Gateways

Front matter

LAI & FLENSBURG WAYS GATE COMPARING DIGITAL COMMUNICATION SYSTEMS IN NORDIC WELFARE STATES comparing digital communication systems in nordic welfare states GATE WAYS

Gateways Comparing Digital Communication Systems in Nordic Welfare States Signe Sophus Lai & Sofie Flensburg © 2023 authors. This work is licensed under the terms of the Creative Commons Attribution- NonCommercial-NoDerivatives 4.0 International Public licence (CC BY-NC-ND 4.0).

To view a copy of the licence, visit creativecommons.org/licenses/by-nc-nd/4.0/ ISBN (PRINT): 978-91-88855-83-1 ISBN (PDF): 978-91-88855-84-8 ISBN (EPUB): 978-91-88855-87-9 This work is Open Access and available to read free of charge and without registration requirement at www.nordicom.gu.se/en/publications/gateways DOI: 10.48335/9789188855848 CITE THIS WORK: Lai, S. S., & Flensburg, S. (2023). Gateways: Comparing digital communication systems in Nordic welfare states. Nordicom, University of Gothenburg. https://doi.org/10.48335/9789188855848 PUBLISHED BY: Nordicom, University of Gothenburg Box 713 SE 405 30 Göteborg, Sweden www.nordicom.gu.se/en GRAPHIC DESIGN BY : Karin Andén COVER BY : Jeppe Sophus Lai COVER IMAGE GENERATED WITH: DALL-E 2 VISUALISATI ONS MADE WITH: RAWGraphs by DensityDesign, Calibro & Inmagik signe sophus l ai & sofie flensburg GATE COMP COMMUNI WA ARING DIGITAL CATION SY YS STEMS IN NORDIC WELFARE STATES

Part I Departures

CHAPTER 1 Follow the data Technological systems, even after prolonged growth and consolidation, do not become autonomous; they acquire momentum. Momentum does not contradict the doctrine of social construction of technology, and it does not support the erroneous belief in technological determinism. The metaphor encompasses both structural factors and contingent events. HUGHES, 1987: 76 In 1973, the Norwegian town of Keller connected to the US-based Arpanet via satellite, placing the Northern corner of Europe as a critical hub in the nascent geography of what would later become the global Internet (Hetland, 1999). Instituting the rise of the digital communication systems studied in this book, this first gateway to the original network of networks, and the following addition of numerous fibre-optic data highways, enabled the Internet to migrate from its American origins to the societal structures of Denmark, Finland, Norway, and Sweden and the everyday lives of Nordic citizens. Supported by a rapid and successful diffusion of digital devices, network connections, and Internet-based services, the Nordics evolved within the course of a few decades from analogue to digital societies – and they are, by now, among the most well-connected regions in the world. In other words, the Internet has gained momentum (Hughes, 1994) as a general-purpose technology (Naughton, 2016) that supports a wide range of mundane activities and societal functions (Edwards, 2002). As Thomas P. Hughes (1987) noted in the quote cited above, significant infrastructural evolutions, like electrification and digitalisation, call for nu - anced understandings of the interplay between technological innovations and societal arrangements. While the breakthrough of the Internet was indeed dependent on the invention of computers, the build-out of network archi - tectures, and improvements in data processing, these technologies were also promoted by government policies, economic interests, and sociocultural

26 SIGNE SOPHUS LAI & SOFIE FLENSBURG norms. As we elaborate further in the book’s analytical chapters, the Nordic welfare states initiated, encouraged, facilitated, and funded a wide range of infrastructure projects, following historically rooted principles of universality and decommodification (Esping-Andersen, 1990). Yet, the tremendous success of these strategies also altered the material conditions that grounded funda - mental welfare ideologies and policy frameworks, leading to a disruption of established modes of governance, institutional logics, and business models. The combination of comprehensive digitalisation processes and strong welfare state traditions means that the Nordic countries are critical cases (Flyvbjerg, 2006) for studying how new technological conditions merge and clash with established institutional structures. The adoption of technologies developed in and for an American context in the welfare systems of Northern Europe offers a unique occasion for studying how institutional structures shape the infrastructural development of the Internet – and vice versa. By studying the structural implications of these evolutionary processes, we gain insight into both the historical path dependencies and the disruptive poten - tials of (digital) technologies. Our analyses thereby touch upon and provide an empirical contribution to heated and long-lasting theoretical discussions about the interplay between technologies and societies. Addressing these classic theoretical enigmas and relating them to our research questions, we in this chapter review and discuss existing literature that addresses the relationship between (new) communication technologies and (old) institutional arrangements.

