Abstract
Belk’s theory of digital extended self has been highly influential in the studies of online behavior and its relationship to sense of self and identity. This study systematically reviews the papers (indexed by Web of Science and Scopus) that cited Belk’s seminal paper between 2013 and 2023. Focusing on studies that used Belk as part of their theoretical framework (N = 25), we classify the substrative contributions of this literature across disciplines into four categories: motivations, personal effects, consumption consequences and relationships with digital apps. Based on a critical examination of this literature, we suggest datafication as an additional dimension to Belk’s five dimensions of digital extended self phenomenon (sharing, dematerialization, re-embodiment, co-construction and distributed memory) and argue that this expanded and restructured model can be fruitful for studying individuals’ and groups’ online behavior in an increasingly datafied society. An agenda for research based on current trends is suggested to demonstrate the potential of the theory.
Keywords
Introduction
How are we to understand our selves in a datafied world? Can the study of self in digital environments—such as social networking sites (SNS; or social media), online games and virtual realities—be meaningfully separated from its study in the offline world? What is the relationship between online self and offline self? Numerous theoretical frameworks including self-discrepancy theory (Higgins, 1987), strategies of self presentation (Arkin, 1981) and self-presentation (Goffman, 1990), have been applied to the study of self in digital environments (e.g., Johnson & Ranzini, 2018). However, most of these theories do not incorporate any element from the digital environment and remain unchanged in application (Leary et al., 2012). Moreover, they tend to under-theorize the boundaries of self and, crucially, the relationship between online and offline self.
In this context, Russell Belk’s theory of Extended Self (1988) is a notable exception. In this paper, Belk offered a theory of self which went beyond the commonplace notion of self as ones’ perceptions, thoughts, and feelings about their mind and body to include other people, places, possessions. Not only this conception became immensely influential in the fields of marketing and consumer behavior (Ahuvia, 2013), Belk offered an “update” for the digital environments which has been influential across disciplinary boundaries. In this updated theory, individuals now also have a Digital Extended Self (hereafter, DES) which includes virtual possessions, places and people. Since its publication in 2013, Belk’s theory of DES has been cited more than 800 times and in fields as diverse as business (Wallace et al., 2017), marketing (Appel et al., 2020; Bartoli et al., 2023), media studies (Holiday et al., 2022; Ozanne et al., 2017), human-computer interaction (Mirbabaie et al., 2021) and psychology (e.g., Koles & Nagy, 2016). Partly due to its wide range of application, there is no overall view of how this theory has contributed to the debates about self in digital environment across disciplines. More specifically, there has been no systematic study to date of how Belk’s theory has been used to understand online behavior and its consequences for individuals’ sense of self and society at large.
This is an important shortcoming because as Cheney-Lippold (2017) and Szulc (2019) have argued scholars researching individuals in the online world are losing sight of their subjects and their sense of self and identity. As digital technology becomes increasingly ubiquitous in life, individuals are increasingly seen as data points in a constant stream of data that records and tracks their every move and interaction (Alaimo & Kallinikos, 2017). Likewise, digital environments are increasingly studied as data platforms that generate and mediate data about individuals in order to repackage and re-use them in myriad other ways (DuFault & Schouten, 2020; van der Vlist & Helmond, 2021). We suggest that by building on Belk’s theory of DES we can gain a more complete picture of individuals, their online interactions and the dynamics of self-formation.
In this paper, we seek to contribute to this goal by offering a systematic review of the cross-disciplinary literature that has incorporated Belk’s DES theory (2013) in their theoretical framework. Our aim is to first, highlight the contribution of existing research to our understanding of individuals and collectives’ self-formation and, second, argue for an updated model of digital extended self which takes both individual’s online possessions (including profiles, posts, comments, likes, etc.) and their digital traces for understanding their online behavior and its consequences.
As we demonstrate, understanding DES is important because it allows us to (1) have a more comprehensive picture of who users of specific platforms are and what they do, (2) gain a better insight about the group dynamics that shape individuals’ online behavior, and (3) critically examine how policies and activities of various social actors (from government agencies and corporations to trolls and hackers) affect these processes. Moreover, Belk’s theory can be used to understand the positive aspects of online consumption (e.g., well-being) as well as negative aspects, in issues ranging from body shaming to addictive behavior and cyberbullying to political trolls and online infiltration campaigns (Bartoli et al., 2023).
The rest of the paper is organized as follows: we begin with a brief overview of Belk’s theory and its background. Then we explain the methodology of our systematic review and, in the following section, present the findings of the most pertinent papers (N = 25) out of a pool of 642 papers. The subsequent discussion section is divided into three parts: first, we discuss the main strengths and weaknesses of the extant literature, then we provide a critique and update of Belk’s theory adding datafication as an additional dimension, and finally we explain how the updated model can help researchers better understand contemporary trends. We discuss the limitations of study before summarizing its contributions in the conclusion.
A Brief Overview of Belk’s Theory of Digital Extended Self
In his seminal 1988 paper, Belk proposed the notion of extended-self as a response to the question “what do our possessions mean for us?” and “how are they related to our sense of self?.” Drawing on James, Simmel, Fromm, and many others, he posited that “knowingly or unknowingly, intentionally or unintentionally, we regard our possessions as parts of ourselves.” Not only our possessions but also important persons and places can be regarded as part of the self if we feel attached to them. However, he maintained a difference between the core and extended self: even though individuals may feel attached to their bodies, internal processes, ideas, and experiences in ways similar to their attachments to certain persons, places, or possessions; the last three are more obviously distanced from our core self and can be more properly called part of our extended self.
