Grappling with surveillance before datafication

The advancement of academic knowledge can be described as a long conversation, where theories and arguments are continuously remodelled, rethought and refined, often in the wake of societal transformations and technological developments. This conversation proceeds through the constant critical scrutiny of previous scholarship, and leaves traces through citations and quotations where new ideas are added to previous achievements. Although Thomas McCoy’s article “Surveillance, privacy and power: Information trumps knowledge” might not be the most-quoted reference in contemporary debates on surveillance, privacy and power, yet, reading it more than 30 years after its original publication, it provokes a number of reflections of relevance for contemporary discussions within the field of media and communications. These questions concern its Foucauldian focus, its early introduction of problematics around “the problem of electronic database surveillance” (McCoy, 1991, p. 33), the role of privacy, and a few other things. In this short commentary, I will continue this conversation, focussing on these three main issues, and relate them to contemporary debates.

First, the article introduces Foucault’s concept of power into the debates about surveillance with a special focus on the consequences for personal integrity and privacy. The article is in fact a follow-up article to a previous introduction of Foucault (McCoy, 1988), where McCoy discusses the relations between Foucault’s relational power concept and the concept of ideology within Cultural Studies, and foremost in the work of Stuart Hall. The point of departure for McCoy’s discussion is Foucault’s power-knowledge nexus (see, e.g., Foucault, 1980), rooted in his “studies of madness, medicine, the prison and sexuality, power relations are functions of what society accepts as knowledge” (McCoy, 1991, p. 33). His argument is that Foucault’s model of power analysis is a more nuanced way of studying power relations than the supposedly dominant “power as coercion” approach. Arguably, this more sophisticated theory of power has since the publication of McCoy’s article become if not dominant, then at least mainstream, in media and communication studies of the culturally oriented kind (which was also the main point of McCoy’s 1988 article). McCoy is by no means the first to introduce Foucault to a wider readership (see, e.g., Dreyfus and Rabinow, 1982), but his text should be seen as a part of the broader “post-structural turn” (although he, in typical 1990s fashion, refers to it as “postmodernism” in the introduction), where continental theory seeps into the discussions in the Anglo-American sphere. However, this occurred early in the debates in Media and Communication Studies and adjacent disciplines, and it notably preceded debates later taken up in the journal Surveillance & Society, which was launched in 2002, more than a decade after McCoy’s article, where many similar debates were pursued–for example in its third issue of volume 1 on “Foucault and panopticism revisited” (Wood, 2002). Its first major manifestation in my own context of Sweden is Annika Sjölander’s PhD thesis on opinion formation in the nuclear waste discourse in a municipality in the north of Sweden from the early 2000s (Sjölander, 2004).

It is also to note that McCoy’s focus is journalism, and those of us who were active at this time in the early 1990s will remember that the introduction of continental, or, perhaps more accurately, French philosophy (Bourdieu, Baudrillard, Foucault, Derrida, etc.), was stronger in the analysis of popular culture, advertising, cinema, and other areas of media and communications research. The most explicit example of Foucauldian theory in media studies is the hugely popular text “The work of representation” by Stuart Hall (1997), in which he thoroughly discusses the relevance of Foucault’s discourse analysis for analysing media content. However, it shall be noted that Hall does not mention McCoy in his text.

Second, and perhaps more importantly for contemporary media and communications debates, McCoy’s article is an example of the early attempts at understanding the changing nature of surveillance towards what would today be termed “surveillance capitalism” (Zuboff, 2015). Although “capitalism” is not one of the parameters discussed in his article (another significant feature of the scholarly debates in the 1990s), and this is significant for most Foucauldian analysis, since his focus on “the problem of electronic database surveillance” (McCoy, 1991, p. 33) paves the way for one part in the equation of contemporary “datafication” and “dataveillance,” that is, the concept referring to the amassment of data across social media and other online platforms (e.g., search engines) for the creation of consumer profiles on the individual media user level for the sake of economic profit. This type of data collection based on users’ navigation in digital space was, of course, not present at the time McCoy wrote his article, but the consequences of this kind of dataveillance–incidentally a concept that appears in the same year as McCoy’s first article, launched in the field of computer science by Roger Clarke (1988), but adopted for the social web by José van Dijck (2014)–are obvious today, and theorized by authors such as Andrejevic (2019), Couldry and Mejias (2019), and others, including myself (e.g., Bolin 2011, 2023).

