Social Media,Delinguistification and Colonization of Lifeworld

Abstract

The article critically investigates, from an interdisciplinary perspective, how the current evolution of social media—like social network sites—interferes with the balance between private, commercial, and public space. We build on the concepts of lifeworld and systems, developed in Habermas’ theory of communicative action. The discussion is supported and enriched by the work of Feenberg and van Dijck, integrating insights from Science and Technology Studies and media studies. Technology philosopher and critical Science and Technology Studies scholar Feenberg introduces technology as a steering “medium” that delinguistifies and possibly colonizes the lifeworld by reinterpreting media sociological perspectives of Habermas, Marcuse, Latour, and Callon. In a similar way, media scholar van Dijck analyses the transition from human connectedness to automated connectivity in the context of social media. We then illustrate the delinguistification and the colonization of lifeworld with a systematic analysis of the contingent evolution of Facebook as one particular case in social media. We focus on three specific artifacts in Facebook, framed as obligatory passage points: EdgeRank, Sponsored Stories, and Gatekeeper. Each of them gives an idea how the private space is subsumed under the commercial space and how the colonization reconfigures the public space in social media like Facebook. In this sense, we complement the political economy analysis of prosumer commodity with the action-theoretical autonomist approach of immaterial labor, highlighting new potential threats of the current social media development.

Introduction

This Special Issue deals with how social media affect the balance between private, commercial, and public space. Our article takes a critical look at how the current evolution of social media—like social network sites (SNS)—interferes with this balance. SNSs are a technology that facilitates social interaction between people in a very particular way (boyd & Ellison, 2008). We employ the concepts of lifeworld and systems, as developed in Habermas’ (1987) theory of communicative action, to theoretically discuss what is happening. This is further illustrated by our case, Facebook. The discussion is supported and enriched by the work of Feenberg (1996) and Van Dijck (2013), integrating insights from Science and Technology Studies (STS) and media studies. In this way, we are able to highlight potential threats of the current social media development, complementing the political economy perspective of exploitation (Fuchs, 2012b) with an action-theoretical autonomist approach.

The structure of this critical analysis of the meaning and the role of SNS (like Facebook) on the everyday social life of people consists of two parts. The first part, “Theoretical Background,” discusses the interdisciplinary theoretical framework based on a literature review. We start from a reinterpretation of Habermas’ (1987) colonization thesis, by including technology as a steering “medium” next to money and power. This builds on the critical theory of technology by Feenberg (1996), which integrates constructivist insights from STS with the media sociological perspectives of Habermas and Marcuse. We then explain how automated connectivity (Van Dijck, 2013) and profitability fit in this framework. Human connectedness is not only stimulated and reused to create an incentive to participate. The process itself seems to disappear from public space. The second part, “Colonization of lifeworld,” illustrates the main arguments in part one with an analysis of the contingent evolution of Facebook. We frame these changes as obligatory passage points (Callon, 1986), which prioritize connectivity in order to grow a loyal user base. Facebook increased profitability by innovating and increasing their advertising means since 2007, for example, by introducing EdgeRank and Sponsored Stories. We also observe how Facebook is designing connectivity into its interface by way of the Gatekeeper tool. The central question is how Facebook as a constantly evolving technology penetrates and possibly colonizes the lifeworld of its users. The developments in this particular SNS can be seen as a prefiguration of the future evolution of social media, and how the balance between private, commercial, and public space are shifting.

The theoretical added value of the article revolves around the question: can we apply Feenberg’s critical theory of technology to critique social media like Facebook, complementing prosumer commodity with changes in communicative action? For this, we more systematically analyze the role of technology as a medium (in Habermas’ sense) with regard to the interplay between systems and lifeworld of social media users. This means that the exploitation of the prosumer commodity (Fuchs, 2012a) is no longer limited to users as workers but also as beings (in)capable of communicative action. ¹ Hence, the ever-changing affordances of Facebook are not a mere expansion of capitalist logic, but also illustrate the shift from connectedness to connectivity. This means that Facebook limits users’ communication to those aspects that drive profit and connectivity. In that way, private space is subsumed as commercial space and as a result, this colonization reconfigures the public space in social media.

