Friendship, markets, and companionate robots for children

Abstract

The aim of this article is to examine how markets enable companionship to be disconnected from the concept of friendship thus enabling an illusion of companionship without the demands of friendship. As friendship is a crucial early relationship for children, this is particularly germane to the world of education. It recognizes the previous lack of philosophical attention to the idea of companionship—a key factor in friendship—and that this omission contributes to a lack of clarity on a variety of issues. Starting with a brief outline of companion friendship, the article examines the idea of the ‘intimate work’ of friendship within the market domain by considering three illustrative examples: firstly, rent-a-friend; secondly paid companionship; ending with companionate robots for children. It then concludes by contending that this is an important issue for children and their development and thus for education.

Introduction

Research increasingly recognizes friendship as one of the most important relationships in a person’s life (Damon 1977; Holder and Coleman 2015). Now, a rapidly growing literature in philosophy has come to debate the problematic relationship between friendship and the commodifying tendencies of markets (Raz 1986; Badhwar 2008; Taylor 2019 amongst many others). As Sandel (2012) claims, we seem to live in an age where almost anything and everything is up for sale … for the right price. One of the key claims to come out of this critique is that markets not only allocate goods according to ability to pay, but also promote particular attitudes to those goods (Kennett and Matthews 2007; Badhwar 2008; Sandel 2012; Elder 2017). Philosophers would almost go as far as to argue that the values of the market are incompatible with those of personal relationships and that attempts to meld the two together will inevitably corrupt or negate valued commitments (Sandel 2012). Although this literature has often been driven by whether or not friendship can be bought or sold and the ethical issues involved in such trade, it has not as yet been raised within the world of children and connectedly, in the educational context on what (if anything) schools should be doing.

Before turning to the main content of the article, it will be helpful to preface it with a brief outline of the contextual features motivating the need for this venture. When it comes to children, a wealth of literature considers early relationships as being more than just social skills—and positions these friendships as an emotional engagement evolving over time characterized by reciprocity, companionship, and affection (Dunn 2004). Within much of this literature, companionship is considered to be one of the most basic dimensions of friendship, referencing the shared activities we engage in with our friends (Healy 2011, 2017). Needless to say, schools play a crucial role as a setting for childhood friendship: children are often in each other’s company for six or more hours every weekday, giving considerable scope and opportunity to engage in the shared activities that enable us to get to know others and to pick out those we like best (Healy 2017, 2019).

Whilst the philosophical attention given to friendship has produced numerous considered tomes of research (for example: Nehamas 2010; Helm 2015; Thomas 2013), the same does not hold for companions or companionship—one of the key features that people look for in friendship. Given the close connection between the two concepts, this omission is perhaps even more puzzling as philosophers frequently refer to companionship in relation to friendship—but do so in such a way as to indicate that the concept is perhaps unworthy of study in its own right, has little to offer that is philosophically interesting or is a substitute word for friendship itself. This is, I want to suggest, a mistake that unhelpfully contributes to a lack of clarity on a variety of issues, some of which this paper will draw out, which are particularly pertinent to education. My intention is not to revisit the arguments on markets and friendships per se, but to undertake a careful analysis of the foundational concept of companionship that is in need of greater nuanced understanding. The position to be argued in this article is simply this—there are two different understandings of companionship which have become mixed up: one aligns with the values we have come to expect in personal friendship; the other aligns with the values more commonly found through commodification. Distinguishing between the two may prove helpful.

It may be useful in this introduction to lay out some of the lines of thought that will be discussed in more detail. Central to my argument is that markets enable companionship to be disconnected from the concept of friendship thus enabling the illusion of companionship without the demands of friendship. The starting point which I am articulating requires us to have some idea of what friendship is before we can consider what it is not so I first discuss a conception of friendship and the existing usage of ‘companion friends’ in philosophy. After this, I examine the idea of the ‘intimate work’ of friendship within the market domain by considering three illustrative examples each of which demonstrates the loosening of the relationship between friendship and companionship: firstly, rent-a-friend (whereby one can pay for the ‘friendship experience’); secondly paid companions (an historical example of buying ‘friendship’); ending with companionate robots for children. I then indicate why this is an important issue for children and their development and thus for education.

