Social pathologies of informational privacy

Abstract

Following the recent practice turn in privacy research, informational privacy is increasingly analyzed with regard to the “appropriate flow of information” within a given practice, which preserves the “contextual integrity” of that practice (Nissenbaum, 2010, p. 149; 2015). Such a practice-theoretical take on privacy emphasizes the normative structure of practices as well as its structural injustices and power asymmetries, rather than focusing on the intentions and moral considerations of individual or institutional actors. Since privacy norms are seen to be institutionalized via the role obligations of the practice's participants, this approach can analyze structural and systematic privacy infringements in terms of “defective role performances and defective social relations” (Roessler & Mokrosinska, 2013, p. 780). Unfortunately, it is still often somewhat unclear what this exactly means within the context of informational privacy, why these performances and relations are defective and for whom. This raises the common objection of a so-called “practice positivism” (Applbaum, 1999, p. 51), that is, the difficulty of practice–theoretical accounts to take a practice-independent standpoint, from which to normatively evaluate the existing practice norms themselves. For example, Nissenbaum herself initially argues for a “presumption in favor of the status quo” with respect to the appropriateness and flow of privacy norms within a practice (Nissenbaum, 2004, p. 127). Such a “practice conservatism” (Nissenbaum, 2010, p. 169) comes dangerously close to committing a naturalistic fallacy, if not undergirded by practice-external criteria (which is ultimately what she does). Merely resorting to existing practice norms to assess what defective role performances amount to, only shifts the question from how to recognize an appropriate flow of information to the question of how to recognize those defective role performances and social relations. Against this backdrop, the central aim of this article is to shed light on this question without resorting to practice-independent first principles or far-reaching universalistic anthropological assumptions. For this, I will analyze the notion of “defective role performances and social relations” in terms of social pathologies.1 Doing so has two advantages: First of all, it can draw on already existing concepts and distinctions, which help to categorize the different levels of analysis that exist in informational privacy research and situate the notion of “defective role performances” within them (Section 1). Second, those concepts and distinctions can serve as a basis for establishing a typology of phenomena with regard to deficient practices of informational privacy (Section 4). Having thus set the scene in Section 1, I can move on to address the notion of “defective role performances and social relations” with respect to informational privacy. In Section 2, I employ the social-ontological recognitional model of privacy (SORM) that I developed in more detail elsewhere (Loh, 2018). The SORM has the advantage to evade some of the pitfalls of practice positivism: It can help explain what “adequate flow of information” and “contextual integrity” amounts to on a more fundamental social-ontological level, without recurring to practice-independent criteria.2 Following the insights gained from the SORM, in Section 3, I analyze pathologies of informational privacy as the structural and systematic failure to recognize all bearers of constitutive roles (BCR) within a datafication practice as standard authorities. This account not only gives an answer to why these practices are defective on a social-ontological level. Also, it helps to explain in more detail how deficiencies that prevent the mutual recognition of standard authority are structured in general. Finally, in Section 4, I distinguish four interrelated types of structural distortions which are often employed in diagnoses of social pathologies, and which are also at play in the context of datafication practices: Epistemic, cognitive, psychomotivational, and power-relational distortions. This article brings together social ontology, practice theory, privacy theory, as well as methodological questions of ideology critique (in the form of theories of social pathologies), and applies them to datafication contexts. As a result, it can provide a theoretical—but still practice-dependent—foundation for empirical analyses of systematic and structural privacy infringements and the underlying issues of datafication. In this respect, the article contributes to the existing privacy research by giving a functional account of these infringements in terms of social pathologies: They are pathological in the sense that they prevent the mutual recognition as standard authorities within datafication practices and thereby seriously weaken the reproducibility of these practices. The task of this section is to motivate the usage of the term “pathology” by highlighting some of its features common to most theories of social pathologies. According to Honneth, one of the main differences between political and social philosophy lies in their perspective on social orders: Whereas political philosophy is mainly concerned with just or at least legitimate social orders, social philosophy primarily analyses “social processes of development, which must be understood as impairments of the possibilities of the ‘good life’ among members of society” (Honneth, 1996, p. 370). For Honneth, these possibilities—and with them, his theory of the good life itself—remain formal in the sense that they are not fleshed out substantially, but rather refer to the “social presuppositions” needed in the continuous process aspiring to the “goal of human self-realization” (Honneth, 1996, p. 388). “Social pathology,” for Honneth, is then the overarching notion for all different types of such impairments to the social presuppositions