Theoretical Background

Following Kaplan and Haenlein (2010), we frame social media as “a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0 and that allow the creation and exchange of user generated content” (p. 61). Web 2.0 was the result of a brainstorm between O’Reilly and MediaLive International in order to find a name for their conference about emerging business models and their technologies (Graham, 2005). The concept of Web 2.0 was created for two reasons. First, to showcase new technologies and second, and more ideologically, to convince its attendants that Internet after the dotcom bubble still had a future (O’Reilly, 2005). Only later on, the term began to contain the possibilities of user participation and empowerment. This went parallel with an increased focus on the notion of consumer as a producer or “prosumer,” as coined by Toffler (1989), despite the fact that prosumption has been around much longer than social media (Napoli, 2010; Ritzer, 2014). However, the economic significance of consumption seemed to increase in Web 2.0 business models.

In the critical political economy literature, this often refers to how users are becoming part of production. This is extensively discussed in the work of Fuchs (2014) who refers to “prosumer commodity.” He defines this commodity as an extension of Marx’ accumulation method. A first commodity, the Facebook platform is created, but this commodity is not sold. Next, users start using the platform for free and become the commodity as they watch advertisements. This commodity is then sold to advertisers.

Fuchs’ perspective represents the classical political economy thinking, which situates prosumption in the sphere of production. Marxist autonomists have a complementary perspective. They define prosumption as part of the consumption sphere, referring to notions of free labor (Terranova, 2004), immaterial labor (Coté & Pybus, 2007; Lazzarato, 1996), and cognitive capitalism (Dyer-Witheford, 2004). They distance themselves from traditional Marxist theory in the sense that they no longer focus on industrialization and the production process as the central and main explanatory factor, yet identify a digital turn. According to Hardt and Negri (2000), we are now in a process called postmodernization:

The process of postmodernization or informatization has been demonstrated through the migration from industry to service jobs (the tertiary), a shift that has taken place in the dominant capitalist countries, and particularly in the United States, since the early 1970s. (p. 285)

Lazzarato (1996) refers to two types of immaterial labor, the first being concerned with the manipulation of cybernetics and computer control. The second entails the creation of subjective standards. The first type of immaterial labor steers the constant interaction between production and consumption where the commodity, consumption of the commodity, and the labor process itself are constantly improved through feedback. The second type of labor provides a different approach to control or influence consumption by generating subjective, shared standards that order the consumption process. This new creation of meaning is found in marketing departments, for setting one product apart from another product in the same sector, for example, by branding (Arvidsson, 2005).

We are particularly interested in the second type of immaterial labor because it requires a common connection or general intellect (Marx, 1857, p. 1442). Immaterial labor draws on this shared understanding or general intellect, which remains autonomous from economic production. This autonomy is seen as an enrichment by the autonomists:

Here one might recognize once again that the instrumental action of economic production has been united with the communicative action of human relations; in this case, however, communication has not been impoverished, but production has been enriched to the level of complexity of human interaction. (Hardt & Negri, 2000, p. 293)

In what follows, we state that by following Habermas’ (1987) colonization thesis, that this subsumption of the general intellect is an impoverishment. The latter is demonstrated by building on Feenberg’s (1996) extension of the theory of communicative action to include technology as a steering medium.

Colonization of Lifeworld

In Habermas’ (2002) theory of communicative action, two worlds exist: lifeworld and the world of systems. The lifeworld is an intuitively and instantaneous resource of background knowledge. Without it, taken-for-granted actions are impossible. For example, we would not use doorbells, because we would not be sure whether people understand they have to open the door if a doorbell rings. Lifeworld is the source of communication and at the same time it is sustained by communication. ² Since it is shared with other humans, it is a necessity to understand each other. It also forms the basis to perform communicative action required to organize our society.

Habermas adapts the notion of lifeworld, originally conceived in phenomenological philosophy by Husserl to designate the background consciousness that functions as our cognitive horizon. Habermas was influenced by Schutz’s (1973) conception of the lifeworld, and he interpreted this definition linguistically. This means that Habermas’ lifeworld is, on the one hand, built from symbolic elements that support communication and on the other hand that communication further supports lifeworld.