The importance of friendship

Few things matter to young children at school as much having a friend to play with (Ahn 2011). As parents of young children will attest, few things upset their child as much as falling out with friends and all of its ensuing distress. At the other end of the age range, the death of friends is credited as being particularly traumatic for the elderly, with research now linking it with having a variety of adverse impacts on their health and life expectancy (for example Ansari and Kang 2019). However, studies show that friendships have particular importance for children as through this relationship they learn how the social world functions and how they need to conform to social expectations in order to be accepted (Dunn 2004).

Friendship in modern times has become a somewhat ‘elastic’ term used across multiple disciplines for numerous types of relationships (Cocking and Kennett 1998; Pahl 2000). It is believed to be of such central importance in our social lives that its moral value is often accepted without doubt (Blum 1980). Despite long being accorded a special place in the depiction of the good life as one of the highest of virtues (Aristotle NE; Cooper 1977; Nehamas 2010), modern philosophy has not always embraced the topic of friendship with anything like the acclaim for the study of governments, politics, or animal rights. Nevertheless, philosophical activity over the past two decades, crossing many areas from internet relationships (Cocking and Matthews 2000; Kristjánsson 2019) to robot friendship (Danaher 2019) to friendship and politics (Deneen 2001) has drawn in some way on the writings of Aristotle (NE). This is in itself unsurprising in that philosophical definitions of friendship often have their origins in the typology suggested by Aristotle (of instrumental, pleasure, and virtue friendships) involving mutual concern, shared activities, affective feelings, mutually recognized and acknowledged goodwill for each other (Aristotle NE). The revival of interest in this model sees the time spent together on joint ventures, the conversations, and shared activities as helping us to examine and adapt to our ‘other self’. Significantly, Aristotle offers a typology of friendship yet the majority of the philosophical research has tended to concentrate on the latter, mostly ignoring the first two categories despite it being clear that for Aristotle these are genuine friendships, albeit not fully complete ones, but lack many elements that result in a deeper, longer-lasting relationship.

Sherman argues that for Aristotle, friendship ‘structures’ the good life (Sherman 1987: 595): the happiness of an individual entails the happiness of others, in other words, friendship is relational. For Aristotle, friendship is one of the greatest of goods, providing opportunities for developing and demonstrating virtuous action. Thus, the ‘good living’ found within friendship is only available to me as a person with friends (Sherman 1993a). This would suggest that a genuine friendship requires both an affective attachment and a mutual engagement in each other’s lives at a deep level. It is this feature that links to my project on companionship: that deep friendship requires companions with whom we share this creation of a jointly pursued life. This weaving together of lives requires spending time together in jointly pursued ventures over a period of time. And it is within this literature that we start our journey in search of companionship in friendship.

‘Companion friends’

So, what would a genuine, full friendship look like?¹ As social animals (Aristotle NE), the person that we are is formed through our relationships with others in such a way that each then becomes co-creator of the other’s virtue by supplying a role model and thus influencing each other’s outlooks and values (Cooper 1980). But to achieve this requires us to be deep-rooted in each other’s lives. The idea here is that these deep friendships have numerous benefits: they offer us a companion in the good times of celebration and flourishing, as well as in the pain and despair of life; someone who walks beside us, even in difficult circumstances, keeping us company; someone who helps us to avoid isolation and loneliness. But it takes on an extra meaning in that it is through ‘keeping company’, this ongoing companionship that we adapt to each other, thus giving a moral and possibly educational aspect to the relationship. But untangling what we mean by this form of companionship is no easy matter and for this, we need to delve a little deeper.

