1 Introduction

In recent discussion, advocates of what has come to be known as ‘relational autonomy’ have argued that assessments of people’s autonomy require evaluation of their social relations: deep, persistent and nourishing relationships are foundational to people’s abilities to realize and maintain their autonomy [1,2,3]. This is most obviously true for societies’ more vulnerable members. For example, children’s relationships to parents, teachers and friends are fundamental to their becoming adults capable of making choices, giving shape to, and taking control of, their lives. But, it is equally true for adults, whose autonomy remains constituted by their roles and relationships with peers, intimates, colleagues and fellow citizens. On this view, autonomy is not accomplished through independence from others, but rather by pursuing an autonomy-supporting interdependence with them.

However, in this paper, we examine some of the ways in which certain relationships are best understood not as constitutive of people’s autonomy, but rather as providing a means for them to cope with, and thus in some measure accept, the absence of autonomy.Footnote 1 Specifically, we look at the ways in which people who work, even in the teeth of deeply heteronomous labour, can nevertheless develop relationships with co-workers that are not meaningful because they provide workers with any more control over their lives, but because they allow them to endure the heteronomy that characterizes, potentially, large parts of those lives: The benefits of some workplace relationships – which we define as the interactions, connections and associations people develop with people at work, whether co-workers or the recipients of one’s labour, and which may or may not develop into deeper affiliations outside of work – act, as we put it, as a form of analgesia.

Once this analgesic function is taken into consideration, it is possible to identify a potential tension between, on the one hand, people’s reasonable desires for increased autonomy at work and, on the other, their reasonable desires to sustain meaningful relationships at work, even as such meaningfulness does not relate to autonomy. When navigating this tension, workers who fail to develop their workplace relationships as instruments of potential power, to be wielded during struggles over working conditions, and when they settle for workplace relationships as a form of analgesia, then they succumb to heteronomy that might otherwise be resisted. There are undoubted risks associated with politicising workplace relationships in this way – which we discuss in Sect. 5 – but where workers are interested in advancing their autonomy, there are also, potentially, considerable benefits.

Ultimately, workplaces are sites within which a range of values are possible. In this paper, we assume both that autonomy is an important good within this range, and that workplaces offer sites where people’s autonomy can be enhanced or denied. Of course, there is more to meaningful work than its being autonomous or autonomy-enhancing. Indeed, it is part of our argument that workplaces can be enjoyable, even where such enjoyment has nothing to do with autonomy: workplaces provide opportunities for social contribution, camaraderie, structure, and purpose to people’s lives, even as they fail to promote their autonomy. While we examine the ways in which work relates to autonomy, and particularly how relationships at work relate to autonomy, our argument in no way depends on autonomy being the only, or even the most important, value that people achieve through their work.Footnote 2

In addition, we restrict our discussion of work to paid employment: We do not consider work that goes on within the household. Our bracketing of this kind of work, in fact, has little to do with the absence of direct payment for household labour. Rather, it is because unpaid domestic workers usually suffer from a dearth of workplace relationships, and the various values, such as belonging, friendship, emotional support, these generate. As a result, the techniques of struggle we describe as propitious for advancing autonomy in the workplace seem less relevant in the case of unpaid domestic labour, and indeed any workplace where typical collegial relationships are few and far between. Moreover, the absence of employers, contracts and the like, might similarly challenge the idea that politicising relationships in homes, in ways appropriate to other kinds of workplace, could offer an appropriate route for securing more autonomy. Nevertheless, given the crucial role domestic labour performs within the economy, and given that parental and household duties can lead to significant sociability between domestic workers operating across different households – support groups for parents, for example – the same tension between, on the one hand, using those relationships therapeutically and, on the other hand, politicising them as a source of power committed to demands for more autonomy, might well still exist. While this is not an implication of our model we investigate in any detail, it is nevertheless one we regard as compatible with it.

The paper proceeds as follows: In Sect. 1, we consider the ways in which relationships help people take control and give shape to their lives, focussing specifically on the ways in which work relationships can promote people’s autonomy. In Sect. 2 we focus on the kinds of heteronomy suffered at work, before identifying and describing, in Sect. 3, the therapeutic functions work-relationships can perform, and the ways relationships thus lessen the negative experience of heteronomy experienced at work, whilst not thereby reducing that heteronomy. In Sect. 4, we identify four ways in which workers’ autonomy can be improved before, in Sect. 5, filling a gap in the literature that neglects consideration of the means by which these improvements can be achieved. Here, we use various case studies of rank-and-file worker activity to evaluate the ways in which workplace relationships can be politicized in struggles for autonomy. However, even as solidarity, and the militant worker activity it nourishes, are important resources for improving workers’ autonomy, we also recognize the ways in which this potentially puts at risk the therapeutic functions which work performs.

2 Relational Autonomy at Work

Autonomy can be broadly understood as ‘the capacity to be one’s own person, to live one’s life according to reasons and motives that are taken as one’s own and not the product of manipulative or distorting external forces’ [1]. Understood negatively, autonomy involves denying other people the right to control the shape of our lives: If our movements and actions are under the control of another, we cannot act autonomously. If our decisions and thought processes are manipulated and distorted by other agents, we are also denied autonomy, even if we do not recognize it. If we are to be autonomous, there are thus decisions regarding our preferences and actions over which we must be authoritative.

But becoming authoritative over our lives is not something we learn how to do on our own. As Jennifer Nedelsky writes, ‘If we ask ourselves what actually enables people to be autonomous, the answer is not isolation, but relationships—with parents, teachers, friends, loved ones’ [2]. We learn speech and take on values from our parents at an early age; we observe the way others act and try to understand their reasons for acting, reasons which can then often be applied to our own lives; we form opinions and values through engaging with our friends, families and through reading things written by others. Autonomy, as Marina Oshana puts it, ‘is a condition of persons constituted, in large part, by the external, social relations people find themselves in’ [3]. People’s autonomy is developed within the constitutive dependencies that fill out their lives, rather than their trying to seek independence from them.

