Privacy in Public: A Democratic Defense

1 Introduction

On January 15, 2015, groups in New York City associated with the ‘Black Lives Matter’ movement organized a protest at Grand Central Station. As we now know, officers of the NYPD monitored that protest, took photos of individual participants and made records of their movements and actions (Joseph 2015), just as they and other agencies did at other protests regarding similar matters, including the ‘Occupy Wall Street’ movement (Schmidt and Moynihan 2012). Going back somewhat further in history, it is well-known that leaders of the Civil Rights movement were subject to intense surveillance by the FBI (see Garrow 2010), as were other members of civil society movements. This is not an exclusively American phenomenon. In the UK, the Metropolitan Police was found in 2009 to have stored details on thousands of activists in a database for years (Lewis and Vallée 2009), including details about their participation in various protests. Police forces in various other democratic countries regularly film peaceful protests and send undercover informants to gather information on participants and their political activities (Fijnaut and Marx 1995; Electronic Privacy Information Center 2019). Unsurprisingly, non-liberal countries have even less compunction about subjecting civil society actors to surveillance (Security Without Borders 2019).

Anecdotal evidence shows that many people feel uneasy about these cases, not merely because of the risk that the data gathered by such surveillance might be abused in a way that interferes with people’s constitutional rights. The systematic gathering of information on peaceful political activities seems to raise ethical and political problems even if such abuse is sufficiently unlikely. This is not merely true when the surveillance intentionally focuses on political activities. Many people also worry that indiscriminate mass surveillance will have a chilling effect on democracy precisely because it will necessarily include surveillance of people’s political activities (Haggerty and Samatas 2010; The New York Times Editorial Board 2013). These worries become even more urgent when we consider the potential of facial recognition technologies and the fact that public spaces also extend to social media platforms.

Traditional philosophical theories of privacy offer us relatively little help in making sense of these worries. The very idea of privacy had traditionally been linked to the idea of a ‘private sphere,’ most centrally the private home and the family, which was seen as especially deserving of protection. But even though most theorists do not restrict the notion of privacy to the private sphere in this sense today, the intuition that privacy rights only protect activities from observation by others that are not already public is still widely shared in both philosophy and the law. In the American context this intuition plays out through the idea of a ‘reasonable expectation of privacy,’ which plays a central role in relevant jurisprudence (Winn 2009), whereas in the European context courts have consistently held that a person has privacy rights if she makes a reasonable effort to protect her activities from public scrutiny.

Once privacy protections are conditioned in this way, political activities like the ones mentioned at the beginning of this section seem to be clearly excluded from them. Not only do they typically take place in public spaces where no reasonable expectation of privacy exists, but they are also consciously designed to attract public attention and thus seem to amount to an effort to forego privacy. If state agencies collect information on such activities, they therefore do not seem to be violating any privacy rights.

On further inspection, things are a bit more complicated. On the one hand, there is a growing body of legal and philosophical literature on the idea of ‘privacy in public’ that challenges the idea that public activities do not deserve privacy protection by definition (Timan et al. 2017; Newell et al. 2018; Moreham 2006; Reidenberg 2014; Nissenbaum 1997; Roessler 2016). As I will show, however, in this literature the idea of privacy in public is even mostly limited to what I will call accidentally public activities — activities such as traveling to work and going shopping, which are de facto impossible to perform in private but which do not depend in any substantive sense on occurring in public view. Political activities, by contrast, are intentionally public. Their very point depends on the public’s being addressed, and the agents therefore typically seek public attention.

On the other hand, courts in different legal systems have frequently restricted the power of the police to monitor political expression based on the idea that such constraint is necessary to protect political rights (see Levinson-Waldman 2017). Just why political rights depend on privacy, however, often remains unexplored in judicial reasoning.

In this paper, I will offer a new and distinctive argument for the claim that privacy protections for intentionally public behavior (such as political speech) are justified because of their contribution to collective autonomy and democratic self-government. My argument will draw on Jürgen Habermas’s theory of deliberative democracy. Habermas holds that democracy depends on the possibility of public deliberation which requires specific forms of communication between citizens that are institutionally secured by what he calls the public sphere. I will extend this idea and show that surveillance of public speech undermines the possibility of specific forms of speech and therefore restricts the liberty of citizens to a potentially problematic degree.

The argument will proceed as follows: In section two, I will survey some existing approaches to the problem of privacy in public. I will argue that many of these privilege the protection of accidentally public activities over that of intentionally public activities. In section three, I will use Helen Nissenbaum’s contextualist defense of privacy in public to argue that we need to pay attention to the specific goods provided by the public sphere and that Habermas’s theory offers us a plausible account of these goods. As Habermas does not discuss the problem of surveillance, I combine the speech-act theoretic foundations of his theory in section four with recent insights in feminist philosophy of language to argue that surveillance is a form of participation in a speech community that undermines the capacity of other participants to speak as they intend to, thus threatening their communicative freedom. I conclude that a democratic understanding of privacy in public provides the best rationale for protecting the privacy of participants in intentionally public activities.

2 Philosophical approaches to privacy in public

Although this paper is primarily interested in the question whether there should be limits on the surveillance of intentionally public speech — which I assume to be the core activity that enables and is enabled by a public sphere — in this section I will consider the issue of privacy in public in a more general sense that relates to all activities conducted in contexts that are epistemically accessible to unspecified others without constraint and therefore public in a more ordinary sense. I argue that the very idea of privacy in public is consistent and that its force depends on the presupposed value of constraints on acquiring and sharing information; I then identify a number of arguments why such constraints could be valuable in the case of public behavior that relate to their value for the agents and for society as a whole.