Three philosophers in particular have attempted to give an account of deep friendship whilst avoiding the metaphysical commitments of Aristotle on friendship: Lawrence Thomas, Dean Cocking, and Jeanette Kennett. Thomas, in labelling deep friendship as companion friendship, claims that one of the distinguishing marks of this relationship is that companion friends have a privileged perspective on each other’s lives (Thomas 1987). According to this explanation, this type of friendship requires spending a considerable amount of time observing each other in a wide variety of roles and contexts—literally being in each other’s company—thus enabling the deepening of the perspective. Dean Cocking and Jeanette Kennett (1998), however, argue that it is possible to have this perspective and not be companion friends at all. They take Reiman’s view that it is a particular form of caring that constitutes the intimacy of friendship (Reiman 1976).

At first it might appear that all three writers utilize the concept of ‘companion friends’ as a descriptor for the best type of friendship (Aristotle’s model of virtue friendship)—with less emphasis on the ‘companion’ part. But it serves to remind us what we believe companions do: they spend time together or engage in joint activities of some kind; have a mutual affirming trust; some self-disclosure—all features the authors refer to as distinctive of Aristotelian friendship (Thomas 1987, 2012, 2013; Cocking and Kennett 1998, 2000).

The point to take from this is that not all that falls within the remit of friendship may be part of what it is to be a companion, nevertheless companionship undoubtedly contributes to the elements we have come to associate with Aristotelian friendship. Within philosophy of friendship, companionship is implied to be a chosen relationship based on a strong affective element—love (Thomas 1987; Keller 2000). The close personal attachment that is found in love, is distinctly intimate combining intimacy, commitment, and attachment, constituted by a particular concern for the beloved for their own sake (Helm 2010). Generally speaking, this intimacy of friendship is argued to be shown by the mutual, personal sharing of information, beliefs, or emotions that are not available to be shared with everyone (Thomas 1987, 1993). One does not share everything with everyone thus encouraging participants to feel that they are in a special relationship. But even so, argue Cocking and Kennett, it is not what we share or who we share it with: ‘It is the fact that I choose to share what I value with you’ (Cocking and Kennett 1998: 517) that defines it. Gordon (2014) however, suggests that it is the willingness of both parties to allow the one to shape the identity of the other that best indicates the existence of intimacy in a relationship.

Similarly, ongoing engagement in shared activities also plays a role in friendship; we would find it strange for someone to claim to be friends with another and not spend any time at all in their company nor have any contact with them. This line of thought is reflected in child psychology in that companion friendship is frequently defined as the amount and quality of time children spend together (for example Parker and Asher 1993; Troutman and Fletcher 2010). Many of these accounts require the maintenance of friendship over time, with best friendship and consistent, frequent encounters judged to be major factors in their survival and growth (Parker and Asher 1993; Troutman and Fletcher 2010). For example, stable best friendships lasting over six months are measured as being higher in companionship, whereas friendships in which competition and/or aggression proved present were judged to be lower in companionship (Bukowski et al. 1994; Troutman and Fletcher 2010). What all can agree on is that without any interactions, friendships are likely to dissolve (Troutman and Fletcher 2010).

If we compare this with our common, everyday understanding of companion friendship, this usually requires some of the following: that a companion interacts with us on a regular basis; that we are responsive to each other; that the relationship between the dyad can be one of being social partners. Companionship could fall within the bounds of being an intimate relationship—but it stops short of full friendship; it is part of what we seek in friendship, but not its entirety. For example, some forms of companionship lack the affectionate regard between friends that is considered crucial in friendship (as we will see in following sections), and which fosters a kind of intimacy (Helm 2010). What we can glean from this is that it is only in friendship that companionship takes an affective shape.

The exact nature of this relationship between friendship and love has had much attention in previous research, drawing in some way on the Aristotelian concept of philia (for example Kapur Badhwar 1993; Thomas 1993; Helm 2010). Whilst it is not possible to give a full account of these rigorous explorations here, the overall agreement is that those with a deep friendship experience feelings of profound affection, intimacy, and attachment within an enduring relationship. One of the ways in which deep friendship differs from other forms of love can be in the reciprocal affection between the two: love per se is still love if unreciprocated (hence the concept of unrequited love) but love between friends (in the Aristotelian model) has to be both reciprocated and mutually acknowledged. In other words, a friend is one who loves and is loved in return.