When Habermas (1987) introduces systems, he wants to integrate an external perspective that focuses on the materiality required to sustain lifeworld: “Problems of material reproduction are not simply filtered out of this perspective; maintenance of the material substratum of the lifeworld is a necessary condition for maintaining its symbolic structures” (p. 151). The support of lifeworld is carried out in systems where media such as money and power steer actors in a way that no longer requires communicative action: “These ‘media’ make it possible for individuals to coordinate their behavior while pursuing individual success in an instrumental attitude toward the world” (Feenberg, 1996, p. 56).

Coordination should be understood as the division of tasks and roles in order to support lifeworld materially, for example, allocating enough farmers to produce food. Coordination facilitates a higher degree of societal complexity, for example, the introduction of money enabled international trade. But this complexity comes at a price. Media coordination delinguistifies: the organization of systems is no longer legitimized through communicative action in the lifeworld. The evolution of modern societies should then be described as the on-going process where

modern societies attain a level of system differentiation at which increasingly autonomous organizations are connected with one another via delinguistified media of communication [ . . . ]. (Habermas, 1987, p. 154)

When the coordination is delegated to media, it is no longer evaluated or questioned through lifeworld but by media’s inherent logics. These inherent logics become the sole evaluative criterion to expand a subsystem beyond its intended boundary. A boundary should be understood as the intended goal a system was created for. If a system expands and changes things beyond this goal, it exceeded its boundaries.

Habermas created a criterion to halt steering media based on his definition of lifeworld. If a system becomes autonomous, it diminishes communicative action, which is not necessarily a bad thing. But it may also expand beyond its intended boundaries and decrease lifeworld. Since this process occurs without communicative action, it is difficult to notice. If this happens, system colonization of lifeworld occurs. For example, Frederiksen describes medical imaging as colonization because it has changed our lifeworld definition of patients. Before medical imaging, people were sick if they felt sick and visited a doctor. “Modern medicine has turned us all into ‘proto-patients’. Without further ado, we accept that we can become patients any minute, even if we do not experience any symptoms at all” (Fredriksen, 2003, p. 291). Now, everyone is a proto-patient because medical routine checks find new diseases before people experience them. In the next part, “Technology as a Medium,” we draw upon Feenberg to incorporate technology as another coordinating medium to handle social complexity.

Technology as a Medium

Technology—according to Habermas—cannot be a medium because it does not have any communicative significance; it merely relieves physical effort without reducing communicative action. However, bearing in mind the constructivist research tradition of STS, we can no longer conceive technology as a neutral non-communicating medium. The central tenet of STS—as opposed to technological determinism—is that technologies are not neutral, but represent the outcome of social, economic, political, and cultural processes. This interdisciplinary field typically investigates how the technological objects are being constructed. STS has increased its scope over the years, starting with scientific knowledge and expanding to artifacts, methods, materials, observations, phenomena, classifications, institutions, interests, histories, and cultures (Hackett, Amsterdamska, Lynch, & Wajcman, 2008).

The Social Construction of Technology (SCOT), a major subfield of STS, observes three stages in the design of technological artifacts (Bijker, 1993): the first stage of interpretative flexibility with multiple articulations of a technology, a second stage of social mechanisms to close interpretative flexibility, and a last stage where closure mechanisms work in a wider social-cultural milieu. This is illustrated by the evolution of the bicycle. During the interpretative flexibility stage, multiple designs of bicycles existed. Each design solved problems for different social groups. Social groups are connected to particular designs through common problems. During the second stage, a technology is closed: alternatives disappear and with it the various social groups they could serve (Pinch & Bijker, 1984).

As a critical STS scholar, Feenberg also does not conceive technology as a neutral non-communicating medium. He refers to Marcuse’s critique that “instrumental reason is historically contingent in ways that leave a mark on modern science and technology” (Feenberg, 1996, p. 46). Technology philosopher, Feenberg, adapts Marcuse’s instrumental reason and applies it to the particular: if instrumental reason is historically contingent, than particular technologies are also contingent referring to interpretative flexibility and closure. A choice is made for the technology that serves the controlling group better than its alternatives (Feenberg, 1996). This is a communicative action albeit delinguistified. It describes a world where a particular technology should benefit its owners more than its users.

