form of documentary film-making methodology, designed to capture the interactive processbased nature of digital self-enactment. Through the construction of biographical films derived from users' SM data, we question current understandings of the management of self in the digital era (Quinton, 2013). Just as Shankar et al.'s (2009) narrative approach allows access to a longitudinal account of identity formation, expression and evolution, our methods reveal interesting insights about our participants' relationship to their digital identities. Additionally, Turner (1974) suggests the need for greater subjectivity in studies of social phenomena and his adoption of Znaniecki's humanistic coefficient into his method of social enquiry resonates with our multi-participant and multistage reflexive methodology.
As our study is focused on the everyday use of SM, we have developed a way in which to unify the digital self which is dispersed across multiple platforms. In doing so, we provided our participants the opportunity to reflect on their own SM use.
Our study is inspired by other forms of narrative research, such as biographical methods (Fillis, 2006;Kerrigan, Brownlie, Hewer, & Daza-LeTouze, 2011) and subjective personal introspection (Gould, 2008;Holbrook, 2005). Patterson (2010) talks about the value of using first-person narratives of others to reflect on the self, and our method takes this stance by distancing our participants from their digital selves, presenting their remediated digital identities to them and allowing them to reflect on these remediated selves. By involving ourselves as academic researchers in designing, overseeing and ultimately analysing the process, as well as using professional film-makers to extract data and create narrative accounts of our participants derived from their SM data, we add this humanistic coefficient which Turner values in the research process. Through the introduction of greater subjectivity, we are not seeking to discover the 'real' digital self, rather to uncover the challenges of digital selfhood experienced by our participants.
The research process proceeded as follows: Six professional film-makers were recruited via a film competition (February to April 2014) and challenged to create films entirely constructed from the SM content of their chosen participant (between May and September 2014) (see Table 2). Applications required the film-makers to justify their choice of participant based on their level of activity on a range of SM platforms. As we were concerned with unifying dispersed digital identities, it was important that participants were active across a range of platforms. By requiring participants to be active on at least three SM platforms, we ensured a richness of data available for analysis. In this way, our participants are not necessarily 'typical' users but are comfortable with operating online and therefore provide good insight into SM usage.
We asked the film-makers to identify the type of narrative that they could tell, based on the SM data available. This allowed for purposeful sampling to occur, where we looked for cases which provided contrasting variety rather than consistency (Eisenhardt, 1991). In terms of the subjects (see Table 2), there were three professionally public individuals, who construct an identity within SM which they use professionally as 'social currency' (two well-known public figures and one SM persona), and three nonprofessionally public individuals.
The film-makers and the participants were interviewed prior to and after the films were made, in order to assess there motivations for participation, identify their reservations about the process and the challenges involved, as well as to understand their perspectives on creating the films and seeing their digital identities in a unified form. In addition, we met with each film-maker to conduct observational analysis when production had commenced (1 day per film-maker), and the subjects were each asked to record a diary throughout the production process. The film-maker interviews and observational analysis allowed insight into how they interpreted their participants' data. Participant interviews and diaries identified concerns participants may have had, how and why they use SM and, significantly, their response to the filmic interpretation of their digital selfhood. Therefore, our corpus includes the finished films, the interviews, observations and participant diaries. The data were analysed using an interpretive approach, reviewed by the researchers and a number of themes emerged regarding digital identity. Such analytical methods allowed interpretation and response by the key research participants, that is, the researchers, the film-makers and the participating subjects inline with Turner's (1974) adoption of the humanistic coefficient. According to Rossman and Rallis (1998, p. 8), qualitative research takes place in the natural world, uses multiple interactive and humanistic methods, is emergent rather than tightly prefigured and is fundamentally interpretive. In addition, the qualitative researcher views the social phenomenon holistically, systematically reflects on who she is in the inquiry, is sensitive to her personal biography and how it shapes the study and uses complex reasoning that is multifaceted and iterative. Rossman and Rallis' (1998) classification highlights the key elements of qualitative inquiry appropriate to the use of a narrative approach to research, where multiple subjectivities contribute to understanding the phenomenon of DP. Adding the professional film-maker to the analytical process allowed additional reflection other than that which would have been possible if the researchers had constructed films from the data themselves.
The data analysis involved thematic analysis (Braun & Clarke, 2006), where our findings were compared to dominant views expressed in the literature to draw out themes. The narrative differences between the different films and the process of constructing these narratives allowed us to understand the challenges underpinning digital identity construction, management and interpretation. The following sections outline the key findings which emerged from the data.