Before embarking on our empirical exploration of the infrastructure, market, and state forces that influence (digital) communication capabilities in Denmark, Finland, Norway, and Sweden, we discuss existing frameworks and approaches across media and communication studies. In doing so, we pay particular attention to comparative typologies used to assess how communication environments are shaped and controlled across contexts. Drawing on established political economist approaches of “following the money” in economic circuits (DeFleur, 1971; Ettema & Whitney, 1994), we argue for an alternate strategy of “following the data” as it travels across boundaries within digital communication systems, regulations, and research on its path between senders and receivers (Flensburg & Lai, 2023). This chapter thereby establishes a theoretical foundation for operationalising the concept of a digital communication system as an epistemic strategy for capturing the complex infrastructural, economic, and political arrangements that structure contemporary digital societies. Epistemic crossroads The key concept we employ in this book – digital communication system – unites classic questions and approaches from political economy and media ecology with contemporary infrastructure and critical data studies.

CHAPTER 1 | FOLLOW THE DATA 27 In developing and conceptualising the notion of a digital communication system, we apply theories about the interplay between societal power and transforming communication environments (Innis, 2007; Mansell, 2004) to the empirical realities identified and discussed by recent studies of Internet and data infrastructures (Hesmondhalgh, 2021; Parks et al., 2015; Plantin & Punathambekar, 2019). In doing so, we contribute to and develop a research agenda focusing on the evolution and institutionalisation of the Internet as a common and increasingly dominant societal infrastructure. As this chapter’s main argument, we unfold how this historical transformation of the underlying material conditions for communicating (and for researching communication) calls for epistemic adjustments of established models for analysing and comparing media and communication structures. The legacies of media systems As briefly mentioned in the introduction, our analyses of digital communica - tion systems stand on the shoulders of decades of research concerned with exploring and comparing national media structures (Brüggemann et al., 2014; Hallin & Mancini, 2004, 2012, 2017; Hardy, 2010; Humphreys, 1996; Humprecht et al., 2022; Mancini, 2020; Siebert et al., 1963; Syvertsen et al., 2014). Focusing on the press, media system analyses investigate how the societal organisation and control of political communication outlets (such as newspapers and broadcast media) are shaped at the intersection between market interests, cultural norms, social development, and state governance.

This research provides valuable typologies for understanding how and why journalism has evolved differently across contexts, while also serving as a key point of reference for studies addressing broader power structures surround - ing national media markets and regulation (see, e.g., Curran et al., 2011; Humprecht et al., 2020; Jungherr et al., 2020; Psychogiopoulou, 2012, 2014; Street, 2010; Tenove, 2020; Van Aelst et al., 2017). According to Daniel C. Hallin and Paolo Mancini in their seminal book, Comparing Media Systems: Three Models of Media and Politics (2004) – and the numerous utilisations of its analytical framework and empirical ap - proaches – Denmark, Finland, Norway, and Sweden are clear-cut examples of so-called democratic corporatist media systems (see Benson et al., 2012; Brüggemann et al., 2014; Enli et al., 2018; Humprecht et al., 2022; Nord, 2008). In comparison with more liberal and polarised contexts (e.g., the US or Mediterranean countries such as France, Italy, and Spain), the large Nordic countries have been characterised by early penetration of mainstream media evolving in parallel with the introduction of democracy, a political culture characterised by collaboration between political stakeholders, and strong involvement of the (welfare) state in the distribution of economic and material resources (Hallin & Mancini, 2004: 143–197). A high consumption of news and flourishing advertisement revenues have been combined with

28 SIGNE SOPHUS LAI & SOFIE FLENSBURG institutionalised state involvement in the form of various support schemes, balancing the fine act of structural regulation while refraining from day-today editorial control (Syvertsen et al., 2014).