Following the advance of the Internet and emergence of new digital environments, Belk updated his notion of extended-self (Belk, 2013). In his new contributions, Belk argued that “the old idea of a core self is an illusion” and defining self in the digital age requires exploring “the relationship between online and offline personas” (p. 478) He identified five dimensions (or novel aspects) of the digital world that influenced the shaping of the DES. They included (1) dematerialization; (2) re-embodiment; (3) sharing; (4) co-construction of self; and (5) distributed memory. Each of these dimensions will be briefly discussed below.
Dematerialization refers to the immateriality of online possessions and representations which, in turn, are inaccessible unless via specific technologies (like PC or smartphone). As photos, diaries, calendars and many other tangible objects are transformed into digital format, Belk (2013, p. 480) argues that the feeling of attachment produced by digital possessions is similar but not quite the same as our feeling toward physical possessions. Belk further highlights the similarity of rituals individuals perform to singularize or decommoditize their virtual possessions to what they do with physical objects. For instance, they may back up, store or curate such possessions just like they were real tangible objects.
The second dimension of Belk’s theory refers to re-embodiment, that is, individuals’ newfound ability to go beyond the material constraints in digital environments and embody themselves through digital texts, photos, videos, and avatars (Belk, 2013, p. 481). Moreover, individuals may experiment in these environments and explore the possibilities for different selves. Belk pays particular attention to avatars, singling them out as the more prominent objects through which we present ourselves as part of our self-extension. Furthermore, avatars allow users to experiment with a range of ideal selves, feeling attached to them and experiencing immersion. The sense of presence/telepresence that appears as a result of immersion in a game means that rather than playing the avatar, the player becomes the avatar (see MacCallum-Stewart & Parsler, 2008, p. 228). As Belk notices, this corrects the primary role Belk (1988) gave to the material body: our virtual body can change our real-world behavior.
The third dimension of Belk’s theory is sharing. In Belk’s view, sharing is not a new phenomenon, but the difference between earlier online interactions and contemporary digital environments lies in the possibility of sharing more and to a broader audience. This difference is not rooted in the internet itself, but rather in the feeling of anonymousness and invisibility available on the Internet. However, individuals are never in full control of the information they share, which means that at any moment what they have shared may become accessible far beyond the intended audience. Since the more information individuals share about themselves, the more their selves become visible, digital environments provide an unprecedented space for presentation (and study) of self. Sharing has also changed the pre-digital norms. What seemed awkward to share beforehand like telling others about a new bought car or decoration of the house or the graduation degree has become totally acceptable among a wide range of individuals.
Individuals rarely act alone in digital environments. Others are often present and involved through interactions. This leads to the fourth dimension of Belk’s theory, namely: co-construction of selves. One way to co-construct self is via comments, likes, etc. Receiving positive comments under posts, for example, helps in affirming one’s sense of self and its value. Besides these exchanges are in many cases shared with others and simulate a real-world conversation between multiple parties and help build an “aggregated DES.” Thus, even though online communications in these environments usually do not need co-presence, joint possessions such as shared folders, channels on YouTube or forums, blogs, comments, etc. give the impression of a rich communication between parties involved and an experience of mutual construction of identity which can be similar to real world interactions.
Belk also points out that this accumulated collection of texts, photos, videos, music, etc. relies on metadata and leaves a rich digital footprint behind. This enables the emergence of distributed memory which is the fifth and final dimension of Belk’s theory. New devices and technologies (such as search engines, timelines, or media archives) provide the capability of producing mass archives of information which function as “individual and collective autobiographical memory cues” (e.g., Nguyen & Belk, 2007). The memory cues involve both factual and non-factual information like events, interactions, emotions which together construct a person or a group’ memory about a shared experience.
In Belk’s view, online memories have several consequences for the shaping of self. On the one hand, it can lead to digital clutter, that is, dispersed storage of whatever can be archived because storing digital information is inexpensive and items can be retrieved using powerful search engines. This makes the DES a disorganized and unruly part of self about which individuals have little awareness, let alone any sense of control. On the other hand, such content will likely remain available even after death, becoming a sort of memorial to one’s self. But rather than being simply a memorial, Belk suggests, it becomes part of an immortal self. But perhaps most importantly, distributed memory enables individuals to extract from this massive resource selected items that put together in a coherent manner represent an integrated narration of self and its trajectory over time.
As mentioned in the introduction, Belk’s conception of DES was widely adapted in the literature and led to a wide range of discussions across different disciplines. In the next section, we explain the methodological steps we took to identify the papers that used Belk’s theory for understanding online behavior and its consequences for shaping selves.
Methodology
In order to systematically review papers that used Belk’s DES theory as the basis of their study, we proceeded as follows: First, the literature search was performed using Web of Science and Scopus to find papers that cited Belk’s 2013 paper. We only considered papers that cited Belk (2013) because this paper is Belk’s most cited paper on DES. While it is possible that some papers relied on the notion of DES without citing Belk (2013), we suspect that such papers do not engage with Belk’s theory of DES as part of their theoretical framework and would have been removed at later stages of this study as described below.