Findings and discussion
This documentary film-making methodology approach provides detailed insight into SM identity performances that involve both challenges and strategies concerning the management of our digital selves. Thus, the following section is split into three key themes; the (non-)bounded self and the concept of SM leakage, discussion of temporal multiple digital identities and, finally, consideration of management strategies for managing contemporaneous and temporal digital identities.

The (non-)bounded self
The findings suggest that the challenge of segregating multiple personas for distinct purposes and addressing specific audiences in a digital age is becoming increasingly difficult. Whilst we can endeavour to present different personas via different platforms (e.g. professional persona on LinkedIn, or social persona on Facebook), it is very difficult to create these multiple identities in the way it was originally theorised. Neither Goffman's (1959) notions of front and back stage performances, nor boyd and Ellison's (2008) argument with regard to the bounded nature of SNS, fully account for our findings. On the contrary, while our participants clearly noted the use of different SNS to communicate differently to different audiences, they found that maintaining a deliberate distinction between such online selves increasingly challenging. We identified a phenomenon we have labelled SM leakage, where elements of DP intended for a specific audience find their way into a wider group. This leakage can occur over time (temporal leakage) where older content was created and shared with a specific group in mind but is accessed at a later date by an unintended audience. Alternatively, content created for one SM platform or audience can be shared beyond the intended group. While computer science scholars are concerned with leakage of personal information via social networks, they view this from the perspective of the hosting organisations not protecting our data sufficiently against possible security threats (cf. Zhang, Sun, Zhu, & Fang, 2010). Our concern is not linked specifically to security, but to wider considerations of self-presentation. Our participants did espouse using different SNS for different purposes; however, it was apparent that these boundaries are very porous and that other audiences than those anticipated when posting have access to the content posted. Although Manick considers his use of these SNS as social rather than professional, he is aware that his DP could have implications for his professional life, even if unintentional. Here, Manick can be seen to operate in a liminoid professional state which may interfere with a relatively stable social state. Alice also feels that it is possible to have different digital selves on different platforms. I'm conscious of how people see me and whether or not that's in a negative professional capacity or a personal capacity. I'm the kind of person that's got a lot of personalities . . . So I think I use social media in a similar way in that I choose which me I want to present on different platforms. [Alice] While some of our participants endeavour to maintain boundaries around their various digital identities, Imogen finds that the disparate SNS impede her desire for a more unified, holistic self. Rather than aiming to keep distinct selves private or separate, she acknowledges the drawbacks of having disparate, disembodied digital selves, which, in isolation, may lead to misunderstandings about the whole person. I feel very disparate online, so I have my Twitter identity, I have my YouTube identity, which all present a different way of how I am. So if people only read my Tweets then they only know me in small bite size chunks. But if they've seen me on YouTube then they know that I say things with a smile and that I'm not very succinct . . . And if they look at me through music then they might think I'm really serious . . . The problem with technology is that it allows you to have all these different options and platforms to play around with and you end up with loads more than you can really take on, and it's easy to start things but it's not easy to finish them . . . [This project] has made me really want to push my website forward and portray this really secure home for all the media and for all the connections. [Imogen] Conscious self-control and self-management (Belk, 2013;Houghton & Joinson, 2010) has been identified as necessary due to SM privacy issues; however, even if the subjects' restrict certain content to a designated circle of online friends, ever changing rules and settings mean that this information may reposted, retweeted or quoted by others. We can liken the practices of liking, sharing and so on to Turner's (1982) notion of metacommentary, where the group tells itself about itself. Boje (2003) refers to this as a form of interpretive re-enactment. Belk (2013) also notes the collectively constructed nature of our digital identities which challenge the singular authorship of our DP.
