In this paper, we attempt to provide a renewed theorization of ‘mediated visibility’. Our goal is to specify the central tenets of what we term as "algorithmic visibility", a general sociotechnical logic which, we suggest, increasingly structures the public domain.
We begin by critically describing one of the most influential and complete conceptualizations of this topic, that of John Thompson (1995, 2005). This analysis raises the question: How have algorithmic social media altered mediated visibility?
To answer it, we employ Thompson’s categories as a conceptual hub. It is proposed that algorithmic social media are underpinned by a new logic of mediated interaction which gives rise to a different kind of mediated visibility, composed of two dynamically linked vectors. First, "how users can see" (algorithmic gatekeeping), which is importantly defined, in contrast to Thompson’s view, by space- (physical and digital closeness) and time- (instantaneity and simultaneity) related data. Second, "how users can be seen", which was inexistent in Thompson’s model and entails multiple modes of surveillance. These two vectors represent, again contra Thompson, a movement of re-centralization of control.
We then propose that these vectors help to redraw the boundaries, and redefine the nature, of the public domain in at least two ways. First, by reconstructing public knowledge. As ordinary citizens and politicians, private companies and states, old and new media industries increasingly rely on algorithmic platforms to get their messages seen, algorithmic visibility impacts on how, when and what kind of information is shared by whom. Second, by reconfiguring recognition, in Axel Honneth’s sense. This new logic affects the realization of solidarity, the formation of communities, and therefore the possibility of social change.
It is finally noted that algorithmic visibility must be embodied by everyday life practices, what gives rise to new forms of resistance, manipulation and inequality.