Hallin and Mancini’s original study, published in 2004, paid limited attention to the Internet, yet multiple studies have used it as a key reference for comparing digitalised or hybrid media systems (see Hallin & Mancini, 2017; Mancini, 2020; Mattoni & Ceccobelli, 2018). Focusing on “Nordic media in the digital era”, Trine Syvertsen and colleagues (2014: 120) argued “that there is a high degree of continuity regarding the empirical realities of how the media systems operate, whether we are speaking of user patterns, institutions, or content”. Media institutions have supposedly adapted to the changes following digitalisation and have “successfully adjusted their strategies to digital and global markets” (Syvertsen et al., 2014: 122–123). Other studies have opposed “the continuing validity of Comparing Media Systems” (Ohlsson, 2015: 62), arguing that the press has been economically weakened, former bonds between the press and the political sphere have eroded, state involvement in the media market has diminished, and public service broadcasters have lost former monopoly positions. Regardless of whether they confirm existing assumptions or point to devia - tions, studies of the political and economic structures surrounding mediated communication in the Nordics continue to focus on legacy news institutions as empirical entry points, reflecting the mediacentrism bias, discussed in the introductory chapter. They thereby neglect to consider the fundamental transformation of the broader communication environment, as pointed out by several critics of Hallin and Mancini’s work (Bennett, 2015; Hardy, 2010; Norris, 2009). Looking at contemporary media and communication structures in “the rear-view mirror” (Ala-Fossi, 2020: 146), media system analyses tend to reproduce “an old notion of the public sphere in which a system of well-connected institutions […] maintain the information gates through which citizens receive news and ideas about who they are and how society is working” (Bennett, 2015: 162). In doing so, they conveniently overlook “the 800-pound gorilla in the room” – new information and telecommunication technologies – making it appear “as though journalism and the media systems are frozen in the mid-twentieth century” (Norris, 2009: 332). Following this line of argument, empirical studies applying the indicators and variables of Comparing Media Systems circumvent questions about how transformations in the broader media and communication environment (such as the introduction of the Internet) affect the continued validity of the original framework (Flensburg, 2020). Few studies actively engage in epistemic discussions about how contemporary, digitalised media systems can be defined and delimited (for exceptions, see Chadwick, 2017; Mattoni & Ceccobelli, 2018). By continuing to focus on newspaper circulation, legacy media regulation (e.g., public service broadcasting policies and news subsidies),

CHAPTER 1 | FOLLOW THE DATA 29 and journalistic autonomy, most studies bypass important structuring forces, such as news media’s competition with and increasing reliance on digital intermediaries (Nielsen & Ganter, 2018), as well as the growing importance of third-party data tracking and advertisement (Helles et al., 2020; Helles & Flyverbom, 2019; Sørensen & Van den Bulck, 2020). These studies thereby neglect to consider the fundamental difference between analogue and digital distribution in the form of new infrastructural dependencies, altered value chains, and regulatory disruption.

