Micro-hegemonic path dependence: How Big Tech power operates through ideological fragmentation and self-reinforcing technologies

Abstract

Big Tech hearings are high-stakes events that draw in various audiences, and have the potential to reveal and influence the ongoing process of power (re)distribution between various actors, including governments and major tech corporations. In this process, public uptake matters. To gain more insight into the uptake of the political discourses that collide during Big Tech hearings, this article analyzes live chats that were produced in association with 35 YouTube live streams of the 2020 Big Tech antitrust hearing. Using a critical approach towards discourse analysis, this article establishes that the different audiences that watched the hearing on YouTube understood the hearing and its discourse in strikingly different ways. As such, this article finds that different YouTube audiences constitute different ‘micro-populations’ (Maly and Varis, 2016) that adhere to different micro-hegemonic beliefs, which draw on different historical events. To elaborate on this finding, this article introduces the concept of ‘micro-hegemonic path dependence’. This concept draws attention to the notion that different niched ideological groups can develop their ‘paths’ divergingly, causing them to arrive at the same event from different contexts, that lead them to understand the same event in unparalleled ways. Though the audiences’ focus on different historic events might appear to negatively affect the potential uptake of Big Tech discourse, their ideological fragmentation in fact allows for the continued self-reinforcement of Big Tech power.

Introduction

Considering the increasing power and pervasiveness of major tech corporations like Google and Amazon, it is hardly surprising that ‘clashes between representatives of the people and members of the global tech elite’ (Beekmans et al., 2023) occur regularly and are highly mediatized. As a consequence, public hearings during which executives from the global tech industry are questioned by politicians draw in large audiences. There is, for instance, ‘no doubt that from the public’s perspective U.S. v. Microsoft was the antitrust case of the 1990s’ (Rubinfeld, 2001: 531). The 2018 Facebook/Cambridge Analytica hearing, on its turn, still features regularly in memes about Meta CEO Mark Zuckerberg’s ‘alien’ behavior (Canales, 2022), as well as in discussions about politicians’ observed lack of technological knowledge (Byers, 2018). The latter hearing also played a prominent role in the Netflix documentary The Great Hack (Noujaim and Amer, 2019), which argues that the hearing triggered ‘35,000 media stories per day’ and that ‘everyone was watching’ while Zuckerberg testified. Though this estimate might be somewhat overstated, it must be clear that these so-called ‘Big Tech hearings’ give public visibility to the collision between Big Tech discourse and the discourse of government, as both tech corporations and governments work towards a new, post-digital equilibrium (Celeste, 2022) in which some societal domains might become subject to the control of Big Tech, while other aspects of everyday life remain largely beyond their reach.

Public Big Tech hearings can thus be understood as ‘tips of the iceberg’ that render the ongoing process of post-digital power (re)distributions partially visible. Yet, these hearings do more than merely foreground the rivalries between major tech corporations and government actors. Due to their aforementioned mediatization, they also have the potential to influence people’s beliefs about the power major tech corporations exercise through their technologies, as well as people’s trust in politicians’ capabilities to take proper action on their behalf in relation to the power of Big Tech. This is significant, as the public at least partially assigns power to both, by casting their votes and by participating in particular societal behaviors, and by deciding (how) to use particular digital platforms, services and affordances, while simultaneously ignoring and counteracting other technologies. From this perspective, it is reasonable to suggest that if the audiences of these ongoing public clashes eventually find the performances of the representatives of major tech corporations more compelling than the performances of elected representatives, it might not be far-fetched to imagine that, at some moment, the idea that major tech corporations are better equipped to deal with certain societal issues – from Google’s ‘water stewardship strategy’ to, for instance, Meta’s efforts to support ‘pop-up vaccine clinics in low income and underserved communities’ (Meta, 2021) – can become ‘common sense’. Of course, the opposite is also true. If public trust in Big Tech weakens while national governments are empowered and invigorated, chances are slim that Google is allowed to become more actively involved in organizing water management, education and safeguarding cultural heritage, just like Meta will probably not be called upon to play a more prominent role in the health sector and other aspects of daily life.