for human self-realization, whether they be further spelled out in terms of “reification,” “alienation,” “processes of instrumentalization and objectification,” “commercialization,” and so forth (Jaeggi and Stahl 2011, p. 697; my translation). Although Honneth's notion of pathology is ultimately tied to the “goal of human self-realization” (Honneth, 1996, p. 388), that is, pathological social structures are “ill” because they ultimately “make[] individuals ill” (Freyenhagen, 2018, p. 412), it is also functional: It captures structural and systematic failures of a social configuration to realize its own immanent norms, that is, the norms that are constitutive for its reproduction (Honneth, 2014, p. 3).3 In other words, the presuppositions Honneth talks about are social means of access to and integration in social structures, rather than universalistic anthropological needs. As a result, his idea of the good life can remain formal, since Honneth talks only about the (social) presuppositions for human self-realization, which in turn would then lead to a good life. Understood in this way, the term “social pathology” not only opens the possibility to analyze social structures as part of “universal conditions of human self-realization” (Honneth, 1996, p. 389), but also to evaluate them from the perspective of their functioning as those conditions. Since pathologies are generated and sustained by the social structures within which they operate, they are structural disorders. At the same time, they are systematic: They are enduring and comprehensive in the sense that they affect all or a substantial portion of the participants of a given structure—oftentimes specific groups. In addition, claims of social pathologies “tend to concern social processes of increasing deterioration” (Freyenhagen, 2019, p. 16; emphasis added): The longer these structural disorders persist, the more severe the effects on the structure as well as the actors will become. In this article, I will move Honneth's understanding of “pathology” even further in the direction of a functional account. Expanding on Stahl's notion of “standard authority” (Stahl, 2021, Chap. 7), I propose a formal and functional account of social pathologies: I employ them as a formal social-ontological criterion for certain processes within social practices that are likely to seriously inhibit their reproducibility.4 In Section 3, I will go into more detail what this means for datafication practices. For the moment, it may suffice to say that these practices will be analyzed as pathological when a certain number of interpretations are systematically ignored and can therefore not contribute to the reproduction of the practice. In a nutshell, this is what “recognition as standard authority” amounts to: The recognition of practice participants performances as competent attempts at interpreting the practice norms and their role obligations. Systematic misrecognitions of standard authority will then ultimately lead to pathological conditions, such as rigidity (what may be called “reification”), detachment (what could be called “alienation”), and finally abandonment.5 Another way of framing the concept of standard authority is to say that it analyzes the impairments Honneth talks about in terms of access: A certain social structure can be called pathological, if—by the very nature of this structure—its norms cannot be adequately accessed anymore by some or all of its participants. “Accessed” in this regard is either meant epistemically, in the sense of being able to give consistent interpretations of those norms (e.g., through role performances), address conflicting interpretations adequately, and so forth. It can also be meant politically, in the sense of voice and contestation options that are limited through structural power asymmetries, which in turn give rise to persistent epistemic injustices, othering, oppression, and so forth.6 How participants relate to the social structure is explicated in this account in a way that has to do with the inner workings of these structures and their stabilization and normalization function. Therefore, the notion of normality that I am interested in is not so much dependent on historically contingent and culturally sedimented norms and social practices.7 Rather, “normality” refers to this more fundamental social-ontological way, in which the participants relate to this structure, regardless of the actual norms and practices. In sum, for the purposes of this article social pathologies are structural and systematic barriers within a given social entity that prevent its participants from accessing (i.e., adequately addressing or effectively contesting) the normative structure of this entity (i.e., its foundational norms, role status, materialities, affordances, etc.). As such, they not only raise the potential for violent conflict or resignation, since some of the participants are structurally and systematically excluded from interpreting and contesting its normative framework. Also, the adaptability of the social entity is seriously impaired: As fewer and fewer participants can interpret its normative structure, this makes incremental change continuously more difficult. These “social processes of increasing deterioration” (Freyenhagen, 2019, p. 16) in turn affect its participants by way of anger or resignation even more. If not “treated,” they will eventually lead to the abandonment of the social structure or its substantial petrification, in which case it is upheld only through widespread domination and coercion. Before I apply these basic assumptions to my SORM, however, I want to briefly highlight on which level of explanation it operates. In other words: What kind of question is the SORM answering within this framework of the wider notion of social pathologies? For this, I find Freyenhagen's “basic grid” (Freyenhagen, 2019, p. 18) very helpful. He distinguishes between four levels of explanation with regard to social