Abstract
A profusion of jobs has arisen in contemporary capitalism involving ‘connective labor’, or the work of emotional recognition. Yet the expansion of this interpersonal work occurs at the same time as its systematization, as pressures of efficiency, measurement and automation reshape the work, generating a ‘colliding intensification’. Existing scholarship offers three different ways of understanding the role of emotions in connective labor – as tool, commodity or vulnerability – depending on their view of systematization as useful, inseparable or dehumanizing. Based on 106 in-depth interviews and 300+ hours of observations, I found that vestiges of all three models lurked in the experience of providing connective labor, yet none fully captured the profound meaning practitioners reported finding in their work. Systems varied on three dimensions, reflecting the relative worth of worker, recipient or the work, extracting value from the forged connections, while the meanings workers derived shaped their perspective on its systematization.
Keywords
Introduction
Some work relies on intimate knowledge, on an emotional understanding between the worker and other people, for its success. Therapists, teachers, coaches, primary care physicians, sex workers, even business managers and high-end sales staff – many depend on their ability to connect to others to make their contribution: clients healing, students learning, employees motivated and engaged, customers satisfied. I dub this work ‘connective labor’, which I define as creating and sharing an emotional recognition of another person, through a process of ‘seeing’ or reflecting them, in order to create value. Yet this work is being transformed by current trends. As employers try to extract greater efficiency or profit from connective labor, they impose new systems, data collection mandates, and technologies that shape its conduct and experience.
Scholarly traditions disagree about whether workers would or should experience these trends as contradictory, and why, as well as how to think about the emotional connections they forge. One tradition argues that workers maneuver within rationalizing systems, sometimes turning those systems into tools with which they seek to establish and further relationship goals (Timmermans and Berg, 2010; Zelizer, 2005). A second contends that systematization is inseparable from the work of connective labor, and that these relationships, and the emotions therein, are commodified as a matter of course by this market work (Bernstein, 2007; Illouz, 2007, 2017). Finally, a third group of scholars argues that while these systems can protect workers from consumer demands, they also serve to exploit or control workers (Hochschild, 2012a; Kang, 2010; Leidner, 1993), and emotional relationships make workers more vulnerable to that exploitation (England, 2005; England and Folbre, 2003).
This article contributes, defines and illustrates the concept of connective labor, while investigating its practitioners’ perspectives on the systematization of their work. Based on interviews with more than 100 people and hundreds of hours of observations, I find that workers spoke of deriving great meaning from their emotional connecting work. Yet the organization of their labor generally imposed a regime of scarcity, so their meaning-making was either buried under what they experienced as an avalanche of need or squeezed by systems that imposed incessant demand for more data to scale up their labor. Systems varied on three dimensions that helped to determine the workers’ experience, and reflected the relative worth of worker, recipient, or connective labor itself. Ultimately, systems extracted value from the connections that workers forged (Tsing, 2012), ‘transformative relations’ that they viewed as more than tools, commodities, and vulnerabilities. The larger meaning practitioners derived from their labor shaped their perspective on its systematization.
Connective Labor and the Systematization of Emotions
‘Connective labor’ is a form of emotional recognition, and as such it borrows from and reworks two substantial streams of theorizing: one in emotions and the other in recognition; space precludes a comprehensive discussion of the voluminous research in both areas. As I define it, connective labor is made up of three components: 1) empathic listening, in which workers cultivate a shared emotional understanding of someone else’s perspective (Clark, 1997); 2) emotion management (Hochschild, 2012a), with which they modify and control their own reactions so as to better elicit the other person’s sharing; and 3) witnessing, with which workers reflect back what they see of the other person. Following Fisher and Tronto (1990), I consider these processes interactive, in which the client or student helps to co-construct the meaning of the encounter.
Features of contemporary capitalism – insecure production, expansive consumption and rampant inequality – generate both an intensifying need for being ‘seen’ and the connective labor jobs that provide it (Bernstein, 2007; Lane, 2015; Pugh, 2015). Rising precarity combines with the new centrality of emotional subjectivity to seemingly require a self expressed through sentiment (Silva, 2013); an entire swath of jobs have arisen that involve bearing witness to that self (Illouz, 2008). Scholars have argued that these trends reflect ‘cognitive capitalism’, an era in which knowledge is produced through ‘the use of cognitive faculties to form relationships and communicate effectively’ (Morini, 2007: 50).