Considering the public visibility of Big Tech hearings and the public’s role in the process of power (re)distributions between major tech corporations and governments, this study asks the following question: How are different audiences interpreting the discourse that is produced during public collisions between representatives of (national) governments and representatives of Big Tech? In order to gain insight into the understanding and uptake of Big Tech discourse by different groups of people in contexts of potential power (re)distributions, this article examines online audience interactions related to the discourse that was produced during the 2020 Big Tech antitrust hearing. This hearing was regarded as particularly significant by various media. The New York Times, for instance, noted that ‘The Captains of the New Gilded Age […] will appear together before Congress for the first time’ (Kang et al., 2020), while other media described the hearing as ‘historic’, and argued the event might significantly impact the power of Big Tech. What also makes this hearing particularly interesting from an analytical perspective, is the fact that it was live streamed by at least 35 different YouTube channels, which regularly enabled YouTube’s live chat function. This function, on its turn, prompted users to ‘say something’. Consequently, at least 157,875 live chat comments were posted next to the hearing’s multiple live streams while the hearing was unfolding. Starting from the persuasion that these live chat comments provide a unique opportunity to reveal how different audiences interpreted the hearing’s discourse while the hearing was unfolding, this article uses Blommaert’s and Verschueren’s critical approaches to discourse analysis (Blommaert, 2005; Verschueren, 2013) to establish how the people that participated in the YouTube live chats responded to the discourse of the hearing in relation to potential power (re)distributions. Following the observation that there are profound interpretative differences between the live streams’ multiple audiences, the concept of ‘micro-hegemonic path dependence’ is introduced as part of the analysis to provide a theory that indicates that present-day media don’t produce a ‘fragmented audience with a common culture’ (Cardiff and Scannell, 1987: 168), but rather fragmented audiences with fragmented cultures that are simultaneously connected and divided through Big Tech infrastructures. ‘Micro-hegemonic path dependence’ entails that different niched ideological groups can arrive at the same event via a wide variety of directing experiences and events, causing them to understand and participate in the same event in unparalleled ways. Using this concept, this paper aims to explain how different groups arrived at the same Big Tech hearing via vastly different historical paths, which cause them to understand Big Tech discourse in unparalleled ways, thus leading to ongoing fragmentation instead of sociocultural convergence. After demonstrating how the concept of micro-hegemonic path dependence can be applied, this article concludes with a discussion that delineates how the consequent communal ideological fragmentation of the various audiences of Big Tech might impact major tech corporations’ future power over society.

Background

A lot has been written about the ‘Internet giants’ that are commonly identified as ‘GAFAM’, which stands for Google, Amazon, Facebook – now known as ‘Meta’ – Apple and Microsoft. For instance, in The Platform Society, Van Dijck et al. (2018) engage with the public values of platforms and their ordering power from the perspective of ‘governance by platforms’. In If...then: Algorithmic Power and Politics, Bucher (2018) describes how the algorithms of major platforms can be seen as socio-technical assemblages that have sociopolitical power, and Burgess et al. (2022) argue in Everyday Data Cultures that these algorithms can be perceived as ‘culture’ and their existence as impacting an impressive range of everyday practices. Other publications have focused on the ideological beliefs that imbue technology companies and their platforms (Barbrook and Cameron, 1996; Srnicek, 2017), on their impact on constitutional ecosystems (Celeste, 2022), on their kinship with historical colonizers (Mejias and Couldry, 2024), and on specific companies and their societal impact (Vaidhyanathan, 2012). Though the focus areas of these publications vary, one of their main points corresponds: major tech corporations are affecting society in ways that merit further attention, both from academia, news media, governments, and their publics.

Part of this attention is channeled into an ever growing record of government investigations and associated public hearings, which are presented as political events that inform potential government actions while in practice offering a mere ‘symbolic measure of accountability’ (Creech and Maddox, 2022). Since a ‘hearing is a spectacle of personal responsibility’, hearings tend to hide the lobbying and ‘structural and cultural conditions’ that play along in the background while focusing attention on the discourse that is produced by the individual witnesses – in this case Big Tech CEOs (Creech and Maddox, 2022). Public Big Tech hearings are not single, historical events, but rather an ever-evolving series of sequels; all disruptively novel, but simultaneously strikingly repetitive and interrelated. The various hearings that took place before the 2020 Big Tech antitrust hearing are, in that sense, components of its historical backdrop. Earlier hearings at least partially produce the ‘path’ (De Munck, 2022) that directs which practices and discourses are deemed ‘normal’ in ensuing Big Tech hearings, as well as the ‘lines of vision’ (Verschueren, 2013: 82) that steer the interpretations and uptake of their audiences. In this sense, historical events serve as referential ‘blueprints’ for their successors (Anderson, 2016). Or, as De Munck argues: ‘certain events set in motion a temporal sequence in which the original event continues to matter’ (2022: 566). This rather loose description of the notion of ‘path dependence’ can be elaborated upon by a more narrow approach labeled ‘increasing returns’. This approach goes beyond the rather shallow notion that ‘history matters’ by incorporating the idea that ‘the probability of further steps along the same path increases with each move down that path. This is because the relative benefits of the current activity compared with other possible options increase over time’ (Pierson, 2000: 252, emphasis in original). The classic example that is used to explain this approach – and nowadays also to critique this approach (Vergne, 2013) – is the design of the QWERTY keyboard. Though the sequence of the letters on the keyboard might not be the most efficient, the cost of switching to an alternative is understood as increasingly high, consequently resulting in a type of ‘lock-in’. Similar ‘lock-ins’ occur not just with technologies, but with various societal practices and phenomena. Building on these arguments, an analysis of the uptake of the 2020 Big Tech antitrust hearing can thus not ignore the idea that previous hearings will at least partially prescribe the potential practices and interpretations of future hearings, and that it might be increasingly difficult to depart from the path that is prompted by their historical understandings and constructions. This applies both to the practical, material dimension of the hearings, and to the practices and ideas of all actors involved, including their interpretations and ideological beliefs: ‘Once established, basic outlooks on politics, ranging from ideologies to understandings of particular aspects of governments or orientations toward political groups or parties, are generally tenacious. They are path dependent’ (Pierson, 2000: 260). Considering the sequential nature of Big Tech hearings and related events, this article incorporates Pierson’s application of ‘increasing returns’ to political processes to examine how the discourses and practices connected to the 2020 Big Tech hearing are dictated by previous events and conditions, and might consequently dictate future events and conditions. According to Pierson, ‘the employment of power may be self-reinforcing’ (Pierson, 2000: 259), and Big Tech hearings might thus be part of a political development that steers the power (re)distribution between major tech corporations and other actors into a particular direction.