Relations, at all levels, are not just the ‘conditions’ under which a freestanding self emerges. It is the very nature of human selves to be in interaction with others. In important ways, people – their identities, personalities and intentions – do not exist apart from such interactions and the relations they underpin. To be sure, people can withdraw from most relations, and can decide to discontinue one relation in order to develop others. But, if these rejected relations have been sufficiently important, even these will remain inescapably a part of who we are. On this view then, relations, ‘including those with collectives of all sorts, become not just potential threats to autonomy but its source’ [2]. As Kimberley Brownlee puts it in her discussion of the centrality of sociality to human well-being: given that ‘much of our sense of purpose is rooted in our social roles… [that] much of the value in our choices comes from our connections with people who share in our deliberations, witness our efforts, and are invested in the results … [and that] we suffer greatly when we are denied avenues to have connections, pursue joint projects, and hold meaningful social roles… any credible account of autonomy must fully acknowledge our sociality’ [5]. Footnote 3

Homes, neighbourhoods and classrooms are all sites within which relationships develop that can either improve or thwart people’s autonomy. Another site, and one that has received considerable attention, are people’s workplaces [6,7,8]. Of course, we arrive at workplaces as already partially formed adults, as people already in important ways formed by our interactions in those other sites. Nevertheless, although it is not always an important site for earlier constitutive relationships, workplaces often, indeed usually, remain important sites for developing the relationships and capacities which nourish people’s autonomy.

The ways in which work interacts with autonomy are complex. First, if people are to feed, clothe and shelter themselves to minimally decent standards, work is unavoidable for most people. When confronting such compulsion, people are unable to shape their lives entirely as they see fit: what one should do in terms of work remains an inescapable question to which they have to provide an answer of sorts [9]. But, second, people can regard work as something they should do in order to reciprocate for benefits received from others’ labour, or to contribute their share of effort toward a general scheme of cooperation that produces various goods they enjoy. That is, people can think that taking from a pool of cooperatively produced goods without making a fair contribution in return shows a lack of respect to those who did, and treats contributing others in an ‘offensively instrumental way’ [10]. On this view, refusals to contribute to the cooperative scheme by working is a failure to act morally. If we are motivated by a sense of duty, and thus for reasons that are our own, then, economic compulsion notwithstanding, working can still be autonomously performed. It is possible to be simultaneously compelled by both a desire to reciprocate other’s labour and the need to earn a wage.

There are four other ways work supports autonomy. First, although people must work, they exercise autonomy – more or less limited – when they choose what work they will do. This we call choice autonomy. Second, there are the meaningful options that income earned at work provides that could not otherwise be accessed. People use their wages to pursue various interests outside of work, as well as enabling them to avoid relying too heavily on others, and all that implies. We call this income autonomy. Third, work can be non-instrumentally linked to autonomy, that is to say, paid work can be constitutive of autonomy by securing us the self-respect that we gain from, amongst other things, the sense of contribution our work provides. Contributing to other’s welfare helps people regard their work as one dimension of a ‘conception of the good… (that) is worth carrying out’ [11]. In this way work contributes to what we call self-respect autonomy [12]. Fourth, there are the opportunities work provides for people to use their ‘natural capacities in an interesting fashion’ [11]. This might involve ‘conceiving and carrying out projects, making decisions, exercising judgment, taking responsibility for decisions, forming goals, planning methods by which to accomplish goals, adjusting goals and methods in light of experience, and other aspects of autonomous agency’ [6, 13]. We call this capacity autonomy.

Workplace relationships play different roles across the four different dimensions of workplace-autonomy. For example, people might choose work, even work which pays less, because it gives them more time to develop relationships with people they care about outside of work. Balancing priorities in this way, making these choices, expresses people’s ideas of a good or worthwhile life. Or, people might choose a certain kind of work precisely because they believe it is rich in the kinds of relationships they value. Conversely, people might abhor certain kinds of work because they instantiate relations of dominance and subordination, rather than egalitarian relations, between workers. Moreover, people might take up, consciously or otherwise, work that exposes them to people from a range of unfamiliar backgrounds, which then becomes a reason to value that kind of workplace. In addition to being valued for its own sake, exposure to difference within the workplace – itself a consequence of antidiscrimination laws introduced to improve the autonomy of certain groups – can also generate valuable epistemic and trust effects for the community as a whole, which might themselves be understood to expand autonomy [14]. These relationships at work are not necessarily limited to co-workers: In some professions, the relationships workers develop with the recipient of their services take on as much, if not more, value than relationships with co-workers. Teachers, nurses and certain kinds of lawyers might regard the relationships they develop with students, patients and clients as the primary reasons for valuing their jobs.Footnote 4

For many people, relationships developed at work can also help satisfy the human ‘need to belong’ [15]. This general need can be broken down into a narrower set of needs pertaining to work. For example, as part of this need to belong we have a need to be accepted into social groups and interpersonal relationships [15]. Workplaces provide one site for such acceptance. We also have the need to make social contributions to others’ lives, to not only receive but to offer love and care to others [5]. When people work, this sense of contribution can thus be understood both in terms of the wider societal needs to which our labours respond, and to the more fine-grained contributions we make as we work alongside co-workers and colleagues. Though perhaps rare, there is, as well, the sense of belonging workers might feel as they collectively influence the decisions and judgements that organize their work, generating a sense of ownership over, and belonging to, the institutions which employ them.

