“The Unique Magic That Happens When Two People Come Together”: Allison Pugh on Building a Society of Connection

In the latest installment of our partnership with the Center for Advanced Study in the Behavioral Sciences, Elizabeth Fetterolf talks with Allison Pugh about Pugh’s new book “The Last Human Job,” which explores the human connections that underlie our work.
“What we are doing by mechanizing encounters is bleeding out the unique and rather mysterious social outcome.”

Allison Pugh is a master interviewer. She has written an influential defense of the method in sociological research and carried out hundreds of interviews over the course of her career. Pugh finds and analyzes the meaning in people’s everyday experiences of work and home, from job insecurity’s effects on intimate relationships to the commodification of childhood. In her most recent book, The Last Human Job: The Work of Connecting in a Disconnected World, Pugh explores the very substance of her primary method; that rich, ineffable moment in an interaction when you, or your interlocutor, feel seen. Pugh develops the concept of “connective labor,” the collaborative work of emotional recognition. This includes three key components—empathetic listening, emotion management (as theorized by Pugh’s doctoral advisor, Arlie Hochschild), and the act of “witnessing,” in which one individual reflects back what they have seen. Drawing on years of interview and observational data, Pugh shows how in sectors like education, healthcare, and therapy, this work is increasingly systemized—a process that she argues makes it ripe for eventual mechanization. In the face of teacher shortages and hype around “chatbot therapists,” Pugh makes a case for connective labor’s value to society and the potential consequences for inequality should it become a scarce commodity.

As intimidating as it was to interview someone so deft at the form, I sat down with Pugh to discuss her new book. Pugh is a force—both in the depth and complexity of her scholarship, and the ease of connection she brings to conversation. Despite the limits of Zoom interviews, I tried my best to reflect back to her what I took from The Last Human Job.

Elizabeth Fetterolf (EF): This project has been many years in the making. How did you initially get the idea for the book?


Allison Pugh (AP): The long-term answer is that I feel like this is the dissertation I should have written. This is what I actually care about. I loved Longing and Belonging, loved the chance to observe kids, and loved what I had to say about that; but this, like getting at what relationship does, is, feels like a lifetime project.

But the more proximate answer is an article I wrote arguing that in-depth interviewing is a practice of seeing the other. It is also a unique emotional interchange with another human being, where we bring our baggage and we meet in the middle and we reflect each other. The article was a defense of the unique magic that happens when two people come together in encounters like that, and an argument against others who I felt were mischaracterizing it as just information exchange.


EF: The argument that talk is cheap.


AP: Right, exactly. And then, in writing that, I was thinking, How do we scale this up? How do we teach it to another person? I actually wrote the article in part for a teaching purpose, because it was a hole in my syllabus in qualitative methods; there was nothing really talking about interviewing as this interpretive, multilayer process. And so, I started to think about what interviewing is like. What else is it like?

It is a clinical practice, it is a personal practice, it is an emotional practice. It involves seeing the other and reflecting the other, and the other has to feel seen for it to work. And so, even though only one person is ostensibly asking the questions, it is still a collaborative moment.


EF: You make a really good case in that article for why interviewing is something that is a little heightened, something a little more than the everyday practices and why that is actually really valuable. Moreover, the project started from this interest in interviewing, and it is also an interview study in which you conduct over one hundred interviews! How did you think about that as you were doing the research?


AP: Yes, I love, love interviewing. I love talking to people in these sometimes intense dyads of collaboration and reflection.

I have so many different people that I talked to in so many different fields. I kept looking at how similar they were, and being struck by how, say, the funeral service director sounded like the home health care aide, sounded like the doctor, the teacher, the therapist.


Developing AI Like Raising Kids

By Alison Gopnik et al.

EF: This book reads to me in so many ways as a love letter to connective labor, but then there is also this paradox at the heart of the book. Seeing the other, or failing to see the other, can also be extremely harmful. Everyone remembers a time when connective labor went so wrong for them. A doctor who brushed them off, a therapist who misunderstood, a coworker who took advantage of their trust.

So how do you balance the real social value of connective labor and then also the potential harms of witnessing?


AP: That is the crux of the problem. It is such a good question. Because the book is a love letter. I do value it. I heard time and time again how people attempted to offer dignity to the other person and found their own humanity. Did their best to motivate someone else to do right by themselves and found their own purpose. So, it is a love letter to that possibility. And yet humans do terrible things to each other.

Even as everyone has had the haunting experience of being wounded, some people obviously have had much more of it than others and are more susceptible to it than others. Not just because of their own sensitivity or years of being disrespected in certain ways, but also because of their positionality. People are not giving them the time they deserve. Not giving them the pain medication they deserve. We know these stories, they have been well-chronicled, and you can’t ignore them.

So, the question is how do you really reckon with the enormous power that this process has to help and to hurt others?