The consistency of the idea of privacy in public has been questioned by theorists who define privacy in access terms (Gavison 2012). If privacy is defined in terms of whether someone has access to information about you, the very fact of doing anything in public (again, meaning in a context where unspecified others are not blocked from acquiring information about you and where such information acquisition does not in itself violate any norms or rights) will lead to an unavoidable loss of privacy, and, consequently, there will be no form of privacy that can be protected by social norms or legal rules (Parent 1983). By contrast, once we think of privacy in control terms, i. e. as the state of being in control over who has access to information, then the idea of privacy in public becomes at least intelligible. ‘Going public’ need not be accompanied by a loss of control — in fact, we can usefully distinguish between wanted and unwanted forms of public visibility. Of course, one can argue that if one decides to make information about oneself public, one merely enjoys control at the moment of decision, but will lack ongoing control once the information is made public. Thus, every public activity will, again, incur a loss of privacy. However, this is too simple a picture. While it is true that once other people are in possession of information about oneself one cannot meaningfully be certain about the ways in which that information will be used and communicated, social, legal and other norms can ensure that some degree of control — and thus privacy — is retained (Inness 1996, p. 50).

In fact, such control is enabled by the fact that courts in many jurisdictions have consistently affirmed a right to privacy in public. In the European context, the German Federal Constitutional Court has held that there are privacy rights for activities that happen outside the home or similar private spaces, as long as the secluded character of the location allows the agents to have a reasonable expectation of privacy (Bundesverfassungsgericht 1999). This limitation was partly removed by the European Court of Human Rights (European Court of Human Rights 2004, par. 75 and 77), which held that the possibility of a private life is incompatible with permanent surveillance of one’s public conduct (for a more detailed overview of European case law, see also Vermeulen 2014). Similarly, the Canadian Supreme Court has recently held that ‘Privacy, as ordinarily understood, is not an all-or-nothing concept, and being in a public or semi-public space does not automatically negate all expectations of privacy with respect to observation or recording.’ (Supreme Court of Canada 2019).

This clearly shows that people can enjoy control over information about epistemically public activities; the question whether surveillance of intentionally public activities is compatible with a right to privacy therefore cannot be solved by referring to the semantics of ‘privacy.’ We need to investigate the normative arguments behind different conceptions of such a right.

It has often been argued that people have an obligation to ignore information about others that is, in principle, easily accessible (‘to avert the gaze’ or to display ‘civil inattention’ (Goffman 1963)), for a number of reasons that are often connected to more general views about the value of privacy (see Roessler 2016; Nagel 1998; Patton 2000). First, privacy in public can be argued to straightforwardly serve the value of protecting people’s individual liberty (Reidenberg 2014; de Bruin 2010). If everyone’s public activities are observed and information about them stored by the government or other powerful parties, this increases the risk that people’s rights will be violated by the observers (for example, when the observers secretly enter the homes of the observed at a time when they are known to be absent) and that such rights violations will remain undetected. Correspondingly, knowledge of such surveillance undermines people’s reasonable belief that they are not at risk of such right violations. Thus, even if such violations do not occur, this knowledge can prompt them to take costly precautions, creating obstacles to autonomous choice in other domains.

Second, the impact of knowing that all of one’s public activities — such as one’s movements, one’s shopping habits, one’s conversations outside of private spaces, and even one’s expressions of emotion — can be constantly captured and potentially communicated to everyone, is clearly enormous not only in a psychological sense, but also in respect to one’s sense of dignity. This impact threatens to destroy any meaningful opportunity to live one’s life autonomously, at least given the present structure of society which makes it impossible to avoid any such disclosure (Westin 1970, p. 31; Roessler 2004, p. 87), as well as one’s sense of having a status as a human being with inherent dignity (Floridi 2016).

Third, authors have drawn on John Stuart Mill’s intuition that democratic societies always run the risk of establishing a ‘tyranny of the majority’ that imposes its particular views regarding the good life or regarding proper behavior on non-consenting minorities, precluding not only ‘experiments in living’ but also any meaningful development of a sense of individuality (Lever 2006). As long as universal tolerance remains an unachieved ideal, intentional non-attention in public seems to form one of the most promising social practices that allow people with wildly different conceptions of the good life to refrain from interfering with each other (Nagel 1998; Mokrosinska 2017).

Fourth, authors have argued more generally that privacy in public is a necessary extension of the private space that people need to develop both their individual autonomy and their intimate relationships with others. Such authors typically reject the misleading spatial connotation of the idea of a ‘private space’ as constrained to particular environments. According to such a view, ‘private space’ refers to having the opportunities for intimacy and for the presentation of different aspects of the self in different contexts (Rachels 1975; Roessler and Mokrosinska 2015; Roessler 2016). As such opportunities tend to be undermined by the unrestricted availability of all information to those with whom one interacts socially, restrictions are necessary. On this view, there is no need, however, for a separate argument for privacy in public as opposed to other forms.

These four arguments refer to the value of privacy in public for the agent whose public behavior is potentially observed or shared with others. It is therefore unsurprising that what is protected by the relevant norms are primarily those forms of public behavior that are merely accidentally public, i. e. those that do not depend on publicity and that the agent would reasonably want to remove from public scrutiny if she could. Agents often lack meaningful opportunities to conduct all relevant activities in non-public contexts; in the least, such a choice can be prohibitively inconvenient and expensive. Our legal norms solve this conundrum, shielding such activities from the risks that usually go along with public actions.

