2. Blameworthiness and echo chambers

Echo chambers are bad. They are bad for many reasons, to enumerate a few: bad for democracy;Footnote ¹ have bad “sociological effects” such as seeing other groups as enemies (Jamieson and Cappella Reference Jamieson and Cappella2008); bad because of propagating epistemic vices (Baumgaertner Reference Baumgaertner2014; Nguyen Reference Nguyen2020). There is a consensus on the existence of negative effects of echo chambers. What is less clear is the answer to the question of who is morally responsible. Are the members of echo chambers blameworthy? Nguyen formulates this question in explicit terms as follows: “Could one behave epistemically virtuously, and yet remain caught within an echo chamber or epistemic bubble? In other words, to what degree is an epistemic agent embedded within such a structure blameworthy, or blameless, for the faultiness of their beliefs?” (Nguyen Reference Nguyen2020: 24) The answer according to the received view is yes.Footnote ² Members of echo chambers are blameworthy, and this is partly because of their bad epistemic practices. For instance, Quassim Cassam (Reference Cassam2016) claims that epistemic vices such as ignoring the contrary evidence are bad, and individuals are blameworthy when they commit epistemic vices such as laziness. Here is how Cassam describes a conspiracy believer named Oliver:

According to Cassam, Oliver's epistemic practices are extremely flawed, and they are flawed partly because of the individual epistemic vices he commits. Due to his “intellectual character,” Oliver deserves to be morally blameworthy. Similarly, members of echo chambers ignore contrary evidence and thus they exercise epistemic vices. It should be noted that epistemic vices do not automatically entail blameworthiness. However, as the above quote exemplifies there is a shared intuition that the epistemic practices exercised in the echo chambers are blameworthy. In The Big Sort, pointing out the harmful sociological effects of echo chambers, Bill Bishop claims that there are all kinds of media input, yet people “self-segregate into their own gated media communities” (Bishop Reference Bishop2009: 81). Bert Baumgaertner (Reference Baumgaertner2014) observes that individual bad epistemic practices lead to the formation of echo chambers and suggests that by adopting impartiality one can reverse the bad epistemic practices associated with echo chambers: “It seems plausible that one can resist the formation of an echo chamber, or at least resist the effects of one, by impartially seeking out people to interact with and allowing oneself to be informed by their opinion.” So it seems that it is the burden of the individual to not commit epistemic vices. Similarly, in “How to Escape Echo Chambers” Nguyen (Reference Nguyen2018) prescribes a Cartesian epistemic reboot for members of echo chambers to eliminate the formation of echo chambers. This strategy briefly suggests that members of echo chambers need to doubt and suspend their faulty beliefs in order to build a healthier network of beliefs. Neither Baumgaertner nor Nguyen explicitly say that all members of echo chambers are blameworthy for their bad epistemic practices, but by suggesting individualistic solutions both accounts imply that echo chambers are bad, bad in part because members are blameworthy for what they believe. Members of echo chambers commit epistemic vices.

However, it is too quick to conclude that all members of echo chambers are blameworthy for their bad epistemic practices. We need to distinguish different cases of being part of echo chambers. Nguyen (Reference Nguyen2020) separates two kinds of membership in an echo chamber: active and passive membership. In the former case, one is intentionally a part of an echo chamber. Whereas, in the latter case, one might be raised in an echo chamber without being able to identify the echo chamber. Nguyen claims that active members of echo chambers are blameworthy for their bad epistemic practices, but passive members might be blameless:

The main difference between active and passive membership is that, unlike passive members, active members intentionally exercise epistemic vices such as laziness, carelessness, and ignoring the contrary evidence. This is what makes active members blameworthy. I agree with Nguyen on the moral responsibility status of passive members. I think motives and reasons play a key role in ascribing moral responsibility. In the lack of intention to be a part of an echo chamber, it is not fair to find passive members guilty of their bad epistemic practices. Yet, I disagree with Nguyen in the case of active membership. I think it is sometimes reasonable for members of echo chambers to perform bad epistemic practices when they intentionally choose to be a part of echo chambers. In the rest of this paper, I argue that friendship might provide strong reasons for an epistemic agent to choose to participate in echo chambers and therefore the members of echo chambers might not be as blameworthy as the received view takes them to be. To flesh out this argument fully, first I need to explain what kind of friendship I have in mind. To that end, I turn to one plausible account in the ethics of belief literature: Sarah Stroud's (Reference Stroud2006) account of friendship.

