1
What is digital sociology?

In social research, as in other fields, the idea has taken root that the digital makes possible new ways of contributing to society. Actual efforts to realize this promise of the digital have proven the initial optimism to be partly misguided. One sobering example is the Samaritan Radar, a social media application that was launched by the Samaritans, an important UK suicide prevention agency, in October 2014. At its launch, the tool was introduced as a way of identifying users at risk of suicide by way of real-time, textual analysis of Twitter data. Once an 'at risk' account has been detected, the Radar would send a message to the followers of the identified account alerting them and 'offering guidance on the best way of reaching out and providing support'.1 Perhaps unsurprisingly, the Samaritans were forced to close the experimental service after a short time, and it was subjected to harsh criticisms in both news and online media. Many argued that notifying people's social media contacts of their supposed malaise without prior consent amounted to a 'privacy violation', while some flagged the risk of stigmatization of individuals already deemed 'at risk'. Yet others questioned the hubristic presumption that a complex and sensitive phenomenon like suicide risk could be detected and managed using simple methods of data analytics.2 Indeed, social researchers could no doubt propose different, better methods to understand and communicate with people in trouble using social media, and providing the impulse to do so could be one positive outcome of this episode.3 However, the Samaritan Radar debacle also sheds light on a wider, rather diffuse phenomenon, namely the remarkably strong expectations, in our societies, that digital technology will make it possible for social research to help solve social problems.

Digital technology presents an important societal phenomenon today, as popular online platforms like Facebook, smart phones and 'intelligent' computational systems have been taken up across the full breadth of society during the last decade or so, from transport to education, from family life to activism, from prison management to wildlife conservation. Whereas the digital used to refer to a fairly special set of practices, those that early adopters, experts, the 'tech savvy' and the young engaged in, today it touches on most aspects of social life. This development has important implications for sociology. But the ongoing digitization of society does not only present an important topic of investigation, it also has the potential to transform the very role that social research itself plays in society. Across society, digital infrastructures, devices and practices are widely seen to offer important, new opportunities for making social research relevant to social life (Back, 2012), for turning knowledge about society into action.4 As I will discuss in this chapter, what distinguishes the digital technologies of today - what sets them apart from the 'Web' and 'information and communication technologies' (ICT) that went before - is their extensive capabilities for monitoring, analysing and informing social life. Today's digital infrastructures, devices and practices collect an abundance of data that can be used to analyse people's interactions and movements, from the SMS exchanges captured by phone companies to the location data amassed through smart phone apps. They also make it possible to translate data analysis into targeted feedback in everyday settings and user activities, from the query terms suggested by search engines, to the personalized updates that are offered by transport and weather apps and other digital services. It is these interactive capacities of digital technology, in combination with its ubiquity in society - the fact that digital technology can appear to be everywhere - that today feed the conviction that the digital makes it possible to re-connect social analysis with social intervention, as in the example of the Samaritan Radar app above.

What makes the digital such a relevant phenomenon for sociology then goes beyond its importance as a research topic. Its contemporary significance must also be understood in terms of the transformations of social research, and of its role in society that it makes possible. These transformations have been described in various ways, but they can be summed up in the belief that social research, through its implementation in computational infrastructures, may gain the capacity to intervene in social life and thereby to address or even solve social problems. The Samaritan Radar project was presented as a way of taking advantage of the widespread uptake of a social media platform like Twitter across society for a progressive purpose, and it did this by outlining a new way of using methods of textual analysis to act on the issue of suicide risk. As such, this project offers a clear demonstration of the belief in the power of the digital to confer onto the analysis of social life the capacity to help solve social problems. However, once the project was underway, multiple challenges to this ambition came into view, such as the risks of privacy violation and stigmatization. Furthermore, as a blogger speculated a few months later, the fact that a suicide prevention agency is monitoring social media might even lead users to practise self-censorship, thereby affecting the very fabric of interaction in these settings.5 As such, the Samaritan Radar episode can also be interpreted as a kind of 'critical' test of progressive hopes invested in the digital. This is partly what makes it such a relevant case from a sociological perspective.

I would like to argue that the digital today does offer fresh opportunities for connecting social analysis and social intervention, but not in the way in which this promise is usually understood. This is because digital societies are marked by far more complex interactions between social life and knowledge - between social research and social action - than tends to be recognized when data analysis is put forward as a way of acting on social problems. The Samaritan Radar episode highlighted some of these more complex dynamics. This tool did not only facilitate interaction - feedback - in the technical sense of sending a notification to an identified user's social media contacts. It also brought into relief more comprehensive forms of interaction between social research and social life. When the monitoring and analysis of everyday activities is used as a basis for intervention into these activities, a complex set of exchanges between knowledge and behaviour is set in motion, as the public debate that followed on the launch of Samaritan Radar also highlighted. When users are identified as a 'suicide risk', this designation may initiate a dynamic in which social concepts - like 'suicide risk' - and social life interreact: once individuals (as well as others) are 'aware of how they are classified', this produces a situation in which these actors are likely to 'modify their behavior accordingly', to quote the description that social theorist and philosopher Ian Hacking (2000, p.32) has provided of what he calls the 'interactivity' between social research and social life. This type of transformative effect, by which the description of a social situation transforms that situation, has long been of special concern to sociologists and philosophers (Thomas and Thomas, 1928; Becker, 1963). As I will discuss in what follows, one of the key questions that arises in digital societies is how computational forms of interaction at work here combine with sociological dynamics of interactivity between knowledge and social life.

In this book, I would then like to propose that the digital opens up new occasions for interaction and interactivity between social life, technology and knowledge, and that these form a central challenge for sociology in a digital age. The proliferation of computational infrastructures, devices and practices across society has given rise to new forms of exchange and mutual adjustment between social research and social life, a development in which social, technical and epistemic processes fuse in ways we need to understand much better than we do now. On a general level, it means that if we are to grasp the significance of the digital for sociology, we must recognize that interaction is not just a notable technical feature of digital technology today. Interactivity - in the broad sense of exchange and traffic between the analysis of activities and those activities themselves - presents a crucial sociological dynamic. Indeed, in the classic view of the early twentieth-century sociologist Max Weber (1905/1968), this is the defining challenge of social science. As Weber famously noted, what distinguishes social enquiry from other forms of research is that it must contend with the fact that the ideas people have about the social world interact with what happens in it. As sociologists have since argued, social research presents a special form of knowledge insofar as it is inherently interactive: social research must expect, and indeed anticipate, that knowledge about social life and social life itself mutually influence one another (Cicourel, 1964). This is also to say that social dynamics of interactivity are not at all new in themselves. However, in today's digital societies, technology must increasingly be factored...