While Nordic scholars largely rely on media system frameworks for assessing the mutual shaping of political systems and communication media (Enli et al., 2018; Jakobsson et al., 2021; Kammer & Hjarvard, 2015; Nord, 2008), alternative approaches are beginning to emerge on a global scale (Gillwald & Simon, 2012; Hunt, 2014; Simon, 2012; Vendil Pallin, 2017). Based on a comparative study of Internet ideologies and data governance, Kieron O’Hara and colleagues (2021: 237) have suggested a typology of four geopolitical Internets: American, European, Chinese, and Russian. These ideal typical Internet regimes are characterised by varying degrees of openness and restrictions; different ideological emphases on state control and innovation freedom; and different network designs. By including dimensions and analytical perspectives that are overlooked by media system scholars, Four Internets (O’Hara et al., 2021) takes an important step towards identifying, characterising, and comparing how digital communication is institutionalised and controlled across contexts. However, much like early attempts to identify media systems (e.g., Siebert et al., 1963), O’Hara and colleagues’ book is an explorative investigation of specific cases and thus does not present a systematic empirical framework for testing, validating, or falsifying existing as well as emerging typologies. In other words, it does not provide an alternative to the dimensions, indicators, and variables of Comparing Media Systems . To sum up, existing studies of national media systems provide an important source of inspiration and a historical backdrop for understanding the rise of digital communication systems. However, research also needs to acknowledge that models and theoretical typologies developed in and for the analogue age are insufficient at best, and misleading at worst, when applied to contempo - rary digital environments. By reproducing variables and indicators without considering their ability to produce valid knowledge about the forces that shape contemporary communication flows, media system analyses become blind to the more fundamental and historical shifts in the ways media and communication are organised and controlled. It is, in other words, time to take a step back and reconsider not only the institutional position of the press, but also the very definition and delimitation of what kind of “system” we study and how we approach it.

30 SIGNE SOPHUS LAI & SOFIE FLENSBURG Institutions & individuals In developing and maturing analytical and methodological approaches that address questions similar to those raised by media system analysists, there is an urgent need for answers that consider the contemporary empirical reali - ties. While acknowledging the legacies of media systems in pointing out the important links between communication media, market power, and policy, we argue that an epistemic turn in the way knowledge is produced is long overdue: away from analysing the structural conditions of (analogue) media institutions as proxies for public enlightenment and democratic debate and towards approaching digital communication environments from the perspec - tive of individuals as ultimate ends of studying communication structures. The conceptual move from (analogue) media to (digital) communication systems epitomises this change of perspective. Building on the broader concept of a communication system (Jensen & Helles, 2022), it represents a (macro) approach to investigating the overarching institutional arrangements that frame mundane communicative activities while also emphasising the historical and material contexts they evolve in. That is, the concept of a communication system recalls “the simple, but crucial fact that each communication technology is a material resource whose distinctive features help to explain the media institutions and communicative practices that have emerged, or which may emerge in the future” (Jensen, 2013: 216). We thus view communication systems, on the one hand, as universally referring to the organisation and control of communication technologies available in any given time and space (Finnemann, 2005). On the other hand, the specific components of these sys - tems (and thereby the methodological and empirical conditions for studying them) are inseparably tied to the concrete communication infrastructures of a given societal and historical context.

The concept of a communication system thereby emphasises how various communication media enable and constrain human activities and societal organisation (Hutchby, 2001), while also creating distinctive material preconditions for developing business models, regulatory frameworks, and other institutional arrangements. We cannot transfer empirical indicators and measures directly from one system to another; instead, we need to consider what the concrete system is made up of and how it can be studied empirically. Going further, we must consider how different gatekeepers of communication rise and fall – become essential and peripheral in the communication system – in close connection to the evolution of different material infrastructures for communicating (Innis, 2007).

Digital communication systems thus refer specifically to contexts where the Internet has evolved into a common and dominating infrastructure for transporting communication content across time and space. Emphasising the material features of Internet distribution, the concept provides a theoretical and analytical lens for understanding the altered conditions for controlling,