Many theorists have circled around the broad issue of emotions in paid work, including Hochschild’s (1979) emotional labor, Hardt and Negri’s (2000) affective labor, and scholars of ‘care work’ and its many variants (e.g. ‘emotional care’ (Cancian, 2000), ‘love labor’ (Lynch, 2007) and the like). While some of these terms strive to serve as umbrella concepts to capture all kinds of feeling deployed in a market context (see Gill and Pratt (2008: 15) re: the ‘bluntness and generality’ of the affective labor term), connective labor is a much more specific concept, capturing the work of being present and receptive to another’s emotional truth, bearing witness to it and conveying that understanding to the other, who then receives it. Emotional witnessing of this kind has been a central concern for psychoanalytic theorists (Benjamin, 2013; Winnicott, 1967), who argue that mirroring in the maternal-infant dyad has repercussions for later psychic health and/or domination in adult intimate partnerships. Like emotional labor, connective labor involves using one’s own emotions with intention – in this case, to sense and reflect the other’s perspective.
Much scholarship on recognition focuses on its political implications (Fraser and Honneth, 2003; Honneth, 1996), with some scholars deciding that the emotional recognition of the other, or what Honneth called ‘love’, is a private matter, beyond the reach of politics. Feminist versions of recognition take intimate recognition more seriously (Connolly, 2010); Butler (2011) argues that recognition involves interpersonal violence, in that it requires that people be configured as legible, insisting that they fit into pre-existing notions and identities. Some scholars argue some of the domination looming here might be attenuated by moving from recognition to ‘acknowledgement’, which does not presume to know the other quite so well and thus acquiesces to uncertainty (Giordano, 2014; Oliver, 2001). While I take this point, I consider misrecognition harms as continually threatened by connective labor in practice, and thus use these terms interchangeably with witnessing, mirroring and reflection.
Connective labor shares some similarities with Foucault’s (1982) concept of ‘pastoral power’, which involves the harnessing of an intimate understanding of another, or ‘guidance through knowledge of people’s secrets’ (Cook and Brunton, 2015: 546), to the aim of salvation, or, in secular contexts, mental and physical well-being. Pastoral power combines two kinds of Foucaultian power: the fostering of self-governing subjects and a disciplining surveillance, making it both individualizing and totalizing, a blend Foucault called ‘tricky’ (1982: 782; as cited in Martin and Waring, 2018; see also Illouz, 2007). Foucault usefully highlights the cultural/emotional dimension to this power – not just the fact of one person’s vulnerability, but how the pastor can make the other feel better about their vulnerability; and how power acts upon the pastor to make them feel responsible. ‘The point is not that everything is bad,’ Foucault wrote (1994: 256), ‘but that everything is dangerous.’
Grounded in this conceptual backdrop, connective labor is the work of emotional recognition taking place in clinics and classrooms, for purposes ranging from care to persuasion to management to control. While it is not limited to the exchange relation, much contemporary connective labor takes place in the market; empirical research has documented the causes and consequences of connections in particular occupations, such as teacher-student rapport (Cornelius-White, 2007), the physician-patient relationship (Kelley et al., 2014) and the ‘therapeutic alliance’ (Poland, 2000), but less about what these might have in common in practice and consequence. 1 My focus here is upon excavating and analyzing the labor of interpersonal recognition, and the impact of systems upon the experience of providing it.
While this work is particularistic, emotional and personal, it is also undergoing massive systematization, or rationalization, in a wide range of fields. As organizations attempt to impose efficiency norms, increase accountability, and scale up the delivery of connective labor, they introduce the use of manuals and checklists; scripts and templates; and even, in some cases, automation and artificial intelligence, with resultant impacts on connective labor. These trends may seem very different, as there is a big technological leap from the checklist to the therapy app. Yet they each depend on deriving abstract principles from connective labor, principles that can be measured, taught and, if need be, fed into an algorithm. Despite the terms ‘customization’ or ‘personalization’, tailoring to meet user preferences involves information and individuals reduced to modular, easily exchangeable units (Mattos et al., 2019). As such, all of these imply standardizing the worker’s part of the encounter, stripping out the personal or the idiosyncratic, those unique qualities that form the grist of workers’ connection to others, and that shape what those connections look and feel like. I gather these trends under the concept of ‘systematization’.