The way I reconciled myself to this is that even though the power of connective labor resides in the individuals, the way to make it work better is to change how organizations produce or enable this work, instead of impede it.

It is not just bad people who wound others; it is actually well-intentioned people in bad settings. And if we can fix those settings, we can spread more of the magic rather than more of the wounding.


EF: Absolutely. You really highlight that in the chapter on organizations that are at least getting it right a little bit more. You argue that it is context that allows people to do this labor, rather than putting responsibility on the resource-constrained primary care provider or the teacher that has too many students in the classroom.


AP: What makes your question so acute is that engineers and people who are really enamored of the AI future—let’s call them technologists—think that AI can solve problems of misrecognition or the harms that humans do to each other.

And that is very appealing. As I report in the book, there is a ton of research already showing that people are more honest with computers, people trust computers to not judge them. And they are already treating apps, AI, technology as a safe space, even though there is a lot of critical work showing us that this is a space full of bias, surveillance, and real social impact on the production of inequalities that is in itself a form of harm.

So I want to first point out that the technologists’ appeal—don’t worry, we can just take humans out and everything will be fine—is quite flawed. Why? It’s about the social product. What I’m talking about here is not just an individual student learning algebra better or at all. Or the risks of a therapist losing their job to an app. Or, the risk of a client experiencing bias from that same app. All of these things are real, have been observed, and I take them seriously, but at the same time, what I’m trying to show with this connective labor angle is that what we are doing by mechanizing or attempting to mechanize these encounters is bleeding out the unique and rather mysterious social outcome, the resonance that happens between two people.

The mystery that is produced is what I end up calling “social intimacy,” to refer to the social product, the social stuff that is produced between individuals.

In this collaborative moment, that is what is lost. When you mechanize something, yes, kids can learn some algebra, but it is not great; or online therapy, or therapy even with a bot, it helps people a little bit, but not completely. But what we are not talking about is how automating these interactions is just hemorrhaging the social. You are letting the social just drain right out of there. And that is a serious loss.


EF: It seems like we are talking about devaluation. For you, there is this sense that when this process is mechanized, something is irreparably lost. But do you feel that is always true with systemization?


AP: No. I believe in vaccines. I believe in science. Systems have been an extraordinary contribution to enable modern society.

And what I watched was these teachers and doctors using systems to cope with problems of scale. They are overwhelmed, so they are using systems to make themselves more efficient. And to some degree that is a good thing.

But I wrote this book because we need to recalibrate the configuration of systems and artisanal human relationships, which I want us to prioritize. I wanted to articulate the value of and prioritize the latter so that when we introduce systems the first question is will that ruin something valuable here? What is the cost?

Rather than what we have now, where an institution adopts a system because they are prioritizing data and then assume that workers can just squeeze the relationship in. There is no priority given to the relationship and that just seems completely backward.


EF: Yes, because it is often so much about efficiency in those situations, which is not prioritizing any one worker, it is prioritizing the bottom line.


AP: But, also, it is even more tragic and messed up because they are prioritizing efficiency but forgetting that the engine of the thing they ostensibly care about—be it caring for older adults or teaching algebra or providing spiritual care, or healing someone’s diabetes—the ostensible task is powered by relationship.

So if they crush relationship with efficiency campaigns, they are crushing the very lifeblood of their activity, the value that they are saying they produce.

EF: In the book, you cite Lucy Suchman, who kind of stands in as the AI skeptic—I think many scholars who take a critical, humanistic approach to the study of technology would put themselves in that camp. AI skeptics are arguing essentially that this discourse about the automation of connective labor is driven by hype. So, a lot of these things that technologists are saying will happen—like automating the work of teachers or nurses—we are nowhere close to that. The technologies are always right around the corner, like self-driving cars. How are you thinking about hype and how are you engaging with both the technologists you are interviewing and then also the tech critical scholars?


AP: Great question. I’m so enjoying this.


EF: I’m so enjoying this too. It’s the connective labor.


AP: I have a couple of things to say. I am kind of ambivalent here. First, I will say that there is more socioemotional AI, more AI in this space, than there was when I started in 2015.

Then, we were talking about an “AI winter.” Now, I think there is more happening in this area. Of course, happening is different from happening effectively. And what Lucy Suchman points out is that much of the hype now is driven by the search for funding. And work across many fields has found that AI hype erases the very human labor that supports the “automation.”

That said, especially with ChatGPT, I do see something different. Sal Khan recently did this Ted Talk about his work to merge ChatGPT technology to make a kind of empathic tutor; that is automated connective labor. There is no question about it. So, we are here on some level. It is not all just hype.

I don’t want to say that it is necessarily working, so I want to distance myself from the hype in that way, but it is more prevalent than it used to be. And I think we need to take it seriously.

We need to be more sophisticated, more complicated about emotions, more emotional not less. We need to talk about emotion as a skill, as a power.