A total number of 1,485 articles, 649 in Web of Science and 836 in Scopus were found on 12th March 2023. Subsequently, papers that did not include either “behav” or “online” in neither the title, nor the abstract nor the keywords were removed from the results for each database. On the SCOPUS database, this was performed using the Boolean modifiers “behav” OR “online” in the “title, abstract, and keywords” feature. On Web of Science, records were retrieved without applying any field. Papers not in English were also excluded. Since an author’s contribution to the literature is best measured by his impact on other researchers (Amjad et al., 2020), papers (co-)authored by Belk were removed to narrow the review to the literature produced by other researchers. Removing duplicate findings from the two databases resulted in a total number of 642 papers.
The full text of 642 papers were then screened. Nearly all of the empirical studies confirmed one or another of Belk’s arguments and observations, but did not do so in a systemic way. Since most of these papers cited Belk (2013) in the introduction or discussion section, we decided to remove papers that did not rely on Belk (2013) as part of their theoretical basis. This criterion removed review papers, design-oriented papers and empirical papers that did not focus on understanding self. In the end, 25 journal papers qualified for detailed review. The qualifying papers were all published in high-quality journals (Q1 or Q2).
The process of search and selection is shown in Figure 1.
Flowchart showing how the final papers for analysis were selected.
Findings
Despite the relatively small size of papers selected for analysis, the range of issues discussed paints a vivid picture of versatility of Belk’s theory. This is why, rather than discussing papers in a chronological order or in terms of disciplinary boundaries, this section is organized around the four main themes that emerges from this literature, namely: (1) why individuals participate in digital environments, (2) how this affects them, and (3) what are the wider social and cultural consequences (4) how digital apps and virtual agents may affect this process
Motivations: Self-Presentation, Bonding and Documentation
Scholars have used Belk’s theory in studies of motivations for a wide range of online behaviors: liking (Ozanne et al., 2017) and sharing media content on Facebook (Johnson & Ranzini, 2018), sharing unboxing videos (Ilich & Hardey, 2020), posts about one’s parenting practices (Holiday et al., 2022). Drawing on Belk’s theory of extended self, self-presentation, bonding (and co-construction), and documentation have been identified by these studies.
Nearly all of the studies on motivations have highlighted that extended self is significant because it contributes to the online presentation of the self. In other words, across different contexts, users’ online behaviors are inextricably linked to their efforts toward impression management and presentation of (extended) self. This is a dimension that has been reflected in many other studies of online presentation of the self, even if they do not use Belk (2013) as part of their theoretical framework (e.g., Ross & Zappavigna, 2020).
Another motivation that researchers have found to be directly related to DES is bonding. Ozanne et al. (2017) demonstrated that liking contributes to bonding with others by showing agreement and maintaining an online relationship. This bonding is not only a part of the construction of self, but also leads to co-construction of identities. The findings of Ozanne et al. (2017) can also be interpreted as evidence for online formation of digital extended self which includes both “likes” and the individuals that users connect to through them (see also Johnson & Ranzini, 2018; McWha et al., 2018; Moncur et al., 2016; Pera et al., 2020).
A third motivation for online behavior, which is closely related to distributed memory in Belk’s theory is highlighted by Ilich and Hardey (2020) in their study of secondary packaging in Instagram posts. They go beyond self-presentation and co-construction to observe that desire for documentation that is, creation of a digital memory is a motivation behind unboxing videos. Although they do not point to creation of a “distributed” memory, they found that “freezing the moment for the future” and “preserving the experience” by posting on social media was a strong theme among the users.
While most of these studies agreed with Belk’s general outlook, Ngai (2023) argues that the importance of human self in studies of online behavior is often overemphasized. Building on a viewpoint rooted in post-humanist feminist theory (Haraway, 2013), she critically engages with Belk’s theory to argue that animals are not merely possessions and “subjectivity is socially mediated through multiple others, including nonhuman others” (p. 5). In her research, she draws on her own experiences of curating an Instagram account for her pet (and “cat research assistant”) and interviews with others in the community of cat owners who do the same for their pets to argue that “the self is always part of and constituted by multiple others” (p. 5). Therefore, she suggests, far more than self-presentation, affects that are shared between “nomadic subjects” including nonhumans are central to experiences of this group of Instagram users.
Together studies on motivation highlight the importance of self-presentation, bonding, and documentation as motives for social media users. Two of these motives (bonding and documentation) are closely related to two dimensions of Belk’s theory (namely, co-construction and distributed memory). Self-presentation, an underlying theme in Belk (2013), is discussed in many studies in the following sections as well. As we shall see, Nagai’s critical approach provides a corrective to the often-presumed importance of self-presentation in online behavior, a presumption that is frequently challenged in the subsequent sections. In the next section we concentrate on how Belk’s theory is used to understand the effects of online behavior at individual level.
Personal Effects: Well-Being and Identity Formation
All of these studies on the personal effects of online behavior rely on Belk’s notion of DES or co-construction to explain how individuals are affected or shaped by their online interactions and behavior. Most of these studies have highlighted change in well-being and identity formation as a consequence of online behavior. Moreover, they generally present a picture of online behavior that is more complicated than interactions motivated by self-presentation.