The hearing and its live streams

The sixth hearing of the investigation into ‘Competition in Digital Markets’ was organized by a subcommittee of the United States’ House Judiciary Committee. After being rescheduled, the hearing took place on July 29, 2020. Its formal objective was to ‘review the effects of market power online’ (Rpt., p. 1, 2022) to inform potential changes in antitrust legislation, and thus to actively shape the path of Big Tech’s power ‘online’. The committee live streamed the entire hearing on YouTube, titling it ‘Online Platforms and Market Power: Examining the Dominance of Amazon, Apple, Facebook and Google’. The stream’s high viewership numbers can at least partially be explained by the aforementioned historical interest in Big Tech hearings, the fact that the committee had been using YouTube as a platform for its public hearings for years, and by the previews that were published by various news media in the weeks before the event. For example, the scheduled attendance of the CEOs of Amazon, Google, Apple and Facebook enticed media to argue that the hearing ‘was drawing attention just for the fact that the four chief executives [...] are appearing together before lawmakers for the first time’ (Lima, 2020). News media also explained the significance of this particular hearing by mentioning that ‘the hearing could launch a new era in antitrust regulation in the tech sector’ (Kelly, 2020), and noted that political opinions regarding the topic of the hearing diverged. In this sense, the event comprised a ‘high-stakes encounter [...], involving a broadcast media logic-driven sense of anticipation’ (Vaccari et al., 2015), as journalists highlighted the uniqueness, key importance and tenseness of the situation.

The hearing’s widely publicized status, combined with the global significance of the four tech companies and the influence of the mechanisms of the ‘attention economy’ (Ørmen and Gregersen, 2022), might also explain why the original YouTube live stream was copied by at least 34 other YouTube channels and ‘repackaged’ into new YouTube live streams. This resulted in a situation in which the same video footage was shown on dozens of ideologically and geographically diverging YouTube channels, which added their own titles, descriptions and placeholder images to the stream. These different YouTube channels regularly enabled YouTube comments and live chats, and sometimes even enhanced the original live stream with additional commentary. In this sense, the original live stream of the hearing was transformed into a variety of parallel ‘multiple existences’ and gained ‘mobility’ (Blommaert, 2010) by contemporaneously traveling to various ‘interactionally produced’ (Benson, 2015) YouTube pages and their audiences. The hearing was, for instance, shown by the British center-left newspaper The Guardian, the international news service Reuters, the American conservative television channel Fox News, the Spanish newspaper La Vanguardia, the technology blog Engadget, and the financial news website Yahoo Finance. The geographical and substantive variety of the live stream’s copies might already begin to reveal that the hearing was deemed relevant from multiple perspectives. It also demonstrates how the hearing’s discourse might be directed towards – and subsequently interpreted from – various centers of normativity (Blommaert, 2010). Following Blommaert’s line of thinking, this occurrence of mobility ‘affects the phenomenology of language, and how we need to think about it in terms of scales, orders of indexicality and polycentricity’ (Blommaert, 2010: 180); meaning that, when the ‘elite discourse’ that is produced by the tech CEOs moves from one social context to another, it becomes discernible how the same utilization of resources invokes different orders of indexicality in relation to the distinct repackagings and their different contexts, which are connected to different centers of normativity, and subsequently summon different interpretations and meanings.

We can already begin to observe the process of stratified reorientation and revaluation in the different frames that were imposed on the hearing’s live streams by the various YouTube channels that copied the original broadcast. The title of the original YouTube live stream – ‘Online Platforms and Market Power: Examining the Dominance of Amazon, Apple, Facebook and Google’ – presents ‘the dominance’ of the tech companies as an agreed upon ‘common sense’ notion (Verschueren, 2013) that requires ‘examining’. Bloomberg’s social media brand, Bloomberg Quicktake: Now also appears to incorporate a condemnatory attitude into its title: ‘Congress Grills Big Tech CEOs’. This tone-of-voice could be linked to the stance of the committee, but may also echo more general interpretations of public hearings as ‘controversial and contentious’ events that focus on ‘grilling’ individuals (Gallagher, 2013). The Washington Post, on the other hand, begins the title of its live stream with ‘Tech CEOs testify at House Hearing’, thereby presenting the relationship between the CEOs and the members of the committee as less hostile and more cooperative. The title at least partially abandons the committee’s original condemnatory stance, as it appears to incorporate journalistic values that are based on the beliefs that news media should be ‘fair, balanced and accurate’ and ‘be clear and not sensationalize’ (Lacy and Rosenstiel, 2015: 21). In this sense, the Washington Post proposes a slightly alternative way to understand the event. Thus, even before the hearing was well and truly underway, the titles and descriptions of the multitude of live streams began to demonstrate how different media seized the opportunity to ‘repackage other players’ frames and/or create their own issues spins’ (Callaghan and Schnell, 2001: 184). Though the differences between the multiple live streams might appear insignificantly small at this moment, they can be understood as omens for the more comprehensive differences between the uptake of their varying audiences, and the different directions in which the (re)distribution of power between Big Tech and other actors might be understood to develop.

