Preparing for this conversation about the class structure of the tech industry, I came across this passage, which opened a brief history of engineering unionism published in 1969,

Are engineers professional workers, who identify with management and find unionism and collective bargaining distasteful and damaging to their professional status and image? Or are they employees who, regardless of the nature of their work, share common problems which they are likely to solve only by combining their economic strength? These basic questions, along with the relationship of organized engineers to technicians and production workers and the use of the strike weapon, have plagued the organization of engineers . . . since the movement appeared.

It was striking to me how closely this expressed the scope of our discussion, despite being published over half a century ago. It also emphasized to me the fact that, while the tech workers’ movement is still only a few years old, there is still useful history that we can study to help inform our work today.

You and I started talking about this subject together after you spoke to a meeting of the Tech Workers Coalition about your Logic magazine pamphlet “The Making of the Tech Worker Movement.” In that piece, you raise a lot of the same questions posed by the excerpt above. To help find an answer there, you draw on some prior work by writers like Barbara Ehrenreich, Erik Olin Wright, and others.

Let’s start with exploring why these particular questions are so urgent and practically important, and what history and political theory can bring to the table.

The passage you quoted poses the following question. How should one understand the social position of nonmanual salaried workers, the kind we typically call “white collar” or “professionals”? In advanced capitalist societies, such workers compose a significant portion of the workforce. And they perform a lot of different kinds of work: they include software engineers, of course, but also teachers, nurses, lawyers, journalists.

What is their class experience? How are these workers classed?

This is actually a very old question. Marxist theorists have been struggling with it since the late nineteenth century, when white-collar workers first began to proliferate with the consolidation of the modern corporation and the rise of modern management. Broadly, Marxists have come to two different conclusions.

The first is that such workers can be absorbed into existing Marxist class categories. Thus André Gorz describes this layer as “the new working class,” while Nicos Poulantzas talks about “the new petty bourgeoisie.” The second is that such workers represent a new class category — for instance, Barbara and John Ehrenreich’s “professional-managerial class.”

When studying this debate, it’s not particularly useful to see one intervention or another as “right” or “wrong,” full stop. They flow from specific conjunctures and have to be placed within their context. Class is a process. Classes are being continuously remade and unmade through the course of capitalist development. The Ehrenreichs and Gorz and Poulantzas and others were observing particular moments in that process. So we can mine their ideas for insights that might help us make sense of contemporary processes of class formation, but we have to do most of the work ourselves, because our moment is different from theirs.

Now, why do this work? Why does it matter? The reason that Marxists have spent more than a century trying to theorize what we might call the middle layers is because there have always been strategic stakes to the debate. There are important political choices that hinge on the question of how to understand the relationship between those in the middle and the working class, as well as how and whether the former can play a role in a socialist project.

I’d argue that there are two main reasons to take up this line of inquiry right now. The first is that some members of the middle layers have grown increasingly militant in recent years. Teachers and nurses have led the way, though many of these individuals are more properly described as working-class. But other sectors of these middle layers have made significant organizing victories, like digital media workers and graduate students. And the tech industry has seen an unprecedented wave of rank-and-file mobilization.

The second reason is that a surprising number of the younger, more downwardly mobile members of the middle layers have become socialists. They aren’t the only ones, and they tend to be overrepresented in mainstream media accounts of the new left movements. But they have contributed to the revival of socialist ideas and the growth of socialist organization. Taken together, these two developments suggest that something interesting is happening with the class dynamics of workers who have traditionally been thought of as white-collar or professionals.

This question is especially relevant in the context of the tech industry. The industry’s workforce contains a multitude of occupations that include gig, warehousing, and contract workers, as well as a significant “middle layer” of technical and white-collar, nonmanagerial employees like software engineers and designers. Obviously, the stratification between these occupations affects the growing tech labor movement. The division not only stunts organization and political momentum, but also distorts the identification of common interests and grievances.

Historically, unions have found the tech industry to be challenging terrain. Employers have always resisted unionization, of course. But a more “common sense” explanation is that workers haven’t felt that unions were necessary. A recent New York Times article posed the challenge that many tech workers “see their bosses as friendly peers,” and that “highly compensated engineers may see themselves as independent operators who have plenty of leverage on their own.”