People always say I sound very forthright and very uncompromising on social media, but when they meet me they're always like 'oh you're such a gentle soul and a nice guy' . . . What I'm hoping is that [the film] will be great to focus on certain outlooks on life, whether that's social, environmental, or political, and that it is nuanced and quite balanced, and it's not too controversial or out there, and perhaps provides a more rounded perspective of certain issues and debates. Because I have been accused at times of being a bit outspoken and landing myself in trouble as a result of that . . . So hopefully [the film will portray] a more representative self and a more rounded self. [Manick] Manick, like other participants, felt that taking part in this project, where his SM data are unified and presented in a more nuanced manner, may allow a more balanced presentation of self than he has achieved so far through SM. Manick recognises the drawbacks of his efforts to use SM selectively. As SM is used by Manick to engage with others around specific social and political issues, inferences are made by his SM connections about his 'true self' based on their access to this data. Peter T. also welcomed the chance to present a more complete self, observing It [the social media film] is a way of counteracting misrepresentation or partial representation of my work via mainstream media . . . These false representations have lead to a deluge of hate mail and death threats. [Peter T.] Although the online environment may allow participants to be more open and expressive, due to issues of unfettered sharing and SM leakage, it can often result in misrepresentation (e.g. Solove, 2007) or an undesired digital self (e.g. Belk, 2013Belk, , 2014, ultimately placing the user in a liminoid state. As a consequence, Manick and Peter T. regarded the films as vehicles to present a more 'ideal' and unified self, which was the case for the majority of the subjects. Prior misinterpretation or selective engagement can result in forcing the DP into a liminoid state, as they are interpreted as breaking social rules. What emerges from our data is a contrast between the struggle to retain the bounded digital selves and the recognition that such disaggregated selves present other challenges of fragmented DP and therefore a desire to achieve a more aggregated DP based on the (distinct) use of various SMS. Such aggregation would counter problems caused by SM leakage.

Temporal multiple online identities
The section above demonstrates how some of our participants aim to maintain boundaries between their various SM platforms. In so doing, they align themselves to notions of selfpresentation (Goffman, 1959), as well as the idea that we can have bounded selves on SM (boyd & Ellison, 2008). However, our findings suggest that the digital selves that we present are not only just multiple contemporaneous identities, where we choose what we present and put forward or hold back in that moment (as per Goffman, 1959), but also temporal multiple digital identities, since we have a past, present and future digital self. Certainly, over time, one might change professional identity, visual and physical appearance, fashion tastes and so on. As a result, the older (past) self may come into conflict with the present (changed) self. Looking to Turner's (1974) dramaturgical theory, his notion of liminality may offer some conceptual clarity; however, Turner assumes a transition from one to the other, where the old state is discarded for the new. Our findings suggest that the ubiquitous nature of our DP makes such transitions difficult, as our discarded self may follow us around. Rather than assuming social time, like Turner, as consisting of a series of events that mark a linear passage of time, DP means that our prior selves can coexist alongside our current selves due to the digital traces that remain, despite transition to a new state. Thus, reacquainting one's self with the past self can be an unnerving experience: When I received the initial script from Louise I felt strange because I hadn't realised she would be using things I'd written on my blog, I thought it would be more photos and status updates, which tend to be less personal in a way. The blog posts she'd chosen were heartfelt poems and observations. I had actually forgotten I'd written them and it was an unsettling feeling knowing they'd be resurfaced again, despite having changed a lot since I wrote them and probably wouldn't have written them now. It was a realisation that I tend to think of things I post online as having a shelf life, today's news tomorrow's chip paper. So reading it again now and knowing it would be re-published felt weird [Alice] This documents the issue concerning temporal shifts or changes in the digital realm. Belk and Yeh (2011), in their paper on tourist photography, also highlight this issue of temporality. They argue that a photograph can be a message from a former self to a future self that is intended to recreate the emotion of the original experience, but this might lead to both positive and negative experiences. In this study, a SM film that brings together a collection of identities can mimic a similar effect. In Alice's case, she has been through considerable life changes and transitions, in particular separation from her long-term partner, as well as a change in sexual orientation. This led to numerous concerns that people from her present would have access to her past self and, as a consequence, perhaps not accept her for who she is now. Therefore, these liminoid transitions leave traces which can make it difficult to fully transition from one state to another. Thus, our DP shifts temporally as well as contemporaneously. One of the participants, Richard, recently survived bowel cancer, but the effects of this life-threatening illness have impacted both his personal and professional usage of SM: I always made my whole Facebook account public and I just followed the general rule that I wouldn't post anything that I didn't want the whole world to see, even though it is a personal account. But after last year [cancer treatment], I felt that I was just too exposed and there are possible repercussions, like if someone or an organisation might think you're going to drop dead in the next six months, then they might not give you a job. So I have now restricted Facebook to be only visible by friends. [Richard] Interestingly, his illness has made Richard aware of the repercussions surrounding SM leakage and how desires regarding privacy may change over time. As a result of temporal changes, Richard now has to consider his pre-ill SM persona as well as his post-ill SM persona. Thus, our concern is about how these temporal and/or contemporaneous identities make sense narratively, and how we manage and find strategies to counter these problems.