Data

Though the YouTube live chat function is enabled by default, YouTube channels are given the option to turn off the chat before, during and after the live broadcast. At the time when the data collection for this research commenced in December 2021, 14 substantially relevant live chats remained. For the purpose of this research, all comments in these live chats were scraped using the Save Live Streaming Chats for YouTube Google Chrome browser extension (tabgraf.com, n.d.). Together, these live chats contain 157,875 comments that were produced by at least 26,545 different YouTube users. 999 YouTube user names appeared in multiple live chats. Some of these names seemed impersonal or common, and could therefore have been used by multiple users. Many other names, however, appeared distinctive and could potentially even coincide with a person’s legal name, which indicates that some users participated in multiple live chats. The latter can be confirmed by comments that point towards experiences with multiple live chats – ‘this chat is a lot better coming from the Fox news stream lol’, for instance. To protect the identity of the YouTube users that participated in the live chats, all user names were deleted from the dataset after a statistical assessment of the data was conducted, and will consequently not be used in this article, just like references to other potentially identifying elements.

To cope with the size of the dataset, the topic modeling tool ‘The MAchine Learning for LanguagE Toolkit’ – ‘MALLET’ – (McCallum, 2002) and the ‘Count Duplicates in a List’ tool (Humbad, 2011) were used for an initial exploration of the data. This led to the observation that most of the 14 live chats displayed both substantive differences and substantive similarities. In all live chats, people, for instance, demonstrated interest in the starting time of the hearing, whereas only people who participated in the Yahoo Finance live chat routinely discussed stock prices. Following this initial exploration, the earlier mentioned differences in the substantive framings of the YouTube channels were combined with the findings from this exploration to select four live chats – the Fox News, Reuters, Yahoo Finance and Engadget live chats – for further analysis. Since this selection still contained 103,032 comments, the selection was further narrowed down by taking the proceedings and content of the hearing into account. Arrays of comments that were produced during the seven opening statements – three of which were performed by different committee members, and four of which were performed by the CEOs – were chosen, together with comments that were produced during seven different hearing interactions, and during two periods when nobody was speaking. This yielded 15,016 comments that were identified as having the capacity to demonstrate how various parts of the hearing sparked responses and uptake in the different live chats. Because this number of comments was still too large for an in-depth critical analysis, the 15,016 selected live chat comments were initially ‘read-along’ – meaning that fragments of the live streams and live chats were replayed – and classified in relation to the question whether they had a clearly identifiable relationship with the live streamed event or not – which helped to distinguish spam comments – and thereafter assigned between one and three codes and a thematic category that reflected the comment’s substantive focus. Comments and interactions that were seen as exemplary for their allocated thematic categories, or deemed of particular relevance with regard to this article’s research objective – to establish how different audiences understand the discourse that is produced during Big Tech hearings in the context of power (re)distributions – were given an additional description and compiled in a separate file to enable further examination using a critical approach towards discourse analysis (Blommaert, 2005; Verschueren, 2013).

Analysis

According to Blommaert, ‘contextualization of discourse’ should be a central element in any critical analysis of discourse, in which an eclectic variety of approaches and insights can be harnessed to lay bare ‘power effects’ (Blommaert, 2005: p. 233). This analysis therefore combines Blommaert’s approach towards indexicality, context and ideology with Verschueren’s more detailed techniques to reveal ideological beliefs in fragments of language (2013). These approaches are supplemented with Pierson’s perspective on the analysis of path dependence in political processes (2000) and Maly’s notion of ‘algorithmic power’ (2022) to focus attention on the relation between power and the temporal and ‘platform’ dimensions of the hearing. Since it is a complex task to examine power (re)distributions over time – indicators for historical links are often limited, and the involved actors, their time horizons and motivations are often variable (Pierson, 2000) – this article takes two words that were observed to occupy a prominent position during the initial exploration of the data as entry points into the divergent interpretations and uptake of the discourse of the hearing: (1) ‘time’ and (2) ‘censorship’. These words are seen as examples that foreground the different paths and ideological beliefs that affect the practices and interpretations of the 2020 Big Tech hearing, rather than exceptional cases that are unique for this particular hearing and its live chats. Inspired by this notion, the last part of this article will engage with the relevance of its analysis beyond the present focus on the 2020 Big Tech hearing.