By contrast, these arguments do not work well for intentionally public behavior, in particular for forms of communication by which agents choose to address an unspecified public. Engaging in such communication is often seen as a deliberate choice to give up privacy, to incur the psychological costs of public scrutiny and to expose oneself to evaluation by the majority. Quite plausibly, protecting people’s intentionally public speech against surveillance is also not necessary for intimacy or individual self-exploration. Once we have to make trade-offs between the protection of different forms of privacy, these arguments will therefore naturally lead us towards giving up the privacy of intentionally public activities first (see also Stahl 2019).

A large part of privacy jurisprudence is ultimately based on one of these arguments – while in the EU’s GDPR the role of human dignity is more prominent (see European Parliament 2016, Article 88) and in the U.S. and Canadian context liberty and autonomy are the more important values, respectively (Levin and Nicholson 2005, pp. 382 ff.), the different justifications of privacy agree in their individualist orientation. It is therefore not surprising that, for example, the European Convention of Human Rights limits the right to privacy explicitly to ‘private and family life’ (European Convention for the Protection of Human Rights and Fundamental Freedoms 1998, Article 8). Courts in many jurisdictions have excluded activities from protection that do not allow for a ‘reasonable expectation of privacy’ (Wilkins 1987), most prominently in the United States, where this standard has historically resulted in courts accepting the ‘third party doctrine’ that entails that, once people share information with others, they lose any reasonable expectation of privacy (for an overview, see Issacharoff and Wirsha 2016). This doctrine, which is rooted, first, in the idea that privacy protects the individual’s choice to share information and, second, in the idea that the possibility of such choices does not necessitate securing further control after sharing, has recently come under attack for its counter-intuitive implications regarding digital technology which almost always involves third-party providers and which thus threatens to destroy the possibility of a reasonable expectation of privacy in its entirety. U.S. Supreme Court Justice Sotomayor thus raised the question whether the third-party doctrine falsely treats ‘secrecy as a prerequisite for privacy’ (United States Supreme Court 2012, par. 957). However, even such criticism reflects continuing adherence to the idea that a ‘reasonable expectation’ of being unobserved is a precondition for the applicability of privacy rights. This rules out privacy protections for intentionally public activities, such as political speech. While courts also sometimes limit state surveillance of political protest (e. g. in European Court of Human Rights 2019), such judgments are often based either on the issue of ‘personal information’, on the assumption of ‘chilling effects’ (e. g. in United States Supreme Court 1963) or on a rather unclear connection between privacy and freedom of expression.

If one wants to take seriously the often-cited intuition that ubiquitous surveillance is a danger to democracy, it is therefore essential to provide a better rationale for protecting the privacy of intentionally public activities.

3 The context of the public sphere

In this section, I will turn to a different approach to the problem of privacy in public that has been developed by Helen Nissenbaum (Nissenbaum 1997, 2010, 2004). While I endorse Nissenbaum’s general approach, I will argue that she needs a substantive conception of the public sphere to develop a theory that leads to meaningful recommendations. For this purpose, I will introduce Jürgen Habermas’s theory of the public sphere. In the next section, I will argue that his theory allows us to understand a thus far neglected dimension of harms that follow from a lack of privacy in public.

Nissenbaum — mainly drawing on examples of commercial data collection — argues that the traditional debate about privacy in public includes a number of mistaken assumptions. The most important of these is the assumption that there is a realm of public information to which no privacy norms apply (Nissenbaum 1997, p. 213). Not only is this assumption not borne out in jurisprudence, but it turns out that what is often called ‘public space’ in a simplifying manner is actually best understood as a multiplicity of contexts (Nissenbaum 1998, p. 215) that are structured by distinct normative expectations concerning the handling of information. Thus, not only is there no single public space, but among all the different contexts that belong to ‘the public’, there is also none which is free of restrictions.

In her subsequent work, Nissenbaum extends this contextualist argument towards a general theory of privacy. Rather than adopting an account according to which privacy protects a single value across all areas of social life, Nissenbaum proposes that we view the main function of privacy as protecting ‘contextual integrity.’ The idea behind this is that we can distinguish between a number of nested contexts in society, each of which comes with a distinct cluster of roles, values, norms and activities (Nissenbaum 2010, p. 133), such as medicine, universities, the family, religion, and the marketplace. In each of these contexts, we find distinct sets of norms regarding the flow of information (Nissenbaum 2010, pp. 145ff., 2004, pp. 120ff.), in particular about what it is appropriate to reveal and how information can be distributed. These norms are not merely empirically accepted; they enjoy prima facie justification because they safeguard the capacity of contexts to produce specific goods. Nissenbaum here draws on Michael Walzer’s theory of complex equality (Walzer 1983). Whereas Walzer holds that specific social contexts constitute the meaning of specific goods and that the meaning of ‘equal distribution’ can only be defined relative to the meaning of the good in question, Nissenbaum makes a parallel argument for the case of information and privacy.

This contextualist argument, Nissenbaum argues convincingly, is able to do without the idea that privacy is bound to a specific sphere of the private to which the sphere of the public is contrasted as one that is free of constraints. Rather, she proposes that we reformulate that distinction as one between different contexts governed by their own norms. She is successful in establishing that this better matches both jurisprudence and citizens’ normative expectations about surveillance in public. However, while she offers a framework for thinking about privacy in public she does not adopt any substantive view about the content of the distinctive goods and values of the public spheres under discussion, that is, political public spheres. If we are to avoid dogmatically assuming that current practices in these spheres are legitimate merely due to having been established, we need some conception of what the distinctive good of the political public is in order to have a critical standard with which to measure both our practice and our intuitions.