The three studies that focus on the role of sharing on social media in improving individuals’ well-being refer to the role of aggregated DES across very different contexts: interactions in cancer survivor networks in Hollenbeck and Patrick (2017), photo sharing among the elderly on SNS in Pera et al. (2020), and pet owners’ sharing of their pet’s “cute” pictures on Instagram in Maddox (2021). In all cases, the authors demonstrate how creation of an aggregated digital extended-self (ranging from photo sharing to virtual gift exchanges, online conversations and comments) contributes to overcoming negative feelings (such as loneliness or being old) and creating joy. Two studies, however, highlight the potential negative consequences as well. Pera et al. (2020) found that not everybody among the elderly demonstrated the same degree of self-acceptance through their shared photos and occasionally people shied away from sharing photos in which they were presented at all. Therefore, they concluded that the aggregated DES shaped by photo sharing can simultaneously function as creating, strengthening, and breaking bonds between the elderly and increase, or decrease their sense of well-being. Maddox (2021) also found that occasionally pet owners become inadvertently involved in situations akin to “digital kidnapping” (the taking and using of a child’s photos without consent). For her, this is most visible in the case of memes that are curated out of users’ posts without their knowledge or consent. Thus, while aggregated DES enables pet-owners to participate in creation of joy online, it can also have negative consequences that are not easily managed.
Research on the consequences of online behavior for self-formation presents an even more complicated picture of online behavior. As we demonstrate, all of these studies suggest that even though self-presentation may be a motivation for online behavior, the consequences of individuals’ online behavior for them are often far more complex and multi-faceted. These studies either (1) undermine the assumption that individuals act strategically and present themselves in any way they deem suitable, or (2) question the very idea of a singular, stable and integrated self.
Underlining the first point, Eagar and Dann (2016) and Kozinets et al. (2017) viewed selfies (and objects included therein) as part of DES and explored the ways in which they are used to construct a public identity. As Eagar and Dann (2016) show “selfies are neither entirely narcissistic nor completely self-mastery [. . .], but rather are a complex interplay of performance, narrative, and audience.” In a similar vein, Kozinets et al. (2017) question the view that selfies are predominantly a reflection of the narcissism of selfie-takers and argue that selfies are part of a social media-assisted identity work that helps to build, asserting, and curating lasting narratives of the self (c.f. Sung et al., 2018). The tensions within self-presentation are also highlighted by Nargundkar et al. (2020) who explored the relationship between consumption and identity through a survey of students participating on SNS. In their view, students DES is constructed as part of their day-to-day interactions with others as they try to “fit in” and “stand out” in online communities. Mirroring the findings of Johnson and Ranzini (2018), they found that identity formation is tied to the self-image, their ideal self and their assimilate-differentiate efforts within a context of social influences. Thus, engagement in self-construction and self-presentation occurs at the same time. As we argue in the discussion, this tension between trying to fit in and the desire to express yourself is one of the least understood aspects of construction of self on social media.
The second point, which represents a more radical critique of the idea of self-presentation as central to online behavior, is supported by studies that suggest individuals’ do not necessarily have or construct. In Procter’s (2021) study, for instance, Second Life users are “always already” in an idealized world and therefore, through the avatar, create an “aspirational or ideal self-image” which become “sites of self-making rather than merely placeholders for selfhood.” In Procter’s view, the avatar-persona is a separate but related part of the self with its own identity. Moreover, contrary to Belk, Procter argues that virtual possessions can be as physically or emotionally influential as material possessions (see also Nargundkar et al., 2020). Highlighting the same issue but in a very different context, McWha et al. (2018) found that many travel writers construct different versions of themselves in offline and online texts across different platforms (blogs, personal websites, SNS, etc.). Moreover, many of them disembodied and re-embodied their online identities by using names (and pseudo-names) in ways similar to avatars. This practice was particularly related to readers’ responses to their writing (through likes, comments, etc.).
Thus, research on personal effects of online behavior complicate the findings or research on motivations and add new dimensions to our understanding. According to this line of research, even though self-presentation may underlie much of our online behavior, the consequences of our online behavior and interactions shape our self and identity. Moreover, DES plays an important role in this process which not only shapes our self but also can affect our well-being. DES also allows us to experiment with and experience multiple identities which together shape a loosely integrated self-concept. In the next section, we focus on the studies that have investigated the role that DES plays in the shaping of self in the context of brand communities and consumption.
Consumption Consequences: Brands and Self-Concept
A distinctive line of research was developed by scholars researching the role of the DES in shaping an online brand image and brand culture and how this, in turn, shapes sense of self.
DES and Brand Culture
Several authors have demonstrated that DES plays a role in the creation of a brand image or culture. Presi et al. (2016) studied the effects of the selfie phenomenon on brand image. Building on Belk’s ideas, they found that consumers’ collage making activity, that is, combining elements of their offline private physical contexts with brand products and symbols in the (digital) world contributes to the shaping of brand image. In a similar vein, Chapman and Dilmperi (2022) found that actors who interact and co-create online content (like sharing photos and commenting or liking) contribute to an experience of belonging to and being part of a brand. Their findings also indicate that sometimes participants in these online communities developed “a feeling of belongingness in the luxury world without having to buy an expensive item” (p. 905). A similar finding was reported by Leban et al. (2020) who explored “lurking” practices of social media users, that is, the collection of “[online] practices not involving active participation in the production of content” (p. 515). These include, collecting, curating, and conversing about luxury brands and related content.