Taken together, these relationships can thus fashion a sense of belonging for workers, both at work and in their wider communities [13]. Under the heading of ‘affiliation’, Martha Nussbaum registers the importance of this particular human need to belong by including it within her list of the capacities considered of central importance to human life [12]. Indeed, for Nussbaum, satisfying these affiliative needs is especially important, playing as they do an ‘architectonic role’ in people’s overall lives: where people are unable to affiliate with others, they become less capable of developing those other capacities on Nussbaum’s list [12]. So, even as their work activity might do little to promote workers’ autonomy, when they are able to feel some sense of belonging, in just one of the ways described above, this can satisfy, to some extent, their affiliative needs. On Nussbaum’s view then, belonging is a prerequisite for developing the cognitive and social abilities that are necessary – if not sufficient – for a minimally decent and minimally autonomous life. In Sect. 3, we come back to the importance of belonging, both for the achievement of autonomy and in terms of the other functions it serves. For now, however, we turn to the problem of workplace heteronomy.

3 Workplace Heteronomy

Some workers might score well across all four of those dimensions of autonomy. That is, they might enjoy considerable choice in occupation, receive large incomes capable of funding all manner of extracurricular activities, perform work that is recognized as contributing to wider society, and exercise their capacities while doing it. To be sure, these different kinds of autonomy can also conflict with one another: workers might benefit from improved income autonomy, i.e. increased wages, if they are willing to submit themselves to onerous work which offers limited opportunities for capacity autonomy. For others, winning promotion may mean increases in income-autonomy and self-respect-autonomy, even as increases in their working hours mean more time spent in an environment where, the promotion notwithstanding, they enjoy only limited control over their work. Additional time at work means less time engaging in activities autonomously chosen outside of work, including spending time with loved ones.

For many, perhaps most people, the primary autonomy-related reason to work are the opportunities facilitated by income earned from employment. In much work, there are limited opportunities for what we described above as capacity autonomy. Take, for example, Taylorism, that ‘scientific management practice’ of the labour process in which workers’ bodily movements are minutely timed and controlled to ensure maximum productive efficiency. These practices have been given a new lease of life in Amazon ‘fulfilment centres’ and call-centres, where management impose algorithms on workers’ movements and actions, to the point where even time spent on bathroom breaks is carefully timed and monitored [16]. The ‘dictators’ who are charged with overseeing these ‘private governments’ act in ways that remove even the idea that work activity might be meaningful for individuals: Productive efficiency is all [17]. In such cases, workers are asked, for many hours of their (working) lives, to commit themselves to activity where their autonomy is severely limited. Heteronomy 8 hours a day (10 with the commute) is traded for some measure of autonomy during some other fraction of workers’ time. Importantly, as well, the savings made as a result of these efficiencies, which explicitly and intentionally deplete workers’ capacity-autonomy, do not necessarily precipitate increased worker remuneration. It is not a natural fact of increased efficiencies that workers themselves will benefit, as these savings might still accrue primarily, even exclusively, to management and ownership.Footnote 5

This problem of reduced work-related skills autonomy extends beyond the infinitesimal micromanagement of workers’ movements to include any work that lacks or, after a given amount of time, comes to lack, sufficient complexity. ‘Eudaimonistically meaningless work’, as Andrea Veltman calls it, is any work which ‘does not develop or exercise human capabilities, permit independent judgment, integrate conception and execution, or otherwise facilitate expressions of agency’ [13]. For example, the work of IRS accountants in David Foster Wallace’s The Pale King – work which is ‘just tricky enough’ to make the accountants have to think about it – might have started off challenging, but very quickly becomes rote, tedious and, eventually, hellish [18]. It is hard to imagine people regarding this work, however hard they might try, as developing or expressing capacities that have anything to do with their autonomy.

While Taylorist factories and the rote work of an IRS agent might be typical and obvious examples of workplace drudgery, heteronomy can be understood in subjective terms as well. Lawyers’ work that might appear to others as stimulating – because it draws on complex cognitive capacities, involves interesting mental puzzles and arguments, and includes the fulfilling of extensive responsibilities– might to many lawyers themselves, be anything but: the challenges appear as mere nuisances, the capacities put to work unimportant, the profession altogether corrupt and self-serving. Just like workers in more obviously disagreeable jobs, such lawyers can come to regard their work solely in terms of the (autonomy-supporting) income it provides and the expanded opportunities this facilitates outside of the office.Footnote 6

There might also be the fact that lawyers, accountants, factory line workers and various other kinds of worker simply do not regard what they do as contributions. Those efforts which are thought to express workers’ reciprocal contributions to their wider communities are regarded by those making such efforts as useless. Alec Guinness, a respected actor – another job many would consider objectively interesting – averred that he found acting a ‘rather silly profession.’ Here again, it need not be the most obviously menial and tedious jobs that are regarded as superfluous and useless by those who do them. David Graeber, in his discussion of ‘bullshit jobs’, asks: ‘Could there be anything more demoralizing than having to wake up in the morning five out of seven days of one’s adult life to perform a task that one secretly believed did not need to be performed—that was simply a waste of time or resources, or that even made the world worse? Would this not be a terrible psychic wound running across our society?’ [19]. Such an understanding of one’s own work, and the relationships it establishes between members of a supposedly cooperatively-producing community, is unlikely to support the bases of self-respect, no matter how well it is remunerated. When the opportunities workers confront are thin on the ground, when the work that is available pays badly, offers little opportunity for developing or expressing complex skills, and is not regarded as contributory by those who do them, then such work is thoroughly heteronomous.

4 Relational Salves in the Workplace

One dimension of workplace relationships that has not received much sustained attention is the way workplace relationships can help workers endure such heteronomy – that is, the relationships people develop at work can act as balm, salve and catharsis, performing what we might call therapeutic functions.Footnote 7 Work environments create fertile ground for camaraderie, community and social sites where experiences can be shared. Employees unhappy with their working conditions can collectively blame, ridicule and satirize management; they can gripe about clients, about the boredom induced by their work, or else about the perceived ‘bullshit-ness’ of the labour. Or, more basically, workers can simply pass the time in each other’s company, creating small islands of sociability amidst the deserts of despised labour.