Feenberg further substantiates his argument by illustrating how technology has the same components as the other two media: “with money it is exchange value, that is, money demands an equivalent; power yields binding decisions which demand obedience; and technology generates what I call, following Bruno Latour (1992), ‘prescriptions’, rules of action which demand compliance” (Feenberg, 1996, p. 59). Prescriptions are communicated through particular technological design decisions. For example, an SNS like Facebook prescribes compliance in the registration webpage, telling users what to do (Figure 1). This means that a Facebook employee could have done the task, but a server does it instead.

Figure 1.

Facebook registration process.

Hence, the speech act or communicative action “hand over personal information” is handed over to the server. In Actor Network Theory (ANT), another major STS subfield, this is called “delegation” (Latour, 1992): the act of delegating an instruction for another human to an object. As depicted in Figure 1, the author of the delegation, Facebook, inscribes an instruction as a strategic action in the object of the registration webpage. Facebook does this with a certain program in mind. A program is the imagined action another actor, the registering user, needs to perform. In this case, the program is “hand over personal information.” Actors may not follow the prescribed program, by neglecting to disclose their email address. Latour (1992) calls this an antiprogram.

The prescribed behavior of users of objects is not always followed; this is why most technologies anticipate multiple antiprograms in order to enforce one desirable outcome. In this case, users are first told to fill in information, if they forget to do so, text boxes appear to remind them. If this still does not prompt the user, the registration process simply comes to a halt: the sign-up button does not take users to the next page and each information field turns highlighted with an exclamation mark as shown in Figure 1.

There are means to decrease the chances of antiprograms occurring through obligatory passage points. An obligatory passage point is a situation where actors are forced to follow the proposed program inscribed in objects because they are driven by a personal interest. Latour explains this through a hydraulic door closer. This object is attached to a door and its program is to close the door after opening it. A person, who opens the door, often has the antiprogram of leaving the door open. But this is impossible with a hydraulic door closer that receives potential energy from every actor who opens that door: “No matter what you feel, think, or do, you have to leave a bit of your energy, literally, at the door” (Latour, 1992, p. 158).

Feenberg’s critical theory of technology based on Habermas, Marcuse, and Latour explicates the delinguistification by technology as a medium. ³ This takes place by prescribing programs, delegation, anticipating antiprograms, and setting up obligatory passage points. From the moment coordination is delegated to a technology, it steers an actor’s behavior without the need for communicative action and authorship seems to disappear. ⁴ This becomes problematic when it presents a(n) (undesirable) prescribed behavior while giving no alternatives. Technology can enforce compliance by resisting to function unless it is used correctly, that is, according to the prescribed program.

According to Habermas, problems arise when systems based on delinguistified media of communication expand beyond their intended goal and thereby colonize the lifeworld. The medium of technology can encroach on the lifeworld by prescribing a shared understanding or expectation, for example, about disclosing personal data in social media. Our premise is that social media in the commercial space are often configured in such a way that their prescribed program is aimed at immaterial labor. In that way, social media technology can lead to colonization of the lifeworld. In order to illustrate this with the case of Facebook in part two, “Colonization of lifeworld,” we first introduce the notions of connectedness and connectivity in social media (Van Dijck, 2013).

From human Connectedness to Automated Connectivity

Van Dijck (2013) implicitly applies Feenberg’s perspective of technology (i.e., social media) as a medium (defined by Habermas) by deconstructing the “social” in social media: “[ . . . ] ‘making the Web social’ in reality means ‘making sociality technical’. Sociality coded by technology renders people’s activities formal, manageable, and manipulable, enabling platforms to engineer the sociality in people’s everyday routines” (p. 12). In that way, connectivity becomes the automated and delinguistified form of human connectedness. “The engineering of sociality (my term) refers to social media platforms trying to exert influence on or directing user behaviour” (p. 178).