The expansion and growth of connective labor, then, occurs at the same time as the pervading rationalization of this work, which strives to impose a certain distance from personal, emotional connections, a rationalization whose symptoms include checklists and manuals, ‘therapy apps’ and ‘virtual nurses’. While some professions – e.g. teachers, physicians – managed to stave off these trends for years by virtue of their perceived distance from market pressures, the neoliberal era has introduced pressures of efficiency, measurement, and assessment into even these corners of work (Reich, 2016; Williamson and Piattoeva, 2019). We might call this co-occurrence of the growth of connective labor and its systematization ‘colliding intensification’. 2
There are three groups of scholars who help us think about colliding intensification. First, some scholars adopt the stance that systems-are-tools. Researchers argue that with standards or ‘commensuration’ comes improved performance and necessary cooperation (Espeland and Stevens, 1998). These researchers contend that when people find themselves within rationalizing systems, they can sometimes deploy those systems for particular ends, such as to coordinate work, to protect or advance professional status or to communicate relations. These arguments can be found among relational economic sociologists (e.g. Bandelj, 2020; Zelizer, 2012), ‘morals and markets’ scholars (e.g. Healy, 2010; Livne, 2019), those who debate ‘discretionary’ power in bureaucracies (e.g. Evans and Harris, 2004), and others who chronicle the advance of standardization, such as in medicine (e.g. Reich, 2016).
This group adds a powerful cultural corrective to notions of systematization as somehow external to the people who develop, oversee or implement them; they argue that while systems carry risks of dehumanization, people imbue even the most standardized of units (e.g. cash, medical guidelines issues from on high) with very human variation (Bowker and Star, 2000; Singh, 2016; Wherry et al., 2019). They also offer sophisticated analyses of systems that document some of this great heterogeneity; Zelizer (2017), for example, usefully notes how we distinguish between blood money, pin money, tips and bribes, and how each is aimed at a different sort of social interaction, while Timmermans and Berg (2010) disentangle types of standardization, advocating greater ‘procedural standardization’ as the path to enhanced coordination of work (although they neatly sidestep the impact of these changes on physicians’ relationships with patients).
The neutrality of their stance towards standardization can create its own mystery: if there were nothing contradictory about the systematization of connective labor, why does dehumanization continue to haunt empirical accounts? Both Zelizer and Timmermans have the same answer – dehumanization is not inevitable, they emphasize, and such fears can reflect a ‘hostile worlds’ ideology, which clouds our view of how systematization’s impact varies radically depending on how it is deployed, by whom and for what ends. Yet the emphasis on people’s capacity to shape systems sometimes downplays their constraints and underestimates alienation, particularly among those whose working conditions make them less able to marshal the discretionary power that can humanize institutions (Remesh, 2008; but see Poster, 2011).
A second group of scholars argues that connective labor and its systematization are co-constitutive. This group includes scholars of contemporary intimacy such as Eva Illouz and Elizabeth Bernstein, who point to the cultural and economic trends that create the intensified needs for emotional ‘authenticity’ and, simultaneously, the rationalization of labor that can meet those needs (Bernstein, 2007; Illouz, 2007, 2008). To take Illouz’s sweeping argument, consumer capitalism and the ‘psy industries’ have generated not just the needs but also the commodified, rationalized nature of how those needs are met; she ends up calling the emotions evoked by this process ‘emodities’ (Illouz, 2017). The co-constitutive tradition usefully explains the origins of colliding intensification and notes that the task of therapists, coaches, counselors and the like – connective labor practitioners – is ultimately to render legible the messy, unique experience of the emotional self. Thus these workers themselves are engines of systematization, the very systematization they sometimes decry (Illouz, 2007).
And yet for this tradition, it seems that the market is all there is, as there is no obvious space between the experience of emotions or relations and the standardized ‘products’ of this work – including the newly legible selves, and their well-being or suffering. Just by virtue of taking place in the market, connective labor transactions become indistinguishable from the market. This move lends emotional commodification an air of inevitability and makes it harder to criticize the extraction of value from relationships. The co-constitutive tradition helps us solve the mystery of why need and system, such seeming opposites, would both intensify in recent years. But as Marx first told us more than 150 years ago, if we pay too much attention to the commodity, we risk being distracted from the labor of the people that produce it.
A close attention to the organization of connective labor is promised by the third tradition, which investigates the politics of interactive service work. Beginning with Hochschild’s (2012a) analysis of emotional labor, this group has since identified a host of different kinds of personal labor, including ‘care work’, ‘aesthetic labor’, ‘body work’, ‘deference work’ and the like (Duffy, 2011; Kang, 2010; Otis, 2011). These scholars attend closely to the labor process, and the workers they focus on are usually low-paid, ‘low-skilled’ women of color – e.g. home health care aides and domestic workers – hence, their acute awareness of exploitation and analyses of gender, race, class and other inequalities. When systems are involved, in these accounts, they are imposed from without, by the state or by employers focused on wringing labor from these workers through such tactics as routinization and surveillance (Diamond, 2009; Glenn, 2010). These scholars help us to see the estrangement of these workers, and its impact on the ‘product’ of their emotional interactions. While systems can sometimes protect workers from overly demanding clients, they can also lead to workers’ alienation (Leidner, 1993). Gender, race and class blend together in connective labor to shape the perils of exploitation for disadvantaged workers, who are often women of color, as well as the misrecognition harms incurred by clients, particularly when they face workers with more privilege; I discuss these issues elsewhere (Pugh, 2021). Yet this tradition is generally so focused on exploitation that there is little conceptual room for the workers to find the labor valuable or affirming.