I am going to propose that we extend Nissenbaum’s argument by supplementing it with Jürgen Habermas’s normative theory of the public sphere. This has several advantages. Not only does Habermas offer a clear answer to the question of what the good supplied by a specifically political public sphere consists in, his theory of the public sphere is also sociologically grounded, tied to a normative theory of democracy, and widely accepted as one of the most advanced analyses of that concept.

On Habermas’s account, the good that is supplied by the public sphere is legitimacy. In his social theory, he argues that the administrative state apparatus is dependent on a supply of legitimation that it cannot itself produce (Habermas 1975). The reason for this is to be found in a basic distinction between two modes of action coordination. Actions can be coordinated either strategically or communicatively. In a strategic mode of action coordination, participants are oriented towards the consequences their actions will have, given the strategic action orientations of others which provide the motivation to select certain options. In a communicative mode of action coordination, by contrast, actions are coordinated through participants’ reaching an intersubjective agreement about what is to be done. The potential for such intersubjective agreement is to be found in the human capacity for communication. Participants in a practice engage in communicative action when they coordinate their interaction through the raising and evaluation of validity claims, that is, claims about the truth of propositions, the rightness of courses of action and the sincerity of expressions of their inner states. While such claims can always be questioned by others, in which case the speaker must supply reasons for them, once all participants are genuinely communicatively motivated they are committed to accept those claims which they endorse after an exchange of reasons as binding on their own behavior.

Genuine communicative interaction in which participants are unreservedly committed to coordinating their interaction through communication can be contrasted with strategic forms of interaction. In the latter forms, communication in a weak sense is also necessary for informing others of one’s orientations and interests, but the communicative exchange does not serve as the source of motivation for participants (Habermas 1984/1981, vol. 1, pp. 95 and 283). It only helps them to adequately calculate the consequences of actions. Habermas holds that modern societies have differentiated themselves into a number of contexts which we can analyze as purely strategically coordinated, such as the state administration and the market economy. In these contexts, linguistic interaction has been replaced by media (power and money) which enable efficient large-scale cooperation free of the inefficiency of establishing communicative consensus (Habermas 1984/1981, vol. 2, p. 281). It is impossible, however, to convert the entirety of society to these modes of action coordination, since strategically coordinated contexts cannot reproduce the elements of the ‘lifeworld’ (Habermas 1984/1981, vol. 2, pp. 119ff.) — that is, meaning, normative consensus and personality structures — which are preconditions for all successful social integration.

In particular, as Habermas lays out in his political theory, the legitimacy of political decisions ultimately depends on their normative acceptability which can only be established through free deliberation. Deliberation is a distinctive mode of communicative action that is tied to the institutional context of the public sphere. The public sphere is a concept that Habermas repeatedly examines throughout his career (most extensively in Habermas 1991/1962, 1996). It denotes a paradigmatic social practice in which citizens are relieved of their roles as private members of the family or private economic actors, along with their participation in administrative decision-making, and engage with each other on an equal basis to determine the normatively correct response to matters of common interest. For Habermas, it is such deliberation, rather than voting, that is the basic practice of democracy and the only possible source of legitimacy (Habermas 1996, p. 296, 1994).

While Habermas readily acknowledges that the notion of an ideal deliberative situation is an idealization (Habermas 1996, p. 19), he holds that this idealization has empirical effects: If people can act under the presumption that they are interacting with others in a genuine communicative manner, this has social effects that allow for the emergence of specific social forms of integration. What is thereby presumed is that everyone engaged in the interaction is committed to a set of communicative norms that establish equality, openness, genuine argumentative commitment and freedom as principles of the debate (Habermas 1996, p. 305). Historically, these presumptions have taken the form of constitutional guarantees of freedom of speech, the institution of a free press and the emergence of forums which have at least aspired to overcome divisions of class and gender.

This conception of the public sphere not only clarifies the specific goods that are at stake in that social context but is also rooted in a view about the pragmatics of language that allows Habermas to generate sociologically interesting insights into the kinds of norm violations that can undermine the functioning of the public sphere. Habermas’s pragmatics draws, in particular, on Austin’s distinction between illocutionary acts (acts that one performs by saying something, such as making a promise or warning someone) and perlocutionary acts (acts that one performs through saying something in light of the extra-linguistic consequences of it being said, such as frightening someone, see Austin 1962, pp. 98 and 102). The possibility of genuine communicative action consists, on Habermas’s view, in the participants’ capacity to perform illocutionary acts of raising validity claims. This capacity both depends on subjective competence and requires a specific uptake by others, which is a precondition for the possibility of such speech acts in the first place. You can perform such speech acts successfully only if others relate to you as a member of a communicative community.

This gives Habermas’s theory a critical dimension which is cashed out by his theory of ‘colonization’ (Habermas 1984/1981, vol. 2, pp. 332ff.) – a notion that is already implicit in his early work on the public sphere (Habermas 1991/1962, p. 160). The public sphere is endangered once administrative and commercial actors intervene in the lifeworld in order to solve the crises of the respective subsystems, either economic or political. When these subsystems begin to distort the cultural integrative processes of the lifeworld through the juridification of family relations, the modern welfare state, or the emergence of public relations and the strategically calculated creation of loyalty in modern politics, then instrumental action orientations begin to replace genuinely communicative intentions (Habermas 1984/1981 vol. 2, p. 323) which in turn undermines the capacity of communicatively integrated spheres to play their distinctive role and leads either to legitimation crises or to pathologies of alienation and social disintegration.