From the moment social media systems extend beyond their intended social goal, the lifeworld is colonized. The risk that this happens is high as

Connectivity quickly evolved into a valuable resource as engineers found ways to code information into algorithms that helped brand a particular form of online sociality and make it profitable in online markets—serving a global market of social networking and user generated content. (Van Dijck, 2013, p. 4)

For example, birthday notifications in the Facebook News Feed prompt users to communicate. This process does more than coercing interactivity; it also erodes the value of connectedness itself and thereby expands beyond the intended goal of kindness to remember someone’s birthday. In that way, a Facebook user will not thank another Facebook friend for remembering their birthday, because Facebook instructed them to send birthday wishes. So the act of remembering one’s birthday is no longer that special because it is taken over by systems. It can even be done fully automatically by, for example, a Facebook application. ⁵ In other cases, users themselves take over the logic of connectivity, when the goal is no longer to socially connect, but to show how many “friends” one has (Bakardjieva, 2014, p. 375).

The Case of Facebook

In the next paragraphs, we illustrate different ways how Facebook employs the medium of technology (in Feenberg’s sense) to transform sociality from connectedness into connectivity (in van Dijck’s sense) for the sake of efficiency and—in the end—more revenue. This is based on a systematic collection of empirical examples derived from desk research in order to illustrate the theoretical arguments of our article. ⁶

The US-based company, Facebook Inc., started as a Harvard College dorm room project in 2004 and began selling stock to the public in 2012. It is now a ubiquitous social media platform visited daily by 968 million users (Facebook, 2015), which generated a global advertising revenue of 3,827 million dollar for the second quarter of 2015. Like in other social media, this surplus is realized through connectivity, which is why Van Dijck (2012) prefers the notion of “connective media.” The technology called Facebook is a Web 2.0 social networking service since it relies on crowd sourcing for its user-generated content (UGC), uses data as central asset, has an interlinkable architecture through its application programming interface (API), is in a perpetual beta, and has a flexible business model (Kaplan & Haenlein, 2010; O’Reilly, 2005).

In line with the socio-technical reasoning of Feenberg, we describe how Facebook is built on the media money and technology to coordinate behavior of actors. ⁷ As an actor network (Latour, 1992), it is an SNS that caters among others to the following actors: users, advertisers, publishers, and application developers. ⁸ These actors have a public or semi-public profile and they articulate a list of connected actors (boyd & Ellison, 2008): friends, pages, and applications. These connections can be viewed through any articulation of the Open Graph: ⁹ the News Feed, Facebook’s advertising interface and access through the API.

We illustrate how Facebook as a system could deeply penetrate the lifeworld of users. We then indicate how it manipulates the lifeworld by highlighting different technological components that interpret, adapt, and massage SNS use, possibly leading to colonization of the lifeworld of users. For this, Facebook has developed various obligatory passage points (Callon, 1986; Latour, 1992). They basically bind together the different actors by one question: “Can we control what users connect to?” This is coincidentally also Facebook’s baseline: “Connect with friends and the world around you on Facebook.” ¹⁰ Each actor in the network encounters the challenge, controlling what users connect to, as an obstacle in a particular way. The obstacles for each of the actors are summarized in Table 1 below. ¹¹

Table 1.

Challenges for actors on Facebook.

Actor	Obstacle	Goal
Facebook	Facebook boredom	Keep users attracted to Facebook
User	Boring/irrelevant content	See relevant and engaging content
Advertiser	Ads are ignored	Show ads to interested users
Publisher/application	Too much content creates attention competition	Appear in users’ News Feeds

We now discuss three different obligatory passage points built into Facebook. (1) Users need to be told and steered to connect to each other and to other pages. This information is then not only used for their own entertainment, but also to increase revenue by targeting advertisements. This refers to the EdgeRank algorithm in the News Feed. (2) Second, it is important that users engage often enough with Facebook so that they deliver data, but also that they consume commercial messages as much as possible. To maximize the latter, Sponsored Stories were deployed as a particular form of Facebook advertising. (3) Third, Facebook designs efficiency into its system in order to maximize connectivity. A valid example of this approach is the Gatekeeper tool for testing design variations and features.

