Privacy: Essentially Contested, a Family Resemblance Concept, or a Family of Conceptions?
Presented at the 2018 Amsterdam Privacy Conference, 7th October, 2018
17 Pages Posted: 30 Oct 2018
Date Written: October 7, 2018
How to understand the concept of ‘privacy’ and related ideas has been very controversial for a number of years. Solove famously referred to privacy as a ‘concept in disarray’. This certainly goes beyond matters of semantics. For instance, understanding the value of privacy depends on how it is conceptualised, while the apparent incoherence of privacy has led some to call for its abandonment as a normative principle.
Different diagnoses of the definitional problem have emerged, in particular that privacy is simply a disparate set of ideas in need of typology, not unification, an essentially-contested concept, or a family resemblance concept. This paper argues that the multiplicity of interpretations of privacy isn’t down to any of these. On Wittgenstein’s interpretation, family resemblance explains agreement, not disagreement, and so applying his theory to privacy does not induce the intended contrast. With regard to privacy as an essentially contested concept, the normative status of privacy itself is disputed, and so it is not clear that privacy can be classed in this way.
This paper will argue that the heterogeneity of views of privacy comes down to something more like the parallel coexistence of different conceptions of privacy in Rawls’ sense (where Rawls talks of different coexisting conceptions of justice). These conceptions coexist at least partly because different definitions of privacy have been drawn up in the context of pre-existing views of its value (or otherwise). A definition will typically have baked within it indications of the value that privacy provides. Disputes over privacy’s value then easily degenerate into semantic disputes.
The varying interpretations of privacy can therefore be put down in part to essential issues with the multiplicity of its application, as many commentators have argued, but also in part to its perception as an appraisive concept that “signifies or accredits some kind of valued achievement”. This paper argues that the concept of privacy is less like the appraisive concepts that Gallie refers to (such as democracy or social justice), but is rather more like a concept which we can in principle agree, but upon which we still may fundamentally disagree (such as tax or the weather).
The paper therefore argues that the problem of defining privacy is not down to any particular issue with the concept itself, but rather the inability or reluctance of commentators to develop a value-free notion of privacy that can be the topic of debate, or to separate out the several different conversations that take place about privacy, where its value (or cost) to different agents or entities is debated. This reluctance typically leads to category error and interlocutors talking past each other. The paper sets out seven levels of privacy debate which are often illicitly intertwined.
Level 1: Conceptions. The basic level is that value-free conceptions of privacy need to be separated and understood. New conceptions often emerge with new technologies (e.g. the smartphone, machine learning) or practices (e.g. social networking), and often (but certainly not always) emerge as a result of some perceived harm or intrusion caused by a new technology. Definitional debates sit at this level.
Level 2: Empirical measurement. Given a conception of privacy, the second level of discussion is whether, as a matter of fact, one has or has not privacy in that particular sense. This idea of ‘the facts on the ground’ is nothing to do with privacy’s value. It may be that I have no privacy in terms of a conception that I simply do not care about (e.g. my face is not veiled).
Level 3: Phenomenology. The phenomenology of privacy is relatively unexplored, yet it emerges strongly in the area of design, where means of making conversations ‘feel’ private on social networking sites have been developed.
Level 4: Preferences. An individual will have certain preferences, that need not be rational or consistent, about whether he or she is private (in a particular respect), and when he/she is visible to his/her network. People also have preferences about others’ privacy – e.g. about whether family members have privacy, or about what privacy is afforded to public figures. At this level, for instance, we might include the sociology of boundary negotiation. Variance among these preferences makes it hard to generalise about them, and level 4 discussions often concern control, with mechanisms such as content and personal data stores aimed at allowing individuals to achieve their privacy preferences, rather than privacy itself.
Level 5: Social norms. Norms, regularities and expectations provide guides for privacy protection and legitimate breaches, and typically vary across culture, gender, class and generation. There are many complex trade-offs to be made, especially as new conceptions of privacy emerge, and the idea of contextual integrity is situated at this level.
Level 6: Law and regulation. Certain types of privacy breach and protection are mandated or forbidden by law. Privacy is not a legal concept (unlike data protection), and is not exhausted as a concept by its regulation. Despite this, a number of analyses of privacy focus on the disparate nature of privacy protections in law, rather than in social life.
Level 7: Ethics and politics. The value of privacy, the roles it plays in our social life, and the threats to other values (e.g. security) are the subjects of discussion at this top level.
The argument of this paper is that the debates that take place at these seven levels need to be separated out to avoid confusion (e.g. to avoid the erroneous assumption that one’s privacy preferences [level 4] must be reasonable [level 5]). The different levels certainly interact, but conceptual hygiene means they must be kept separate.
Suggested Citation: Suggested Citation