Thank your for your subscribe
Oops something went wrong. Please check your entry

Digital labor and the limits of critical thinking

In the analysis and understanding of the Internet, the role of sociology needs to be defended: facing a widespread tendency to economization, it offers resources and references that allow to change perspectives and view practices from another angle. This difference is healthy. But let's face it: in the way it conceives the Internet, the sociological debate is itself beset with contradictions. Digital labor provides an illustrative case.

April 2017
lire en français
lire en français
Executive Summary

Since its pioneering days, the Internet has undergone significant changes that have impacted researchers, essayists, scholars and public intellectuals. In the 1990s, the Web was analyzed based on the works of Deleuze and Guattari. Nowadays, we refer to more diverse authors: Adorno, Horkheimer, Foucault, Dallas Smith. The concept of free labor, encapsulated by the expression “digital labor,” appeared in this precise context. In a critical perspective, it could be summarized as: “Internet users who work for free for platforms that keep getting richer.”

Digital labor describes the way clicks, exchanges, interactions, content creation, Internet communities and users also produce value – a value that is subsequently extracted and monetized by platforms. But the term was applied to other fields, such as collaborative economy, even when contracts and payments were involved. However, the generalization of the concept of “free labor” to service economy is somewhat confusing and probably undermines the very interest of this notion.

Critical thinking and its limitations

Changes of intellectual framework in the perception of the Internet cannot be understood without taking into account the morphological transformation in the way we use Internet. The Web quite simply reconfigured the demographics of its audience, triggering a major morphological transformation. The 1990s narratives applied to a small and well-defined fraction of the population, both from a social, territorial, cultural and gender point of view. That’s where the imaginary of the cooperative community came from. This model could only be deceived by the transformation of the audience, with very different uses and practices than those predicted by the original model.

It’s always somewhat disturbing when a positive narrative turns into a critical one: implicitly, this suggests that new audiences (working class, women, social and geographical diversity etc.) don’t have the inventive creativity or genius of pioneers. The emergence of this critical narrative reflects a form of subconscious class ethnocentrism, with social disregard for the shabby contributions of small bloggers, Facebook likers and mediocre contributors. This implicit disqualification contributes to the transformation of participatory productions of skilled, white male Internet users of the 90s into selfish and narcissistic commercial contributions of working classes today. This question is never raised, at least by those who convey the critical narrative of the Internet.

However, one could argue that the worm was in the fruit from the very beginning. What changes is the way we regard these practices. People who posted content on Flickr ten years ago were seen as extremely creative, while those who post selfies on Facebook are viewed, with a hint of disdain, as “alienated subjectivities, dominated by the market, led by vain narcissism,” etc. Very similar practices and behaviors are labeled quite differently!

Incidentally, as a theoretical digression, while the first theorists of the Internet had described the community in terms of empowerment, referring to Deleuze and Guattari, they also used concepts devised by Italian Marxism, including operaismo (workerism), from such authors as Toni Negri or Yann Moulier Boutang – the latter being one of the best interpreters of this period, with his idea of positive externalities.

In simple terms, these studies suggest that capitalism is lagging behind the productive energies of the so-called “multitudes.” Therefore, it can only chase the externalities produced by the interactions between these multitudes in trade, languages, production etc. The connected society produces a value that exceeds the capture capabilities offered by a rent capitalism that constantly seeks to make profit from the exchanges between Internet users.

However, this complex analysis resulted in a much less subtle interpretation model in which capitalism encompasses the entire society, leaving no place for externality: the entire society is subject to the capital. Dallas Smith played an important role in the development of this idea, which borrows from Horkheimer’s criticism of entertainment in the 1950s. In any situation, whether work or leisure, there is no exteriority in the production of overall value by capitalism. In this perspective, whatever they do, Internet users work for the capital. Even on Wikipedia! Google uses Wikipedia for its small text boxes... and therefore, all wikipedians work to enrich Google.

This has become a central component of the concept of digital labor as we define it today. But the critical vision in which it takes place is so comprehensive that it ends up being ineffective. At the very least, it makes things more difficult for the sociologist, for a theoretical reason: “people think they are writing an article for Wikipedia but in reality, they are working for Google” is a true and relevant statement from the standpoint of political economy. But for a sociologist, it’s completely useless. Describing actors that “believe they are doing x but are actually doing y” doesn’t help explain the experience of individuals nor even the ignorance of their own practices. Where is the point of alienation ? The description of the overall operating mechanism is relevant, but it doesn’t explain the motivations of these individuals, which is one of the crucial tasks of sociology.

In addition, the story is complicated by another problem: narratives produced by theorists of the Internet have performative effects on the way we produce services, conceive them, etc. Hence, reducing all user interactions to their economic aspect only transforms Internet users into calculators. Therefore, sociology has one responsibility: to support the idea that the players’ experience is not based only on economic calculation.

Interest and calculation

Which model did the Internet pioneers use to explain the commitment of Web communities? The main idea is that their contribution didn’t have a precise purpose. Most community studies confirm this fact.