EdgeRank

A first way to address the challenge, “Can we control what users connect to?” is EdgeRank, an algorithm that tries to predict what users may find interesting based on their previous experience, kind of message, and freshness of the message. ¹² The EdgeRank algorithm is not freely available, but general aspects on its heuristic can be found online (Widman, 2012). Widman presents EdgeRank as a summation function of the factors: affinity, weight, and time decay. This model was updated by Constine (2014) to include post performance. The most important factor in this sum is affinity: the amount of interactions between a user and the creator of a post that enters EdgeRank. The affinity score is increased by reactions of a user towards the creator and his or her utterances. Weight is measured by looking at the form of the post: the more interaction it creates, the higher the weight factor. This means that movies are rated higher than pictures and this list continues in decreasing order for webpages, status updates, comments, and likes. Time decay is a factor to represent the freshness of a post; the more recent, the higher the score. Lastly, post performance compares how the post is received among other users; the more users interact, the higher the score.

With the factors affinity, weight, time decay, and post performance, the algorithm de facto creates a threat of invisibility:

Thus, EdgeRank, by functioning as a disciplinary technique, creates subjects that endlessly modify their behaviour to approximate the normal. Because interaction functions as a measure for interestingness, practices of liking, commenting and participation become processes through which the subject may approximate this desired normality. (Bucher, 2012, p. 1176)

In that way, EdgeRank, as an obligatory passage point, addresses the obstacles of all actors (in Table 1). It has become obligatory as there is only a News Feed based on the EdgeRank algorithm, while in the past one could also use a News Feed with all items being posted.

However, Widman (2012) and Constine (2014) omit a fifth factor in their presentation of EdgeRank: money. Money directly offered to Facebook is an extra factor that coerces EdgeRank to show content more prominently on behalf of the advertiser, publisher or user that pays. This is a little more complex as each entry (i.e., edge) promoted with money enters a real-time bidding system with other entries that target the same user profile. The more an actor pays, the higher his or her chance of getting a successful translation and thus convincing EdgeRank to publish their story. This leads to colonization of the users’ lifeworld, as the system that delinguistifies sociality transgresses its intended boundaries. The initial goal of EdgeRank is affinity, that is, to show relevant and engaging content to the user (see Table 1). However, by adding direct money input to the equation, this original goal is extended, while this is not made clear to users (Bilton, 2013). The threat of invisibility (Bucher, 2012) is now directly coupled with a paid solution. By creating attention scarcity, Facebook can now sell attention which Facebook advertisers and page owners are likely to buy to increase the chance of reaching users. However, the latter happens in a delinguistified way and therefore does not enter the public space to be discussed and possibly questioned.

Nevertheless, Mark D’Arcy, Worldwide Chief Creative Officer of Facebook, said there are limits to money as a factor in EdgeRank. ¹³ Facebook has to maintain a balance between affinity and profitability, and therefore users can indicate bad or spam-like advertisements. This feedback lowers a score similar to the affinity score, which increases the price for advertisers to show the advertisement. But it remains a form of interaction where users are only able to provide feedback on the outcome of the process and not the process itself. In fact, users are now evaluating content of ads for Facebook, which is yet another valorization.

Advertising and Sponsored Stories

By the introduction of News Feed based on EdgeRank, connectedness is rendered programmable and becomes connectivity: “Not only is the News Feed a means of constant surveillance of one’s friends, but it provides members with incentive to log on to the site more frequently” (Cohen, 2008). This connectivity is then tied to advertising. Showing relevant and engaging content to the user (see Table 1) increases stickiness and entices users to become loyal members. This is required to sell more attention to advertisers and to increase the chance that users will consume advertising, which is traditionally shown separately (Cohen, 2008). In addition, prompted by the attention scarcity, advertisers are coaxed with more options to personalize content and create messages that are difficult to ignore. Ideally, these promotional messages are very personalized and leverage a personal relation.

The Sponsored Story (SPS) is an emblematic example of a second obligatory passage point, which colonizes the lifeworld. With SPS, advertisers can pay to highlight in a user’s News Feed any UGC of his/her Facebook friends related to a specific page, app, or other item of the advertiser. These SPSs can be targeted based on a user’s interest and profile details. Facebook does not offer SPS any more as a separate advertising format since April 2014. Yet, the basic idea behind SPS is still part of Facebook ads, as they state that “[ . . . ] social context—stories about social actions your friends have taken, such as liking a page or checking in to a restaurant—is now eligible to appear next to all ads shown to friends on Facebook” (Facebook, 2014).