As we have already suggested above, developing friendships at work, and the sense of belonging this facilitates, even when the work itself does not offer significant opportunities for all aspects of autonomy – low paid, low on recognition, low on skill – can still be understood as facilitating a person’s autonomy through the affiliation it supports. Even in those conversations between workers griping about work, they can develop their understanding and empathy, and provide opportunities to both give and receive support from others. Such skills do not have to develop as an aspect of work itself. In the teeth of even the worst kinds of work, workers’ relationships to one another can nevertheless foster some degree of autonomy. For example, the coal pits of the Welsh valleys were sites of not only of toilsome labour performed under the tyranny of pit-bosses, but were also places where tremendous camaraderie, self-respect and deep community life thrived, both down the mines and in the pit-villages [20]. Even though these types of workers often have little control over whom it is they work alongside, the sociability that exists is often deep and intense. Indeed, the difficulty of the shared work is, conceivably, a spur to especially deep senses of belonging, irrespective of the personal differences that might persist.

Through such friendships developed at work, during break-times and after-hours, often with people with whom we would otherwise have little chance of interacting, workers gain new perspectives and values, develop support networks, find people who can develop our interests and to whose life we can contribute: In other words, nurture relationships that support workers’ autonomy. For example, Terry Jones explains the value of her colleagues working in difficult conditions as a ‘factory girl’ in the 1960s:

The camaraderie and sense of fun was brilliant. There was always someone having a cigarette or their ears pierced in the loos. The language was foul – “pit language” my mother called it. But for the most part, the girls were like family. They taught me the facts of life – about sex, love, relationships, which I’d never learnt at school. And they’d be there for you if you got dumped or had a crisis [21].

Jones’ social needs are thus (partially) answered even in a situation where she otherwise lacked meaningful control over her labour. Even though the factory owners were unable or unwilling to make the work itself more conducive to developing relationships of ‘trust, collaboration (or) respect’ between workers, the workers themselves found the time and space, at work, to achieve those goods to some degree [22]. Jones’ relationships with her colleagues, which were characterised by a degree of equality, but also traversed relations of seniority to some extent, were crucial to Jones’ sense of belonging in the workplace, and to her achieving affiliation with others.

Indeed, to be unemployed is, amongst other things, to be cut off from environments where friendships and other intimate relationships have typically flourished [23].Footnote 8 One of the redeeming features of work then, is that even in working environments where feeling of frustration and helplessness are endemic, even though work is fundamentally compelled, the grind can still be shared with others [24]. If Sisyphus could have a chat and laugh with other people while pushing his boulder up that hill, then maybe Camus’ counselling us to “imagine Sisyphus happy,“ to regard his “struggle itself towards the heights” as “enough to fill a man’s heart,“ would seem more plausible [25]. But such grind-sharing need not be understood as autonomy-supporting. Even if Jones had not developed especially deep relationships with her co-workers – if the ‘sense of fun’ did not develop into an autonomy-enhancing education of sorts – the value of ‘fun’ within the workplace persists, as something that helps employees get through the rote tasks of factory work. Indeed, employees cared less about altering the work itself precisely because of those workplace relationships’ analgesic effect. Heteronomy at work is, in the relationships developed during time before, during, between or after shifts end, being managed and mitigated.

This more ‘sociable’ myth of Sisyphus makes clear that relationships at work can be valued quite apart from whatever autonomy-supporting affiliation they nourish. Let us imagine that the Sisypheans’ bosses gave that group of boulder-pushers a couple of options. Option one: they can move to a shorter hill. However, if the pushers make this choice, they will be unable to communicate with each other. Option two: they can talk to each other, but must move to a higher hill. Whatever decision the pushers make, the calculus used to decide will not refer to the value of improving the autonomy of the work – which is held constant – but only to the perceived value of increased sociability during utterly heteronomous labour.

These pushers are in a situation between the rock of Joseph Raz’s ‘man in the pit’ and the hard place of his ‘hounded woman’. The first is someone who, having fallen into a hole, is unable to climb out or get help, but who has enough food to stay alive. However, his options are severely narrowed, confined ‘to whether to eat now or a little later, whether to sleep now or a little later, whether to scratch his left ear or not.’ The hounded woman shares an island with ‘a fierce carnivorous animal which perpetually hunts for her’. In contrast to the man in the pit, she must draw on ‘her mental stamina, her intellectual ingenuity, her will power and her physical resources’ if she hopes to remain alive: ‘She never has a chance to do or even to think of anything other than how to escape from the beast’. In Raz’s examples, the man in the pit only has ‘trivial options to choose from. His options are all short-term and negligible in their significance and effects.’ The hounded woman has the opposite predicament: ‘All her choices are potentially horrendous in their consequences. If she ever puts a foot wrong, she will be devoured by the beast’ [26]. Neither of these people is even minimally autonomous.

The pushers find themselves in a situation that is the worst of both these worlds. They must, like the woman, draw on their physical resources and stamina to push their boulder up a hill. But it is not effort of a kind that might be described as requiring mental ingenuity or skill, but is brute labour of a basic kind. And, like the man in the pit, they have no options that could be described as anything other than trivial. They decide the speed with which the rock ascends the hill, and maybe the pace they jog back to the foot of it, but there is no other room for manoeuvre. In sum, they will be as bored as the man in the pit, and as exhausted as the hounded woman.

However, just as the socialized myth of Sisyphus creates the possibility for the pushers to find their situation more bearable, adding companions to the man’s pit or the hounded woman’s fleeing, while not increasing their autonomy – since no significant lever of control has been added to the scenario – could still help to make the situation more bearable: the pit dwellers can converse with one another and the hounded woman can have a deep, brain-nourishing nap while her companion stands guard. In other words, the value of companionship, across all three examples – pushers, pit-dwellers, the hounded-couple – is in the way it helps those people endure the absence of autonomy.