For our purposes, companion friendship involves this form of love (Thomas 1993) and is one of very few interpersonal relationships in which the parties interact both intensely and frequently. There is widespread agreement in the literature that deep friends are thus motivated by this active, caring concern for each other’s happiness, enabling friends to create a ‘shared narrative’ through a life lived with joint activities, interpreting such events through a similar system of jointly accepted/deliberated values (Sherman 1987). Sharing our lives with another enriches us in ways unavailable to us as individuals (see Kapur Badhwar 1993). Such pleasurable interactions, the complete delight in spending time together, helps us to refine both our sense of self and the value of our projects (Sherman 1993b).

But the real strength of this argument is that to be a companion in the Aristotelian sense goes beyond being just physically present, to suggest something more: ‘the friendship of good men is good, being augmented by their companionship; and they are thought to become better too by their activities and by improving each other; for from each other they take the mould of the characteristics they approve’ (Aristotle NE: 1172a). In other words, companionship is a necessary component of the fullest friendships in that it creates this site for the development of the virtues. In other words, in such circumstances, we are not only responsive to the other’s company but we also become in some way ‘patterned’ by each other. Essentially, such companion friendship can be a form of character education foregrounding their moral virtues. Distinguishing the main foci of philosophical endeavours allows us to now turn to where markets meet personal relationships and explore why this is seen by philosophers as so problematic.

Markets and the intimate work of friendship

If our modern understanding of friendship includes that it must be entered into freely,² this alone seems to suggest that monetary issues would nullify any possibility of overlap between markets and friendship. Nevertheless, this seemingly simplistic view has been greatly expanded on and developed. Some philosophers claim unequivocally that the colonizing tendency of market values is not only at odds with or radically opposed to the values of friendship, but that they negatively debase the attitudes needed for the bonds of friendship to survive and flourish (for example Schwarzenbach 1996; Sandel 2013). The essential point of these arguments is that using a market as a way of distributing these goods changes the way we think about them. For example, if someone offered to buy my friendship, it would indicate that they really did not understand what it means to have a friend. But what exactly is wrong with buying friendship? To prevent initial misunderstandings about the way such concepts have been subsumed within empirical educational discourses, necessitates a brief review of the main relevant charges.

Firstly, the concept of ‘intimate work’ has come to be used in the social sciences for cases where there is paid provision of services which entail some form of intimacy in a dyad or range of persons (for example Schoenbaum 2015). This includes roles that might have previously been fulfilled by family members which are now ‘outsourced’ to other non-family members (nannies, maids, daycare workers, hairstylists, etc.). Markets then arise to offer an extraordinary range of possibilities of support or relief—as long as you have wealth to cover it, there will be someone willing to supply the service—even friendship (to be developed later).

Secondly, this phenomenon can be explored in three distinct ways: firstly, and the main focus of this paper, is where companionship can be disconnected from the concept of friendship: we can have the illusion of companionship without the commitments or demands of friendship. Secondly, when money enters into the equation, there is no guarantee that the relationship is authentic (Sandel 2012). We want our friends to genuinely like us and want to be with us; we do not want inauthentic gestures that may be interpreted as genuine, but which prove to be deceptions. This can be particularly damaging for children who may fail to understand such duplicities and find this deeply distressful. Thirdly, current advocates worry that people are changing their ways of interacting, particularly when it comes to difficult or possibly disagreeable issues and would much prefer to communicate disagreeable news electronically/via social media (Gardner and Davis 2013); others prefer to bypass the need to converse or interact with anyone at all unless they absolutely have to—and that this may have implications for friendship itself. To develop this further, in the following subsections, I describe three examples of companionship outside of friendship. In the first subsection, the other becomes an object for a short period of time (they can be picked up and then put down again once the activity ceases); in the second, the other becomes an ‘object’ on a more long-term basis; in the third, the other is quite literally an object. Let us examine these in turn.