It is clear that the issue of multiple temporal selves challenges notions of bounded selves. Where these lines are more blurred (or non-existent), the challenge of multiple temporal selves is less stark. For example, in the case of Peter T., his personal became public at a young age when he started to campaign on various civil rights issues; therefore, he does not have reservations about his DP from a temporal perspective.

Managing contemporaneous and temporal online identities
The subjects manage their DP in several ways depending on their type of SM usage. Some of these techniques had already been enforced by particular subjects prior to the study but became apparent as a valuable insight into managing one's digital self during the research process. Kaplan and Haenlein (2010) identify different types of SM (self-preservation, selfdisclosure, collaborative projects, SNS, sharing communities). However, arguably, these categories are more usefully divided into SM where avatars and pseudonyms are deployed, versus those where users present a form of authentic self, such as SNS. So, there is a separation in terms of our digital identity between the places where we can engage in fantasy consumption practices (Denegri-Knott & Molesworth, 2013;Stone, 1996;Turkle, 1995) to escape our offline realities and the SM places where we can present our authentic selves in different ways. Marwick and boyd (2011a) highlight the difficulty in segregating multiple authentic identities and instead propose techniques to try and overcome this issue, for example, by using multiple accounts, the creation of pseudonyms, or creating fake profiles to obscure true identities. The Baristacrat has an authentic online identity, but his online alter ego provides a lot more freedom to engage in his interest and hobby: I've got my personal one [online identity], which is for my friends and keeping in touch . . . And then I've created the Baristacrat one, which is for me to review cafés and write about that and have all the social media connected to that . . . I can engage with them in a different way that's not so invasive on my personal side . . . It just makes it a little bit more comfortable that they don't actually know who you are, your personal life, and I just feel free to post my location online. [The Baristacrat] The example of The Baristacrat's pseudo self shows a very stark idea of creating multiple selves in the virtual world and is perhaps another solution to managing the multiple self. The creation of online personas with clear functions facilitates the deletion of the account when it is no longer required, therefore the ability to move from one state to another. If one has a particularly managed or professional identity then an online persona or alter ego can help in maintaining that, but being yourself through multiple personalities might be problematic due to issues concerning leakage and temporal shifts. The use of pseudonyms allows for a comfortable range of contemporaneous and temporal selves to coexist without challenging or compromising the others. The Baristacrat can also coexist alongside possibly incompatible other contemporaneous selves, again, without compromising the others.
However, our other participants wished to present their 'authentic' selves in the digital world, as part of their overall identity project. In these cases, the desire to preserve boundaries is challenged and here, looking to the world of celebrity may allow us insight into how multiple selves can be managed online. As noted earlier, the development and management of celebrity brands is more complex than the management of consumer brands. However, many celebrities have developed a level of expertise in balancing the management of their private or public/professional lives.
According to Marwick and boyd (2011b), networked media (e.g. SNS, gossip websites, fan sites, blogs etc.) have changed celebrity culture, predominantly the ways in which people relate to celebrity images, how celebrities are produced, and how celebrity is practiced. Marwick and boyd (2011b) argue that celebrity is a learned practice. In their analysis of celebrities on Twitter, they found that celebrity practitioners reveal what appears to be personal information to create a sense of intimacy and closeness between participant and follower, they publicly acknowledge their fans and use language and cultural references to create affiliations with their followers. They carefully construct their online persona to include both public details interspersed with private details, to give the impression of candid, uncensored looks at the people behind the personas.