This description of colonization adds sociological depth to Nissenbaum’s idea that the violation of contextual norms endangers the integration of specific contexts. However, Habermas’s high-level social theory does not yet answer the question of what specific role informational norms play. It merely creates room for the hypothesis that privacy in public might be more than merely necessary for the safeguarding of liberty. Once one realizes that the driving forces behind contemporary forms of systematic surveillance of public activities are almost exclusively strategically motivated actors – the police, intelligence services and corporations – we can ask more specifically about whether it might be an essential dimension of the value of privacy in public that it can serve to protect the conditions of genuinely communicative action against colonization from the administrative or economic subsystems.

While Habermas’s theory extends Nissenbaum’s account to provide a plausible conception of what is at stake in the specific context of deliberative, political publics, I will argue in the next section that the harm done to the public sphere by strategically motivated surveillance can only emerge if we engage in analysis of the speech-act theoretic foundations of communicative action and the role that social context plays in the capacities and roles of both speakers and listeners in public spheres.

4 How surveillance can undermine the public sphere

In the last section, I examined Habermas’s conception of the public sphere in order to identify the specific goods that are at stake in the context of the public. In this section, I will argue that certain forms of surveillance and observation both diminish the public sphere’s ability to supply these goods and that they do so in a freedom-undermining way. While I have so far operated with a relatively vague concept of surveillance, the argument that follows will be concerned with surveillance in a more distinctive sense. In contrast to mere information-gathering or monitoring, I take surveillance to be a form of monitoring that is motivated by strategic interests and, in particular, with monitoring the behavior of people and the relation of this behavior to the monitoring agent’s interests (Marx 2015; Dandeker 1990, p. 37). As such, surveillance is intrinsically tied to strategic action contexts. Furthermore, in the examples that I discuss, I will mostly be concerned with surveillance exercised by distinctively strategically motivated actors—such as administrative or corporate actors—and less with the increasingly important phenomenon of mutual, ‘synoptic’ (Mathiesen 1997) surveillance. Even though I take my arguments also to have implications for the latter phenomenon, I do not explore them here.

It has to be noted that Habermas never considers the question whether privacy in public, that is, the existence of rules which protect public spheres from surveillance, is a precondition of functioning of those public spheres. There are two reasons for this. First, when Habermas talks about privacy, he takes that concept to denote the opportunity of retreating from public view into a private sphere where certain forms of self-discovery are possible. When he argues that ‘the constitutional protection of “privacy” promotes the integrity of private life spheres [and] circumscribes an untouchable zone of personal integrity and independent judgment’ (Habermas 1996, p. 368), he thereby identifies a precondition of political participation, but one which is secured outside of the public sphere (for an extension of this argument regarding the ‘stunting of social and personal development’, see Parsons 2015, p. 7).

Second, when Habermas considers the range of interventions that undermine the integrity of the public sphere, he focuses exclusively on the role of speakers. In particular, he identifies those interventions as the main problem through which strategic intentions migrate covertly into seemingly communicative contexts, such as through public relations or the strategic engineering of political consensus (Habermas 1984/1981, vol. 2, p. 346). In these cases, speakers engage in ‘latently strategic interaction’ (Cook 2004, p. 64), i. e. forms of communication ‘in which the speaker inconspicuously employs illocutionary results for perlocutionary purposes’ (Habermas 1984/1981, vol. 1, p. 305). He never considers, however, whether surveillance — which is clearly an activity through which agents are mostly involved in the role of the hearer — can also have undermining effects.

It is relatively obvious why latently strategic speech undermines the functioning of communicatively integrated spheres. If there are speakers who merely pretend to be motivated by communicative intentions in order to strategically manipulate others to accept some point of view, this risks more than undermining — once discovered — the argumentative force of their own communicative acts. Once participants know that there is a risk that rises above a certain threshold (Cooke 1997, p. 87) that others will participate insincerely in communicative practices, this will also lead to a generalized suspicion towards others’ attempts to enter into deliberation, thereby making the success of such deliberation unlikely.

Similarly, one can easily extend Habermas’s arguments to understand what is problematic with ‘coercive surveillance’ (Parsons et al. 2015, p. 218) practices in which information is gathered for the purpose of sanctioning speakers for their views. Entertaining the possibility that one’s political speech may lead to losing one’s job, being imprisoned or socially excluded may have a coercive effect that directly conflicts with Habermas’s requirement that deliberation must be (as far as possible) free from coercion (Habermas 1990, p. 198). The ‘chilling effects’ of surveillance that result from this mechanism on discourse have been explored empirically (Penney 2016, 2017). It is important to note, however, that the main discourse-undermining force associated with them is not to be located in the surveillance practices themselves, but in the psychological effects of either the actual risk of sanctions established by actors outside of the discourse, or of the speaker’s beliefs about the likelihood of such sanctions (Penney 2016, pp. 126 ff.). While very often, refraining from surveillance is the only realistic way to avoid generating chilling effects, the root cause of these effects are the coercive psychological consequences of sanction threats, not changes in the structure of discourse.