CHAPTER 1 | FOLLOW THE DATA 31 monetising, and regulating communication. It entails an openness towards the types of institutions that have emerged – or might emerge in the future – by insisting on investigating the underlying material resources before identifying the actors that control them. That is, if we are to make sense of and be able to critically discuss the business models and disruptive effects of global Big Tech corporations or the continuous influence of welfare state institutions and logics, we must first map out the grounds on which these power battles are being fought. To do so, we need to develop a strategy for identifying how hu - man agency is enabled and constrained through the organisation and control of basic societal resources. The capability approach, introduced briefly in the introduction to this book, offers a strategy for doing exactly that. A capability approach Pioneered by Amartya Sen (1995) and Martha Nussbaum (2000), the capability approach pushes a research agenda that is fundamentally concerned with what individuals are capable of and why – including how institutional and material arrangements in different societal settings enable and constrain these capabili - ties. While the units of normative judgement are individuals and, as properties of individuals, their capabilities, it is a central principle of the approach that individuals are not independent from the structural contexts of their everyday lives. Capabilities are thus “not just abilities residing inside a person but also the freedoms or opportunities created by a combination of personal abilities and the political, social, and economic environment” (Nussbaum, 2011: 20). Closely related to the literature on digital divides (Mansell, 2002, 2017b; Mihelj et al., 2019; Norris, 2001), the capability approach emphasises that peoples’ capabilities cannot be assessed by simply measuring their access to news media or high-speed Internet connections, their skills and knowledge, or their ability to purchase devices and services. Instead, it considers how these activities are enabled and constrained by a wide range of factors (Robeyns, 2005) – be they personal (e.g., relating to physical constraints, literacy, and gender), social (including norms, policies, discriminating practices, and power relations), or environmental (as in geographical, infrastructural, and other physical arrangements) (Robeyns, 2016).

The smartphone provides an example with which we can explain the rela - tionship between individuals and these factors. The interest that individuals have in smartphones is generally not in their technical nature or material makeup of processors and batteries, but rather the capabilities they offer, for example, connecting with others across time and space through text messages. The extent that smartphones can contribute to this functioning, however, depends on personal factors (e.g., literacy and ability to use the device), social factors (e.g., existing cultural traditions and policies that foster different digital literacies), and environmental factors (the availability of a physical infrastructure, e.g., a telephone network or a mobile broadband connection) (Lai, 2021: 25–26).

32 SIGNE SOPHUS LAI & SOFIE FLENSBURG The capability approach – and our application of it – is thus normatively (or ethically) individualistic, but ontologically non-individualistic (Robeyns, 2003: 65). For the purpose of analysing digital communication systems, this approach first and foremost serves as a practical tool for readjusting the research scope by not giving any institutions, sectors, or legislative schemes priority status as objects of analysis. Instead, we ask how individuals’ digital agency is enabled and constrained by the material and social arrangements that prevail in a given setting, and we scope our analysis according to these empirical realities. That is, we start our investigation from a (hypothetical) individual residing in a particular context and ask how her abilities to communicate are framed by the material environment and institutional arrangements that prevail. In doing so, we move beyond assumptions about how specific types of technologies, institutions, and regulations influence human capabilities and, instead, open up for assessing infrastructures, market actors, and modes of governance that de facto influence how communication environments are shaped. Refraining from analysing personal (micro) structures, our adaptation of the capability approach focuses on the social and environmental (macro) structures that frame human communication within different contexts. While such analyses cannot provide comprehensive and sufficiently nuanced accounts of how individuals’ capabilities are enabled and constrained in practice, they do provide a crucial backdrop for understanding and comparing individual media repertories (Hasebrink & Domeyer, 2012) and user patterns (Miller et al., 2016). By not empirically starting from individuals as a way of approaching societal structures, but rather moving in the reverse direction, we contribute to ongoing efforts to understand how varying structural conditions frame life in digitalised and datafied societies.

In contrast to previous theoretical and philosophical applications of the capability approach in media and communication studies (Couldry, 2019; Hesmondhalgh, 2017; Jensen, 2021), our use of the approach is closer to the empirical interventions conducted in the field of global development studies (Kleine, 2013). The mappings and comparisons of the large Nordic countries provide empirical grounds for discussing how the Internet influences basic capabilities by enquiring into the structural implications of (individual) digital communication. As a means for opening up for a broader discussion of the legitimacy and desirability of the current state of digital communication systems, we identify the underlying resources that support mundane uses of the Internet and map out the ways they are controlled. Through this approach, we can cut across historical and bureaucratic classifications (such as the definition of “media”) that constrain our ability to understand and monitor emergent power structures and control mechanisms. The approach also motivates us to ask ourselves the basic “so what” question of why our findings are important, and more specifically how the current state of digital communication systems impacts peoples’ everyday

CHAPTER 1 | FOLLOW THE DATA 33 lives. That is, while applying a macro perspective in our studies of digital communication structures and focusing on infrastructures, markets, and states as analytical dimensions, we return to the individual as a guiding motive for our investigations. Individuals are, in effect, both a means and an end for studying digital communication systems.