The advent of advertising based on social context (like SPS) is economically motivated and tested. SPS increased the Click Through Rate (CTR) with 46%, the number of users who click on an ad (Constine, 2011b). This means that 46% more users click on an ad appearing as an SPS compared to a regular Facebook ad. This resulted in an 18% lower cost per fan (someone who follows or likes a brand page).

This type of promotional entry forces two actors together: user and advertiser. It does so by camouflaging advertising as UGC. Users like to read UGC on their News Feed and dislike ads, which are shown in the right-hand column of News Feed. This type of mixing editorial with promotional content is typically forbidden in older media (like newspapers and television), where advertising needs to be clearly separated or indicated. However, in the digital environment of social media, we see more and more seepage between both, with formats like advertorials and content marketing. By pushing camouflaged ads in News Feed, users are proscribed to see them, even more so as users cannot block these ads with ad blockers (which excludes even further antiprograms). ¹⁴ The colonization of lifeworld takes place because the initial goal of sharing experiences between friends online in the News Feed is extended towards advertising, without being visible to the users.

Gatekeeper

Lastly, we illustrate how the rational criterion of this medium, efficiency of connectivity is deeply ingrained in Facebook. Our critique is not limited to the manipulation of algorithms in EdgeRank and the use of ads like SPS, but it extends to the design of Facebook itself.

Typically, for a Web 2.0 service, Facebook has the status of perpetual beta (Kaplan & Haenlein, 2010). This means that they do not introduce large updates on end-user equipment. Instead, small features are changed continuously on the servers of the platform (O’Reilly, 2005). This notion of the perpetual beta is integrated as a tool called Gatekeeper that tests small design variations and features on Facebook: “Every day, we run hundreds of tests on Facebook, most of which are rolled out to a random sample of people to test their impact” (Boz, 2012). The developers define a goal they wish to test with their proposed changes and this is measured and compared to alternative designs: “We ran a test which instead reduced the number of people we showed per page by 60% but gave each more space and a larger button to engage with, and we saw a 70% increase in friend requests” (Boz, 2012). The use of a tool like Gatekeeper is a typical way of how sociality can be made more efficient through connectivity. This tool is a third type of obligatory passage point as users are unaware of being tested on design issues. They have no choice but to be steered towards the most efficient interface in a delinguistified way.

The delinguistification of design changes can then readily lead to colonization of the lifeworld. This is demonstrated by the emotional contagion study in 2014 where A/B testing was applied to the selection of UGC shown to Facebook users (Kramer, Guillory, & Hancock, 2014). This was a highly controversial study (with a Facebook data scientist as first author), which has been intensely debated (e.g., boyd, 2014; Fiske & Hauser, 2014; Kahn, Vayena, & Mastroianni, 2014; Meyer, 2014). For this research, EdgeRank manipulated UGC shown to test subjects on Facebook. EdgeRank was either selecting negative or positive content. The UGC content of the test subject was then analyzed to see if they transferred this mood. Here the delinguistified manipulation directly influenced users’ lifeworld by presenting them with “happier” information, which results in happier users who react more positive to advertising. In that way, the changes in the design and interface expand beyond the initial goal of increasing efficiency and user-friendliness, by encroaching on the private communication itself and which emotions are shown. What is more, if these researchers had not published their results, this manipulation would not have entered the public space. This means that system colonization is difficult to criticize because it never seems to occur in our lifeworld, because it is so hard or impossible to perceive.

Findings and Discussion

We have defined EdgeRank, Sponsored Stories, and Gatekeeper as obligatory passage points that force and steer connections between users and actors who wish to pay for attention, in order to address the question: “Can we control what users connect to?” Our aim was to understand when these systems become problematic for the users of Facebook, by applying Feenberg’s critical theory of technology. More, in particular, we analyzed if and how these particular instances of connectivity lead to system colonization of lifeworld. First, this can only be the case if that technology steers behavior in a delinguistified way. Second, these connectivity systems need to extend beyond their intended boundaries and thereby decrease lifeworld.