In the early 2000s, economists joined the debate and brought in the theory of rational expectations, divided between defenders of intrinsic motivation and extrinsic motivation. When the drivers of contribution and interaction are based on fun, passion or excitement, we refer to them as internal passions.In other words, the direction and value of an activity is not independent from its goal. Orientation, reward and justification are closely involved. They are “intrinsic.” However, if action is driven by the prospect of obtaining a high grade or making money, then motivation becomes extrinsic: this is what economists are interested in, generally speaking. In the early 2000s, Tirole and his co-author Josh Lerner published a landmark article on the subject: “The simple economics of open source.” They tried to solve a puzzle: what exactly motivates people who code all week long for work and keep on coding during their evenings and week-ends, out of sheer passion? Extrinsic and intrinsic motivation overlap around a same skill. These reasons rely on activity itself and not on something external. But economists don’t like intrinsic motivations! Tirole and Lerner prefer extrinsic motivations, suggesting that “volunteer” developers benefit from their activity: they acquire a reputation in their community groups that they will eventually “sell” on the labor market. Intrinsic motivation is transferred into extrinsic motivation.

I strongly recommend a book on this subject by Geoffrey Brennan and Philip Petitt, theorist of republican philosophy: The Economy of Esteem. The notion of esteem allows them to describe something which is both intrinsic and extrinsic, embodied by subjectivity, passions and affects. It can be capitalized, in part, in the form of reputation, recognition or merit. Merit is the pivot point of this interpretation. But Brennan and Petitt also note that esteem is not subject to a finalized calculation. If our actions are deliberately guided by our desire to gain esteem, the purpose of this search destroys the meaning of the esteem we gain. These authors highlight the non-teleological purpose of esteem: real esteem is only acquired without seeking it.

This feature of esteem sums up the fragile compromise that supports the description of many activities within web communities. After studying many digital communities (wikipedians, amateur cooks, cultural or political enthusiasts), I discovered that they all share three common features.

First, communities always invent an artifact to represent this intrinsic/extrinsic material and visualize it through merits. Every Internet community has its own ranking system, karma points, list of best posts... in short, an aggregated representation of judgments, awards and merit that support the internal hierarchy. This system supports the incredible violence, often described by observers, that divides people in two classes: the deserving and the undeserving. We are witnessing the quick and sustainable implementation of meritocratic trainings, in which we tend to overlook the initial status of individuals and focus on the input and care they provide the community and that will bring them prestige. It is fascinating to observe the development of these metrics and of their small artifacts within communities.

Second feature: in doing so, communities have contributed to the individualization of expression, which is the main driver of the social dynamics behind the Internet. We are willing to ignore the statutes and diplomas of individuals (even if these resources will be useful to them) but in exchange, they must reinvent themselves, be expressive. The condition of expression in a community is a balance between assertions of identity and the constant need to mitigate the risks this individualization may pose to the community by producing too much celebrity, notoriety or importance. Hackers even adopted the following ethics: within a community, important people are forced to show humility. For example, on Wikipedia, the following rule defines the wikidove spirit: “Don’t bite newcomers.” The greater we are in the community, the more we are responsible for taking care of new entrants. This ethical stand partially counteracts the fact that capital acquired within a community via an externalization into metrics which is then exported towards other markets could produce both fame and strategy and thereby distort the community. Gabriella Coleman also observed this phenomenon in the Anonymous community.

Third feature, these communities produce common good; and common good is never the sum of individual actions. The miracle of Google’s algorithm is that it calculates the sum of all individual actions projected on the Web through hyperlinks. However the produced result is far more valuable than the sum of its parts. This decoupling between individual actions and the common good is operated through some kind of mysterious transubstantiation! And one could argue – this position is supported by those who evoke digital labor – that Internet users should be compensated for an activity whose value will be divided into smaller parts and redistributed as dividends... But this overlooks the fact that people’s individual data are absolutely worthless. Value derives from the fact the aggregate data produces something completely new. The economic calculation that serves as basis shows that our clicks produce value for Google, but not for the individual. Value emerges from this shift from the individual to the collective.

Three shifts

We have been grappling with this idea for nearly fifteen years. This model has shifted and I will focus here on three axes of this shift that explain how we transitioned from the valuation of esteem and merit to the idea that, in fact, we are calculating – not only as far as platforms are concerned but also, increasingly, on the side of Internet users. In order to make a sociology of communities, we need to gauge this calculator strategy and understand how it became part of the imagination, subjectivity and actions of Internet users.

The first shift was suggested to me by the economist Michel Gensollen. He formulated a strong idea by showing that the market on the Internet took advantage of the positive externality of non-market sites. Typically, in the 90s, people were surfing because there were many personal sites and non-commercial blogs. Free content is what attracted them most... So companies had an interest in running after them! But this balance of power has completely changed, in a double sense. Today, as far as websites and blogs are concerned, rankings are mostly dominated by commercial sites. Bloggers are strongly encouraged to blog on the NYT rather then typepad etc. In other words, to take advantage of the positive externality of commercial sites in order to reach more readers. The balance of power is therefore reversed.