Structuring forces of digital communication To unfold how the theoretical conceptualisation of a (digital) communication system frames our analytical approach to studying the structural forces that enable and constrain human capabilities, Figure 1.1 illustrates the analogue communication system prevailing prior to the Internet (Flensburg, 2021; see also Bar & Sandvig, 2008). Creating a historical backdrop for our analyses, the figure emphasises four infrastructures, post, print, telephony, and broadcast (in the innermost squares), surrounded by distinct markets (the middle squares) and different forms of state regulation (outermost squares). The horizontal and vertical axes represent different modes of communication, namely asynchronous and synchronous distribution and one-to-one and one-to-many communication (see also Jensen & Helles, 2017). As illustrated in Figure 1.1, the analogue communication system was characterised by each communication technology being “distinguished […] by the number of message receivers (one, a few or many); by the nature of interactivity, if any, between sender and receiver; and by the difference between synchronicity and asynchronicity” (Braman, 2004: 158).

ONE-TO-ONE ONE-TO-MANY ASYNCHRONOUS SYNCHRONOUS Postal regulation Press regulation The press Postal service Post Print Telcos Telephony Broadcast Broadcasting corporations Broadcast regulation Telecommunications regulation Infrastuctures Markets States Source: adapted from Flensburg, 2021 (inspired by Bar & Sandvig, 2008) FIGURE 1.1 The analogue communication system

34 SIGNE SOPHUS LAI & SOFIE FLENSBURG The figure illustrates how the technological distinctions of each type of communication infrastructure supported the development of equally distinct media and communication sectors that were characterised by a clear division of labour between key gatekeepers: The postal system handled asynchronous one-to-one communication, the press mediated asynchronous one-to-many communication, telecommunication operators carried synchronous one-to-one communication, and broadcasters distributed synchronous one-to-many communication (Bar & Sandvig, 2008). By owning, developing, and supplying vital infrastructural resources for each mode of communication (e.g., postal routes, printing presses, copper wire networks, and electromagnetic frequencies), these gatekeepers gained key societal positions that were legitimised and mandated by various legislative frameworks. News media and broadcasting policies were broadly concerned with establishing the conditions for institutions circulating and mediating public information and debate, while postal policies and telecommunication regulation focused on laying out the rules for providers of one-to-one communication between individual senders and receivers. The regulations and policies depicted in the outer squares of Figure 1.1 were, in other words, developed for and directed towards the gatekeeping institutions, reflecting the professional codes and norms of each sector as well as the physical features of the underlying technology.

Figure 1.1 thereby illustrates how the institutional order of the pre-digital communication system rested on a material basis of analogue communication technologies that have been disrupted by the rise of the Internet as a common and converging infrastructure. Alongside the gradual shift from analogue to digital distribution, old gatekeepers (e.g., newspaper publishers, broadcasters, and postal service operators) are losing their former privileged positions as suppliers of different forms of mediated communication and are being pushed to the periphery of the communication system. In parallel, new actors are gaining significant positions by ensuring network peering (i.e., allowing operators of access and backbone networks to exchange content) or supplying operating systems and other forms of fundamental software infrastructure. The activities, market assets, and infrastructural power of these digital gatekeepers cut across and go beyond the sectoral divides of the analogue communication system, thereby challenging established institutional arrangements and regulations. While the material core of the analogue communication system is irrevoca - bly dissolved, its institutional logics continue to live on in regulatory frame - works and policies. Regulators, policymakers, and even industry stakeholders and researchers tend to start from the outer layers of the model, discussing how regulatory frameworks and legacy business models originating in the analogue system can be amended to fit the new digital reality. By not fully acknowledging the extent and implications of the material transformation, they fall short in addressing the fundamental challenges that digitalisation raises, and they allow the new gatekeepers to fly “under the radar” of public