pathologies: On the symptomatic level, theories describe and analyze actual instances of pathologies and how they manifest in the social world. Within the realm of informational privacy, this level is concerned with most of the sociological, media-theoretical, and psychological studies and findings. For example, analyses of consumer manipulation by tech companies such as Dark Pattern or Addictive Designs, all kinds of explanations for the ambivalent reactions to the Snowden and Manning leaks and the prevalence of the privacy paradox (Dienlin and Trepte, 2015), or critical accounts of the notion of “data ownership” (Prainsack, 2019) and “data sovereignty” are located on this level. The diagnostic level, in contrast, is primarily focused on explicating what exactly makes these symptoms a social pathology. Theories on this level are mainly concerned with what constitutes adequate/functioning versus defective social relations in general, often in terms of self-realization, flourishing, adequate flow of information, consistency within the normative structure, coherence, and so forth. This is also the level that the SORM is addressing. This level of explanation is to be distinguished from the etiological level, on which theories are primarily engaged with the more fundamental causes of these defective relations, such as the workings of socio-economic megastructures. With respect to informational privacy, there is a growing literature that reflects on the mechanics of a new “surveillance capitalism” (Zuboff, 2019) and its repercussions (Cohen, 2019; Sevignani, 2015; Webb, 2019). Finally, on a therapeutic level, theories are focused on measures to counter or revert either some of the symptoms or the causes. Most of the time, the exploration of potential remedies is preceded by an analysis of symptomatic deficiencies. As a result, this level is rarely addressed all by itself. In the previous sections, I referred more generally to “social structures,” “social entities” or “social relations” as being defective or becoming pathological. At this point, however, it is important to be more specific: In order to make use of the SORM, I analyze these social phenomena in terms of social practices. This means that in my usage, pathologies do not amount to “diseases of society” (Honneth, 2014; emphasis added), but are rather confined to structural and systematic deficiencies of types of practices – primarily datafication practices.8 This has several advantages: First, such a focus does not have to presuppose that structural and systematic barriers to access always affect society as a whole. Rather, it can analyze pathologies on a more fine-grained level as pertaining to certain practices or types of practices. Second, it does not need to argue for macrolevel deficiencies, which inevitably bring with it the difficulties of normative arguments with regard to societal structures (Freyenhagen, 2018, p. 422). Third, a focus on social practices does not favor or preclude any macro-social analyses. As illustrated above, my account operates on the diagnostic level that is independent of explanations on the etiological level. While I cannot go here into the details of specific accounts within the broader field of practice theory (Bourdieu, 1976; Celikates, 2018; Schatzki, 1996; Spaargaren et al., 2016), I quickly want to highlight six main features of the concept of a social practice, which I have motivated in more details elsewhere (Loh, 2018; 2019, Chap. 2.1): According to this definition, social practices are (1) repeated, coordinated patterns of action that (2) follow certain rules, which (3) are at least partially internalized as “Knowing-Hows” (Ryle, 1949, Chap. II) and routinely performed. These practice norms will (4) surface and become thematic, if the expectations of individual actors with regard to the performances of other actors are frustrated due to differing norm interpretations. This mismatch in expectations of behavior evokes “reactive attitudes” (Strawson, 1962, p. 66), which are typically expressed in the form of criticism with regard to the other actor's role performances. The mutual expectations of behavior are thus (5) “generalized” (Mead, 1934) in social roles and their associated role privileges and obligations. These role obligations confer the first-order deontic status of “role bearer” on the participants, in the sense that they incur certain role privileges and are expected to discharge their role obligations. In addition to this first-order deontic status, the fact of conflicting interpretations points (6) to the second-order deontic status of a “standard authority” (Stahl, 2021, Chap. 7). Such conflicts may arise either in the form of differing performances, or in the form of explicit criticism of the performances of others. The notion of standard authority, as developed by Titus Stahl, is a social-ontological explanation about how norm interpretations typically work within social practices. It claims that, beyond the respective role obligations (first-order deontic status), actors within a given practice incur a kind of “meta-obligation to respond to possible criticisms” of their own role performances (second-order deontic status).9 Since role performances are themselves nothing but actor's interpretations of their role obligations (first-order deontic status), criticisms of these performances are in essence interpretations of these interpretations. Recognizing each other as standard authorities means that the other actors within the practice are recognized as having the authority to perform these kinds of meta-interpretations (second-order deontic status). This recognition is mutual. Therefore, all actors within the practice are authorized to interpret the norms of that practice, and not just the norms of their own role. “Authorized” then means that it gives the actors performing such criticism a normative force: The addressees have an obligation to respond to these criticisms. This authority is “standard” in the sense that it is independent of the authority their role conveys them. For example, even though within the practice of the court room the judge has a very different role authority than the defendant, both hold the same standard authority to criticize each other's role performances. They obtain this authority simply by virtue of being actors within a given practice. “Standard” in this sense is a shortcut for a basic deontic power to interpret and criticize the practice norms and the behavior of the other actors, which is equally distributed among all actors of a given practice. The notion of standard authority therefore is not a kind of natural or transcendental right,10 but rather a social-ontological explanation of how social practices reproduce and adapt through interpretations-as-performances.11 As such, it does not entail the “right to rule” (Raz 1986) in the sense of having one's interpretations and criticisms always complied with. Rather, individual criticisms will typically change practice norms incrementally by a general change in interpretations, if and when other practice participants gradually pick up on these criticisms. Instead of a “right to rule” one could call it a “right to access” in the sense of epistemically assessing, performatively interpreting and contesting practice norms, as well as criticizing other's interpretations. By way of example consider the practice of doctoring: A doctor D asks her patient P to describe the history of the medical condition that brings him to her office. Hereby, she is discharging some of her role obligations to thoroughly examine and question patients (first-order deontic status). If P follows suit, he implicitly accepts D's interpretations of her own role as a doctor. However, by responding: “What? You can't tell me anything!,” he exercises his second-order deontic status of standard authority and challenges D's interpretations. D now incurs the meta-obligation to justify her performance, maybe by saying: “Do you want me to help you or not? I am your doctor; I need to know your medical history in order to properly diagnose your illness.” In doing so, doctor D acknowledges patient P's standard authority. As we will see in the next section, this acknowledgment can be insincere. In some cases, the meta-obligation is not even discharged at all, however superficial. Were P to add to his criticism: “Doctors shouldn't know anything about patient's medical records anyway,” he would exercise his standard authority on a second level, explicitly aiming at the normative structure of the doctoring practice per se, and not solely at the particular role performances of D. Also, these interpretations can be belittled, ignored or suppressed by the other practice participants, resulting in a misrecognition of standard authority. The notion of social practices serves two purposes in this line of argumentation: On the one hand, it identifies who should recognize whom as a standard authority.12 Until now, I talked more generically about “actors” of a practice. With the basic outline of “standard authority” in place, I am able to specify this: “Actors” are bearers of constitutive roles (BCR) within a practice. By “constitutive” I mean all sets of role types, without which the practice could not be instantiated. For example, the practice of doctoring may differ considerably in its instantiations. In order to be considered as “doctoring”, however, it needs at least the roles of doctor and patient. Such roles are constitutive in the sense that they define a certain practice (Searle, 1969, Chap. 2). On the other hand, the concept of social practices opens up the possibility to identify contexts of informational privacy by analyzing them as datafication practices.13 A datafication practice is a practice that is defined by at least two constitutive role types: The role of data-collecting subject and the role of datafied subject. Typically, datafication practices do not stand on their own, as they require the datafied subject to do something within another practice. Think about the example of the doctor, where the datafication practice is part of a larger doctoring practice. Other examples include interacting on social media, buying or selling online, walking through a CCTV-covered area, and so forth. As a result, datafication practices are almost always subpractices of other larger practices. In a wide sense of “datafication,” all social practices entail a datafication sub-practice. As repeated, coordinated patterns of action (cf. the first of the six features I mentioned earlier), they instantiate “symbolic interactions” (Blumer, 1969) that require perceiving, interpreting, and reacting to performances of other actors. Datafication practices entail not only norms about which data the actors of the main practice may gather—Nissenbaum's “appropriate flow of information”—but also how they should go about doing this. While the doctor may touch her patient in order to examine him, commuters on a train typically may not do this to gather information about their environment. Also, excessive staring, intentionally listening in on conversations, etc. are at least frowned upon or even directly challenged. In a narrower sense, however, “datafication” means a systematic collection, storage, processing, and dissemination of data not only for the purpose of making sense of one's environment, but to categorize, infer other data, make prognoses, and so forth. Studies and findings on the symptomatic level are typically concerned with this narrower sense and especially interested in the digital means to collect, process and visualize data. In order to give a more precise explanation for these datafication mechanisms and highlight the specific concerns of digital datafication, I will limit the investigation to this narrower sense of datafication. Nonetheless, datafication practices in this narrow sense are also mostly subpractices to other (oftentimes digital) practices. In