Buying friendship: rent-a-friend

The emergence of a company called Rent-a-Friend offers an initial example of where markets and friendship have come to ‘mingle’. Based in New Jersey, USA, Scott Rosenbaum used the prevalence of rent-a-friend companies in Japan as the model for his proposed business (Rosenbaum 2009). Advertising itself as strictly a platonic friendship site, and not a dating site, the website allows the ‘renting’ of local friends around the world for a variety of purposes: for example, a companion to go to an event with; someone to go to the theatre with; someone to hang out with; a local tour guide, etc.

Whilst this is a company via which people are willing to exchange acts of friendship for money, the buyer can be under no illusion that this is a proper friendship: conversations between the two cannot denote the emotional depth we seek in friendship; any affection would have to be false between essentially two strangers. In other words, doing things together is dependent on a monetary transaction. In Aristotelian terms, there is unlikely to be a history of reciprocal interactions between the two (although it is possible that they may find similarities in their lives); they may have general goodwill towards each other but it is probably transient or short-lived. Indeed, the site specifically claims to offer ‘the friendship experience’: arguably all the fun of friendship without the hassle and hard work. It is unclear exactly what experience of friendship is being offered. A counter-argument might well argue that if one hired the same person on each occasion, then perhaps such a build-up of a history of shared interactions might eventually create a companion friendship. But the ‘companion’ is still being paid to do this.

Perhaps a further step would be outsourcing the friendship itself: paying someone to do the friendship activities you really do not want to do (Danaher 2019). Take Sybil and Alfred— Sybil wants to go to the cinema to see some ‘chick-flick’ with Alfred; Alfred wants to stay home and watch the last episode of his favourite programme on TV. Usually, they take it in turns to decide what to do—but Alfred really does not want to go to the cinema and hates chick-flicks. So, he pays someone else to go with Sybil: he stays home to watch TV and Sybil has company for the film.

My riposte to the above would be that this arrangement misses out on part of what it was that Sybil sought: she wanted to go with him. The activity is not a shared or joint one (in Aristotelian terms)—for it is the time spent together on such joint endeavours that is symptomatic of deep friendship because of the implications for character education. In essence, the intimacy we seek in deep friendships requires the feeling of experiencing some things together (Cohen 1999). Whilst you certainly could hire someone to keep you company and perform friend-like tasks, the relationship is somehow diminished by not being an authentic expression of friendship. Broadly, it encourages the deliberate screening out of information, personality traits, or values that might have curtailed the very existence of the relationship or prove disagreeable to the client. In short, the hired friend has turned themselves into an object, selling their company, malleable to the whims and desires of the payer. In such circumstances, they have no reason to be truthful or honest about who they really are, and every reason to lie and deceive so as to please the patron. Any feedback or any agreeableness might then prove inauthentic (and thus untrustworthy) based purely on the need to have their time compensated for financially or to get a ‘good report’—important for future engagement. Such an inauthentic friendship could only have self-interest at heart and is unlikely to result in the ‘patterning’ spoken of earlier. But the buying of emotional relationships such as friendship (as in Rent-a-Friend) also represents a different kind of economic consumption: a shift in the way in which we think about personal relatedness which leads us into difficulties. So in the next exemplar, let us look a little more closely at the concept itself and move from short-term to long-term companionship.

Paid companions

The term ‘companion’ is often used with multiple meanings but at a basic level, a companion is someone who keeps company with another (Walker 2009). The term originated around the sixth century from the Latin com panis meaning ‘with bread’—in other words, those who share bread with us (Boisvert and Boisvert 1997). Around 1300, a further definition arose from the Old French, expanding the meaning to ‘a mate, friend or partner’.³ Both of these can still be traced in modern understandings of critical companionship found in nursing and other care domains, where this form of companionship is defined as requiring ‘a high degree of inclusion, equality, transparency and authenticity’ (Greggans and Conlon 2009: 113).