Furthermore, with few exceptions (see Hochschild, 2012b; Lopez, 2006; Sherman, 2007), scholars in all three groups do not spend a great deal of time on the actual connections forged between worker and client. The systems-as-tools scholars seem to suggest relationships are one among a range of instrumental tasks the systems enable them to accomplish, while the co-constitutive scholars might view emotional connection as part of the lure of commodification, the feelings of relief and warmth part of the consumer’s reward. The politics of interactive service work researchers sometimes acknowledge feelings of connection in interpersonal work. Yet research in this latter group often portrays emotional attachment between worker and client as either a myth that serves to obscure how care relies upon and reproduces gendered and racialized inequalities, a distraction from their own working conditions, or a vulnerability (Uttal and Tuominen, 1999), the root of the ‘emotional hostage effect’ (England, 2005: 390) that employers exploit to bind workers to bad jobs.
The intensified need for connective labor and its intensified rationalization are twin trends that impinge directly on the practitioners who do this work. Scholarly traditions disagree about how to view trend, work and worker. How do these workers view their witnessing work? How do they understand the co-occurrence of connective labor and the campaign to systematize it? What is the effect of rationalization on their emotional experience of reflecting the other?
To answer these questions, from 2015–20 I conducted 106 interviews and more than 300 hours of observations of connective labor practitioners in therapy, teaching and medicine, as well as other occupations (see Pugh, forthcoming, for methodology details). While this research took place in the United States, the systematization of connective labor is a global phenomenon, wherever ‘seeing the other’ takes place within a context of increasing measurement, commodification and control.
I found that emotions may indeed be tools, commodities or vulnerabilities, particularly for disadvantaged workers, but that they also extend beyond these characterizations. Connective labor practitioners do more than teach or give therapy; their witnessing work often creates profound emotional meaning, including a sense of purpose in their own work, the human dignity that arises when someone is effectively seen by another person, and a greater understanding about self and other. Many report feeling overwhelmed by the need they meet in their jobs, which makes welcome the introduction of some systems but not others. Systems vary along three dimensions that ended up consequential for the meaning the practitioners value: whether the systems are aimed at the connective labor or other tasks, whether they are designed to augment or replace that work, and whether they are introduced in regimes of scarcity or plenty. The following three cases were chosen to illustrate these findings.
Emotions and Systems in Connective Labor
‘You’re Getting Immediate Joy, and You Can Be Present.’
Simon Jarret 3 was a white physician working in a community clinic that served low-income people in the San Francisco Bay Area. For Simon, connective labor was useful in his work, a vital tool for the quality of his medicine. ‘If I ask the right questions, I can, you know, this patient I’m just meeting, I can ask about anxiety, stress and sometimes they break down in tears, and I’m like, “okay, this is really what’s going on”,’ he said. Knowing ‘really what’s going on’ helps him to tackle their medical issue, which often pales in comparison. ‘Then I can reassure you, like, okay, “what’s your concern about this minor muscular-skeletal pain?” And you know, sometimes they’re like, “oh, yeah, I know, I know, that’s my tension”.’
Yet connecting to patients was also valuable on a more basic level, beyond its functionality. Not an effusive man, Simon had a low-key style, friendly and gentle in a palette more muted than intense, yet his narrative was threaded throughout with the word ‘joy’. ‘[The relationship] is what you build, because we’re humans,’ he said. ‘I mean, [there’s] an inherent pleasure in that, right?’ He derived considerable personal meaning simply from connecting to patients in his work.
But systematization affected his capacity to do this work. He was good at this emotional reading, he said, but doing it all day was exhausting, ‘a 15-minute, you know, hamster wheel’. On top of the sheer quantity of people he saw, as part of working in a community clinic with scarce resources, his clinical work faced another serious impediment: the electronic health record (EHR). ‘It’d be one thing if it was just the joy of interacting with patients and then there wasn’t also this “oh, my gosh, I’ve got to do a hundred different tasks as a result of this visit and get it documented and entered into the electronic health system”,’ he said.