In what follows, a different argument will be explored that focuses on the effects of surveillance on the internal structure of deliberative practices that are under surveillance by agents of either the state bureaucracy or commercial enterprises that merely listen, observe and record information without actively participating in the communicative exchange. According to Habermas, we will have to classify such actors as purely strategically motivated due to their role in the media-steered subsystems of the state and the market. But Habermas never considers the effects of strategically motivated listening on communicative contexts.

In order to examine these effects, I will move beyond Habermas’s theory towards other, more recent attempts to analyze the structures of communication on the basis of speech-act theory, mainly by feminist philosophers of language. Much of this work focuses on an idea developed by John Austin, namely the idea that illocutionary acts, in order to be successfully performed, depend on certain kinds of ‘uptake’ by their audience.

For example, Austin writes that

[u]nless a certain effect is achieved, the illocutionary act will not have been happily, successfully performed. […] I cannot be said to have warned an audience unless it hears what I say and takes what I say in a certain sense. (1962, p. 115)

This straightforwardly entails that people can be blocked from performing certain illocutionary acts successfully if the social conditions are such that they cannot secure the necessary uptake. As part of an unrelated argument, Donald Davidson provides the classic example for this by imagining a situation in which a stage actor attempts to warn the audience that there is a fire by shouting ‘Fire!’ but fails to actually warn them because they consider it part of the play and thus do not provide the necessary uptake (Davidson 1984, p. 7). This gives us a first clue as to how the interaction with listeners shapes the capacities of speakers. Rae Langton has famously taken up this argument with the idea of ‘illocutionary disablement’ (Langton 1993, p. 315) which amounts to a form of silencing. She argues that an actor can be silenced,

not in the sense of rendering his spoken words inaudible or written marks illegible, but in the sense of depriving those sounds and marks of illocutionary force: of preventing those utterances from counting as the actions they were intended to be. (Langton 1993, p. 316)

While Langton discusses silencing mainly in the context of situations where people are denied the authority to perform certain speech acts, I hold that the more general insight of this argument can also be applied to the phenomenon of strategically motivated passive participation as it arises in situations of state or commercial surveillance of public speech. Before I am able to do so, however, a number of issues need to be further examined.

Imagine that a group of people have gathered to protest government inaction in the face of a threatening environmental crisis. As part of their public protest, a speaker lays out a moral argument for the need to change the government’s policies. Alongside the protesters and passersby who stop to listen to the speech, there is a small group of plainclothes police officers in the audience who were assigned the task of making a list of attendees and of the points they make in their speeches. While these officers may or may not also be listening to what is being said as citizens and evaluating the arguments on their merits, they are clearly also listening in another capacity, as bearers of a functional role. In this role, they refrain from taking up the arguments as arguments and treat them (or play their part in a larger collective epistemic procedure by which the arguments are treated) as evidence for what the speaker believes and as establishing some probability of her acting in certain ways in the future. In other words, while the speaker’s intention is presumably to raise a normative validity claim, the uptake that her statements receive is one by which they are treated as testimony about her mental state and dispositions.

While it might seem at first glance that this already establishes that the speaker is silenced by being denied the desired uptake, this is not the case. First, in almost all cases, there will be a part of the audience that still treats the arguments as validity claims (the fellow protesters and passersby). Second, and perhaps more importantly, according to the dominant view in speech act theory the officers in the example do not fail to provide the desired uptake. According to this view, in order for an illocutionary act to receive uptake it is sufficient that the audience correctly recognizes the speaker’s intention to perform such an act (Hornsby and Langton 1998; Tanesini 2018). The problem in Davidson’s theater example, on this account, is that the audience does not recognize the speaker’s illocutionary intention, whereas in the present example the officers most likely do recognize that the speaker has the intention of raising a validity claim, although they do not respond to that intention in the desired way. To use an even more obvious example: If you warn me that the Illuminati are after me but I consider you deluded and take your warning only as evidence of your mental state, then one might say that you were in fact able to perform the speech act of warning, albeit in some sense unsuccessfully, and therefore were not silenced. You merely failed to be taken seriously. Not being taken seriously, however, is not the same as not being able to perform a speech act.

For these reasons, surveillance does not directly result in silencing. However, I will argue that it may have the effect of indirectly silencing speakers in virtue of the unintended but recognizable perlocutionary effects that it attaches to speech acts. The issue in the example is not that the speaker is unable to perform certain speech acts but that she cannot avoid doing so without also doing something else. While the speaker succeeds in raising a validity claim, she can do so only by simultaneously also supplying the strategically motivated listeners evidence about her mental states, which is taken up by them in unintended ways. Informally, one could say that the speaker intends to merely perform the act of raising a validity claim, but that option has been removed by the surveillance agent. It has been replaced with another option that consists in performing two actions at once: raising a validity claim (and performing the associated speech act) and providing evidence about her mental states. This remains the case even when we admit that the sincerely communicatively motivated part of the audience provides the speaker with the desired uptake and takes her seriously, and even when the speaker is unaware of the surveillance. The second element in this option bundle — providing evidence about one’s mental states — is not an illocutionary act that the speaker performs, however, since she has no intention to that effect (in the standard case) that could be correctly recognized by the audience. That others acquire knowledge about the speaker’s mental states is rather an unintended perlocutionary effect of the speech act.

As such, this is not yet problematic. In almost any speech situation, what we say will affect our audience in various intended or unintended ways, and typically we cannot avoid a variety of perlocutionary effects. Indeed, most communicatively motivated members of the audience of a political speech will typically also acquire certain beliefs about the thoughts of the speaker, although inducing such beliefs is not intended by the speaker.