Historically, the term companion embraced a paid employment that was very different from that of companion friendship. There is surprisingly little written about the history of paid companionship but the exemplars are often interchangeable with that of ‘lady’s companion’. In the 18th and 19th centuries, the opportunities for employment for middle-class women with little financial means were limited. The paid companion role was often fulfilled by such ‘redundant’ women, whether single or widowed and was one of the few respectable ones available that did not entail teaching or being a governess: ‘Paid female companions were the hired friends of other women … they were expected to provide their mistresses with company and entertainment in addition to serving as a confidant and chaperone’ (Hoffer 2011: 107). Key to the role was that the paid companion was not seen as being at the same social/financial level as the mistress, yet was more than just a servant. The indistinctness of this according to Hoffer (2011), blurred the role of ‘upstairs’ and ‘downstairs’: they may have had the run of ‘upstairs’ but there was no question of them eating or living side by side with the master and mistress.

Paid companions had as wide a range of duties as there were examples of persons holding the role: to keep company, to provide pleasure, amusement, security, and to serve as a chaperone when needed by her mistress; she had to be of ‘good breeding’ to be capable of providing the company to a lady and similarly to have particular accomplishments—she would read to her mistress or play music, run errands, and play the part of a confidante (Hoffer 2011). These considerations meant the paid companion was completely dependent on the continuing goodwill of her mistress, with no guarantee of support in her own frail health or age-related incapacities: in that sense, she was simply a companion without friendship. So we now have the outer boundary for our concept where companionship becomes divorced entirely from friendship.

Children and companionate robots

So far, I have considered rent-a-friend and paid companionship—but why should these be an issue for education? And it is here that we start to draw our previous arguments together to use as a springboard into the world of children, and thus education, by focussing on the use of companionate robots in schools and considering the type of companionship promoted here. For those unfamiliar with social companion robotics, a few words to set the scene are in order. There is a strong likelihood that long term, in the coming years, computer companions will be more or less a certainty in our world for young and elderly alike. More relevant for present purposes, this means that those who are now children will inevitably have to deal with issues arising from this form of artificial intelligence in their everyday lives, and the lives of their own children. Nevertheless, there is rarely much thought given by schools to the wider issues involved before introducing such objects into the classroom: how should children be introduced to companionate robots? How should children think of their robotic toys? How should they (the robots) be treated by children? What sort of companionship is being promoted?

A recent surge of literature on robotics and their many uses in both social and medical care has centred around whether or not it is possible to befriend a robot. Indeed the ethical and moral standing of robots in law has led to some unusual and unforeseen examples: Saudi Arabia, for example, gave a robot (Sophia) citizenship in 2017.⁴ For these reasons, The Committee of Legal Affairs for the European Court of Human Rights has called on the Commission to take measures to control the growth of robots, particularly companionate robots, since 2017 (European Court of Human Rights 1999–2014; De Swarte et al. 2019). I want to sidestep this argument (in the main) and start from the premise that children are increasingly adept at using computers and computerized toys, the development of which has moved on from simply creating machines that work independently, to robotics integrated with aspects of everyday life. Children increasingly grow up in a world surrounded by robots, particularly at school: the teaching of science, technology, engineering and mathematics (STEM) and robotics on the curriculum; the use of robots as mentors for children with autism (Nao); mediators for children who cannot go to school (VGo for remote students); robotics as peer tutors/learners. As such, these robots bestride our educational worlds (as learning tools) and our home lives (as edutainment/pleasure machines).⁵

It is important to note that the current title given to these objects is usually ‘companionate robots’: companionate, as in companion-like: it is not part of their current capability to replace human companionship. This changing use of the term ‘companion’ adds an important element to our understanding of the shifting nature of the role. Take, for example, suggestions that we refer to pet animals as ‘companion animals’ (Bulsara et al. 2007): these go beyond calls to treat animals well, to relabel and shift the vocabulary used to indicate that they are more than objects or property. This enables us to better understand the role that some animals play in our lives and explain the deep grief that we frequently experience on their loss (Bulsara et al. 2007). In other words, keeping company, being a companion, is not necessarily a role solely played by humans. And this may have implications for how we might come to think about companionate robots further down the line.