This insight is instructive, but it doesn’t offer the full picture. Independent bargaining can only effect so much and has little effect over conditions that are generalized across workplaces. Exhaustion is widespread among white-collar workers in tech, for example. A study conducted by Blind, a chat application that allows employees to communicate with their coworkers anonymously, last year found that nearly 60 percent of tech workers felt “burned out.” Concerns over discrimination in hiring and career advancement, as well the limited say that workers have over their work, are also common.

Employers have long lamented the leverage a tight labor market gives to white-collar engineers and other employees. An industry spokesperson writing in the 1960s called programmers “the Cosa Nostra” of the computer field, whose ability to use the threat of quitting unless certain conditions were met posed an existential threat to the future of tech. The business literature in these early days illustrates employers’ desperation to bring order to an “uncooperative,” “stubborn,” and “lazy” workforce. A review of one early management text reassured its readers that “There is a vast amount of evidence to indicate that writing — a large part of programming is writing after all, albeit in a special language for a very restricted audience — can be planned, scheduled and controlled.”

In the decades since, the practices of management in the tech industry have been refined and altered. I had a conversation with a former supervisor of mine a while back, who described how, in the ’90s, “a single software engineer could take down a company if they wanted to.” That’s no longer true today.

This begins to get at the process, as you described, where classes undergo a process of being “remade” and transformed, and where employer pressure from above generates conditions that make militancy and organization possible, even among those occupations located “in the middle.” This creates the potential for unity and solidarity across a starkly divided workforce.

You’re getting at something that’s particularly important to emphasize for people who are not familiar with the tech industry, which is that not every white-collar tech worker is a senior software engineer at Google. There are many office roles within tech firms that are not highly paid, especially those that are perceived to be less technical. And there are many tech firms with difficult working conditions, where “crunch” hours are common.

Finally, increasing numbers of white-collar tech workers are subcontracted, from content moderation and data labeling at the lower end to product design and software development at the upper end. Such workers typically earn lower pay, receive fewer to no benefits, and endure precarious employment.

Even so, I don’t think it would be accurate to characterize most white-collar tech workers as proletarians. There are proletarian aspects to their class experience — and these aspects vary widely by kind and intensity — but the relations of production lived by an Amazon product designer are considerably different than those lived by an Amazon warehouse worker.

This is where I find Erik Olin Wright’s theory of contradictory class locations useful. Wright proposed that the class condition of the middle layers is a mix of both proletarian and bourgeois (or petty-bourgeois) elements.

Take a software engineer at a major tech company like Google. On the one hand, they are likely to enjoy a high salary and a significant degree of autonomy in their workplace. They are also likely to have substantial labor market power, which means they can usually find another job without too much trouble. Finally, they may hold considerable equity in the firm through stock-based compensation.

These are the bourgeois (or petty-bourgeois) elements in their class position. Management often places great emphasis on these elements in order to foster a sense of identification between employees and executives — which is why, per the New York Times quote you cited, many of these workers “see their bosses as friendly peers.” Moreover, these elements, as interpreted through the lens of dominant industry frameworks such as the “Californian Ideology,” have traditionally encouraged such workers to view themselves as entrepreneurs and fellow (or future) owners rather than workers.

But the bourgeois dimension of this stratum’s class experience exists alongside, and in tension with, a proletarian dimension. Even within the relatively privileged case of the Google software engineer, one finds proletarian elements. Foremost among these is the exclusion from real decision-making — namely, the investment and production decisions of the firm. The power to determine what the firm makes, how it is made, and whom it is made for is held by management. In this sense, the Google software engineer is a worker like any other.

Of course, this has always been the case. But it began to be perceived more acutely in the period following Donald Trump’s election in 2016, as concerns about tech executives’ willingness to build tools to help implement the new administration’s agenda — particularly around immigration and the so-called “Muslim registry” — fueled a new mood of moral urgency around technology’s capacity to cause harm. As workers began trying to mitigate these harms by claiming more control over what their firms were building — by organizing against contracts with the Pentagon or ICE (Immigrations and Customs Enforcement), for example — they met resistance from management and discovered that they were, in significant respects, workers.