‘Reclaiming my time’ as a pro-Trump, micro-hegemonic practice

It might seem counterintuitive to think that a seemingly straightforward notion like ‘time’ can conjure different meanings for different groups of people. The word ‘time’ has a prominent presence in every live chat and can easily be connected to general comments that were posted in the live chats about the starting time of the hearing – ‘What time will it start’ – references to wasting time – ‘lame waste of time’ – and peevish responses to the hearing’s one-hour delay – ‘none of this crap ever starts on time’. In the Fox News live chat, however, the most prominent use of the word appears as part of a phrase that might be interpreted as ‘hearing jargon’: ‘reclaiming my time’. This sentence was posted frequently by members of the Fox News audience throughout the hearing, but was almost absent from most of the other live chats and from the hearing itself, and was not used in the titles or descriptions of the four YouTube channels. The phrase can be used by committee members to ‘reclaim their time’ when a person is ‘lying, misleading, or drawing out the time by not answering the questions directly’ (Urban Dictionary, 2020). In recent history, the phrase was most famously used by the Democratic representative Maxine Waters during a 2017 hearing, in which Secretary of the Treasury Steve Mnuchin appeared to be stalling to ensure Waters’ time would run out before he could answer her question (Romano, 2017). Waters’ utilization of the ‘procedural interruption’ became a viral meme and a symbol for the empowerment of black women and other groups whose time had been historically ‘wasted’ by dominant societal actors (Emba, 2021; Wingfield, 2019). In this sense, the phrase was transformed from a legislative discursive tool into a ‘catchphrase for commentary on the Trump administration, racism, misogyny, and other progressive issues’ (Romano, 2017).

Further analysis of the Fox News live chat, however, demonstrates that its participants are not indexing Waters’ anti-Trump use of the words. In fact, they hardly refer to Waters at all, and when they do, their comments index pro-Trump, anti-feminist sentiments without displaying any direct awareness of the 2017 hearing: ‘THE GOLDEN GHOULS starring Hillary Clintron (sic), Maxine Waters, Elizabeth Warren and Nancy Pelosi. Tonight after Oprah’. There is, however, a different hearing that does play a significant role in the Fox News live chat. One day before the Big Tech antitrust hearing, on July 28, 2020, Attorney General William Barr testified before the House Judiciary Committee during a public hearing titled ‘Oversight of the Department of Justice’. Though the hearing’s title sounds more mundane and less negatively preconceived than that of the Big Tech hearing, news media described the hearing as a ‘full-scale assault against Attorney General William Barr’ (Herb, 2020) and the committee’s Democrats as its main ‘aggressors’ (Fandos and Savage, 2020). Just like the 2020 Big Tech antitrust hearing, the hearing featuring Barr was livestreamed on a variety of YouTube channels, including that of Fox News. The content of the Fox News live chat that was produced in conjunction with the live stream of the Big Tech hearing is riddled with comments that refer to the earlier Barr hearing. Some of these comments make explicit connections between the treatment of Barr and ‘reclaiming my time’: ‘I want every Republican to ask a question and then “reclaim your time” before they can respond...censor them like the Democrats censored Barr’, for instance. Through comments like these, it becomes clear that the reason for the high prevalence of ‘reclaiming my time’ in the Fox News live chat can be traced to the audience’s view on the performances of Democrats during the Barr hearing. The repetitive posting of ‘reclaiming my time’ has been recontextualized and has acquired an ideological meaning that deviates from the phrase’s popular online use. Instead of signaling anti-Trumpism, in the context of the Fox News live chat, the phrase signals disapproval of the Democratic representatives, and ‘being in the know’ when it comes to the Democrats’ poor behavior. This departure from the ‘normal’ anti-Trump use indicates that the Fox News live chat participants comprise a group of people whose beliefs and experiences differ from the beliefs of other groups. Furthermore, the Fox News live chat version of ‘reclaiming my time’ also demonstrates that its participants might be relatively stable, as a significant proportion of the people who are commenting on the Big Tech hearing is apparently aware of specific details of the Barr hearing, and adhere to Fox News’ perspective on the event. In this sense, the people that participate in the Fox News live chat can be understood as a ‘micro-population’ (Maly and Varis, 2016), which means that they are a ‘niched ideological group’ (Maly, 2023) that adheres to a set of ‘common sense’ ideas and beliefs that are not recognized as ‘common sense’ by other societal groups. The use of ‘reclaiming my time’ should thus be understood as a ‘micro-hegemonic’ ‘identity discourse’ or ‘emblematic feature’ (Blommaert and Varis, 2011) that people can operate to point towards their membership of an identity category – in this case their orientation towards a group of people that regularly watches hearings via the Fox News YouTube channel, supports Trump as opposed to the Democrat Party, and believes Democrats are inclined to treat hearing witnesses unfairly.