Taken together, these studies suggest that there are three main mechanisms through which consumers’ extended self contributes to brand culture: (1) the blending of users “private physical contexts” with brand products and images, (2) cocreation of online content (like sharing photos and commenting or liking) and (3) lurking practices that contribute to an experience of belonging to and being part of a culture. They also highlight the influence of brand culture on the shaping of self which, as we show in the next section, is the focus of another strand of research.
DES, Consumption and Self-Concept
Another line of research, has explored the relationship between consumption and self in the opposite direction, that is, how consumption and brand culture shapes behavior and sense of self via DES. Two studies, in very different contexts, has shown that brand community shapes sense of self. Schembri and Latimer (2016) studied a network of professionals using Adobe products as an online brand community and showed that such social networks have a salient role in constructing and co-constructing brand cultures by, among other means, enabling self-construction and the development of emotional relationships. Their aggregated DES which primarily consisted of “public display of awards, achievements and client portfolio” and “sharing biographies and personal narratives” in the digital environment contributed to the formation of a brand culture, which turn shaped their own identity. Similarly, Sobande et al.’s (2020) study of black women on social media demonstrated that by resisting and avoiding the consumption of mainstream media outlets and participating independently in social media platforms, Black women create online content and commentaries that yield a “sense of inclusion and potentially somewhat global collectiveness.” These studies rely on Belk’s theory of DES (especially notions of re-embodiment and co-construction) to explain the influence of online community culture on shaping self.
Investigating clusters of individuals on SNS, Wallace et al. (2017) conducted a survey of Facebook users who “Liked” brands and found that perception of self-congruence with a “Liked” brand increases with social tie strength, which is in turn positively associated with perceived attitude homophily. In other words, the more people felt that others shared their view about a brand (product, store, etc.) the more they felt close to them and the more likely they were to like that brand on SNS. Similarly, Ruane and Wallace (2015) demonstrated that in social networks, it is more likely for consumers who are influenced by others to choose brands that express themselves, or seek tribal membership. Thus, SNS provides individuals with an environment in which they construct their aggregated extended self by liking brands as part of connecting with similar-minded others.
In such cases, the weight of aggregated DES induces individuals to try to fit in and maintain a consistent self-concept. In a recent study which did not focus on brand products, Groening et al. (2021) explored the relationship between online wishlists and consumers’ likelihood of purchase. Building on Belk’s (2013) and Denegri-Knott and Molesworth (2010) about virtual consumption, they argue that the need to maintain a consistent extended-self explains why owners of public wishlists were more likely to make a purchase compared to those who kept their wishlists private. They also point out that online wishlists provide a unique liminal space between ownership and non-ownership that enables consumers to interact with products without the need to actually purchase them. Thus, research on the relationship between consumption, brand culture and self highlight the importance of DES in shaping online behavior and sense of self. It also shows how aggregated DES can limit the behavior of individual and push them toward conformity.
Beyond Social Relations: Digital Apps and Virtual Agents as Part of DES
A more recent line of research focuses on individuals’ relationships with digital apps and virtual agents (VAs). What separates this literature from other studies is that it focuses on digital apps or virtual agents as possessions or quasi-persons that individuals may use or interact with and how our relationship with them changes when they begin to see them as part of their extended self.
A few authors have studied the relationship that develop between apps and their users from the perspective of digital extended self. Relying on Belk’s framework, brand personality and brand-self congruence literature, Wang et al. (2016) explored the effect of personality traits of users and websites on emotional attachment of users to social media websites. They found that a match between users’ personalities and the digital artifacts’ (such as websites) [perceived] personalities increases the chances of affective connection between the user and the website as a result of the artifact becoming a part of the user’s extended self. In contrast, Scholz and Duffy (2018) treated Augmented Reality (AR) apps as the product to be consumed and focused on understanding how they can foster a close and intimate consumer-brand relationship in consumers’ intimate spaces. They documented how as a result of using the apps in the familiar, casual space of consumers’ homes, the boundary between consumers’ self and the app collapsed and led to a consumer/app fusion that “feels personal” and creates an open fluid consumer-centric space allowing self-expression and self-experimentation. This integration of AR apps into consumers’ extended self during enjoyable activities makes apps invisible and more likely to be used in the long term. Perhaps the main finding of this literature is that digital apps are not merely platforms through which individuals’ extended self takes shape; they are constitutive of DES and an important part of it in their experience of digital interactions.
Another line of research focuses on individuals’ relationships with virtual agents. Two separate studies focusing on the phenomenon in experimental settings have found that once individuals’ extended self includes VAs (Virtual Assistants in Mirbabaie et al. (2021) and recommendation systems in Trzebiński and Marciniak (2022), they begin to trust VAs more and view them as socially equal. Their findings also revealed that, in similar contexts, VAs do not differ from humans as team members and perhaps, can be considered as socially equal. While research in this area is very primitive, it seems that by mimicking the interactions of users with other humans, AI agents can earn their trust and participate in shaping an aggregated DES.
Discussion
In the previous section, we reviewed the contributions that used Belk (2013) as the basis of their theoretical framework. In this section we first summarize our findings and discuss the main contributions of this literature in terms of their engagement with dimensions of Belk’s theory and argue that its potential as a framework for understanding individuals’ behavior in digital environments and its consequences is still underexplored. Then we propose datafication as an additional dimension to Belk’s original theory and provide a conceptual structure for the theory in our extended version. Finally, we discuss the advantages of using this model for research in the context of contemporary and future trends.