Companionate robot toys are aimed at fulfilling particular psychological needs such as communication, interaction, and companionship. In brief, their purpose is to provide opportunities whereby these needs can be met by facilitating human–robot interactions. These robots are (in the main) human-designed objects and can be switched off and put away when play has finished. But in all of this we must remember that robots for children are not entirely new. Children first started to encounter what we now call companionate robotic toys in the 1970s/1980s: K9 (in Dr Who), Furbies, Tamagotchis, My Real Babies, Baby Alive, KittyCat, etc. Many reading this article will have seen or played with one of these, often seeing them as a halfway step between the ‘real thing’ and a static toy—so it is not surprising that young children may get confused over their status. Indeed, the more modern ones are deliberately designed to look harmless, unthreatening, and to have appealing features so that human–robot interactions go as smoothly as possible, encouraging this object/subject dichotomy. But by making robots look and act ‘human’ they are indirectly also influencing human behaviour towards them.

Attributing human characteristics to inanimate objects or non-humans is often a way of trying to make sense of our environment, often ascribing complex emotions and social interactions, allowing us to make mental representations of social intentions particularly when controlled through voice commands (Kanda et al. 2008; Montalvo 2017). Children have always anthropomorphized the toys in their possession. Research seems to suggest that the tendency to anthropomorphize objects, including robotic toys, starts early in childhood development (Bumby and Dautenhahn 1999). Yet the dolls or teddies of the past were unchanging, passive objects compared to the technological expectations of modern toys centred round movement and interaction. But in anthropomorphizing the robot, we also impute similar human-like motivations and reasoning structures like our own, thus we become more likely to see the other as a moral being (Dautenhahn et al. 2005; Van Oost and Reed 2011). And this is where schools need to tread very carefully. When added to their increased mobility over other toys and anthropomorphism, this can create the illusion of agency, and the possibility for inferring possible mental states.

Perceptive readers will immediately chime in ‘but these do not purport to be friendships, but are merely “loose relationships”—and asymmetrical ones at that’. And that would be correct: the examples demonstrate a lack of balance between the parties: an unequal distribution of power, or control, or in the level of commitment, or dominance. The point I want to bring out here is that symmetry is symptomatic of companion friendship, whereas asymmetry is much more symptomatic of the companionship. However, the symmetry (or asymmetry) of a relationship is not always easily characterized nor directly open to examination, nevertheless I will try to unpick further what these particular examples add to this quest.

Let us start with rent-a-friend: a hired friend, according to Michael Sandel, is not the same as ‘the real thing’ (Sandel 2013). To turn a human relationship into a transaction, he claims, degrades it and ‘crowds out’ non-market values. In other words, markets change the character of the goods. Yet it is abundantly clear that the majority of buyers are not expecting to have purchased a ‘real friend’ so such examples do not imply that this is how we gain friends. Nevertheless the example signals something interesting about how we have come to outsource the intimate labour associated with friendship in that we are now willing to buy in the activities previously found in companion friendship (someone to hang out with or go to the cinema with), and bypass the process of actually making friends, compromising with friends over which activities we choose to do together, or working towards creating the shared life of friendship that Aristotle spoke of (Aristotle NE). We are treating friendship as something that is, in some way, disposable.

One of the distinguishing features of Aristotelian friendship is a form of equality between both partners: quite simply, neither should be in a position of complete authority or power over the other (Thomas 2013). There will be some elements of difference between friends as absolute equality is rare—but the overall picture is that a certain level of equality is needed to encourage the development of trust and mutual disclosure. The key point here is that where there is a large gap between persons, there can be consequences for how the relationship survives or flourishes (Healy 2011). But when companionship is divorced from conceptions of friendship a very different picture emerges. For example, the paid companion is in a deeply asymmetrical relationship whereby power is exercised by the powerful over those with fewer social or financial resources.⁶ One has a privileged perspective to be accommodated; the other is powerless—and ultimately disposable. Whilst it is no doubt true that this form of paid companionship rarely exists in modern times, it does raise some ‘grey’ areas: does buying companionship differ from paying for a babysitter or a dog sitter? What about the nanny or the carer for the elderly or disabled? To some extent, they are still dependent on the whims/goodwill of the payee and are more akin to services rendered for a fee.