The problem in the surveillance case is not this general phenomenon but the specific social dynamic that it generates. This dynamic has two parts. The first part is one in which surveillance undermines communicative speaker roles. Put briefly, once a speaker knows that her speech act will generate a certain perlocutionary effect on a part of the audience that is strategically relevant to her, it makes it rational for her to view her own speech no longer exclusively under the description of producing validity claims but also as part of a strategic interaction with the relevant part of the audience. In other words, knowledge about (the risk of) producing such effects rationally induces the speaker to reconceive her conversational role in a way that introduces a new set of motivations as role-appropriate, a change which then undermines the idealizing assumption that all participants are communicatively motivated which is necessary for participating in a deliberative situation. Not all foreseeable perlocutionary effects have this consequence—when I do not anticipate that those effects will be strategically relevant to me in that they inform further interactions with those whom they affect in a non-communicative mode, they provide no rational motivation to step out of the communicative role-understanding. However, in many cases of governmental or commercial surveillance of deliberation, speakers who are aware of the surveillance are rationally justified in anticipating that there will be a risk that the records which their speech creates (which are typically persistent and available to an a priori undefined range of agents) will inform further non-communicative interactions that affect their interests.

What does it mean for a speaker to react in this way to surveillance? If she is partly strategically motivated in terms of how she presents herself, she cannot openly attempt to get surveillance agents to believe that she has certain beliefs or intentions. Such an attempt would most likely be self-defeating. Rather, she must couch her seemingly sincerely communicatively motivated arguments in terms that make it likely that the surveillance agents will draw inferences that she thinks are in her interest. In other words, the most rational way for a speaker to influence her strategic position is to engage in latently strategic action.

Of course, this is not to claim that this always happens. Speakers can (and perhaps often should) resist the temptation to consider the strategic effects of their speech. This is where the second part of the dynamic comes in — the part that concerns the effects of the possibility of surveillance on an initially communicatively motivated audience. Assume that the audience knows that there is a risk of surveillance, and they assume that this fact is epistemically accessible to the speaker. On the basis of the above argument, they must conclude both that it is rational for the speaker to consider communicative and strategic considerations when raising validity claims and that this is therefore (at least somewhat) likely to happen. This will undermine their trust that, in any given case, the speaker’s actual speech acts are purely communicatively motivated, i. e. that the speaker is ‘unreservedly pursuing illocutionary aims’ (Habermas 1984/1981, vol. 1, p. 305). Rather, they must take into account the possibility that the speaker is engaged in latently strategic action. This possibility, however, will undermine the communicative force of the speech act (Cooke 1997, p. 87): Once I suspect that someone is raising a validity claim merely to latently pursue perlocutionary aims, the relevant ‘speech acts are robbed of their illocutionary binding and bonding power’ (Habermas 1998, p. 332). It is important to note that this differs from the coercion argument discussed above, insofar it does not assume that discourse-external risks do not allow speakers to freely participate, but rather that their objectively changed role within the discourse makes reasons available to them for changing their behavior that others then can rationally ascribe to them, and that this change undermines their intended speaker role.

The degree to which the conditions for illocutionary success are undercut by suspicions on the side of the audience that speech acts are latently strategically motivated will depend on the features of the situation — in particular, on what is at stake in the strategic interaction between the speaker and the surveillance agents, how likely it is that surveillance is taking place, and whether the speaker and the audience know about this. At least in some cases, however, this suspicion can lead to a situation where the audience is justified in provisionally treating the speaker’s claims as strategic moves.

At this point, it might be helpful to compare the surveillance case with two other situations that also involve the interaction between strategic actors and public deliberation. A first situation is one in which government agencies or corporations transparently enter into public deliberative processes to gauge the degree of public support for various policies they might implement. This case is structurally similar to the surveillance case, except for the major difference that, what is of interest is typically how many people have certain attitudes towards a policy but not who those people are. While such consultations can have discourse-undermining effects if the stakes are high enough to motivate people to misrepresent their actual beliefs to further their personal interests, typically they do not give speakers a reason to try to strategically manipulate their own contribution in order to generate information about themselves as individuals. A second situation is one in which government official or corporate representatives actively engage with the public to argue for policies they support. Even though this case also involves strategic actors listening, this case is less troubling, as even the strategic participants must recognize that they can sway the opinions of their interlocutors only by presenting reasoned arguments. In this situation, the speaker position of the genuinely communicatively participants is not undermined.

Both cases contrast with the surveillance case insofar the latter introduces specific, powerful changes in the role of the originally communicatively motivated speakers. I will now argue that if this danger is realized, the speakers do not receive the intended uptake by the audience.

Of course, it is still true that their original, communicative intention will continue to be recognized by the audience. However, while this has traditionally been held to be sufficient for a speech act’s receiving uptake, authors such as Rebecca Kukla have more recently argued for speech act theoretic models that do not support this conclusion. According to Kukla,

[t]he uptake of a speech act is others’ enacted recognition of its impact on social space. Intentions in speaking are part of the story that gives a speech act the performative force it has, but they are not privileged or definitive; The speaker may only discover, in how her utterance is taken up, what sort of speech act it really was. (2014, p. 444)

Kukla argues that the force of a speech act is not exclusively constituted by the audience’s recognition of the speaker’s intention (which comes after the speech act as a separate event). This force is rather constituted by ‘responsive action’ on the part of the audience (Kukla 2014, p. 454) which can even constitute the speech act as one that the speaker did not intend even if the audience correctly recognizes the intention (Kukla 2014, p. 445; see also Cooke 1997, p. 61 for a similar interpretation of Habermas). This leads to a new understanding of silencing (see also Tanesini 2018) that is helpful for analyzing the case of surveillance. If, due to the presence of strategically motivated listeners, a discourse situation has changed such that even the communicatively motivated parts of the audience have reason to treat speech acts as potentially latently strategically motivated, the link between those speech acts and their intended normative effects becomes weakened or even severed. This undermines the ability of speakers (even if they reject any strategic conception of their communicative role) to perform the intervention into the space of reasons that they want to affect, thereby silencing them.