Discussion of Findings
As our review demonstrates, scholars from various disciplines have used Belk’s theory in a wide variety of contexts (see Table 1). One strand has used DES to understand motivations for online behavior through the notion of DES, mainly from a self-presentation perspective. Another line of research, largely critical of the self-presentation assumption, has highlighted the consequences of DES shaped through online interactions for individuals, namely, their well-being and identity formation. A third line of research, with a general focus of marketing and consumer culture, has showed how DES shapes identities of groups and communities. Finally, research has also demonstrated that once humans consider VAs and digital apps as part of their extended self, their relationships with them changes.
Final Papers Included in the Analysis.
However, from the perspective of dimensions of Belk’s theory, very few studies have engaged with all aspects of Belk’s theory similarly. Two papers only adopted the concept of DES and only one paper explored four aspects, the highest number of dimensions considered in any paper (see Table 1, cf. Güngör & Ozansoy Çadırcı, 2022). Sharing is, by far, the most extensively researched dimension of Belk’s theory. While this might seem to be the result of social media features or a reflection of current academic trends (e.g., John, 2024), one may argue that, as far as Belk’s theory is concerned, this near-ubiquitousness reflects the fundamental role that sharing plays in the process of formation of the self in online environments. Nevertheless, users’ activity on social media usually involves interactions (including sharing) with others, and as various studies have demonstrated, sharing in online environments enables feedback and participation of others (by means of likes, comments, etc.) in the co-construction of self. It is therefore not surprising that many studies primarily explored these two dimensions.
In contrast to sharing and co-construction which were explored extensively, other aspects of Belk’s theory have received little attention. Very few papers studied dematerialization, re-embodiment or distributed memory. While this apparent imbalance may be regarded simply as a reflection of researchers’ interests, the extensive attention that scholars have paid to sharing and co-construction may have roots in a more fundamental issue. Both sharing and co-construction are aspects of Belk’s theory in which individuals play a more visible role and can be more easily explored and documented. While it is possible that researchers view dematerialization as taken for-granted and distributed memory and re-embodiment are aspects in which individuals do not necessarily play an active role and are probably less conscious about. This suggests that studies have been limited to aspects of individuals’ behavior or experience that are more directly related to their traditional active sense of self, rather than exploring what happens to their extended self, outside of their intention or “personal” experience and in-between the online and offline worlds.
There is further indication that researchers are still studying individuals’ online behavior along the lines of the traditional sense of self and pay little or no attention to the notion of self in its fully extended sense. Few studies have incorporated digital possessions (posts, photos, etc.) into their analysis. McWha et al. (2018), for instance, rely only on interview data and fail to incorporate the content of travel writers’ blogs into their understanding of how their identities change through their writings (cf. Holiday et al., 2022). This points to a methodological shortcoming in much of the extant research on the DES. Despite the apparent diversity (which ranges from data crawling to quantitative surveys to qualitative ethnographies), it rarely ventures beyond the perspective or intention of individuals and often fails to address group-level phenomena (Cf. De Leyn et al., 2022). There is, therefore, ample space for researching individuals’ extended self as well as groups’ aggregated extended self and their relationship to online behavior, particularly from mixed-method and data analytical approaches (See, e.g., Lee et al., 2020).
Paying attention to individuals’ and groups’ extended selves can be a particularly promising line of research because, at present, there seems to be a tension between two aspects (or framings) of individuals’ online behavior: on the one hand, there is evidence to suggest that individuals find ways to reveal and express themselves more freely in the online environment (e.g., Bartoli et al., 2023; Eagar & Dann, 2016). On the other hand, much of the present research highlights the ways in which individuals are constantly fashioning and managing their online self-presentation, often seeking acceptance and influence among peer groups (Ozanne et al., 2017). The question remains, however, in what circumstances do individuals feel free to express themselves or alternatively, pressure to conform to group norms and expectations. At the same time, as the difference between brand culture and black consumption literature (Sobande et al., 2020) shows, it is necessary to understand norms as a group-level phenomenon that is reshaped and reworked through interactions (Moncur et al., 2016; González-Bailón & Lelkes, 2023).
More broadly, one can argue that the dichotomy between self-expression and impression management is too simplistic: in reality, people chose a wide variety of coping strategies to reduce the burden of social pressures felt while enjoying some of the freedom that digital environments provide (for a discussion of different privacy management strategies arising from these concerns see Kang & Shin, 2021). Moreover, increasingly individuals are becoming aware of the vast digital footprint that online behavior leaves behind and seek to manage it. This suggests that Belk’s DES theory can provide a basis for future research in this area (see below).
Taking a broader perspective on online behavior can be crucially important for making sense of the dynamics of a wide range of online phenomena (from spontaneous reactions to targeted campaigns) and how they take shape. Research has already suggested that using social media may lead to a somewhat paradoxical situation in which individuals may have the opportunity to liberate themselves from (restrictive local) ties but nevertheless remain within a closed circle of individuals with similar beliefs. Studying DES can equip scholars with conceptual tools necessary for understanding the dynamics of these relationships. It also helps to understand when and why individuals may choose to participate anonymously (e.g., O’Leary & Murphy, 2019) and how it affects the dynamics of the situation. This can be particularly valuable for studying the role of digital worlds in reflecting, creating, or amplifying social problems that are affecting societies today.