CHAPTER 1 | FOLLOW THE DATA 35 monitoring and regulation (Flensburg, 2021). In effect, our change of epis - temic strategy is illustrated by shifting the focus to the centre of the figure and outlining the physical resources that are taking the place of former (analogue) infrastructures, their ownership, and the ways they are (not) regulated. In the course of developing and substantiating the concept of a digital communication system, we first establish and elaborate on infrastructures as the material foundation for communication, then we consider how power is accumulated through the ownership and control of these critical resources, and finally we discuss the role of governments and state authorities in or - ganising, distributing, and regulating access to and control of communica - tion infrastructures. In doing so, we approach the evolution, organisation, and control of digital communication systems from three theoretical entry points – emphasising, respectively, infrastructures, markets, and states as key structuring forces of communication – and discuss how they relate to one another. This ultimately aids us in operationalising them as overarching analytical dimensions that can ground the selection of empirical indicators and measures.

Infrastructures When using the concept of infrastructures, we refer to the material resources and systems such as sewers, electrical grids, traffic systems, and various types of distribution networks that together support a wide range of mundane activities and basic societal functions (Edwards, 2002). Communication infrastructures thus broadly comprise technologies that enable us to send, receive, and store information across time and space, as illustrated in the core of Figure 1.1. Similar to traffic networks and other large-scale supply systems, these infrastructures span local routes, uniting individual households, regional exchange hubs, national freeways, and global super-highways. Like cartographers drawing up vast traffic networks, the goal of infrastructure researchers is not so much to understand the cargo being transported (in the case of the Internet, the bits of data containing various forms of communi - cation content), but rather to make sense of the conditions for travelling. Applying an infrastructural perspective thus entails “turning away from the symbolic and investigating the structural – this is the Internet not as ‘what people say with it’ but as ‘how it works’” (Sandvig, 2013: 90). Directing attention to the physical manifestations of (digital) communica - tion, a growing cohort of media and communication researchers have in recent years taken a “turn to infrastructure” (Hesmondhalgh, 2021; Musiani et al., 2016; Plantin & Punathambekar, 2019). Examples include studies of vast data centres, where the abstract idea of “the cloud” (Peters, 2016) materialises in endless arrays of server racks (Velkova, 2016), minute software packages of mobile apps (Gerlitz et al., 2019), extensive networks of fibre-optic subma - rine cables (Starosielski, 2015), or cemeteries of technological waste (Parks,

36 SIGNE SOPHUS LAI & SOFIE FLENSBURG 2015; Thylstrup, 2019). Across the broad and diverse field of infrastructure studies, scholars adhering to this concept share a determination to investigate “the basic, the boring, the mundane, and all the mischievous work done behind the scenes” (Peters, 2016: 33). In fact, many researchers define the very concept of infrastructure as technologies residing “in a naturalized background, as ordinary and unremarkable to us as trees, daylight, and dirt” (Edwards, 2002: 185), only becoming “visible upon breakdown” (Star & Ruhleder, 1996: 113).

Coming out of science and technology studies, a large fraction of the field investigates the situatedness and constant maintenance of technological systems, arguing that “analytically, infrastructures appear only as a relational property” (Star & Ruhleder, 1996: 113). In this view, infrastructures are not “things” in and of themselves, but only become “real infrastructure in relation to organized practices” (Star, 1999: 380). Following this line of argument, studies ask not what, but when an infrastructure is, and they seek their answer in ethnographic case studies of particular infrastructural settings (see, e.g., Harris, 2015; Parks, 2015; Star, 1999; Velkova, 2020). This theoretical approach also entails the consideration of infrastructures as more than physical systems, namely as “the whole array of organizational forms, practices, and institutions that accompany, make possible, and inflect the development of new technology, their related practices, and their distributions” (Bowker et al., 2009: 103).