Third, McCoy aligns with a longstanding discussion about information and privacy, himself dating this as far back as the late 19^th century and an article in a very early volume of the Harvard Law Review about “The right to privacy” (Warren and Brandais, 1890). The discussion on privacy is, indeed, the main point of McCoy’s argument, and he argues that “information invades privacy” and creates specific regimes of truth about citizens. Now, it is true that Warren and Brandais do discuss privacy intrusion from, for example, the press into details of private persons, but the bulk of their argument is about the legal protection of private property, including intellectual property (they are, after all, legal scholars). Hence, they approach privacy from a legal, rather than a social or cultural, standpoint.

This part of the argument is, I would suggest, the most dated one. The main problem is that it presupposes that privacy is a natural phenomenon, or an unquestionable given. However, there is a rich literature on the rise of privacy as a phenomenon, one of the main reference points being Philippe Ariès and Georges Duby’s (1987–1991) massive five-volume set of the history of private life. In an argument on “the privacy parenthesis,” British-Danish scholar Tom Pettitt (2013) has pointed out that the idea of privacy differs globally and has shifted in meaning over time. He argues that privacy coincides with the printed word, and that the parenthesis is now closing with the advent of digital, changeable text which becomes more fluid and changeable in the digital world. Privacy, it is thus argued, appears with the rise of bourgeois society, industrialization, new ideologies, and other things following from Western mass industrial society–not least the printing press and its ability to fixate words into standardized form. With digital media, as texts are becoming fluid, changeable, and open for restructuring, updating, and transformation, this fixed form disappears. This coincides with a more pervasive mapping of all individuals, for the sake of producing the digital consumer, and, therefore, privacy has come to be revalued as a phenomenon, albeit in a different way than it was before privacy became a norm. Seen from this perspective, the changed nature of privacy coincides with more broad societal changes, and to mourn its disappearance is at best nostalgic.

One of the peculiarities of the article as a whole is the treatment of its main concepts: information and knowledge. Information is mentioned 115 times in the article, but nowhere do we get a definition of what information is. This is all the stranger, since this is the main point of the article, as stated in the subtitle “Information trumps knowledge,” and it is only towards the very end of the article, in the conclusions, that McCoy discusses the relation between information and the other key term “knowledge.” Here, on the next to last page of the article, in the last three paragraphs of the conclusion, McCoy refers to Machlup (1983), who distinguishes between the two entities by claiming that information is “atomic and fragmented, whereas knowledge is structured and coherent; information is perishable and fleeting, while knowledge lasts; information carries a current of messages, while knowledge results in a pattern” (McCoy, 1991, p. 45). From the viewpoint of today, with the continuous amassment of data that are far from “perishable” (the internet never forgets, as is often concluded), these distinctions might not hold to scrutiny, and do not contribute much to what “information” ontologically is, in the context of the argument.

One could, in fact, argue that when McCoy discusses “information” in relation to possible privacy concerns, he is implicitly referring to representational information about individuals built on sociological variables such as gender, ethnicity, age, income, etc.–the way in which governments collect dossiers about their citizens. Today we would need to add various forms of non-representational data or information, that is, operational data (Dyson, 2012). Operational data are the data that do things (rather than just represent them). Most data are in fact never read by humans, but they are a part of the consumer profiling that algorithmically directs individually targeted messages and are decisive for the ways in which navigation on the interactive web occurs. Operational data, then, are data that trigger machine action. “Information” in the digital world is thus so much more than the sociological aspects of the individual social subject. Most often it is not the individual social subject but the consumer pattern that is addressed, the digital consumer that is a certain aspect of an individual’s behaviour, sometimes labelled the “data double” (Haggerty and Ericson, 2000). Who you are is less important than what you consume, so to speak. This is also why “you are not important” is both an argument pushed by the platform companies, and a widespread belief and defence mechanism and justification for admitting companies to track the movements among media users, as has been proven in many empirical studies (Deuze, 2012; Bolin, 2018).

Arguably, McCoy could not possibly foresee the sophisticated algorithmically based tracking of citizens and consumers of datafied society. And while his observation that “[p]ersonal information could be correlated, integrated and shared by participating governmental departments and business organizations” (McCoy, 1991, p. 34) points in the direction of increasingly more sophisticated means of dataveillance, the scale and depth of contemporary datafication processes were too hard to imagine at the time. Nonetheless, the questions raised in his article have been recurring over the years since its publication and can be considered as one thread among the many that make up the entangled fabric of the debate on contemporary data-driven surveillance society.