It is clear from the above discussion that Belk’s theory still holds an enormous potential for exploring different aspects of individuals’ behaviors in digital environments and its consequences using a variety of methods. It can be used to understand identity formation as a group-level phenomena and the ways in which social media features enable (or disable) specific forms of interactions that have consequences for individuals as well as groups (one example being the controversies surrounding Facebook’s dislike feature). But, as our analysis has indicated, this requires attention to aspects of Belk’s theory in which individuals play a less active conscious part. In order to enable this, we believe, Belk’s theory must be expanded and restructured as suggested in the next section.
Expanding and Restructuring Belk’s Theory
Despite the richness of Belk’s theoretical framework, we believe that he has neglected one important aspect of the DES, that is, datafication. Although the term has been used by various scholars to draw attention to the increasing role of data in the contemporary digital world (e.g., Lycett, 2013; Mejias & Couldry, 2019; Southerton, 2020), we use it here to refer to the “processes of rendering information into machine-readable quantifiable data for the purpose of aggregation, analysis, and anticipation of human behavior and social interaction” (Koenen et al., 2021, p. 137).
In digital environments, datafication and dematerialization are two sides of the same coin yet refer to distinct processes. While dematerialization refers to the immateriality of online possessions and representations, datafication indicates the generation and storage of data tokens referring to dematerialized digital objects (Kallinikos et al., 2010). For instance, as soon as a digital photo is uploaded to social media, data about its uploader, file size, image size, time of uploading, etc. are generated and stored. Moreover, over time the database stores data about who accessed the photo and the time of access. Some authors have referred to this phenomenon as “computing the everyday” and social media as data platforms (Alaimo & Kallinikos, 2017).
Datafication is a fundamental aspect of the DES because it enables re-embodiment and sharing by allowing access to digital possessions. Moreover, it is necessary for the traceability of individuals’ digital footprints and, as such, can fundamentally shape our experiences (or otherwise) with distributed memory. In other words, the way in which we encounter digital memory cues about our past depends on how data about those dematerialized events and experiences are recorded, kept, and retrieved across different platforms and environments (Mayer-Schönberger, 2009).
In addition, datafication is what makes social media platforms (and other digital environments) “magnets for regulation.” As Mueller (2015) has argued “the mundane human activities” that take place in these environments (including negative behaviors) have always existed but were never as transparent as they are today, due to the “storable, searchable records” that these environments afford. This hyper-transparency and “objectification of social interaction,” leads to demands for regulation (especially in response to abhorrent behavior) which often targets these platforms rather than the actual offenders. On the other hand, the controversies surrounding “the right to be forgotten” often concern the way these records shape the opportunities and restrictions that the way in which data is stored offers to individuals (e.g., Abril et al., 2012; Koops, 2011). Seen from the perspective of Belk’s theory, digital footprints are part of individuals’ extended selves that play a part in the online construction of their identity (whether they agree to it or not).
Adding datafication to Belk’s theory moves us toward a more comprehensive picture of how individuals’ extended self takes shape in digital environments (see Figure 2). Furthermore, one can structure Belk’s theory within a schematic structure that clarifies the relationships between different elements. Dematerialization and Datafication are the basic features of participation in digital environments that rather than merely generating representations, enable Re-embodiments of individuals to take shape in online and offline worlds. Then, through Sharing practices this extended self becomes (partially) accessible to others, enabling individuals to interact with each other by acting upon each other’s DES (like someone’s posts) and thus participating in the Co-construction of (individual and collective) identities. This activity, spread among different digital environments and different time frames, leaves behind a digital footprint that constitutes a Distributed Memory available to individuals as future reference points or money cues.
A structured and expanded model of Belk’s digital extended self theory.
Placing the Expanded Theory in Contemporary Trends
Several technological trends are transforming people’s experience in the digital world. First, the use of Artificial Intelligence (AI) in games and social media environments (such as recommendation systems or chatbots) has been on the rise (Appel et al., 2020; Sands et al., 2022; Trzebiński & Marciniak, 2022; Yuan et al., 2013). Recommendation Systems shape individuals’ interactions and experiences in many different ways, for example, by managing their timeline, recommending connections, or providing memory cues, depending on theirs (Kunst et al., 2022). Recommender systems which are designed as chatbots participate in digital environments more actively because they are often indistinguishable from individuals in their activities (whether “posting,” sharing, or liking) and thus can have an important role in individuals’ experiences of online co-construction of self and distributed memory (see Seering et al., 2019). How individuals’ interactions with each other change as a result of the activity of these non-human agents and how this shapes their (sense of) extended self can be a very important area of research in the next decade. The impact of these agents and the concerns they have caused are already an important area of policy in several countries (Chavalarias, 2016). This is an area where our extended model can be hopefully deployed to produce insights.