First, EdgeRank and Gatekeeper are systems that evaluate and discriminate through the criteria of efficiency and profitability. News Feed elements and design features are discarded or kept if and only if they increase connectivity or generate profit. In both cases, this evaluation occurs in a delinguistified manner, completely oblique for each actor in the network except Facebook. ¹⁵ In particular, users no longer understand the reasons that explain how and why they connect. They are steered like Truman in the movie “The Truman Show” (Paramount Pictures, 1998), but the fact that they are in this show is no part of their lifeworld. Facebook tries to disappear as a technology. The fact that Facebook is modified in real time via trial and error further decreases the control of users. The process of delinguistification is incremental and on-going because Facebook is in a perpetual beta. Algorithms make visibility choices based on efficiency, which also leads to choices regarding the overall design and the launch of new features. Each update creates a new phase of interpretative flexibility, at the same time means closure of the previous one. This delinguistifies the former update and changes. By doing so, the attention is moved to the currently proposed changes. For example, users are no longer protesting against News Feed itself, they are wondering what will change when new privacy and data use policies are implemented. In other words, incremental updates seem to enforce stability, which further diminishes the possibility for critical reflection as they technologically force an end to a public debate (if any debate ever occurred, or could occur, at all). It could be argued that this is no longer interpretative flexibility since this whole process is blackboxed. However, we can still expect that people at Facebook have a discussion about alternative updates (update A or B). Users, on the other hand, see one update (A1) following another update (A1.1). Each update is a new alternative to the existing configuration prior to the update. In this example, updates are linear and only their linearity hints at interpretative flexibility. We therefore call this incremental variant linear interpretative flexibility. Here, users have a limited agency to reject and leave or to accept and stay with regard to the proposed update.

Second, we have illustrated how the obligatory passage points on themselves (EdgeRank, Facebook advertising, and Gatekeeper) do not lead to colonization of lifeworld, but they do increase the risk. The analysis has indicated how particularly the factor of money and real-time bidding in EdgeRank has led to expansion beyond its original goal. The same thing happened with the introduction of social context in SPS advertising and with interface design changes stretching into private communication.

Conclusion

The application of Feenberg’s reinterpretation of Habermas was fruitful because it generates a deeper understanding of how social media are changing and complements the critical perspective on the risks for people and society.

We have used Facebook as a typical case of how social media are changing. They are first developed to focus on efficient communication in the private space and then successfully create obligatory passage points. Next, they need to prioritize revenue and profitability over efficient communication. This shifting balance is related to their underlying, evolving business models. Social media platforms in general, receive investments and are told to focus on growth, which translates in finding means for more efficient communication. After the investment phase and when, for example, going public on the stock market, these technologies have to become self-sufficient by drawing in new actors to increase revenue. In this regard, the promise of better communication in the private space is only a lure to create a commercial space in the private sphere, and hence colonizing the lifeworld.

The proliferating delinguistification as well as the colonization of lifeworld adds new reasons to be critical about social media. So our analysis goes beyond the traditional political economy’s focus on prosumption in the sphere of production (Fuchs, 2012c), as it also links up with the notion of immaterial labor. The two types of labor, defined by Lazzarato (1996), are present in the Facebook analysis. The first type (manipulation of computer control) we find in the continuous processing of information and user feedback via the EdgeRank algorithm and the Gatekeeper tool. This optimizes the connectivity among users, and hence the interaction between production and consumption. The second type (generating shared standards to influence consumption) is visible in the way that the obligatory passage points delinguistify and steer interpersonal communication. As a result, capital has subsumed social communication (connectedness) and changed it into a more profitable form (connectivity). Hardt and Negri (2000) stated that this unison of communicative action with economic production would only increase production to the level of complexity of human interaction, without an impoverishment of communication. We have illustrated that this subsumption did increase economic production but only by limiting communication. There is an impoverishment signified as colonization.

Future research needs to further investigate and empirically substantiate if and how the lifeworld is changed by system colonization from the users’ perspective. Do users themselves experience their use of the SNS as an obligatory passage point? Is this something they require to an extent large enough to tolerate constant design changes and encroaching advertisements? If they do not experience this, how can we best bring this within the public space and thereby enhance user empowerment and mitigate disempowerment (Pierson, 2012)?