Returning to the notion of ‘path dependence’, the observed connection between the Big Tech antitrust hearing and the Barr hearing indicates that the practices and beliefs of the Fox News micro-population are at least partially inspired by previous experiences. The fact that only the participants of the Fox News live chat are using ‘reclaiming my time’ as an anti-Democrat identity discourse, however, also demonstrates that there is more to it than the persuasion that ‘certain events set in motion a temporal sequence in which the original event continues to matter’ (De Munck, 2022: 566). The Barr hearing clearly matters to the people in the Fox News live chat, and only to the people in the Fox News live chat. In this sense, the ideologically driven fragmented meanings of a seemingly straightforward word like ‘time’ already begin to demonstrate that the concept of ‘path dependence’ should be expanded to include a micro-hegemonic dimension. Building on this, the concept of ‘micro-hegemonic path dependence’ can thus be used to highlight that not all practices, interpretations and uptake of an event are steered by a singular ‘original event’ (De Munck, 2022), but that different niched ideological groups follow their ‘paths’ towards the same political event in relation to different historical experiences. While certain aspects and recontextualizations of the 2017 Mnuchin hearing are of key importance for some left-wing audiences, and components of the 2018 Facebook hearing were deemed highly significant by various media, the most recent event of interest on the path of the Fox News audience was the ‘combative Barr hearing’ (Herb, 2020) – which, on its turn, did not make a big impression on the ‘paths’ of the three other audiences whose comments were analyzed. For the Fox News micro-population, the mere presence of Democratic representatives invokes the posting of ‘reclaiming my time’, thus demonstrating that historical micro-hegemonic experiences play a role in the fragmentation and polycentricity of interpretations, and will thus also produce differences in uptake.

Censorship as the actual topic of the hearing

Though this short discussion of ‘reclaiming my time’ demonstrates that the interpretations of a contemporary event can be directed by the ideologically driven experiences and interpretations of a historical event, the relevance of this observation can just as easily be criticized from the same perspective. The ‘increasing returns’ approach towards path dependence entails that the ‘long movement down a particular path will increase the costs of switching to some previously forgone alternative’ (Pierson, 2000: p. 261), which means that ideas and practices that derive from the Barr hearing can by no means be interpreted as components of an ingrained path based on a mere analysis of the 2020 Big Tech antitrust hearing’s live chats. The two hearings are only 1 day apart. The Barr hearing might be momentarily ‘top of mind’ because of a Fox News news story, and subsequently quickly forgotten. The use of ‘reclaiming my time’, however, is far from the only striking distinction between the four live chats. To demonstrate how micro-hegemonic path dependence can lead to more durable, fixed ideological fragmentation – and how this might impact interpretations of the discourse and power of Big Tech – the next part of this analysis engages with the different ways in which the four audiences understand the general meaning and objectives of the hearing, amongst other things by paying close attention to the operation of ‘censorship’.

In the Reuters, Engadget and Yahoo Finance live chats, antitrust issues and monopolistic behaviors are regularly mentioned while the hearing is unfolding. When a question about the topic of the hearing is posted in the Engadget live chat – ‘what’s this about? Monopolies?’ – for instance, another user answers: ‘not sure, maybe anti-monopoly stuff (they’re getting too big)’. Similar questions are raised in the Yahoo Finance live chat – ‘is this an antitrust hearing or something else?’ – and in the Reuters live chat, the topic is even explicitly linked to the perceived irrelevance of Amazon CEO Jeff Bezos’ opening statement: ‘What does this have to do with breaking Antitrust laws?’. Though it is evident that the people in the Reuters, Yahoo Finance and Engadget live chats interpret the topic of the hearing in relation to their own contexts – and thus comment on the live stream of the hearing in different ways – discussions about the perceived actual topic of the hearing in the Fox News live chat are even more divergent. While members of the Fox News micro-population appear to be just as likely to view the opening statements of the tech CEOs as irrelevant, and the performances of most politicians as inadequate, their criticism mostly emanates from the topic of ‘censorship’, rather than from ‘antitrust’: ‘Wah wah wah. Boohooo. What does this crap have to do with censoring?’. On closer inspection, in the Fox News live chat, the notion of censorship imbues all sorts of responses. From people making apparent general comments about the hearing and the investigation – ‘Just stop censoring people, it’s that simple’ – to people making suggestions regarding what should be done with the CEOs – ‘CENSOR tHEM’ – and commenting on the actions of the politicians that are involved: ‘we the people hold the chairman in contempt of our freedom of speech. Vote them all out!!!!!!!!’. Even the discursive tool of ‘reclaiming time’ is understood as a censoring practice. Furthermore, many of the responses that include the notion of censorship mention the perceived ideological dimension to what the Fox News micro-population understands as ‘Big Tech censorship’, for instance by arguing that ‘Big Tech’ favors communists and Marxists – ‘What is the purpose of this when the Marxists are never censored on these platforms?!!!’ – and is deliberately antagonizing ‘conservatives’: ‘we care about your sensory of conservatives’. Some people even ‘perform’ digital censorship through written reenactments – ‘I’m being censored’ and ‘Message Deleted By Google’ – and interpret glitches in the live stream and the apparent lack of findability of the Fox News version of the live stream as a sign of censorship: ‘YouTube shadow hiding this video……. use your phone and type in fox. NOTHING no live video anywhere’. A user even notes that it is impossible to type the words ‘Donald Trump’ in the YouTube live chat: ‘Hahahaha! YouTube is sensoring us in real time! Every time I mention our president’s name, my comments are blocked’. Considering the amount of Trump spam in the Fox News live chat, while also observing various other comments that contain Trump’s full name – ‘I love Donald Trump, god bless president trump’ – there’s plenty of evidence to suggest that the words ‘Donald Trump’ are not being blocked. Furthermore, it is highly unlikely that Google would replace an inadmissible YouTube comment with the words ‘Message Deleted By Google’, just like it is highly unlikely that if Google was replacing inadmissible comments with its own message, this message would only appear once. For the purpose of this article, however, questions regarding the practical occurrence of ‘Big Tech censorship’ are subtopical at best. What really matters, is the observation that, for the majority of the Fox News live chat participants, every dimension of the live streamed hearing can be interpreted from the perspective of ‘censorship’. While a glitch in the live stream can index a bad internet connection in the Engadget live chat – and thus a lack of technological skills on the side of the government – in the Fox News live chat, technical hiccups scream malicious ‘silencing’ intent.

