Apples, oranges, and self

ABSTRACT

Extending Thompson’s discussion of the extended self, I contrast his view if the I Self (in William James’ terms) with my focus on James’ Me Self. I go on to discuss the Aggregate Self, the Digital Self, and various concepts of plural selves and cultural differences. Thompson’s introduction of distributed self fits nicely here and further expands the discussion of self. I introduce the concept of the Extended Object and briefly examine thanobots and AI and the way they may affect our concepts of self in the future.

KEYWORDS:

• Extended self
• ontologies
• I/Me self
• extended object
• AI

I am truly delighted to see that Craig Thompson (2024) has chosen to address, challenge, and extend my papers on the Extended Self. Finally, someone has gone beyond obligatory citations based on vague familiarity with these papers and has really grappled with the underlying assumptions and suggested directions for future possibilities. May many more such critiques and discussions follow. They are welcome and long overdue.

What I want to do here is to continue the discussion, not so much by challenging Craig’s assumptions or forecasts but by clarifying the sense in which we may be talking about apples versus oranges in our two conceptions of the self. Both selves exist, but they are two quite different things. They have been discussed in prior literature as ‘I’ (Craig’s focus) versus ‘me’ (my focus). They involve profoundly different ontologies of the self. I discuss this distinction in greater detail in this response.

But before getting into those details, let me unpack the basic differences between Craig’s ‘I’ approach and my ‘Me’ approach. While I have attached our names to them for clarity, neither of us invented these concepts. William James (1890) distinguished between ‘I’ as subject and ‘Me’ as object. He also included other people and objects as part of me. I have done so as well (Belk, 1988). Wittgenstein (1958) did not include people or objects as part of his ‘self-as-object’ (Me), but he is close to James (1890) in his conception of ‘self-as-subject’ (I).

Thompson too is largely concerned with the ‘I’ self – the thinker of thoughts, the subject, the agentic self. In contrast, I am largely concerned with the ‘Me’ self – the material self, the object self, the mirror self. By the mirror self, I refer to Cooley’s (1902) ‘looking glass self’, or the self as seen through and reflected by others. Both the I self and the Me self are a necessary part of all self-conscious beings. Although Thompson and I focus on one or the other, they are normally quite interdependent. The union of the two comes in what Baldwin (1897, pp. 372–394) recognised as the link between me and my desires, which are ever-changing in the Lacanian quest for self (Belk et al., 2003). The I self attempts to realise these desires. But because of the agency of other people and objects, the I self isn’t in full control of this realisation of the evolving Me self. Moreover, agency can be slippery. As Gell (1998) argues, while normally a car owner is the agent and the car is the patient, when the car breaks down at night these roles may be reversed.

As part of my later conceptualisations of the digital self (Belk, 2012, 2013, 2014b, 2015a), I outlined the ‘co-constructed nature of the digital self’. This feature involves the feedback of others via social an digital affordances and includes people (and now some AI devices) that tag us in online group photos and often add meta-data through online comments. These actions help to jointly construct our extended self (Me).

With the distinction between the I self and the Me self in mind, we can see that Thompson’s concern with ontology, or world view, is a key focus of his comments which are directed at the I self and its agentic efforts to construct the self. My concern, focused on the Me extended self, is different from his. However, in formulating the Extended Self (Belk, 1988) I did outline several ways in which objects and people can become a part of self, including the subject-based acts of creating them, purchasing them, and coming to know them (c.f., Sartre, 1943). That is, behind the object Me there is often the subject I. But not always. Besides the me-focused co-constructed self, possessions may also be received as gifts that may be cathected into the extended self through the psychic energy imparted to them by the giver, the receiver, their relationship, and the ritual act of gift prestation (Belk, 1996).

Another process through which the agency of Thompson’s imperial I self loses some of its ontological control is that of sharing. In sharing (Belk, 2010), there is a loss of control as ‘I’ becomes ‘we’ and ‘mine’ becomes ‘ours’. In acts of sharing, we lose control of possessions and gain expanded access to self-expressive goods such that ‘I am what I have’ (and desire) expands to ‘I am what I can access’ (and desire to access) (Belk, 2014c). Together with the plural de facto ownership of these shared goods and the self-aggrandisement they provide, ‘sharing in’ (Belk, 2010) also creates or reinforces a sense of shared self among participants.