sum, the norms that specify the “What” and “How” of datafication belong to the datafication (sub)practice, and are expressed in the role obligations of the data-collecting subjects and datafied subjects. Accordingly, the SORM claims that a practice is informationally private, if and only if within the corresponding datafication practice(s) all BCR recognize each other as standard authority.14 Informational privacy is therefore not so much spelled out in terms of the instantiation of certain privacy norms in practices per se, but rather by assessing the social-ontological structure of datafication practices. Since what “private” means will differ considerably across practices, the “contextual integrity” (Nissenbaum, 2010, p. 149) of a practice is established and maintained by the fact that all BCR in the corresponding datafication subpractice recognize each other as standard authority. As I have argued above, this means that all data-collecting and datafied subjects mutually recognize each other's authority to criticize the role performances of the others, and by virtue of this the very structure of the datafication practice. In an informationally private practice, all datafied subjects have the authority to question, contest, and demand justification for the systematic collection, storage, processing, and dissemination of their data due to their standard authority in the corresponding datafication subpractice. By virtue of this authority, they effectively interpret the practice, and can thereby reproduce but also change it. This does not mean that their authority has to be always complied with in order to call a datafication practice “private.” What amounts to an “appropriate flow of information” within the practice, still depends on the generally accepted practice norms. The SORM only explicates what it means to be “generally accepted” by way of the notion of standard authority. From this it becomes clear that the notion of standard authority does not refer to an ideal discourse situation, in which all participants are included in an egalitarian fashion (Habermas, 1986; Habermas, 2001).15 Nonetheless, as the doctoring example shows, oftentimes conflicting interpretations will be made explicit (Brandom, 1998; Celikates, 2018). In pathological practices, it is often this very ability of making it explicit, which is distorted by misrecognizing a certain portion of the practice participants as standard authority. These pathologies amount to the structural and systematic failure of at least some bearers of constitutive roles (BCR) within a datafication subpractice to mutually recognize each other and all other bearers of constitutive roles as standard authorities. This recognition entails, as I have laid out in the last section, (a) the authority to contest the role performances of the other BCR within the datafication subpractice; and (b) demand justification for their interpretations of their own role obligations. “Structural” and “systematic” refer, as mentioned in Section 1, to enduring and comprehensive deficiencies that typically affect all or a substantial portion of actors—or, as we can now specify—one or more types of BCR within the datafication subpractice. These deficiencies are generated and sustained by the normative structure of the subpractice. By transferring Stahl's account of standard authority to the realm of informational privacy, the SORM gives an answer on the diagnostic level, that is, to the question “What is a social pathology of informational privacy?”. Nonetheless, we still have to assess why these structural and systematic misrecognitions of standard authority constitute a social pathology. The answer is tied to the inner workings of social practices: They depend on this mutual recognition (a) for their ability to stabilize mutual expectations of behavior; and (b) their ability to incrementally adapt the content of these expectations. In social practices, participants constantly criticize each other and thereby interpret, ascertain, and normalize role behavior (Berger & Luckmann, 1967; Giddens, 1984). These interpretations-as-performances reproduce the practice, in the sense of replicating but also incrementally changing it. In this regard, the mutual recognition as standard authority marks the transition from regular to rule-bound behavior: It captures the normative element of social practices that is responsible for their social integration (Habermas, 1986, Vol. 2; Lockwood, 1964), and therefore distinguishes them from coincidentally aligned behavior on the one hand, and coordination by mere coercion on the other. Without this recognition, social practices as defined in Section 2 could not exist, in the sense that they would lose their reproducibility in the long run, that is, they could not be internally stabilized and could not adapt to changing norm interpretations.16 Since role performances would not be recognized as competent attempts at interpreting the practice rules, it is hard to see how individual beliefs about practice norms and role obligations can transform into commonly held beliefs that this is the “right way to φ”. Without such beliefs (in the form of internalized Knowing-Hows), no common practice that normatively binds its participants would emerge (Hart, 1994, p. 89). The only form of stabilization through social integration that could be achieved without assuming a mutual recognition as standard authority, is through a coercive or coordinating force from outside the practice. Moreover, the constant back and forth between interpretations-qua-performances not only explains how role obligations evolve and normalize, but also how they change over time. In sum, mutual recognition as standard authority is a necessary condition for reproducibility (in the sense of integration, stability, and adaptability) and thereby for the functioning of datafication practices. For these reasons, a structural and systematic misrecognition of (mostly a group or type of) BCR a