Introduced in 1972, the EHR was in near universal use in Australia, the UK and the Netherlands by 2005 (Jha et al., 2008); it is only in the last ten years that it has become standard procedure for most practicing physicians in the US (Office of the National Coordinator for Health Information Technology, 2019). More than 50 percent of US primary care practitioners report burnout (Peckham, 2016), and there is evidence that links it to use of the EHR (Shanafelt et al., 2016), particularly with the lengthy documentation required in the US context (Downing et al., 2018). While some (often younger) primary care physicians acknowledge the benefits of its introduction (e.g. providing access to current standards of treatment, easing coordination between doctors, labs and other professionals), clinicians told me that they went from approaching the patient as a whole, in which the individual components of a check-up were an indivisible ritual, to an itemizing of increments as small as rabbit pellets. Along the way, the patient’s particularities, and what the physician did to assess them, were standardized (Freidson, 1988). The impact of this system was devastating for someone who values relationship, Simon said. ‘I mean, all of it is really, I would just say, is tangential to the actual role of the physician, as healer and diagnostician.’
The EHR was a system introduced in spite of the physician’s connective labor, not in order to improve or replace that labor; as such, physicians who do connective labor are forced to squeeze it into ever smaller segments while the artifacts of data primacy expand. ‘Imagine it’s like you had, you know, psychological dog experiments where, like, you’re getting shocks throughout that,’ Simon said. ‘Like you’re having all these things that you don’t actually value that someone else’s controlling and you have to do, like, “oh, you need to bill this” or the computer system is really a distraction and unwanted and something outside of your control.’ Simon’s bitterness reflects his experience of its presence as a ‘hostile world’ colliding with what gives his work meaning and value (Zelizer, 2005).
Simon is a good example of many people I spoke to, clearly gifted practitioners able to do an emotional reading of others to get beyond words to find deeper meaning. Many of these connective laborers found their work humbling, and they talked about being honored by the privilege of bearing witness to people in their vulnerabilities. But many also felt overwhelmed by the way their work was organized, squeezing more and more needy people into ever narrower units. Some systems – such as checklists or memory prompts – were welcomed, evidence of the degree to which they alleviated some of their terrible burden, but most practitioners experienced those involving data collection and analytics as significant impediments to the work they valued. The deep meaning that they gleaned from their connective labor was not reducible and was part of what shaped their perception of those systems as, in effect, like ‘hostile worlds’.
‘We Need to Do the Social and Emotional Stuff because, for Some Kids? The Test Doesn’t Matter.’
Pamela Moore was an exuberant African-American woman in a middle-school classroom in a large California city, teaching seventh graders as an ‘intern and teacher of record’ during the day and taking classes at night to finish her certification. She exuded a felt charisma, and her warmth and energy radiated throughout our interview.
The school had many children who were ‘going through a lot’, she said. Almost all of the students were non-white, with 86 percent socioeconomically disadvantaged; a quarter of the school were English language-learners. She told a story of a pair of brothers ‘who just totally broke me down’, showing up in smelly clothes because someone had actually been murdered in their home and no one had cleaned up the crime scene. ‘So, there’s blood, and gore, in this child’s home and he and his brother had to try to clean it up.’ She experienced a lot of what she called ‘vicarious trauma’ teaching there. ‘I can understand why my instructional coach probably thought I was fragile’, she said. ‘I needed to cry.’
The relationships she forged were certainly instrumental tools for Pamela, keeping kids coming back to school, letting her know what motivated or interested them, helping her see when their lives were overwhelming their capacity to learn. At the same time, however, they were also powerfully meaningful for her personally, inspiring her intensive efforts to connect with her students.
Systems suffused her work above and beyond her connective labor – e.g. the curriculum that she consulted and deployed, the certification that she sought to be allowed to take a long-term job, the grades she assigned. She complained that data needs were getting in the way of getting to know her students. ‘You’ve got a curriculum, you’ve got 180 days, you have a certain number of academic instructional hours, you have testing, testing, testing, testing. It’s hard because we need to do the social and emotional stuff, because for some kids? The test doesn’t matter,’ she said. ‘I have a little boy right now, it’s not working, there’s attention issues, his diet, he’s got a girlfriend, he doesn’t care about school anymore.’
At the same time, other systems applied directly to connective labor, such as the approach she adopted to quickly and efficiently ascertain her students’ moods every morning. I check in with my students, we do something called an SEL [social-emotional learning] check-in. I say ‘Okay, on a fist-to-five, fist being the worst day ever, like apocalypse, and five being “woohoo, I’m going to Disneyland,” just quietly let me know, are you a two, a one, where are you today?’ And I let them know where I am.