Complications to singular-self ontologies due to non-singular selves

In instances of the co-construction of self, gift receipt, and sharing in, we can see elements of what I have called the Aggregate Self, and what James (1890) described as other people as part of our self. In certain cultures, a related concept is that of the ‘dividual’ or ‘partible self’ (Appau et al., 2020). The concept of ‘dividuals’ comes from McKim Marriott’s (1976) work in southern India and Marilyn Strathern’s (1988, 1999) work in Melanesia. These related concepts of the self are different from the Western concept of the individual as a singular entity. Instead, the Melanesian person is partible and disbursed over their relations with particular other persons (Gell, 1998; Strathern 1988). This is achieved largely through transfers and gifts of objects and animals. In the case of Indian dividuals, the self is instead permeable. Here too, the self is the sum of relations with others (Fowler, 2004), but these relations and gifts are likely absorbed in the manner of Self Expansion (Connell & Schau, 2013) or ‘inclusion of the other in the self’ (Aron et al., 1991, 1992). These and other alternative conceptions of the self and other persons are considered in greater detail in Belk (2014a). The term ‘dividuals’ was later used by Deleuze (1992), but with different meanings (see Hietanen et al., 2022) that are not considered in the present discussion.

The point is that much of Thompson’s critique of the ontology of the Extended Self (singular) is disrupted by these constructs of interpenetrating selves as well as the co-constructed selves noted earlier. These selves may also be subsumed by the notion of the aggregate sense of self (Belk, 1988). The one welcome exception to Thompson’s focus on the singular self is his discussion of networked distributed self as seen, for example, with the distributed ownership and joint control of crypto currencies. This possibility of the networked or distributed self has been discussed by Enfield (2017), and even extends in some cases to distributed agency by the dead as with legal wills, post-mortem virtual selves, and forensic anthropology’s investigations of human remains (Crossland, 2017). This latter case illustrates one possibility for the Me self to persist after the I self is gone (see also Belk, 2015b, 2018; Belk & Humayun, 2018, and the discussion of the extended object in the following section).

Distributed agency has also been analysed by Humayun and Belk (2016, 2021) in the context of bitcoin. Unlike fiat currency where a government can at least partly direct their country’s economy through monetary policy, there is no one with centralised authority to act on behalf of a Bitcoin economy or the Bitcoin brand. Moreover, even if Bitcoin’s anonymous creator(s) were to surface, they would have no power over the value or stability of the coin and no PR platform from which to issue press releases. Once the coin was set up with provisions for creating (mining) new coins, all underwritten by the Bitcoin blockchain, it became a self-sustaining currency with guaranteed proof of ownership. As with the distributed owners of Bitcoin, the distributed self is not shaped by the individual alone. Parents, siblings, friends, teachers, employers, partners, children, influencers, online bullies, and trolls can all effect both our I self and our Me self.

What we call ‘self’ is really ‘selves’. Besides distributed selves and multiple selves shared between individuals, there are also multiple selves within individuals. These may be situational selves, as when we act differently at different times, and in different places and circumstances (Belk, 1975). They may also be dialogic selves, as when we have an internal dialogue about whether to spend or save, indulge ourselves or show self-restraint, and volunteer or remain silent (Ahuvia, 2005; Bahl & Milne, 2010). This is something Thompson correctly recognises in pointing to Tian and Belk (2005). Based on our research, Tian and Belk (2005) observed that we might well have a photo of our family on our desk or in our locker at work, but it is unlikely that we would have a photo of our boss on the wall at home or on our cellphone. Nevertheless, the workplace and the work self reach into the home through electronic tethers. This trend was exacerbated by COVID mandates to work from home, and their continuing legacy. But Thompson seems unconcerned with multiple selves and the problems they create for an imperial individual self ontology.

Instead, he sees my one-time flirtation with assemblage theory and Actor Network Theory as a promising development that opens the gate to an increasingly object-oriented ontology in which we outsource more and more of our consumption choices to technology (Hochschild, 2012). While Thompson follows these perspectives to the distributed self, discussed above, I would like to trace them to a different ontology entirely. There is a counterpart of the extended self to be found in the concept of the extended object.

The ultimate flat ontology: the extended object

Picking up on Thompson’s call for a more open and networked conception of the self extended into other people and things, we may take the role of objects one step further and consider that objects may be extended somewhat similarly to the way people are extended. I have written about this possibility in several recent papers (Belk, 2015b, 2018; Belk & Humayun, 2018). The idea is that objects can acquire a biography based on their uses by and associations with humans. Igor Kopytoff (1986) roughed out the idea that things can acquire a narrative by being singularised or made unique and meaningful by those who spend time with them. Kopytoff made the key determinant of whether a thing is a singular possession or a commodity, its saleability.

But singularised possessions are not just on-again and off-again commodities. They can even take on the aura of sacred objects. A home, a pet, a car, a kayak, or a piece of furniture are all examples of objects that may be treated as secular sacred objects by those who own them (Belk et al., 1989). For example, Epp and Price (2010) followed the life course of a large family table. Grant McCracken (1986) found that parents kept children’s [singularised] clothing and toys in the attic for a time so their ‘hot memories’ could ‘cool off’ before disposing of these once special objects. And Hirschman et al. (2012) found that garages serve as similar transitional spaces for unused objects before disposing of these relics of suburban America (Arnold et al., 2013, pp. 44–47).