Moreover, one relatively underexplored aspect of how social media shapes individual and aggregated selves is the role played by features available to users, for example, like/dislike, comment, re-share, feed, timeline, etc. (van Dijck, 2013). The role played by these features can be explored using our extended version of Belk’s theory in various ways: first, they enable individuals’ explorations and interactions in online environments (dematerialization). Second, the use of features forms part of users’ extended and aggregated digital selves (e.g., Hong et al., 2017). Third, they have implications for individuals’ identity through notions of co-construction and re-embodiment (Cf. Webb, 2001; Bessière et al., 2007). Fourth, these features often make it possible to track the records of individuals’ past self (what Belk would call the distributed memory of their activities) enabling them (and others) to trace the formation of their identity (datafication).
The role played by features of social media platforms comes into sharper view when one compares the dynamics on these platforms with those on online gaming platforms. Even though the extant literature on the construction of extended selves on online gaming platforms is very small, there are indications that features that are more commonly found on these platforms (such as avatars or various forms of exchange of digital objects) shape a different dynamic from that found on social media platforms. This lack of comparative studies focusing on the role of different features in various digital environments promises to be a fruitful avenue for further research.
Along the same lines of reasoning, one can explore individuals’ online experiences is virtual reality environments (VRE) which effectively offer alternative worlds with their own rules of participation for individuals to interact with each other as well as own possessions or sell or buy goods (or services; Kruzan & Won, 2019). While such possibilities are not without precedent, the extent to which individuals engage with VR environments can have important consequences for the shaping of their extended selves both online and offline (Nargundkar et al., 2020; Procter, 2021). As Scholz and Duffy (2018) have shown in their study of AR apps, the more these VREs become merged with the real world, the more they are likely to become an important part of individuals’ sense of self. In extreme cases, individuals may begin to become wholly immersed in the digital world to the extent that the most important aspects of their identity belong to the virtual world (see Evans, 2012). In other words, the virtual world becomes their real world. We believe our model offers a helpful framework for analyzing the multitude of variables that shape individuals’ experiences and their extended self in these worlds and beyond.
Thus, our extended model can be used to study the dark side of individuals’ DES as well. As Smith (2018) has argued, people in Western societies generally assume that data is normal, necessary and enabling. But research has already indicated the negative effects that individuals’ digital footprint can have on their life opportunities (Jeanningros & McFall, 2020). Search engines like Google are increasingly incorporating this data into their search algorithms making individuals’ online behavior (such as their tweets) more visible to anyone with a passing interest (Bilić, 2016). In turn, individuals who are aware of this trend are modifying their online self-disclosure to ensure their life chances. De Leyn et al. (2022), for instance, discuss how young individuals from marginalized ethno-religious backgrounds use the intensified visibility of their relational maintenance behaviors on social media in order to arrange status and social ties. Our reconstructed model can redirect current research from a focus on individuals’ online behavior to exploration of group-level interactions and how they shape their (individual and aggregated) extended selves. This will affect how we understand (conscious and unconscious) online self-disclosure (cf. El Ouirdi et al., 2015) and how practices like fraping (Moncur et al., 2016) may have negative effects on individuals. Moreover, our model can help explore the role of “data phantoms” generated through individuals digital footprint that can seriously damage their health through the so-called “precision medicine” delivery mechanisms (see Green & Svendsen, 2021).
Limitations
Our study was limited in several respects. Firstly, our focus on online behavior and its influence on individuals’ self-concept justified our choice of keywords “online” and “bahav” in the systematic review. While the result (N = 642) covered the majority of papers citing Belk (N = 836 on Scopus), it is possible that some relevant papers did not use these terms (possibly on theoretical grounds). It was beyond the scope of this systematic review to find and include such papers, but future reviews using different terminology can shed a different light on this literature. More importantly, there is a potential risk of bias in our findings because the reviewed papers were overwhelmingly based on studies of anglophone world and other groups and populations were underrepresented. We hope that future studies balance this situation.
Secondly, at a more conceptual level, our contribution—following Belk—is based on self as “an individual’s self-concept.” We acknowledge that there are other conceptualizations of the self (Leary et al., 2012) that can be useful in our understanding of self in the digital world. Future research can build upon alternative conceptualizations to expand our understanding of this phenomenon. Thirdly, our discussion of future directions is primitive and based on current understanding. As new digital technologies emerge at speed, it would be hardly surprising if our analysis comes to be seen outdated in a short period of time. Nevertheless, we believe this to be a challenge that social research on digital technologies must grapple with and, we hope, understanding DES can be helpful in this endeavor.
Conclusion
Individuals’ online and offline lives are increasingly intertwined in contemporary society, and we believe that a holistic approach such as Belk’s can make important contributions to our understanding of the emergent issues of datafied society. This paper set out to demonstrate the power of Belk’s theory to illuminate multiple aspects of individuals’ online behavior, and to argue that there are still many untapped opportunities for investigating how individuals’ DES takes shape in contemporary society. Through a systematic review of literature, we underscored the contributions and shortcomings of the literature that has built on Belk’s theory. We also extended and structured Belk’s DES theory and suggested that datafication can be recognized as a neglected dimension that connects DES and its digital trace. Placing this framework in the context of contemporary trends, we argued that dynamics of interactions of individuals, formation of aggregated DES, and other group-level phenomena can be studied more holistically using this extended model. Future research can shed light on more complex aspects of online behavior and provide further insights into individuals’ DES and its consequences.
Footnotes
Acknowledgements
The authors wish to thank Dr. Saeed Nariman for his encouragement and helpful comments on the earlier versions of this paper. They are also grateful to Prof. Russel Belk for his thorough and supportive feedback.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.