One day, a student raised her hand and said ‘“I’m not feeling it right now, I’m feeling kind of depressed”,’ Pamela recalled. ‘So, I got the lab started and checked in with her, and she had been cutting herself.’ Pamela was devastated. She brought the girl to the nurse’s office, but the memory was another searing moment of the suffering she witnessed among her students.
Pamela’s ‘SEL check-in’ is a fairly popular teacher short-cut, one that brings a palpable efficiency to detecting mood among students quickly, even though it likely misses any who refuse to share in such a public way. Proponents say that it builds student awareness of their own feelings as the first step towards being able to regulate them, that it normalizes all kinds of feelings, and that it builds community. Yet it is certainly an instrument of standardization. While the fist-to-five check-in ‘normalizes’ feelings, it does so by ensuring that they are made legible to others, both qualitatively – ‘I’m feeling kind of depressed’ – and quantitatively – ‘to be honest, today I’m a two’. The messy inarticulate particularity of feeling is rendered smooth and modular, communicable through a common language, and ‘measurable’ by common metrics. Critics consider SEL curriculum as a way schools rationalize student emotions, exclude violent or intense emotions as inappropriate and those who feel them as emotionally illiterate, and create neoliberal subjects, better prepared to take on individual responsibility for socially structured inequalities (Wilce and Fenigsen, 2016). One could certainly argue that the fist-to-five check-in produces ‘emodities’.
Furthermore, the fist-to-five reflects an emphasis on efficiency imposed from above. Classroom time is compressed, with the school assessed in part by how the students perform on statewide tests of a mandated curriculum that is not about recognizing or regulating feelings. So Pamela needs a fast way to elicit emotional information – and to communicate that she sees her students – while subordinating these needs to meet these dictates. I saw connective labor workers deploy technologies to the same end in other settings.
At the same time, however, Pamela’s work is more than simply extractive or standardizing. As a tool for systematizing connective labor, the fist-to-five system enabled her to elicit emotions from students in trouble, but it is the profound meaning she made of their relationships that led her to treat their responses as not the end but the beginning of the conversation, and the feelings themselves not simply as tools, commodities or vulnerabilities, but as signs of her larger purpose.
‘You Genuinely Had an Impact on People, and You Saw How Things Were for Them.’
Veronica Agostini was a white woman with a fresh BA in Psychology working as a ‘coach’ at an online start-up trying to cure social anxiety. When she joined, the company offered an app incorporating the principles of cognitive behavioral therapy (CBT) into a 12-week online program, one that later became a service firms could offer their employees as part of a benefits package. ‘And it was basically someone had taken, like, a CBT workbook and just put it on the website,’ Veronica recalled, describing the scripted text templates she used with clients at first. ‘So it wasn’t like the coach was therapy. The coach was there to keep the clients accountable, to just be someone who’s, like, checking in on them and, like, encouraging them to go on.’
While the plan may have been to relegate coaches to the sidelines, Veronica soon found her role expanding in weekly phone calls with the clients. They may not have been doing ‘therapy’ but, as she described her work, it started to sound increasingly familiar. ‘Basically what happened is you talk to someone one time, and then, you know, if they were interested and you did a good job on the first call, you would talk to them every week for sometimes, like, 12 weeks. I talked to some people for months,’ she recalled. ‘I think a lot of what many people got out of it was just having someone there who was, like, nonjudgmental, a good listener. It was “half an hour a week I can talk to this person and they understand what I’m going through and they’re not judging.” And that was what a lot of people liked about it.’
Some aspects of the job made it clear it was not therapy. Often people would just drop the app and ‘ghost’ the coaches, behavior more akin to closing a self-help book than it was to the sometimes lengthy experience of ceasing a therapy relationship. There were also rules in place to keep the client coaches from acting as mental health professionals. ‘So you were allowed to reflect what was in the program, but you could never give advice,’ said Veronica. ‘You obviously could never make a diagnosis. Um, you couldn’t, you know, tell them what to do or anything like that.’
But the rules dividing their work from therapy seemed more for the coaches than for the clients, some of whom were happy to think of the coaches as private counselors. ‘Lots of people would just want to talk, and they wouldn’t be going to a therapist,’ Veronica said. ‘So this was like, “100 bucks a month?” Like, “I can scrape that together, like, why not?”.’