But these are all cases where the person eventually chooses to de-singularise objects and turn them into saleable commodities or worthless junk (Gregson & Crewe, 2003). Their former sacred or singular meanings were imparted by humans to whom they were dear. I am instead talking about cases in which objects that, due to their histories with people, have accrued meanings that persist after disposition. Examples might include a piece of jewellery once owned by Paris Hilton, a music box owned by Winston Churchill, or the estate of a dead loved one. Museums may intervene to maintain objects and homes owned by famous or historically important people like John Fitzgerald Kennedy (Newman et al., 2011).

Steven Conn (2010) once raised the question, ‘Do Museums Still Need Objects?’ in our digital era when we can easily download a virtually perfect copy. He concluded that they do, as did André Malraux (1947/1967) in an earlier era. The reasons that Conn gave for museums needing objects included inspiring awe by presenting the real thing, as evidenced by people crying at being able to touch a piece of moon rock in the Smithsonian, making a pilgrimage to the shrine at Lourdes, and queuing to see the Mona Lisa at the Louvre.

But contemporary museum curation demands that the museum must do more than offer a collection of marvellous objects like the Wunderkammern of old (Davenne, 2011/2012; Mauriès, 2002). The curators must now also create meanings and tell stories about these wonderful objects while being sensitive to the context, the indigenous people from whom they may have been acquired, and the political climate in which they are being displayed. In Christensen’s (2012) terms, museums today are presenting ideas as much as things. He gives the example of the 19th century house of Henry and Matilda Gates in Fayetteville, Arkansas which is now a museum. Rather than simply exhibit Victorian furnishings and period rooms, each room is dedicated to an idea. So, the visitor learns that Henry owned a dry goods store, was an abolitionist, and proudly displayed the emancipation proclamation in his store window. Matilda was a women’s suffrage leader, an advocate for Native American sovereignty, and a supporter of separation of church and state. Their youngest daughter married L. Frank Baum who wrote The Wonderful Wizard of Oz, The Art of Decorating Dry Goods Windows, and The Life and Times of Santa Claus. In addition, their house was used as a sanctuary on the Underground Railroad helping escaped slaves move northward. This is a much more complete story than simply showing a house full of old stuff or celebrating an elite white man. The extended objects now convey accumulated stories and ideas about their former owners.

The preservation of memories and stories of ancestors is something that McCracken’s (1988) ‘Lois Roget’ also tried to impart to her children, to whom she hoped to bequeath various pieces of antique furniture along with the stories of their ancestors who once owned them. In this way, the furniture would also serve as mnemonics to cue the stories as they entered the lives of a new generation. But her children did not want the old-fashioned furniture or the old-fashioned stories. They were modern, mobile, and much more concerned with new technologies. A second best choice may be to find an appreciative heir outside of the family who appreciates some of the meaning of these things (Lastovicka & Fernandez, 2005). Those of us who have lost parents and had to dispose of their estates know the pain of disposing of once-treasured possessions whose personal meanings to the deceased are abandoned as they are sold to new owners who will be oblivious to their stories. Or, more commonly, the objects may be unceremoniously junked like the scene at the end of Citizen Kane when his treasures, including his beloved childhood sled ‘Rosebud’ that had preserved his one happy memory, were tossed on the bonfire of the vanities.

Our connections to the extended object are a subject that has received considerable research attention, since death is less final for the symbolic Me self than it is for the agentic I self (Arnould et al., 2009; Curasi, 1999; Curasi et al., 2003; Guillard, 2017; Karanika & Hogg, 2013; Price et al., 2000; Winterich et al., 2017). As Ekerdt (2018, p. 29) observes,

The standard view, which is difficult to doubt, is that possessions cohere around a person or household and these persons’ subjectivity gives meaning to the objects. At the same time, it is possible to maintain things have a ‘material agency’ that shapes older adult’ behavior in ways unintended by human subjectivity.

What Ekerdt means by this is that possessions, like the people in one’s extended self, can be a blessing and assist the older adult in various ways, but they can also be a curse when they annoy, hinder, or overwhelm us. They can become a burden. And if older people wait too long to downsize overwhelming accumulated possessions, doing so may be beyond their capacity.