3. Friendship

According to one prominent view of friendship, friendship involves epistemic partialityFootnote ³ (Annis Reference Annis1987; Blum Reference Blum1980; Stocker Reference Stocker1976, Reference Stocker1981; Stroud Reference Stroud2006). Michael Stocker (Reference Stocker1976) emphasizes the special motives and emotions friendship involves (Stroud Reference Stroud2006: 502). Building on Stocker's account, Sarah Stroud claims that it is part of the very nature of friendship that it “imposes special demands on our actions” (Reference Stroud2006: 502). These demands include charitably interpreting our friends' actions, giving more credence to our friends, being loyal to our friends, etc. (Stroud Reference Stroud2006). When friendship and epistemic ideals clash, Stroud (Reference Stroud2006) argues that we have strong reasons to suspend our epistemic ideals and be epistemically partial towards our friends. We can summarize the main thrust of her argument as follows:

A detailed analysis of this argument is not required for my purposes. But the support is fairly clear and intuitive. A good life requires friendship, and friendship sometimes requires epistemic partiality (e.g., charitably interpreting our friends' actions as opposed to strangers, giving more credence to our friends). Consider an example from Stroud's discussion:

From different parties' points of view, Sam's actions are interpreted differently. Sam's friends see his actions as perfectly normal, even laudable, while the affected parties might label the same actions as harassment. The upshot of this kind of case is that in our daily life, we as epistemic agents are not living up to the ideals of main epistemological theories. For instance, sometimes we do not search for firm evidence and justification when we are interpreting the actions of our friends. Stroud's account explains why we stop being epistemically virtuous in the case of friendship and concludes: “If satisfying a certain set of demands or ideals would preclude friendship, it follows that we all have very strong reasons not to live up to those demands or ideals.” We need friends; therefore it is sometimes permissible to choose friendship over being epistemically virtuous.

One important point is that Stroud does not claim that we should choose friendship when epistemic ideals and friendship clash. Rather, she claims that we have strong reasons to choose friendship over epistemic ideals. However, I think this has a normative import. If one has strong reasons to choose friendship over epistemic ideals, then one is sometimes permissible to do so. I will qualify this claim in the next section after exploring the role of friendship in the formation of echo chambers and its implications for moral responsibility.

4. Echo chambers and friendship

Returning to echo chambers, I have two aims in this section: 1. to show that echo chambers usually involve friendship; 2. to argue that sometimes friendships give us strong reasons not to live up to epistemic demands or ideals even when the friendships are formed in echo chambers.

Echo chambers usually involve friendship. This is fairly intuitive. It would be too strong and likely be false to claim that all echo chambers involve friendship. Think about some online communities on Reddit. The users on some harmful subreddits such as /incel interact by sharing memes that target certain groups of people (in this case, especially women). The interaction between members of groups is mostly through memes. Since most of the users do not share their actual identity, no one knows who the posters truly are. In this kind of case, it is hard to specify whether the members qualify as friends. Granted that not all echo chambers involve friendship, nevertheless, I think a weaker claim holds: Many echo chambers involve friendship.

But before going any further, I need to answer the following question: What is friendship? Stroud's account does not help define the boundaries of friendship. But the vast literature on friendship centers around the following conditions: mutual care, intimacy, shared activity, and trust (Helm Reference Helm2021). The status of necessity for each condition is contested.Footnote ⁴ However, we can arrive at an operative definition of friendship following Bennett Helm's description:

According to Helm's description, friends have a similar interpreting viewpoint of worldly affairs; share joint actions such as playing tennis; give more credence to each other than to strangers; and invest emotionally in the relationship. Each of these aspects is satisfied in the echo chambers that involve friendships. Consider the case from the movie Goodbye Lenin. Alex's mother wakes up from a long coma and she is in a physically bad condition. Alex and his mother live in East Germany, but the wall has already been demolished. Alex's mother hates West Germany, so if she learns that the wall is destroyed she might suffer from a fatal shock. So to protect his mother Alex must fuel the hatred towards West Germany while hiding the truth about the wall. Even though his mother starts to observe some signs of the ‘truth’, she reinterprets them according to Alex's story.

Consider a real-life example: CrossFit. The members of CrossFit communities share a joint athletic recreational activity. The members discuss the functioning of muscles and the body in general. They share dietary programs and advise and support each other. In “How We Gather” Casper ter Kuile describes communities like CrossFit and SoulCycle as “spaces traditionally meant for exercise have become the locations of shared, transformative experience.”Footnote ⁵ What makes some CrossFit and SoulCycle groups echo chambers is that once the group identity is established, the outsiders (e.g., people who do not exercise) appear to be “less” human. Once the individuals are bonded with the ties of friendships, the commitments of friendships follow. This aspect of friendship is succinctly described by Stroud as the following:

Jamieson and Capella, in a similar vein, claim that members of echo chambers form “community bonds” and cultivate a group identity, they share a “sense of belonging” and become loyal to the group (Jamieson and Cappella Reference Jamieson and Cappella2008: 179–80). This suggests that echo chambers are suitable places where friendships are cultivated. The members spend an immense amount of time together, they share joint activities, they trust each other, and emotionally invest in the group.