EF: One thing I love in the book is that you talk about Cognitive Behavorial Therapy (CBT) as this mechanization process. CBT is not necessarily what many people would associate with automation—people are thinking about AI therapy chatbots, for example—but CBT actually really highlights a lot of these tensions around the systemization of therapy. How did you start thinking about CBT as an engine of this broader systemization in your therapy case study?


AP: What I say about CBT, and my conclusions about therapy in general, capture the argument of the whole book.

That is, it is not like we go from this full-on artisanal human-to-human walking through the daffodils, and you flip a switch and all of a sudden one of us is a robot. Like that is not actually what I’m saying.


EF: No, that is never happening.


AP: Right, exactly. Instead, data analytics and the industrial model systematizes this work in incremental ways little by little, and the end result of that is a form of mechanization.

I know many CBT therapists and I know people who have had CBT therapy and have gotten much better. I know it works for some things—I don’t want to come out as like, this big anti-CBT person. There’s quite a bit of a debate in academic psychology and counseling about the degree of systemization and manualization and it is funny because it does echo some of the debates in sociology between qualitative and quantitative research. And the way that sociology has emerged from those debates has been to say, “Both are valid; it depends on the question.” And I can hear that when I talk to therapists about different methods.

So, these debates that were roiling in psychology at the time of my research were perhaps asking the same questions I was. How much value do we put in the relationship between therapist and client, and how much do we put in these manualized, systemized approaches? I think the case of CBT, which involves some scripting, with particular sequences therapists follow—it epitomizes a theme of the whole book, which is that there are these baby steps toward systematization, which become pit stops on the way toward mechanization in the future.


EF: Your point about the debate within sociology is so interesting. I think this is a question of epistemic value—these are two different epistemic frames, and the resolution of the sociology question is oh, well, it depends on the question. Yet, there are power differentials between these two frames in terms of their status and credibility within the discipline. That has not gone away.


AP: Yes.


EF: And there are these same divides in psychology. CBT is considered much more evidence-based than going on gut feelings of practitioners or the words of clients. It makes a significant difference for how seriously people take these two different modes.


AP: It is certainly true for how seriously institutions take them, and that comes with power.

But I think we shouldn’t minimize the—can I call it a silent majority? [Both Allison and Elizabeth wince, then laugh] I will say the silent plurality.


EF: Yes, I like that.


AP: That kind of, you know, values the idiosyncratic, the particular, the artisanal, the emotional. There are a lot of people out there who value that and are trying to figure out how to get this kind of care. I am interested in the tension between the artisanal and the systematic, and that tension animates the whole book.


EF: So, this is maybe a little bit of a detour, but I wanted to talk about your intervention into the care-work literature in the article that you wrote for Signs. You outline the long and complicated relationship that care-work scholars have had to emotion; it’s a sticky concept. Scholars have spent a lot of time trying to convince people that care is not just this good, gooey feeling—it is also a site of extraction, of exploitation, a way to obscure labor and to excuse poor working conditions. You acknowledge all of that, and yet you make this argument that we should not turn away from emotions, but in fact turn toward them in all of their complexity and messiness.

What motivated you to make this intervention and what do you hope for the future of feminist care-work research?


AP: What motivated me? Well, the critiques are valid. In the article, I talk about the disability care-work scholars who highlight the paternalism of care and how they said, Never mind emotion, we just want personal service. And then the critical race scholars saying, Every time we talk about emotion, we end up with these typologies that put women of color over here in custodial, dirty work, denigrating what they do and elevating what white women do, so we just need to get rid of those typologies. Get rid of the emotions. It is leading us down a bad path.

I thought that both critiques were accurate. But at the same time, I thought that the conclusions were right, but the recommendations were wrong. We need to be more sophisticated, more complicated about emotions, more emotional not less. This really applies to care work: we need to talk about emotion as a skill, as a power.

I hope that people read that and see it as an invitation to complicate and expand upon what we know about what emotion does, how care workers use it. I want a lot of other questions answered about emotions. I want a more variegated map. A more complicated landscape will help address and get us beyond these very valid critiques by the critical race and the critical disability scholars.

EF: At the end of the book, you talk about a social movement for connective labor. How do envision that?


AP: It has got to happen on a cultural level. That is number one.

Then I’m thinking about the realms of business, commerce, and organizations, and I point to practices that employers can adopt to enable better connective labor, and regulations to make AI more transparent, but also to protect relationships at work. Is it too much to say that there is a war going on? Probably. You could say there is a battle. A battle for our souls.

But that battle has not yet been won. People are still worried about what I call a depersonalization crisis, about loneliness and our social losses, thinking about how to prioritize connection. The people I meet and talk to, they are excited and interested in technology, but they are also worried about it. So, the battle is not over. icon

This article was commissioned by Caitlin Zaloom.

Featured image: Allison Pugh, courtesy of Princeton University Press.