The micro-hegemonic paths towards ‘Big Tech censorship’

‘Censorship’ can broadly be described as ‘social and political injunctions on forms of speech and representation’ (Moore, 2016), and has gained increasing prominence during the last decade. In an article about what he deems to be a ‘free speech panic’, Davies notes that the ‘language of “free speech” and “censorship” is old, but the fervor of this panic is new’ and difficult to reconcile with the emergence of phenomena like social media and citizen journalism (Davies, 2018). According to Davies, ‘censorship’ has become a ‘political totem’ and a ‘cherished value for conservatives’, fueled by the perception that ‘the identity of conservatism is under siege’. While Davies describes how universities are seen as bastions of censorship and online platforms, by contrast, as the enablers of free speech – thus echoing Zuckerberg’s corporate discourse of giving people ‘voice and free expression’ (Meta, 2019) – more recently, Leidig comprehensively expounded the United States’ far right ‘outcry against Big Tech’s conservative censorship’ (Leidig, 2023: 172), which began to develop around the same time as Trump’s turn against digital media (Leidig, 2023). Though it might appear sensible to argue that ‘Big Tech’ has given people more practical options to spread their ideas and organize interaction (Celeste, 2022), at the time of the 2020 Big Tech hearing, people with far right convictions were adhering to the belief that companies like Facebook, Twitter – now known as ‘X’ – and Google were actively attempting to rid society of their views as they worked towards ‘left-wing cultural dominance’ (Leidig, 2023: 186).

Just like the use of ‘reclaiming my time’, the micro-hegemonic utilization of ‘censorship’ did not materialize out of thin air. Actors like Fox News and Donald Trump are at least partially responsible for the beliefs of this particular niched ideological group, and their historical consideration of ‘censorship’ is subsequently vital to the group’s interpretation of Big Tech discourse through the lens of ‘censorship’. Though there are obviously endless events and developments that seep into the construction of this micro-population and its meaning-making processes – and it is, like in any analysis, difficult to determine how far back relevant events and processes should be traced (Pierson, 2000) – it can, for instance, be noted that 2 months before the 2020 antitrust hearing, Trump ‘issued an executive order aimed at limiting the legal protection offered by Section 230’ (Siripurapu, 2020) – a legal section that deals with issues of liability. ‘The move came after Twitter appended fact checks to several of his tweets regarding voting by mail. The president has long feuded with big tech companies, arguing they are trying to “rig the election” against him and are masquerading as neutral while suppressing content they disagree with’ (Siripurapu, 2020). While other media described Twitter’s actions as ‘slapping a fact check label’ (Clark, 2020), or as ‘placing a fact-checking warning’ (Allyn, 2020), Fox News argued that ‘Twitter has once again taken action against President Trump, this time censoring a tweet for “glorifying violence”’. The idea that what Twitter was doing in 2020 was ‘censorship’ was not a universally accepted interpretation of the events, and is still – even after Trump was banned from various platforms in 2021 – debated. Like the Barr hearing did for the deviating occurrence of the word ‘time’, Twitter’s actions against Trump, coupled with Fox News’ anomalous description of the events, at least partially explain the particular use of ‘censorship’ in the Fox News live chat. While the interpretive paths of most audiences might not even run along Twitter’s decision to append fact checks to Trump’s tweets – remember there is not even a representative of Twitter present at the hearing – for the Fox News micro-population, this is arguably the most significant event leading up to the 2020 Big Tech antitrust hearing. This determination also explains why people in the Fox News live chat regularly ask where then Twitter CEO Jack Dorsey is – ‘WHERES THE COWARD JACK DORSEY, THE CENSORSHIP KING’ – while people in the other chats are more interested in the absence of other actors, like former Microsoft CEO Bill Gates. Combined with Trump’s historical anti-media stance and the ‘free speech panic’ (Davies, 2018), Fox News’ framing of the events involving Twitter has such micro-hegemonic strength that even the non-presence of a Twitter representative and the non-mentioning of censorship can be interpreted as signs of censorship by the Fox News micro-population: ‘Democrats know this filibuster about monopoly goes nowhere. Get back to CENSORSHIP!’.