Contemporary Sisypheans might understand their work-relationships in ways that are similar in kind, if not degree, to this socialized myth of Sisyphus.Footnote 9 For example, working in call-centres offers little room for autonomy. For example, in Phil Taylor’s study of call centres, one worker commented on the impact of workplace surveillance and the recording of her calls:

It does have an effect as you know that if you don’t do things by the book they could be listening and could pull you up on it, you’ve got to be on your guard all the time, you can’t just be yourself [28].

For the call centre worker, doing her work means sticking to a rigid set of instructions without much room for initiative. Just as the man in the pit chooses when to itch his nose, the call-centre worker gets to choose which part of the assigned script to use. Just as the man in the pit can sit in this corner or that one, call centre workers can decide at what level to pitch their voice when interacting with callers. To be sure, problems might emerge that will require some mental acuity on the part of the call-centre worker, but procedures are usually in place to restrict what workers are able to do, even in terms of smoothing out (only apparently) unforeseen problems.

But, as with Jones in her Dagenham factory, the tedium of the job need not eliminate all opportunities for affiliation and belonging. Presuming that employment in a call-centre allows for sufficient shared time and provides some shared communal space for workers, then they can use breaks for interacting with colleagues. There thus remain possibilities for developing meaningful work relationships even when the work itself is heteronomous.Footnote 10 It is this which goes missing from Schwarz’s account of meaningful work, since she collapses her definition of meaningful into work arrangements that ‘allow all persons to act as autonomous agents while performing their jobs’ [6]. This unnecessarily narrows the meaning of meaningfulness as it relates to work: relationships that help workers stick out their jobs and thus continue to enjoy whatever limited autonomy-enhancing benefits are accrued via wages and other goods, can also be regarded as meaningful, even as the work itself might not. Part of the value of those relationships that help us persist with heteronomous labour is that they offer a form of analgesia, which makes it possible for people to return, reenergized to a degree, to the kind of work which lacks value insofar as autonomy is concerned: Making work bearable is not to make it autonomous, and the hours of boredom-inducing heteronomy (compounded by the dead-time of commutes (self-identifying reference removed)) still squats over workers’ daily realities. Indeed, the relationships developed at work and elsewhere – at home, with friends and intimate and other associates – can still only be developed within the teeth of that non-autonomous working life, in time not allotted to work [27].

Even when workplaces are propitious for friendships and other intimate relations, and even when the affiliations they precipitate serves autonomy in some limited ways, this does not alter the fact that for large parts of a person’s day, she may well be engaging in wholly heteronomous activity. The friendship a person enjoys between 12:05 – and 12:55, might still produce some benefits between the hours of 09:00 and 12:00, a time consisting of only mindless, rote tasks: At 10:30 one worker, Adam, looks down at his watch and says to himself ‘just another hour and a half until lunch with David.’ David is the light at the end of Adam’s tunnel, but the darkness of the tunnel, and the time spent in it, remains unchanged.

5 Increasing Workers’ Autonomy

Many theorists who discuss heteronomous work find it appalling that so many people must spend so much of their time, for so little pay, as ‘greeters, burger-flippers, assemblers and clerks’ in workplaces that are ‘authoritarian and hierarchical’. Such facts on the ground, these theorists argue, are incompatible with justice and demand remedies [29]. What remedies are needed will depend on what it is about the situation that demands improvement. There are four, non-exclusive options: First, work can be better remunerated so that, whatever the content of the work, there are improved opportunities beyond the workplace. This is an improvement in workers’ income autonomy. Second, work-activity itself can be improved so as to become more autonomous, either because the labour process draws on more complex capacities, or else because it becomes a recognized basis of contribution and thus a grounds for self-respect. These improvements might involve increasing workers’ control over the definition and distribution of their work, reducing or even eliminating the workplace as a site of ‘private government’ [17]. These changes improve workers’ capacity autonomy and self-respect autonomy, respectively. Third, society can be organized either in ways that people can more readily sustain minimum standards of living without the income provided by work or else in ways that less work is required for workers to sustain minimum standards of living. This would result in an increase in workers’ choice-autonomy, even as such choices refer to activity outside of the labour market. Fourth, workers can leave to find employment that better supports autonomy in line with options one, two and three above.

We can flesh out these different options by using Veltman’s example of workers at a KFC. In these restaurants, employees’ ‘nearly every move behind the counter and in the kitchen is predetermined’, to such an extent that a person’s labour is reduced to ‘count(ing) to seven as she shakes a skillet’ and then ‘count(ing) to ten as she rolls chicken in batter.’ Work is organized in this way ‘on the assumption that it is best for quality control, and ultimately for company profit, that fast food workers be relieved of the need to think or make judgments about cooking’ [7].

There are a number of things that might happen here to improve these workers’ autonomy. First, work procedures might be changed – including expanding workers’ participation in the decisions that pertain to how those procedures are defined, organized and distributed – to allow for increased in autonomy, specifically in workers’ capacity autonomy. The hope here is to try and rescue the content of the work. How much autonomy will be possible is ultimately constrained by the function the work must serve. This is, after all, a for-profit restaurant, and so certain things will have to remain the same: The chicken will still need rolling in batter, the skillet will still need to be shaken some number of times, there will still need to be the construction of burgers and fries and cokes and the dispensing of sauces. Taken together, these actions will need to be combined in such a way that they ultimately produce a profit, a profit that means this restaurant survives in competition with others. Employees might then take control of this process, without such control improving the content of the work: They define the tasks to some degree, combine the components of a successful order as they see fit, but the tasks remain boring and rote and mindless. Given health and animal-rights concerns, it is also far from clear that workers serving fast-food will or should regard the work as contributory, even if people are willing to pay for it. In any event, even if we concede that the organization of the work might become more autonomous under worker’s collective control, that need not mean such autonomy will be regarded as especially valuable.