This long detour through speech act theory has been necessary for arriving at a nuanced and empirically adequate way to understand how surveillance can lead to a form of colonization in the sense described by Habermas. Habermas argues that the intrusion of the market and the political system into the communicatively structured lifeworld undermines the normative capacities of that mode of interaction via the restructuring of social relations through the dominance of strategically defined roles (of the consumer or the welfare state client, in Habermas’s examples). Similarly, surveillance by strategically motivated actors can restructure deliberative relations when audiences can assume that it is rational for speakers to latently endorse strategic motives. The intrusion of strategic motivations undermines the binding force of communication and thereby the legitimizing force of political discourse for, among other things, the surveillance policies themselves. This is not only a legitimation problem; by forcing such role-understandings onto participants, uninvited surveillance also diminishes the collective freedom of deliberating groups (see also Stahl 2016). Furthermore, as the contributions by feminist philosophers of language show, the way in which speakers are silenced through these colonization processes also allows us to understand surveillance as a violation of those forms of individual liberty that are constitutive of the public realm.

5 Conclusion

What, then, is wrong with surveillance of the political public sphere? I have argued that surveillance practices where agents monitor public speech, identify speakers and treat their speech as testimony about their beliefs, is particularly problematic for the public sphere in a way in which other forms of interaction between governmental and corporate agents and the public are not. We only capture the full harm done by such practices — and thus the full justification for imposing legal limitations on them — once we shift our focus beyond their effects on individuals and focus on the collective value of intentionally public activities. The argument presented in this article is that surveillance of contributions to public deliberation harms collective practices in various ways. First, it introduces strategic action orientations into a communicative situation and thereby eliminates the possibility of a fully communicative speech situation, constraining the collective liberty of the participants. Second, by making it reasonable for participants to attribute latently strategic motives to each other, it undermines the trust necessary for unconstrained deliberation. Third, in severe cases, it leads to a situation in which speakers fail to secure the necessary uptake for their intended speech acts and can thereby be constrained in their ability to perform these acts. All of these harms additionally lead to the diminishment of the public sphere’s capacity to critically evaluate political decision-making, to inform formal democratic procedures and to generate legitimacy.

While the argument has to be further developed to be directly applicable to policy and jurisprudence, one can at least draw some immediate conclusions. First, in contrast to other arguments which focus on the way in which people’s personal liberties at least partially depend on them not being under surveillance in public space, this argument suggests a particular need for protection of those public interactions which are necessary for a functioning democracy. It makes a distinction between the surveillance of democratic political debate and other forms, such as the large-scale monitoring of commercial interactions, of consumption practices or of traffic. While the latter types of surveillance might be problematic for other reasons, they pose a less direct danger to democratic legitimacy.

As the argument focuses on the preservation of a capacity present in the public sphere, not its exercise, it suggests, second, that the practice of public deliberation is prima facie worthy of special legal protection. There are, of course, forms of political speech—such as racist hate speech or the advocacy of terrorism—that merely appear to be forms of communicative action, but in fact violate the requirements of deliberative openness to such a degree as to lose any special claim to protection. In such cases, where the opportunities offered by the public sphere are abused to plan its demise, the need to protect the foundations of democracy clearly overrides the prima facie argument against surveillance. At the same time, one should be conscious that the deliberative practices of subordinate groups in particular have historically rarely conformed to idealized descriptions. To apply the argument to actually existing democratic debates, lawmakers and courts should therefore adopt a rather wide definition of the public sphere and be aware of the context of unequal power in which many forms of political deliberation take place.

Third, the main risk that the argument presented in the article identifies is not one which concerns particular instances of surveillance. The occasional surveillance of public events or forms of surveillance in which information is only generated for short-term use may not rise to the level of endangering the functioning of the public sphere. Rather, the argument concerns the proliferation of systematic surveillance of public deliberation combined with long-term data retention and its effect on public interaction. Rather than focusing on individual cases, one ought to adjudicate the permissibility of surveillance from the perspective of the total sum of surveillance practices in a society (see the concept of a ‘total sum of surveillance’ (‘Überwachungsgesamtrechnung’) adopted as a standard by the German Constitutional Court in Bundesverfassungsgericht (2010); see Roßnagel 2010).

Finally, the argument points to a fundamental normative dilemma of surveillance practices in the public sphere: If the justification of surveillance policies depends on their democratic legitimation, the fact that their enactment endangers the very conditions of the possibility of such legitimation suggests that there is a limit to what forms of public surveillance are compatible with democracy, even if they are supported by majorities.

The democratic defense of privacy presented here is a part of a larger overall argument for privacy in public. It is limited to a subset of public activities — namely those that constitute the public sphere in the Habermasian sense — but it suggests that those practices should command the particular attention of lawmakers, the courts and a concerned public.