Other aspects of the work echoed therapy as well. The client coaches were reflecting and witnessing clients who shared profoundly intimate details; they were trying to help the clients get better, and they suffered from burnout. ‘I think in the beginning I – it really was, like, hard,’ Veronica said. ‘Um, it was so hard to set boundaries and I would think about the clients, like, after I went home. I was always just like, “Um, I don’t know, like, how to do this”.’ She laughed in disbelief. ‘And then these people are telling me, like, really heavy things all day. People will tell you a lot when it’s basically anonymous, you know? And I’ve had people who were dealing with really serious trauma.’
Initially, the company gave no training at all on how to handle these issues, introducing ‘motivational interviewing’ training only later on, and indeed met coach concerns with what Veronica called ‘a pretty callous attitude, where it was sort of like, “you can quit if you want to”. Which is just like, “No, people are here because they care”.’ While emotions kept many tethered to the job, Veronica ended up quitting due to overwork, when the firm scaled up her client roster to untenable levels, from seven to twelve daily.
Nonetheless, Veronica said she found the job moving, humbling, and powerful. ‘You genuinely had an impact on people and you talk to them every day, and you saw how things were for them,’ she recalled, saying she hoped to train as a therapist. ‘I loved talking to all these people. I loved feeling like I had, like, an impact.’
Yet despite both the profound meaning of the work, and the emotional wallop it delivered to the coaches, Veronica’s own language joins in minimizing her effect. She ‘loved feeling like she had an impact’ but quickly followed that with ‘Um, even though it was really them that was doing – you know, they were doing all the work, you know, I wasn’t really doing anything. I was just, like, cheering them on and helping them, like, work through some, like, hard things sometimes.’
Thus the company maintained Veronica was not doing therapy, clients came and went abruptly as if they were not doing therapy, and Veronica asserted that she was not doing therapy. Their collective insistence renders invisible the work she did to witness the clients’ needs and struggles, and the price she paid for it in the emotional residue their trauma left behind.
There is a lot in Veronica’s story that suggests commodification. CBT is a routinized form of therapy, with fairly standardized stages and predictable assignments for the client, and as such it is particularly given to scripting, even automation (Stark, 2016). Its systematization in the form of the app moves the mental health assistance on offer even further into the range of a standardized commodity. If the emotions are commodities, what does that mean for Veronica? She is perhaps akin to a farmer or roughneck, those manual laborers at work on an oil drill, performing the difficult work of extracting these products. In this scenario, Veronica uses her connective labor – aided by the mechanized CBT processes – to wrest more positive emotional states from the raw material of depression, anxiety or shame.
Yet while these processes of connective labor certainly take place in the marketplace, we acquiesce, I believe, to a powerful reduction if we consider as relevant only that which is explicitly bought and sold. Even if one could buy ‘freedom from social anxiety’ – as both company and clients seem to think you can – reducing that transaction to a commodity accomplishes a de facto erasure of the connective labor that went into it, the particularities of the relationship that surrounded it, and the secondhand emotional suffering that the coaches endured. Because it relies on the standardization of emotions, the commodity language seems to render invisible the work to produce those emotions, as if all Veronica is doing is extractive as opposed to interactive. It requires that the personal, the particular, be shrunk, but does so not just to the outcome (the feelings) but also to the work (the connective labor) and the worker (the practitioner). This reduction not only renders Veronica invisible, it also pushes aside the powerful meaning that she makes from the experience, the profound and humbling experience of bearing witness to another. Just because the firm benefits from her invisibilization (by not having to train, hire or pay for skilled workers, and by extracting the value that her witnessing helps create), the client participates in it (by treating Veronica as an extension of the app) and even Veronica submits to it, does not actually mean that we need to subscribe to that erasure ourselves.
In the case of the social anxiety app, emotions are at once commodities (standardized things bought and sold), tools (used in Veronica’s interactions to keep people motivated) and vulnerabilities (keeping people at work even when they were burnt out). But it is the compelling meaning of the work that shaped her experience of the rationalization of connective labor – as its own purpose, aided by training and impeded by the scripting and speed-up of her work.
Discussion
Despite differences of gender, race, education, and other dimensions, these three case studies have important similarities: notably, a sense of the deep meaningfulness of the work, a prevailing sense of time scarcity and a felt ambivalence about the introduction of systems.
Amidst these commonalities, three themes serve to distinguish among these cases to highlight what systems do to connective labor. First, some of the systems target connective labor itself (such as the fist-to-five), while others target ‘everything but’, i.e. the other components of the worker’s job (such as the EHR), thereby forcing practitioners to squeeze the connective labor in on the side and along the way generating enormous stress. Second, some systems center human interaction, and seek to augment it (such as the motivational interviewing training regimen for the client coaches), while other systems subordinate it to other priorities such as billing (the EHR) or seek to replace it altogether (such as the scripted templates for the anxiety app text responses). Finally, some systems are imposed in such a culture of scarcity (such as the community clinic) or speed-up (the social anxiety app) that it compresses workers’ ability to pause, to see the other and convey that which they have seen.
