The strong emotional context
Before the pandemic, it was already shown that MDI propagates on social media platforms by playing on the emotional reactions of the online audience (Zollo et al. 2015; Khaldarova and Pantti 2016), aiming to deliberately stir powerful emotions in their readers. Some researchers called this feature of MDI a form of “empathic optimisation” (Bakir and McStay 2018, p. 155). Emotional manipulation in news items (especially click-bait) is an efficient way of capturing user’s attention since emotion-stirring news are usually more interacted with than the neutral ones (Bakir and McStay 2018, p. 155). It may seem that only MDI is emotionally loaded, whereas genuine news sound more sober and neutral. But this would be a misleading view of how information travels on social media platforms. Emotional reactions do not belong to misleading information alone, rather these are a normal side-effect of the emotional infrastructure already embedded in most social media platforms.
Users of social media platforms are allowed a palette of actions and reactions: some are seemingly neutral (commenting, sharing and posting) while others have a clear emotional valence: liking and using other emoticons to endorse or dislike a post. These emotionally charged reactions are easier to perform than the neutral ones: it takes a split second to click like on a post, but some more time to comment on it or even share it. Most of these emotional reactions have dedicated buttons which can be clicked mindlessly, yielding the interaction seamless. In the wake of the Covid-19 pandemic, Facebook even added a new emoticon expressing a reaction of solidarity (see Fig. 1). The assumption was that now, more than ever, users needed to express emotions online with a richer palette. However the simplistic way of expressing such emotions did not change, it was part of the interaction design from the beginning.
The emotional infrastructure of social media was not something requested by users but designed from the start. Major social media platforms are oriented towards maximising the user’s engagement, i.e. how much time one spends on a specific platform and how much attention is consumed (Whiting and Williams 2013). These kinds of interactions actively promote the user’s “attention bulimia” i.e. a behaviour oriented towards “maximizing the number of likes” (Del Vicario et al. 2016, p. 1) and presumably other positive reactions. Most buttons for emotional reactions are of positive emotions (like, love, hug, laugh) while in recent years Facebook added some more nuanced emotions such as angry, sad or cry. But the overwhelming effect of these emotional reactions is to make other users feel liked by their social network hence, to make the platform a place where one wants to keep returning to for emotional gratification.
For many users confined to their homes by the pandemic, social media platforms became a window to the world, as the television set used to be in older days and the easiest way of relating to others. In such times of distress and uncertainty, users posted more frequently than usual (Cinelli et al. 2020) but some of the information posted was not meant to inform others, but rather to express one’s concerns and emotions related to the pandemic. Posts were also meant to get reactions from one’s friends in an attempt to confirm that the others were also feeling the same way as one does. Posting about the pandemic became a strategic way of gauging other’s emotions on the crisis situation and gathering some feeling of consensus from one’s social network. The consensus sought on social media was of an emotional nature which may be at odds with an epistemic consensus about the nature of the facts at hand.
The weak epistemic context
During the pandemic, several epistemologists and philosophers of science stepped up and tried to educate the general public on what sources to trust as experts and how to discern facts from fiction about the pandemic—in podcasts, opinion pieces and on their social media accounts (Weinberg 2020). While this effort is laudable, it needs to be complemented with another approach, taking into account the wider epistemic context in which information travels on social media. This is a particularly weak epistemic context in which information is not always shared to inform. Social media platforms are not places where one usually goes to be informed. At least in regular, day to day situations, users turn to social media platforms to relate, to communicate and to be entertained (Fuchs 2014). The weak epistemic context of social media is ruled by serendipity (Reviglio 2017), meaning that many users get to be informed by accident.
In a crisis situation, users tend to change how they use the platform and shifting towards the communication of vital information such as imminent risks or their location and also seeking to be informed by latest developments from people from the local site of the disaster. The entertainment function tends to become secondary in emergencies (Zeng et al. 2016). In the 2020 pandemic situation, the difference was that the crisis was global and that the duration was rather long. This time, the uncertainty that accompanies a crisis situation was extended over months. As epistemic agents, online users tried to make sense of what was going on with them, what they could expect and to assess the personal risks, over a longer period.
The pandemic was an extended crisis situation compounded with social alienation on top. This made users feel lost and overwhelmed by problems one could not understand. Hence the desire—legitimate to a point—for everyone to be an expert so that they could at least understand what was happening to them. People did not want to be experts in epidemiology, quarantine measures, and home remedies for viruses because of a sudden surge of intellectual curiosity. They needed a way of coping that was also understandable to them. Meanwhile, the official discourse of “trust the experts” and “please don’t share information you do not understand” incapacitated them as epistemic agents. Requiring users to do nothing and just comply went against the general desire to do something, as a way to take control. Given the increase in posts on the pandemic by regular users, it may seem that many have tried to become experts overnight in epidemiology, viruses and vaccines. The comic below (Fig. 2) illustrates the frequent situation emerging during the pandemic of members of the lay public hijacking the role of the expert.