Companion(ate) robots move the conversation even further down the road of commodification. Here, the companion has literally become the object. But can robots ever be companions? At a basic level, they might fit the mould of ‘paid companion’ or a rented friend experience in that they ‘keep company’ (defined loosely) and have been bought or supplied by someone. But it marks a move from ‘ownership’ of an object to a ‘relationship’ with an object particularly taken with the tendency of young children to anthropomorphize. However even here, it requires more than the behaviour (acting as a companion) to embrace a type of psychological model of emotional bonding as an embodied agent (hence the description of companion-like).

The choice of entering into a friendship of one’s own volition, without force or payment, seems to be a necessary fundamental feature of the Aristotelian model. In other words, there is something wrong about attempts to buy your friends. Some might try to argue that very young children often do precisely this: they think nothing of trying to bribe someone with a biscuit or similar gift to be their ‘best friend’ or to play with them in the absence of other companions. In such situations, it is not the friendship itself that is being bought/sold, but the opportunity to become friends by spending time together. At best, the lack of authenticity in these relationships could only simulate the emotional responsiveness associated with friendship (false friendship); at worst, the imitation and simulation of a decidedly human relationship may cause one to believe in the existence of a non-existing relationship (mistaken/false friendship), with all the attending one-sided features we associate with friendship proper. And it is this factor that continues to gnaw on the bond between friendship and companionship.

Conclusion

Companionship, understood as someone to keep us company, is usually one of the things sought in friendship—both for adults and children. Yet when friendship is subject to market values, it is not just the attitudes that may change—the nature of companionship radically alters. For this reason, any viable conception of companionship has not only to be able to face these differences in friendship, but also to challenge us to think about wider social implications brought about by new developments in social living. I do not pretend to have given a full account of companionship—to this end, I have concentrated on ‘companion friends’ (where companionship is a component of friendship) and given three examples of companionship where the financial aspect disconnects it from friendship altogether. In the first, the relationship is short term and functionary; in the second, the companionship is longer term but lacking key features of deep friendship; in the third, the relationship is with an object. In this short section, I want to draw these arguments together and show more explicitly why this is such an important topic for education and make suggestions as to a way forward.

Schools can be very efficient sites for developing and maintaining friendship; the physical aspect of giving a time and space in which to mingle and meet prospective friends and engage in shared activities opens opportunities.⁷ Moreover, most schools tend to have some form of social or moral teaching somewhere on the curriculum that reinforces this; that is, teaching about different types of relationships. For example, a policy document in England (DfE 2019) to promote the positive principles about relationships both on and offline undoubtedly recognizes the importance of friendship for children; this has the potential to be beneficial in helping children to navigate their way into social life.

There are no simple answers to the challenge of social development in schools that will fit all circumstances. However, a first step might be to ensure that children know that positive relationships can go beyond friendship: that there are a wide variety of ways in which people might acquire ‘company’ and for a variety of purposes (both short and long term). Children will inevitably have had some experience of the momentary lack or loss of friendship (for example, when their friend is absent through illness) and the desire for a playmate entices them to widen their social sphere. This then offers us a way in to discussions about companionship—including companionate ‘objects’. Most young children will have had experience of a pet or favourite toy that they turn to for play purposes and/or will have seen some form of robot or android either on television or in numerous films aimed at children.⁸ Despite the frequency with which robots can be found in school settings, we must bear in mind that interacting with objects such as robots on a regular basis is a very different thing to learning about them on the curriculum—all of which could turn out to have very different implications for the moral and social development of children.

Finally, in case you think I am wary of robots or robotics, let me assure you, I am not. Used properly and with thought, I am sure they can be a wonderful addition for educational purposes. But educationalists need to both prepare for, and start to articulate, within educational policy, how and why this should come about. And a good way to start this conversation, particularly with younger children, is around companionship and what it is to keep company. This in itself may spark a more nuanced and imaginative approach to their thinking on friendship, companionship, and various forms of ‘illusory companionship’.

Acknowledgements

Many thanks to Patricia White (UCL IOE), Mary Richardson (UCL IOE) and Lewis Stockwell (University of Hertfordshire) for their kind readings and comments on earlier versions.

References

Footnotes