In Wright’s formulation, those who inhabit a contradictory class location are pulled in two directions. They can focus on the ways in which they are bourgeois, and identify with the capitalist class; or they can focus on the ways in which they are proletarian, and forge alliances with the working class. What we’ve seen in the past few years is that a number of white-collar tech workers, through the process of engaging in collective action at the workplace, have foregrounded the proletarian elements of their class experience. They have come to see themselves as workers. They speak the language of class and use the techniques of labor organizing. Through struggle, they have pushed their contradictory class location into a more proletarian alignment.

I think it’s right that probably most “technical” white-collar workers are located in a contradictory class position. While I don’t think we should overstate any particular “bourgeois” elements (the vast majority of equity or stock options given to workers of any given tech firm outside of a major employer don’t often amount to much, if anything at all), our experience is unquestionably distinct and relatively privileged. But those advantages are extremely limited by the power of employers, who have distinct class interests against both “technical” and nontechnical workers within any strata.

Technical workers are able to leverage the conditions of a tight labor market to our advantage to secure jobs with generally higher salaries, benefits, and conditions. That reduces the urgency and appeal of more collective strategies for pursuing material gains and security. Simultaneously, however, lacking formal power within our workplace, technical workers remain subordinate to our employers’ authority.

This dilemma is not unique to modern tech workers. This has been more or less the general experience of American engineers and “technical” workers since the growth of modern industry. Unlike more traditional “professional” occupations, such as doctors and lawyers, engineers and other “technical” workers are a mass occupation, subject to the same employers’ authority as any other worker. Similarly, technical workers also lack the ability to control competition on the labor market through credentialization.

Various conditions and experiences pull technical workers between the poles of “independent professional,” on the one hand, and exploited laborer, on the other. Engineers and other technical workers have been able to find common cause with the labor movement and view collective bargaining and militant strike action as consistent with their interests in the past, especially when faced with greater material uncertainty.

Throughout the Great Depression, engineers and other technical workers joined unions in vast numbers. Following World War II, approximately 10 percent of American engineers were unionized. But by the 1960s, membership in engineering unions declined dramatically, as economic prospects improved for nonunion engineers and the union movement in general began its long period of retreat.

As you pointed out, the contemporary advancement of collective action among white-collar tech workers has revealed the power inequality in the workplace. That inequality is at the root of many of the notorious injustices that characterize the modern tech industry, including the moral bankruptcy of employers who profit from military and police violence, the rampant sexism and racial discrimination that plagues the industry, and the hyper-exploitation of workers at practically every level.

Recognition of that inequality creates the potential for solidarity to be formed between the different strata of workers across the industry who, while faced with radically different conditions, do share a common interest in organizing. While many challenges exist in organizing that common ground, many possibilities exist as well. The resources and support of unions such as the Communication Workers of America and the Office and Professional Employees International Union can certainly help new organizing initiatives in the tech industry.

I’d like to pull on a thread you introduced about professionalization, which might give us a way to conclude. You drew a distinction between “more traditional ‘professional’ occupations, such as doctors and lawyers” and “engineers and other ‘technical’ workers.” This brings us to the specificity of tech.

We can talk about the middle layers in general, but this is a capacious category with enormous internal differentiation. The processes and practices of class vary widely within it. And to understand the distinct contexts where class takes place, it helps to look at history — we have to understand how occupations have been constituted, and the struggles that have shaped their formation, in order to understand their particular class mechanics.

In this regard, the most important thing to know about tech is that computer programmers never fully professionalized. Programmers first emerged in large numbers in the 1950s and 1960s. These were the decades when computers were becoming mainstream, migrating from military labs into corporate offices. Demand for programmers soared as a result — and men poured in. Previously, programming was primarily done by women. It was considered a mechanical, uncreative job and was poorly paid. With the mainstreaming of computing, women were pushed out, and programming became a predominantly male occupation, now redefined as creative, intellectual, and prestigious.