A discussion on conditions of trust and expertise makes sense in a regular epistemic context when agents try to acquire knowledge about a domain they know nothing about, having to choose which experts to trust (Goldman and O’Connor 2019). However, acquiring knowledge was probably not the main goal of social media users who started posting scientific information which they did not understand. Rather, many users tried to build some understanding of the situation, to make sense of the events. In these cases, shared understanding in a circle or network of friends seemed to be more important for users than gaining access to expertise. The scientific information was posted by lay users to back one’s personal opinions, to urge for a certain course of action, or to gather consensus. Given these weak epistemic uses of information targeted at emotional fulfilment and networking purposes, the regular content-focused measures would probably be less effective than predicted on such users.
The strong normative context
Related to the previous point and stemming from it, most factual information shared around on social media had some normative implications which often shadowed any knowledge claim. Descriptive information was used for prescriptive or evaluative aims. Scientific expertise was co-opted strategically to enforce one’s own pre-existing evaluative opinions. Typical MDI claims are not merely descriptive claims of a state of affairs in the world, but often embedded in a normative context be those prescriptive or evaluative claims, both types are meant to change attitudes of the online users. MDI was shared because it prescribed actions or led to evaluations of the state of affairs which users already agreed with. Hence debunking the facts would have solved only half of the puzzle, since the user’s motivation to believe these normative claims would have not been dealt with.
One example of the strong normative context for MDI, also involving a clear “politicisation” of MDI during the pandemic (Howard 2020), concerns one of the most popular types of claims analysed by EU vs Disinfo (2020) in which the EU was depicted as powerless and scattered in dealing with the pandemic. This claim was traced back to Russian-backed agencies which aimed to make users believe that, ultimately, Russia was stronger than the EU (Howard 2020). Such claims can be debunked by showing that there were coordinated measures taken by the EU, however, the implicit claim that other states dealt better with the pandemic than the EU is hard to debunk since it is not explicitly stated. This is just one type of difficulty with MDI which cannot be tackled with a content-focused approach: implicit evaluative claims in which one term of the comparison is not named.
Some evaluative claims can be checked (if these involve relational predicates which are measurable such as “better than” or “more efficient than”) however others, incidentally the politicised evaluative claims, are harder to assess. In Russian-backed claims against the EU, the name of “Russia” is not mentioned anywhere in the text of the “news”, since the aim is to erode the trust in EU from its citizens. If these citizens happen to be in Eastern Europe, this erosion of trust could lead to an anti-EU generalised feeling, and ultimately bottom-up pressures to exit the EU. These kinds of campaigns cannot be easily fact-checked since the effect is achieved by playing the long game. What looks like news about the pandemic is a dog-whistle about something else.
The strong normative context is visible also when using scientific expertise co-opted to back up prescriptive claims otherwise untenable. One example is an unpublished paper by Blocken et al. titled “Towards aerodynamically equivalent COVID19 1.5 m social distancing for walking and running” (2020) in which an animated image showed a simulation of how joggers coughing will spread particles of droplets when running at a distance of 1.5 m from each other. The paper became viral on social media despite not being peer-reviewed nor published on a pre-print website. The visual animation showing the spread of droplets was understandable by every lay member of the public, without needing to have specialised knowledge of actual aerodynamics, and presumably made the paper so popular among non-scientists who used it to make prescriptive claims by non-scientists. While the authors hypothesised that it might be unsafe to run close to another—and that even 1.5 m distance might not be enough for jogging—the social media audience took this as a reason to shame the runners in their neighbourhoods (Koebler 2020). Meanwhile, the first author of the study posted a document on his website answering certain questions about the studyFootnote 3 and refused to draw any epidemiological conclusions, urging for other’s expertise. But social media users did not shy from becoming experts and drawing the conclusions themselves, as the information in the Blocken et al. paper was just ammunition in a larger informational battle about what others should do.
Even if the scientific claims of regular users are checked, their aim remains to prescribe actions for others and to evaluate the world in a way that will be endorsed by one’s community of friends. For these purposes, other pieces of news will be co-opted if the first ones were flagged as hoaxes. Content-based approaches are then ineffective against this strong desire of social media users to emit evaluative or prescriptive claims about the world and strategically use science-looking sources to back these up. One should address the very desire of regular users to evaluate the world from the little soap-boxes that social media affords.