Secondly, the employees can push for higher wages and additional benefits to increase the instrumental link between their work and their autonomy, leading to increases in income autonomy. In the US, fast-food workers pushing for rights of unionization and a $15 minimum wage can be understood as pushes for more autonomy, without thereby – at least in the short-term – changing the minute-to-minute content of the work itself. With extra income and better benefits, workers will have more money and security to engage in the things they do care about. They might also, perhaps, be able to move closer to where they work, or buy a car, to reduce the time otherwise spent commuting via underperforming public transport. The work remains dull, and is not considered any more meaningful, but autonomy has increased.

Thirdly, assuming we want a fast-food industry, another option might be to fully automate the work. As a result, there will be no human beings engaged in counting, rolling or shaking, because it is done by machines. In a cashless economy there might not even need to be people at the till – a card is swiped, an order made, sufficiently nimble machines are activated and minutes later the box-fresh order appears. In a situation when 3D-printers are constructing houses, a fully automated fast-food restaurant does not seem beyond the realm of feasibility. Of course, such a change will lead to a massive reduction in work opportunities, which will eliminate masses of paycheques: For the 12 million people employed as fast-food employees around the world, this would devastate their autonomy. Such technological replacement threatens autonomy when there are no means by which the benefits of time saved and efficiency achieved can be effectively distributed to include workers. Therefore, if technological advances across all industries are going to have any positive effects in terms of workers’ autonomy, systems need to be put in place that allow for them to share in these advances.Footnote 11 Andre Gorz’s earlier defence of unconditional basic income was grounded in precisely the expansions in free time – and autonomy – concerted efforts at mass automation would facilitate [30].

Whatever option is chosen, simply positing a demand, or providing a set of imperatives to which institutions are invited to conform, does not help develop the means by which workers can claim that demand, irrespective of whether that demand be for more meaningful work, greater control, more remuneration, or an improved social safety-net separated from a work expectation [8, 31]. Part of the non-ideal circumstances workers confront is the fact that employers and governments are failing to take appropriate action. Making the case for more meaningful work, or well-paid but alienating work over a reduced working-week, or access to sufficient income absent employment, is going to require, in the absence of institutional supports and the political will to develop them, workers own concerted efforts.

One source of power, relevant to generating these concerted efforts, is the relationships between workers, developed both during working hours and outside them, which can be usefully employed as a way of pushing for these changes. If Frederick Douglass is right that ‘power concedes nothing without a demand’, then workers, potentially, need to combine in order to make sure that their demands for more autonomy vis-à-vis their work – in whatever form such a demand takes – are irresistible [32]. However, if we accept the idea that workplace relationships provide important sites and resources for workers striving for more autonomy, then those relationships must aspire to do more work than affiliation and analgesia. If workers care about significantly reducing the heteronomy of their labours, then workplace relationships cannot only provide light at the end of a working day’s tunnel, but must also aim at changing the tunnel itself.

6 Politicising Sociability

As we have seen, sociability at work performs a number of functions vis-à-vis autonomy. First, relationships developed at work can facilitate, even if only in quite limited ways, people’s autonomy, through the affiliation they help workers achieve. Second, workplace relationships can also make heteronomous work more bearable, quite apart from whether or not they make workers more autonomous [33]: Part of the value of such relationships is in how they help workers endure the lack of autonomy experienced during their working lives [7]. However, there is another dimension of workplace sociability, not yet considered, which workers can develop as part of struggles for more autonomous work activity, for increased remuneration, for shorter working days, and/or for enhanced public welfare supports. In what follows, we consider this power as a form of politicized sociability, which exists in possible tension with some of the other functions workplace relationships perform, but which can, when employed effectively, help workers confront employers, and various other agents, who might attempt to reduce – or contain their demands for – autonomy.

How one believes relationships at work – and beyond – can help increase worker autonomy will depend, in part, on how one views economic realities and the strength of certain normative requirements.Footnote 12 More specifically, they will depend, first, on how one believes workers can achieve more autonomy within contemporary labour markets. For those persuaded by more libertarian economic and political visions of the economy, workers will likely have to develop autonomy-supporting strategies primarily as individuals, avoiding the kinds of collective action that, however understandably attractive, will simply interfere with the correct workings of a properly functioning marketplace. However, even on this view, questions regarding how work-based relationships serve or fail to serve workers’ autonomy, remain important. Even for those who do not regard collective solidarity between workers – and the actions thereby supported – as an appropriate use of workplace sociability, workers who care about autonomy might still need to understand work relationships as potentially relevant for it.

On such a view, work-based relationships might be best understood in terms of helping workers respond to changes wrought by capitalism’s inherent drive to ‘creative destruction’. People will need extensive networks of family, friends and neighbours to support them during difficult times, helping them to adjust to the ‘invisible hands’ moving the economy. In the interest of economic adaptability, they should cultivate enough sociability to remain productive, but stay on their guard about getting too attached to any given workplace in case it must shut down, a victim of the ‘emetic’ flushing economies, on this view, must periodically undergo [34]. These relationships can be understood as preserving some limited spaces within which workers can take some control over their lives, however unstable those lives might become.

For those who do not share this view of economic reality, and who might be sympathetic to turning workplace relationships into tools for enhancing workers’ autonomy (along those axes and towards those ends discussed above), the work these relationships can potentially perform is differently demanding. In order to feel capable of confronting employers – and, historically, this has also meant taking on the forces of employers, and sometimes even unions, combined with those of the state – workers require a sense that they possess enough collective power to win such confrontations.

Building this power is difficult. There have been a number of institutional forms that have, historically, helped. For example, workers have combined into work-based interest groups, like unions, to demand more from their employers and/or from the state. There have also been parties of various political stripes committed to the cause of labour. Margaret Kohn describes the early 20th century People’s Houses where workers met for both political reasons and straightforward companionship [35]. There are also the grassroots community groups that operate on issues related to housing and welfare rights. All these have provided places within which workers developed a collective sense of belonging, a shared culture – sometimes an entire ‘proletarian public sphere’, often rooted in a sense of community facilitated by the propinquity of work and neighborhood – and political agency guided by their interests.