These dimensions appear to overlap with how much employers, at the moment of their intervention, value connective labor, value the worker, or value the recipient. Connective labor that is valued is visible to employers; yet, ironically, with that attention comes the introduction of new systems for training, evaluation and enhancement, which can encourage workers to consider relationships as tools. When employers value the worker, interventions attempt to lighten the burdens or augment their efforts, as opposed to systems that replace or script the work, which can lead to emotional connections experienced as commodities. Furthermore, some practitioners found certain systems affirming, particularly those they actively chose to deploy, that helped them conduct their connective labor more effectively, and that enabled them to fit their work within time constraints imposed from without. These distinctions map loosely onto social inequalities like gender, race and class, and also vary depending on whether the practitioner has more or less advantage than the people they serve. Yet privilege was not enough to stave off systematization entirely: it did not protect advantaged practitioners like Simon from the systematization of his connective work, while advantaged clients like Veronica’s were still getting mediated therapy.
Ultimately, systems enabled employers to extract value from connective labor, in what Tsing (2012: 519) called a form of piracy, with ‘sponsors of scalability stealing from the work of transformative relations’. The target of their extraction was none other than the emotional connection that these practitioners forged, the very thing that they said made their work meaningful. Rationalizing regimes siphoned efficiency or profit from feeling, even when employers did not train, recognize or value the emotions underlying the service involved, and even when systems impeded those emotions in the first place.
Conclusion
Contemporary capitalism, through both the insecurity of production and the expansion of consumption, has generated an era of ‘colliding intensification’ – intensified needs for connective labor, as well as its intensified systematization – taking place in the domain of emotion. It reflects the (culturally and historically specific) meaningfulness of feeling, and its penetration by the market. Emotions have been not just colonized by work and by consumer culture; they are made legible to others by a standardization process that seems to create commodities, tools and vulnerabilities.
Nonetheless, their connective labor occupied what some workers say is a unique space – a world of secrets, that they get to be there for – and this privilege is part of what makes these practices meaningful for many workers, modern priests practicing ‘pastoral power’, allowed into an inner sanctum not of sin but of feeling. Yet that honor does not protect it from rationalization. As we saw in the last example of the therapy app, under circumstances of extreme systematization, the work can simultaneously feel like a privilege and yet also be rendered invisible, even to the workers themselves.
Three existing approaches offer different understandings of what workers experience at the intersection of emotions and rationality and why. Ultimately, we can find evidence of each of them in the case studies provided: workers deploy rational systems in part to make emotions legible, but also to establish their relations with others – with particular responsibilities and unique emotional costs and benefits – relations that are shaped by organization of their work.
This research highlights the profound and irreducible meaning many workers find in these relations, an account of how systems vary in their perceived impact depending on three dimensions, and a reckoning of the systemic extraction of value from ‘transformative relations’ that workers see as meaningful. In order to develop these contributions, this discussion has adopted a ‘lumping’ approach to both connective labor and to its systematization; future research could usefully explore how connective labor varies by gender, class and other inequalities, how divergent contexts shape its conduct, and the impacts of different kinds of systematization.
We are only just beginning to know what connective labor does and how it works, the complex resonance between people, the meeting of their particular emotional selves. Current trends are transforming the conditions of this work without much knowledge about what it is, what enables it, and what else it does. By centering the meaning that workers can make from forging emotional connections with others, we can understand both the lure and the risk of this systematization.
Footnotes
Acknowledgements
Thanks to Rosalind Gill and Lisa Blackman for insightful commentary and editing. Thanks also to Jennifer Petersen, Jennifer Cyd Rubenstein, Denise Walsh and five anonymous TCS reviewers for feedback on earlier drafts, and Arlie Hochschild, Eva Illouz, Viviana Zelizer and Rachel Sherman for their comments on portions of the argument. Genevieve Charles, Lena Gloeckler, Jaime Hartless and Allister Pilar Plater provided excellent research and editing. I’m grateful for feedback from audiences at the University of Southern California, the University of Minnesota Institute for Advanced Study, the Conference on Real Work in the Virtual World in Helsinki, Finland, and the New School for Social Research. This project was supported by National Science Foundation Grant No. 1755419, as well as funding from the American Council of Learned Societies, the Center for Advanced Study in the Behavioral Sciences, the Berggruen Institute, and the American Sociological Association.