However, and this is where demographics and therefore sociology become crucial, markets are interested in non-market activities on social networks, with mass public, working-class demographics, both heterogeneous and varied. In short: Facebook. These spaces in which people engage in non-market activities (namely: organizing their sociability, gaining reconnaissance and reputation, building their lives and identity) have become the target of the advertising market because this is precisely where non-market externalities can prove beneficial for the market. In reality, something crucial is at play because it is a completely different audience demographics that is at stake.

The second shift is due to the fact that the model described by Brennan and Petitt is obviously very fragile. We don’t know how this “esteem” holds, both inside and outside the subject. There is a high risk of transforming esteem into a capital and hence, a motivational resource. Clearly, all the devices that helped Internet users calculate their reputation have transformed merit and esteem (both internal and external values) into increasingly external values. Ultimately, individuals have been transformed into strategists of their own reputation. This slight displacement is what fuels the criticism against the strategic action of Internet users, while their predecessors were presumably both disinterested and deserving.

Third shift: calculators, who were the embodiment of the mathematical and machine-like collective intelligence of Internet users, have become receptacles within which economic interests preempt collective intelligence. The way they are represented has been radically changed. When it first appeared, Google’s PageRank was the very embodiment of the pioneering spirit. Its computing power reproduced exactly the way collective authority was calculated based on the reputation sent by Internet users through their hyperlinks. Today, nobody would say this and we will probably mourn the disappearance of PageRank. But it has become a driving force in the transformation of Google’s economic interests, a tool that acts on Internet users to transform them – and calculate them. The metrics that helped the community represent the internal distribution of its hierarchies was reinterpreted as operations serving the commercial interests of platforms that extract the value of the work performed by Internet users.

In conclusion, as a theory of exploitation, the concept of digital labor is apparently very accurate and there is no reason to challenge it. It works just fine for any economist who doesn’t need to take into account individuals and subjects – someone who doesn’t need internal criticism. But humanities cannot overlook an analysis of the experience of Internet users. Unless we present a theory of alienation that explains why this experience is an illusion, an artifact produced by an exploitation mechanism that would make it unnecessary to ask people what they think since the reasons they give are illusory. How is it that people who firmly believe they are working for the sharing economy, who are promoting self-development, self-esteem and recognized merit… are actually working for the commercial sector? This brings us back to the questions that I mentioned at the beginning of this intervention.

One of the thinkers who would help us overcome the limitations of this criticism is Michel Foucault. And we could even pit two different Foucaults one against another.

The first Foucault would point out the subjectivation process by which we produce subjectivities. Narcissism, ignorance and alienation all contribute to creating a value that is subsequently captured by businesses. Foucault developed an idea often used to confirm this explanation: something slips within the web-surfer’s subjectivity in order to build it, conform it and rationalize it... This is pointed out in all the debates about algorithms, filter bubbles, censorship. Some even describe Facebook as a concentration camp! They evoke a prison logic, control, closure... devices that take control of the subjects, muzzle them from the inside.

But on the other hand, Internet users eagerly continue to work for free. They don’t feel locked up inside one single platform – they have 25 open windows at once! The motivation they give to their actions is in line with forms of experience that have little in common with the somewhat unrealistic descriptions that are made of them. Those who claim we have less information than in the past have forgotten what the pre-Internet era looked like! Given the reality of facts, it is fair to say that these descriptions are so strong and so marked by a model of manipulation and alienation that occurred well before the Internet that they reproduce an old pattern. Hence, the need to use a second Foucault, the one who wrote between 1977 and 1979 about biopolitics and, more interestingly for the issue at hand, “governmentality.” The latter, unlike power, doesn’t apply to players but to the rules of the game itself. With governmentality, we produce an environment and in this perspective, players are simple statistics oscillations. We no longer focus on the criminal. Instead, we try to regulate the possible effects of crime on society in order to optimize the way it works, keeping in mind that everyone is more or less a criminal.

Given the prospects opened by the second Foucault, one can consider that the policy of platforms is to produce environments in which individuals are granted all necessary freedom, without necessarily prescribing anything. As a result, platforms would intensify our desire for freedom and autonomy. They would produce an environment, rather than a subject. This is exactly what is happening to the Internet and the economy of platforms.

A real criticism would be to ask how platforms can produce environments within which people can genuinely interact, exchange and share – without the platform exerting any limitation, constrain and discipline over their actions. The platform only takes a share of the value produced by the collective effect of the environment in this organization. The policy of these platforms could therefore be summarized as such: create the conditions for more autonomy and freedom, adapted to forms of expressive individualization within society; avoid any semantical or significant impact; avoid any identification of calculation; do not provide any guidance whatsoever. Simply take advantage of the collective effect of user interactions.

This article was written after a presentation in Monique Dagnaud and Olivier Alexandre’s EHESS seminar, The Californian Model. Other pieces will follow.

Dominique Cardon
Associate Professor of Sociology at Medalab, Sciences Po, Associate Researcher at Centre d'études des mouvements sociaux, EHESS