However, even as the status and compensation associated with programming increased, programmers never became proper professionals in the way that doctors, teachers, and lawyers did. The reasons are complex and are explored at length in Nathan Ensmenger’s excellent book, The Computer Boys Take Over. But the bottom line is that the kind of institutions that implement and oversee professionalization never materialized for programming: there is no professional association and no body of law that defines what you need to know, and what you need to do, in order to be considered a programmer.

There are computer science departments, of course, and many programmers study computer science in college. But a computer science degree isn’t really a professional credential in the way that a law degree is; as one colleague once put it to me, it’s a bit like studying architecture and then going to work as a carpenter. There are also plenty of programmers who are self-taught. But I’d argue that all programmers are at least partially self-taught: self-education is a big part of the job.

The reason all of this is relevant for class dynamics within tech is that the absence of professionalization has created an enduring ambiguity about what exactly a programmer is. Is a programmer a professional? An entrepreneur? An artisan? A worker? These are not merely theoretical questions; they are questions of real consequence for class struggle, questions that are in fact posed and answered through the activity of class struggle.

From the start, tech’s ownership class was ambivalent about professionalization — it might make programmers easier to manage, but it also could also foster a sense of occupational cohesion that might encourage a bid for greater autonomy and even unionization. If they were ambivalent about the idea of programmers as professionals, however, tech executives were absolutely committed to ensuring that programmers did not see themselves as workers.

Various managerial techniques were designed to prevent that identity from developing, techniques that continue to this day: an informal office environment, a team-based and relatively horizontal work culture, stock options, generous on-site amenities. These aren’t just about keeping employees happy; they work to conceal class divisions by blurring the lines between management and the rank and file. They are attempts to occupy the empty ground left by the absence of professionalization by cultivating an identity that benefits the ownership class.

This is precisely the ground that white-collar organizers are now contesting. Their insistence that tech workers are workers is an argument not only about how employees of these firms should view their structural position within the workplace. It is also an argument about where their real source of power is located.

Their power doesn’t come from some imagined partnership with management — such a partnership doesn’t actually exist, as the slightest degree of contestation reveals. Neither does it come from their social influence as professionals, as the institutions required to generate this influence were never built. Their power comes from their leverage over the spaces where profit is made. As workers, they propel the productive process; as workers, they can bring that process to a halt.

Along those lines, I think that the tech worker movement has validated two points: one, that even technically skilled white-collar workers have a common cause with workers of other occupations against their employers; and two, that it’s possible to organize these workers on the shop floor. Practically demonstrating these facts was a crucial first step, because it presents a real alternative to simply relying on the individual advantages that come with having “professional” technical skills.

But now it seems we’re coming to the end of the initial phase of the tech worker movement. So the big question is what should come next.

A recent piece by Carmen Molinari at Organizing.work points to the kinds of conversations we should be having. In it, she argues that a hyper focus on organizing around ethical issues in tech constrains the potential for growing movements in the workplace. I largely agree with that argument, but regardless of how one feels about it, I think we’re at a moment now where we need to discuss how we can effectively extend our movement and deepen its reach beyond the initial core of activists and organizers. Doing so will require thinking carefully about which issues and grievances are most deeply felt among people in the different occupations and sections of the industry.

The rapid rate of growth of this movement has perhaps made it difficult to carefully assess what’s taken place and what’s ahead. A great deal of practical experience has been gained among workers and organizers in just a few years, but there’s been little room to share what’s been learned. Developing a means of exchanging these lessons and experiences, and of putting workers in touch with one another between the disparate networks and organizations that have been established, will be essential. The research project Collective Action in Tech recently announced a publishing initiative along these lines and put out a request for contributions.

While there’s a lot of hard work ahead, there’s exciting potential in the future of the tech workers’ movement. It’s already showed that the industry is not “unorganizable,” and that workers at every level can be organized to build power to give themselves a collective voice at work. Extending these early successes is especially important, since tech workers are the strongest line of defense against the threat posed by the large tech employers.