The two professionally public authentic participants (Imogen and Peter T.) are very savvy about working with media and they understand how to make the private public When I was eighteen or nineteen when I signed my first record deal, I was always made to believe that I must be terribly secretive, and I shouldn't really talk to my fans . . . and I might get stalkers . . . The more I questioned that, the more I opened myself up and got comfortable in just being myself, and having a very small line about who I am publically and who I am privately. It just makes my life so much more enjoyable because it means I can live my life and live my music in tandem with one another and not have to have them separate in separate boxes. [Imogen]

Self-reflection
The above section considers SM management strategies for the pseudo and authentic selves. However, what is also apparent in our study is that the possibility for selfreflection offered by reviewing SM (in this instance through the format of a film) is seen as valuable in understanding DP. The intervention of the professional film-makers can be seen as an example of what Turner (1974) refers to as a dramatic breach. Turner would see this as leading to 'a phase of mounting crisis' (Turner, 1974, p. 38), which could lead to a change in social relations. In a similar way to Shankar et al. (2009) study, this opportunity for reflection allowed our participants and film-makers to reflect on their DP and on their use of SM. Our subjects, though all active and SM savvy users, recognised the benefits in engaging in reflection on past activity to develop retrospective editing practices and to gain value from this reflection. For Peter T., this resulted in a sense of continuity and achievement: One personal impact of the film was to remind me of the vast number of issues and campaigns that I've been involved with. I tend to go from one to the next and it becomes almost a non-stop blur of campaigning . . . So it's a useful reminder and also when I'm feeling exhausted, it helps give me a sense that despite my tiredness it's worth carrying on because together with others, I'm helping shine a spotlight on human rights violations [Peter T.] Here we see Peter T. experiencing reflexivity, as Myerhoff (1982, p. 105)
As Schwarz (2010) observes, we have entered an unparalleled era of self-portraiture. Together with blogs (Cohen, 2005) and web pages (Schau & Gilly, 2003), this has arguably led to greater self-reflection on how the social nature of DP as content liked, shared and commented on can play the role of meta-texts (Turner, 1974). Subsequently, this has resulted in some SM users (and researchers) becoming concerned with actively managing identity and reputation and warning against unfettered sharing (Joinson, Houghton, Vasalou, & Marder, 2011). After viewing her film, this was a concern for Alice, and she consequently amended the security settings of her online profiles so that her new SM friends do not have access to her past self: After watching a draft of the film and being reminded of my former self I had the sudden realization that all of these photos are in the public domain, they're for any one of my Facebook 'friends' old, but more importantly new to see. This made me feel uncomfortable. I only really wanted the old photos there for me to see, and perhaps the people who knew me then. I have changed a lot in the last 5 years and I didn't want new friends and acquaintances to have access to this old me. These photos that do not represent who I am now. I went onto Facebook and made more of my old albums private. [Alice] Here, retrospective editing is clearly in play to deal with issues surrounding temporality in the digital age. Additionally, interpretation of meta-texts to aid development of a desired DP was also evident in Richard's self-reflexive experience: I sort of got more awareness of the impressions people have got about me from my social media activity. I think it's made me think about whether I should be using social media differently . . . I started to look like I was a bit of a buffoon . . . Especially the first part of the film, it was all sort of clowning around and it was like someone who you can't then take really seriously . . . It doesn't feel appropriate anymore and even though I'm still doing that a lot, I feel like it's past its sell by date [since the cancer treatment]. I don't really want to come across as that person anymore . . .. [Richard] Together with acknowledging the importance of how one might be perceived by others, Richard also felt a desire to change his persona on SM. Therefore, alongside the retrospective editing, the self-reflexive nature of the whole production process did lead to behaviour change for some subjects' with regard to future SM representation. This was also clear with Manick, who felt that he needed to alter the tone of his voice: What I think has come across is that I do have control of what I say so maybe I can be much more considered about what I say, much more nuanced, and that in a sense you are constructing a sense of identity. So those are aspects that I will take with me about how do I as a social being and also as an individual, that engagement is not separate, but to look at those conversations that happen between one's interaction with the world and other people, and one's private life so-to-speak. And I like that because I think the only way to regulate stuff is actually knowing when to show considered responses to things that I disagree with hugely, and that one is an adult of fifty two and not a teenager of fourteen. [Manick] Thus, for the majority of the subjects, the final films and, for that matter, the whole production process served as a strong self-reflexive exercise.

Conclusion
This paper has considered the complexity of DP and, in doing so, has highlighted two new concepts: SM leakage and multiple temporal selves. Deepening our understanding of the nature of DP is important in light of the increasing focus put on aggregated SM data in the move to big data analysis. Our findings suggest a dynamism in our DP that needs to be factored into such big data analytical processes and we argue for greater focus on understanding our unified DP rather than the reification of our dispersed and aggregated selves. As consumers, we do not consume as aggregations of data points, but as complex social beings, and methods of understanding DP which account for this are essential in developing marketing management practices.
Aside from the implications for practice, our study establishes a series of significant theoretical contributions. First, while the turn to performative theory in SM research is welcome, we highlight the conceptual limitations of using pre-SM theory in an off the shelf manner. Our updating of the dramaturgical theories of Turner and Goffman incorporates an important understanding of temporal issues not yet accounted for. While Goffman recognises that people have multiple contemporaneous selves, he does not consider temporality. Conversely, Turner (1960) does consider the movement from state to state, and although he rejects notions of progression from one to the other, there is an assumption in his theorisation that we leave behind a state when taking up a new one. This assumption is also evident in studies that have examined identity narratives (Shankar et al., 2009). Our findings require both a combination of Goffman (multiple selves) and Turner (liminality in moving from state to state), as well as the need to move beyond both to understand DP. In fact, we may be simultaneously in a liminoid state with respect to one area of our life, while not in relation to others. This is where the desire for the bounded DP comes in, as SM users attempt to bind their activities within certain platforms in order to manage these different states. However, we found this to be challenging due to changing user policies from SM platforms, as well as changes in practices among SM users themselves. The resulting SM leakage acts to challenge these attempts at creating secure boundaries.