Infrastructures of path dependent ‘censorship’?

Besides the fact that ‘censorship’ is barely discussed during the 2020 Big Tech antitrust hearing, there is one even more evident element that could entice the Fox News micro-population to deviate from their censorship-path: the way in which YouTube is used as a platform to stream, watch and discuss the hearing. In his 2022 research agenda, Maly argues that any analysis of digital discourse and power in post-digital society should address ‘the impact of “non-human” activity in contemporary meaning making processes’ (Maly, 2022: 3). The Fox News micro-population has not gathered in, for instance, a football stadium to watch and discuss the hearing. They are watching the hearing ‘together’ online, on a platform that is owned by one of the tech corporations whose CEO is being questioned as part of the hearing. Multiple YouTube users highlight the apparent contradiction the involvement of YouTube creates: ‘ironic that we are watching it on YouTube’. The occurrence of this ‘ironic’ situation can initially be understood through a more customary application of path dependence, focusing on technological and economic development. As mentioned before, the House Judiciary Committee has been using YouTube to live stream its hearings for years. Departing from this practice, even if this would seem substantially and morally sensible, would probably come at the cost of spending time, money and energy on setting up and learning how to use a different streaming platform, exacerbated by the inherent loss of online audiences, archival material, data, and indexed ‘prominence’ – causing the committee to become less visible online. Similar mechanisms influence the behavior of the YouTube users and the YouTube channels that copied the original live stream. Media like Fox News and Reuters cannot simply abandon YouTube as a platform for the spread of videos, nor as an infrastructure for ‘cheap’ public engagement through the copying of content, as this would negatively impact their ability to compete in the attention economy. Clearly, like power in general (Pierson, 2000), the employment of ‘algorithmic power’ (Maly, 2022) might be self-reinforcing as well, and the paths of different micro-populations are not only diverging because of their varying experiences and identity discourses, but also because the algorithms and infrastructures of Big Tech facilitate and encourage their further fragmentation. In the Fox News live chat, the ideological incentive ‘to identify and stick with’ (Pierson, 2000: 154) the belief in Big Tech censorship is sufficiently strong that even the existence of a findable and technically functioning YouTube live stream, combined with a comment section and live chat that encourages people to ‘say something’ – which includes the opportunity to post pro-Trump spam and to criticize Big Tech – cannot change the course of the Fox News micro-population’s path, and their associated micro-hegemonic beliefs.

Conclusion: all ideologically fragmented roads lead to Silicon Valley

This article originally set out to use a critical approach towards discourse analysis (Blommaert, 2005; Verschueren, 2013) to uncover how different audiences interpret the discourse that is produced during public collisions between representatives of (national) governments and representatives of Big Tech, as the understanding and potential uptake of Big Tech discourse were deemed to be relevant in relation to the further development of power (re)distributions between Big Tech and other actors. Evidently, the people that congregated in the live chats discussed various topics in relation to the 2020 Big Tech antitrust hearing. For instance, seeing Democratic representatives in the live stream triggered people in the Fox News live chat to ‘reclaim their time’, while listening to the opening statements and ensuing interactions enticed the same audience to believe that the, in their eyes, actual topic of the hearing – ‘censorship’ – was being censored. In this sense, the hearing’s discourse is obviously generating responses, and may thus, from a practical point of view, be seen as producing uptake. However, it is also clear that much of the uptake of the 2020 Big Tech antitrust hearing indexes and reproduces discourses of other events and developments, and that the actual words that were uttered during the Big Tech hearing only have a limited influence.

Looking at the profound interpretative differences between the audiences of the hearing, the uptake of the discourse that was produced during this particular hearing appears small, and impossible to understand and appreciate without taking micro-hegemonic path dependence into account. Different micro-populations arrived at the different live streams of the 2020 Big Tech hearing via different paths, which provided them with strikingly different ‘lines of vision’ through which they could understand the event. However, despite all of these differences, it should be noted that all of these micro-populations did arrive on YouTube, which people in various live chats found ‘ironic’. This potentially unifying factor, however, in fact explains how Big Tech’s power is allowed to function through a combination of ideological fragmentation and self-reinforcing algorithmic power. While the increasing costs of abandoning the technologies of Big Tech induce all actors involved to continue their use of these technologies – YouTube, in this case – the algorithmics, affordances and impact of the ‘attention economy’ simultaneously bring about the further development of fragmented audiences with fragmented cultures, who stick to their own distinct ideologically driven paths. As such, it should be clear that, even though the words of the tech CEOs did not evoke a lot of direct reproduction, as a series of sequels, Big Tech hearings and related events and processes are slowly producing a more profound type of uptake that further encapsulates Big Tech’s power over society. As Big Tech combines self-reinforcing algorithmic power with the power of self-reinforcing fragmented ideological beliefs, it becomes increasingly difficult to imagine that an attractive alternative for the current ideological fragmentation within the boundaries of Big Tech will emerge. There’s no need for Big Tech to ‘divide and conquer’, as society is already divided, and major tech corporations can thus quite easily continue on the path taken, on which all roads lead to and through their technologies while remaining algorithmically separated.