Within recent analyses offered by contemporary labour organizers and strategists, where any kind of a working-class political infrastructure is lacking, an important dimension of working-class power building has been the need to more directly politicize relationships already existing in the workplace, i.e. without the mediation provided by the various institutional forms contained within such an infrastructure. In Eric Blanc’s discussion of the ‘red state revolts’, which occurred in the education sectors in Oklahoma, Arizona and West Virginia in 2018, work-relationships developed alongside online conversations between the workforce of different schools. Facebook groups were particularly useful as a starting point because they enabled discussions within spaces beyond the surveillance of employers and superintendents, where issues related to health insurance, school funding and remuneration could be discussed in relative privacy. These conversations were then taken back into the workplace, where colleagues could continue them during breaks, in the hall, or at the end of the day. One teacher described the evolution of these dialogues as follows: “For weeks, there were these constant get-togethers in hallways and in mailrooms, where we’d discuss what the hell was going on. Lots of folks were initially scared, but I kept on repeating that it was worth the risk. People started asking each other: ‘Would you go out if I go out?’ There were some who initially said no, but they changed their minds—eventually” [36]. In this example then, workplace relationships – lubricated by politicizing online interactions – were broadened and nourished into something, explicitly, political and confrontational, which in turn developed into a formal, and formidable, association.

To be clear, there was no overarching ideological or even partisan affiliation that unified these teachers and support staff. Despite the red-baiting engaged in by state forces and right-wing media – especially in Arizona [36]– these teachers were composed of a small number of self-identified socialists – a preponderance of whom ended up organizing the eventual strike – with the majority of participants split between liberal and conservative affiliations, or else none at all. Some were registered Democrats, some Republicans, and some were even Trump supporters. But across this diversity, workplace relationships were developed into, what turned out to be, quite formidable political associations. These relationships, which eventually fed into prolonged state-wide strike action across all three states, were used to extract considerable concessions from state government (pay rises, increased school funding, changes to health insurance) for all public employees in the education sector. Thus, from having politicised already existing workplace sociability, these striking teachers and other school employees (janitors, administrators, etc.) developed a rare form of power, one which contrasted with the usual situation of, as Eric Blanc put it, ‘working-class resignation’ [36]. On this view, the relationships that developed between school employees were not, in this instance, important because they were constitutive of any relational autonomy. Nor were they valuable in helping teachers endure the absence of autonomy. Rather, those relationships served as the forge and expression of a power that, when applied correctly, helped generate gains in terms of better pay and working conditions, i.e. significant improvements in workers’ autonomy.

The function of certain mediating institutions in the history of the working class can be understood, precisely, in the way they managed to effectively combine sociability and political agency. Kohn describes the ways in which the People’s Houses of Europe offered sites of ‘encounter, proximity, concentration, simultaneity and symbolism’, which marshalled workers’ desire for a glass of wine or beer after work, into an identification with socialism, and all that entailed with regards to their own sense of power [35]. However, such a cocktail of politics and sociability is now rarely, if ever, on the cards, and certainly not to that previous level of intensity. Absent these mediating institutions, and the cultures and communities that supported them, understanding and developing workplace relationships as a source of power can be a challenging and risky endeavour.

Marjorie Murphy’s [37] description of the challenges confronted by the American Federation of Teachers (AFT) – one of the few desegregated unions at that time – during negotiations for school improvements, offers an example where precisely these tensions played out amidst the racial politics of 1960’s America. In New York, AFT teachers asked ‘for more control over the curriculum, professional preparation time, paraprofessionals in the classroom, and the power to remove disruptive children’ [37]. Each of these demands would have represented gains in teacher autonomy. After a two-week strike, the United Federation of Teachers (UFT)– the AFT’s largest local – won these demands. However, African American community leaders, inspired in part by black nationalism, reacted against the victory, emphasising a need for more Afro-centric curricula and local community control of schools. They also argued teachers’ demands for powers to remove students and increased salaries were, respectively, racist and selfish [37]. In Newark, during similar disputes, striking teachers found themselves harassed, assaulted and even hospitalized during contract disputes [37]. Amidst their struggles to increase resources for themselves and their students, to improve wages, and to gain paid recognition for the preparatory parts of their job, teachers (Jewish, African American and white) saw friendships dashed, careers ended or curtailed, and previous gains undermined.Footnote 13 Struggles for autonomy, then, are complicated, difficult, and sometimes dangerous. The losses incurred can be reckoned both in terms of reductions in teacher autonomy and as reductions to the therapeutic value those relationships would, absent the waging of those struggles, have continued to serve in those workplaces.

Jane McAlevey’s account of ‘deep organizing’ in Nevada similarly focuses on the relational components that underpin successful fights for worker autonomy [38, 39]. An important part of McAlevey’s approach to organizing is the drawing up of what she calls ‘Power Structure Analyses’ (PSAs), in which workers map out their ‘resources’ both at work and within their local communities [39]. These resources are, specifically, the ‘personal relationships, social networks and knowledge of their community’ that workers could mobilize in any push for workplace gains [38]. On this view of organizing, workers are encouraged to regard their relationships and community networks (including co-workers, friends, fellow congregants in faith communities, and fellow members in tenant associations and neighbourhood groups) both as nodes within an overall power structure and ‘resources (these workers) didn’t even know they had,’ to be used against those in positions of power [38]. In addition, the victories workers went on to win with the help of these ties, deepened the relationships between workers and their communities, which then added to their overall pool of resources, ‘eventually enabling the workers to challenge the political and economic power structure that dominated their lives’ [38]. In other words, once these relationships, thoroughly apolitical to begin with, started to be understood as sources of power, to be wielded as such during contract negotiations and labour activism (up to an including strike action), workers increased the collective control they enjoyed over their lives. Although this required workers adopting, to at least some degree, an instrumental view of these relationships, doing so, ultimately, increased their autonomy.