While these boundaries are perceived between activities on different SM platforms, our participants were also concerned with boundaries between past DP and present or future DP. Turner (1960Turner ( , 1974 notes the significance of conflict in his theories of performance and, indeed, our participants did experience conflict, although this was often within their identity as past, present and future selves interacted. While Turner views the liminoid as a temporary transitioning state (other than in the case of those choosing social positions that are viewed as liminoid), we suggest that our findings regarding the prevalence of multiple temporal selves imply a more established liminality for DP. The ubiquity of our past DP through the dispersal of SM accounts means that we may never fully transition from one state to another, as traces of ourselves are more readily available than in the past.
As we are profiled via our SM data, and as such data are often interpreted as representative of our offline selves, providing theoretical insight into the nature of our DP is of increasing importance. We have introduced the idea of SM leakage, a modification of the notion of data leakage, where attempts at keeping different DP separate from one another through the use of different SNS or security settings, or the meta-textual interactions of other SM users fail. This may imply a greater emphasis on the liminoid state as established rather than transitional, thus requiring us to find ways to manage these liminoid states. We also highlight the relevance of temporal as well as contemporaneous digital selves, which may contribute to instances of SM leakage. In order to understand the significance of these findings, we need to review them in relation to the current dominant thinking on DP.
First, it is useful to refer back to Kaplan and Haenlein's (2010) six categories of SM usage. As previously outlined, these categories are more usefully divided into SM categories where avatars and pseudonyms are deployed versus those where users present a form of authentic self. Regarding the latter, this paper contributes theoretically to Goffman's (1959) presentation of the self as well as Belk's (2013Belk's ( , 2014) extended digital self, through establishing that digital users operating as their authentic selves have temporal as well as contemporaneous multiple selves. The dominant understanding of different SM platforms as bounded (boyd & Ellison, 2008;Bullingham & Vasconcelos, 2013) is challenged by our findings. Our research methods reveal the perception of such boundaries, yet the actual boundaries were much more fluid than anticipated by our participants. In particular, the subjects, when reflecting on their digital biographies noted the challenges of managing such temporally determined digital identities in relation to shifts in their offline personal and professional identities. Although Belk (2013) highlights the importance of fluidity in the digital space, he does not take into consideration SM users' multiple temporal identities.
Recognition of temporality in relation to developing DP requires SM users to formulate strategies to facilitate these temporal shifts. These may include adopting micro-celebrity practices and accepting that there is a very minimal divide between the public and private persona in digital spaces, thus learning how to professionally manage one's public identity. For one subject, the creation of an online pseudo identity, which provides additional online freedom and can be deleted at a moment's notice, is paramount. Other strategies concern the impact of self-reflection, in particular, retrospective editing in order to 'manage' the multiple selves, as well as dispensing with the notion of the separation of the selves. Thus, acknowledging the temporal as well as contemporaneous self cannot be viewed entirely negatively, as the possibility to reflect on the overall self through the filmic medium did allow participants to understand themselves from the perspective of others, as well as from a temporal perspective.
Moving forward, there is an increasing need to develop further innovative methods to aid the evaluation and management of our digital selves, to ensure accurate representation of self and develop methods of coping with multiple temporal as well as contemporaneous selves.

Notes on contributors
Finola Kerrigan is Reader in Marketing and Consumption at Birmingham Business School where she researches and teaches marketing and consumption. Finola has published her work on marketing and consumption in the arts, branding and social media in a range of international journals, edited and single-authored publications.
Andrew Hart currently works for Merkle Periscopix, a Digital Marketing Agency in London, alongside being a consultative research member of the ReelLives project at Birmingham Business School. His academic background resides in film studies, having examined film from a cultural, textual and production perspective. He has subsequently co-authored a book chapter in the area published in The Routledge Companion to Arts Marketing (2014), as well as a journal article (2016) in the International Journal of Research in Marketing. His research has been presented at various academic marketing conferences around the world, whilst working on a number of other projects for a range of high-quality international journals.