I noted earlier that the Me self of identity may live on after the I self has died, and even the I self may continue to have some agency after death. That the dead are brought to memory by the objects they leave behind has been recognised for some time (Belk, 1988). The same is true of the digital artefacts they leave behind (Belk, 2013), creating a role for the executor of the deceased’s digital estate so that these digital objects present them in the most favourable light (Carroll & Romano, 2011; Pineto-Padoch, 2022). Online memorials have also been treated in prior research, but recently innovations like QR codes etched into tombstones have emerged. These QR codes can be scanned to learn more about the dead through photographs, videos, messages, and autobiographical material (Hutchings, 2017; Kasket, 2019). To the extent that these artefacts were created before death, they can be seen as part of the digital Me that continues to reflect the self, much like our academic publications may if we are fortunate and someone still reads them. Echoing a Black Mirror story called ‘Be Right Back’ (Brooker & Jones, 2018) and a therapeutic 2D chatbot called Replika (Possati, 2023), it is now possible to scrape the deceased’s online messages, social media activity, and other digital traces, and create an interactive chatbot or robot that responds as the deceased might have (Henrickson, 2023; Sisto, 2020, 2021). As Thompson hints, with such developments, and thanks to emerging AI developments, not only might the agentic wishes of the dead be carried out post-mortem, but the I self may continue its spontaneous actions, if only through the pseudo-self of a thanabot.

Coda

Thompson concludes with a discussion of humans creating an artificial general intelligence (AGI), and he cites a 2003 Nick Bostrom article about his worry that we are already approaching life in ‘a multiverse of simulated realities’. Well, maybe. But 11 years later Bostrom (2014) wrote a book about superintelligence. In that book, it is the superintelligent AI that is trying to figure out if it is living in a simulated reality. Here, the human’s task is to try to determine the AI’s theory of mind and find the best way to control the AI before it behaves in a way that is dangerous to humans. Bostrom lays out ten alternative ways in which this might be done. These are things that AI scientists and regulators are currently worrying about.

Thompson concludes that thanks to AI, ‘“consumer research”, as we currently understand it, may soon be radically transformed’ (2024, p. 555). More than that, I am suggesting that consumers as we know them may soon be radically transformed. Although the rate of human genetic improvement may be too slow compared to the rate of improvement in self-programming AI (Belk, 2017), there are other possible human enhancement technologies that may be much faster. In the future, we may well have enhanced transhuman consumers who use biochemicals, nanotechnology, prostheses, and neuro-enhancements in an attempt to extend their lives, enhance their mental abilities, increase their speed and strength, and perhaps even achieve secular immortality (Belk, 2022b; Lima & Belk, 2022).

But even without enhancement technologies and biological immortality, there may be ways for the Me self and even the I self to continue to have an active presence after death. This is not the digital self only. Nor am I talking about artificial intelligence (Bakpayev & Belk, 2020), artificial emotions (Belk, 2022a), or artificial life (Belk et al., 2020), all of which are directed at enhancing an external non-human entity rather than creating an enhanced human being. Instead, I am referring to the possibility that we humans could become God-like superhumans (Lima & Belk, 2022), as Bostrom (2014) forecast.

Or is this mere hubris to make us feel more secure in a world where the I self has less and less control of the digital Me self? Already we receive tweets and emails from algorithms. Soon algorithms will be composing our messages and posts for us. And in the future, startling as it would initially be, we may receive perfectly timed birthday greetings from a dead friend or relative with a personal note that is exactly what they would have said.

In fact, this is already possible. After Eugenia Kuyda and her dear friend Roman Mazurenko moved to San Francisco to start their online tech companies, Roman went back to Moscow for a wedding. But in Moscow he was hit by a car and killed. Genia was devastated. All she had of Roman were thousands of messages they had exchanged online. Since her company created chatbots, she had her software engineers use these messages and others gathered from friends and relatives to create a Roman bot. She exchanged messages with the bot and was convinced that it replicated what Roman would have said perfectly. Her company, Luka, has since released the companion digital bot Replika that interacts with the user, gets them to talk about themselves, and gradually becomes more and more like the user (Possati, 2023). However, there are now firms in the deathtech industry that make it possible to order thanabots (also called digital zombies—Bassett, 2015) on the order of Roman. Thus, in 2020 Kanye West presented his then wife Kim Kardashian with a (non-interactive) thanobot and hologram of her dead father for her 40th birthday (Bassett, 2022, p. 156). Whether these post-mortem chatbots alter our conception of if and when the self dies, remains an open question. We may also have to rethink whether the agentic I self really dies with the body. And this brings us back to William James’ discussion of the spiritual self (see Cooper, 1992). But that is another story for another day.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Russell Belk

Russell Belk is Royal Society of Canada Fellow, York University Distinguished Research Professor, and Kraft Foods Canada Chair in Marketing in the Schulich School of Business. He studies extended self, possession meanings, collecting, gift-giving, situational effects, sharing, digital consumption, and materialism, and is primarily qualitative, conceptual, visual, and cultural.