Crucially, our argument is not that workplace sociability should simply become an instrument wielded for the sake of producing more autonomy for workers, as if we might simply replace the goods of affiliation and therapeutic friendships with militant comradeship. Instead, our argument is only that, for workers who have an interest in increasing their autonomy – at work and beyond – the sociability they develop in their workplaces should be partially oriented toward those tasks that can help facilitate such increases. This imperative to politicise workplace sociability might be more urgent for some workers than others. Part of Alex Gourevitch’s radical justification of strikes – specifically, for ‘low-skill, high labor supply workers in sectors like service, transportation, agriculture, and basic industry’ – is that such workers confront significant pressure on already low wages, lack union support much of the time, suffer from precarious contracts, and will be more adversely affected by downward trends in the economy – inflation and the like. For such workers, protecting existing levels of autonomy, and enhancing them, might require more readiness to engage in militant forms of worker activity, than ‘higher skilled, low-supply workers’ operating in other sectors, where negotiated forms of bargaining might be more appropriate [40]. But even in those less advantaged sectors, this orientation is not a constant feature of work relationships, but is rather something that can, at critical moments, be effectively turned to political and autonomy-expanding purposes. It seems perfectly possible, indeed likely, that if politics came to dominate such relationships, workplaces would become suffocating places.

From their front-line accounts of labour activism, Blanc and McAlevey call, respectively, for ‘militant minorities’ and ‘skilled organizers’. For both Blanc and McAlevey, such people are needed to develop workers’ political orientations, precisely because these orientations do not naturally emerge in workplaces. Indeed, workplaces have long been consciously designed by employers precisely to prevent the development of the connections and collective agency that might lead to confrontational forms of activity [41]. In the terms developed in this paper then, these militant minorities and organizers act to build upon the social connections already nourished in the workplace – and already doing work answering people’s social needs – in order that they may become levers capable of generating worker power. This power might need to be exercised through difficult and time-consuming actions, such as resisting employer harassment, standing firm during contract negotiations, the ‘spade’ work of organizing co-workers and preparing, sometimes, for strike action and other forms of industrial action. For workers convinced of the efficacy of such action, and who are interested in increasing their autonomy, workplace relationships are, inescapably, important sources of power.

There are, potentially, significant opportunity costs consequent to developing this kind of power. For example, if one must orient oneself toward thinking of one’s relationships as sources of power to be mobilized for political purposes related to the workplace, then there is less time for pursuing new relationships or attending to other, pre-existing relationships (at least in the short-term – with more autonomy won through workplace action, this might improve in the long run). In addition, some otherwise pleasant relationships, those that supply entertainment and camaraderie in the workplace, might end because of disagreements related to workplace action. Co-workers may think, or come to think, the demands their colleagues make are unreasonable. They may not want to rock the boat in a way that puts their jobs at risk. As a result, such workers might come out against organizing efforts, even going so far as to cross picket-lines during strike action. In the above example of teacher militancy, for example, African American community leaders encouraged the crossing of teacher picket lines precisely because of profound differences over the strategy and perceived purposes of those strikes [37]. While worker solidarity does not require unity across all political or ideological questions, there remains the need for some baseline agreements on strategy, as well as agreement regarding the proper way of handling disagreements. If collective action results in failures, it may be especially hard for colleagues to move beyond recriminations against co-workers who did not join in the struggle at decisive moments, or who steered it in directions with which they did not agree. Politicizing sociability in this way, while potentially effective as a means of developing collective agency needed for enhanced autonomy, can also put meaningful relationships – and all the goods thereby served – at risk.

7 Conclusion

Many academics, perhaps most especially philosophers, have relatively little experience of work as it exists for the vast majority of people. Our work might be hard, but it is also intellectually challenging and rewarding. The idea, then, that the people with whom one works might provide the primary reason to stick out, for many years, some otherwise dull, dangerous or despised job, is not something that readily occurs to us. Nevertheless, ‘other people’, far from being Sartre’s hell, are in fact necessary for keeping many people out of it. In this paper, we have focussed on the tension between this part of worker’s workplace relationships – what we have described as the ‘relational salve’ or ‘analgesia’ they supply – and their role as resources, as a source of power, to be drawn on in struggles for more worker autonomy.

Any discussion that makes the case for more worker autonomy, presupposes possibilities for change. A world where so many people are destined to ‘suffer what they must’ at work is a depressing prospect. Such changes will come in the form of, for example, work that is more propitious for the exercise of capacities, reduced working hours, increased remuneration, or expanded access to other sources of income. Relationships that are developed at work are a possible, and historically important, means through which such changes can be effected. When these relational resources are not treated as a source of power through which autonomy can be expanded and used to change people’s working environments, they might still help make more bearable situations defined by heteronomy. However, if people come to regard their work as lacking opportunities to exercise their capacities, or as using up too much of their time, or as being too poorly remunerated, and thus as incompatible with autonomy, then the relationships developed at work are also an important source of power for doing something about that, and can be developed and nurtured with just such power in mind.

We conclude with a final reference to the socialised Sisyphean labour of rock-rolling. When these Sisypheans ascend their hills together, they cannot be expected to regard, only and always, their relationships with fellow pushers as an instrument for obtaining future autonomy. Sometimes they will only want to shoot the breeze, or to share their misery, to grouch and find solace in one another’s company. Sociability in any form is unlikely to survive politicisation pitched always to levels of intense militancy. Nevertheless, there is the possibility that, where no part of their sociability is dedicated to creating the means whereby their situation can be improved, their relationships with one another will only ever facilitate collective resignation to a shared, and heteronomous, fate.