Sometimes friendships give us strong reasons not to live up to epistemic demands or ideals even when the friendships are formed in echo chambers. This is also straightforward. Once we accept both Stroud's claim that friendships involve epistemic partiality and the above claim that echo chambers usually involve friendship, it follows that friendship gives us strong reasons not to live up to epistemic demands or ideals even when the friendships are formed in echo chambers. The argument runs as follows:

I defended (A1) above by way of examples. I cannot undertake an exhaustive list of echo chambers but appealing to the strong intuition here would serve my purposes. The intuition is that cults, fitness communities, political audiences, hooligan groups, etc., that qualify as echo chambers are likely to involve friendships. For (A2), I appealed Stroud's account. I find this account plausible, and for my purposes, I do not need to defend Stroud's account against its attackers.

One might object that the jump from (A3) to (C1) is not warranted. According to this objection, there might be an imprecise threshold for which epistemic partiality demanded by friendship could be allowed. In other words, the harmful sociological effect of choosing to participate in an echo chamber might outweigh the benefit an epistemic agent gets from friendship. Individuals search for a good life, and to that end, they end up committing to echo chambers because of the friendships they entertain. But echo chambers are harmful to social life. If the social harm of participating in an echo chamber offsets the benefits of having a friendship, one should choose epistemic ideals rather than friendship.

I do not think this objection threatens the argument I presented. Regardless of the strength of this objection, here one should note that my claim is not that people should always choose echo chambers over epistemic ideals when echo chambers involve friendships. I only suggest that when epistemic ideals and echo chambers clash, people have strong reasons to choose to participate in echo chambers. Borrowing Regina Rini's (Reference Rini2017) phrase in her discussion on fake news, I claim that it is individually reasonable for epistemic agents to actively participate in echo chambers. Accordingly, to clarify the normative force of this argument, I added a further claim (5) that these strong reasons make it sometimes permissible to engage in bad epistemic practices. It is sometimes permissible because, as stated in the above objection, the harmful effects of participating in echo chambers might outweigh the benefits of friendship, in those cases individuals would still be blameworthy for their active participation in echo chambers.

My discussion here runs parallel to Regina Rini's (Reference Rini2017) discussion. Building her argument on recent epistemology of testimony literature, she writes that “shared partisan affiliation encourages testimony recipients to grant more credibility to testifiers than would otherwise be warranted” (Rini Reference Rini2017). So, according to Rini, it is individually reasonable for epistemic agents to give credibility to fellow partisans and distribute false information via their social media accounts. That is why Rini suggests that to eliminate the transmission of fake news, we need to focus on institutional change rather than individual bad epistemic. I agree with Rini that it is reasonable to expect individual epistemic agents to share fake news content. Similarly, I claim that it is individually reasonable for the members of echo chambers to continue their activities in the echo chambers when those echo chambers involve friendships.

For similar reasons, both Thomas Kelly (Reference Kelly2008) and Kevin Dorst (Reference Dorst2022) claim that belief polarization is rooted not in irrational epistemic behavior such as not responding to the evidence but in selective scrutiny and self-doubt – which are perfectly rational practices. As Dorst (Reference Dorst2022) writes:

According to Dorst, people end up adopting distorted beliefs due to a process not different than the adoption of healthier beliefs. Each individual with their own limited cognitive resources adopts a stance and “spend their time trying to explain away problems with it” (Dorst Reference Dorst2022). Agreeing with this picture, I would add that aiming for a good life plays an important role in adopting that stance in the first place.

Similar to Dorst, Baumgaertner and Justwan (Reference Baumgaertner and Justwan2022) recently claimed that individuals in echo chambers exert what they called a “preference for belief.” They observe that “in the absence of defeaters, individuals tend to (and are warranted/justified to) believe what someone in their epistemic community invites them to believe.” This tendency, they claim, makes the echo chambers resilient to individualistic interventions such as promoting open-mindedness. So they suggest that to prevent polarization more investigation needs to be done on the initial formation of the beliefs in a given echo chamber.Footnote ⁶ To me, this highlights the importance of the link between echo chambers and friendship. Baumgaertner and Justwan (Reference Baumgaertner and Justwan2022) do not question why individuals have the tendency to believe what their peers say in epistemic communities. Individuals might have this tendency partly because they entertain friendships in echo chambers which gives them strong reasons to be epistemically partial.

Now, if we accept the conclusion (C2), this leaves us with the following puzzle: Either (i) there is a ground on which we might consider the members of echo chambers less blameworthy or (ii) we should reject Stroud's account of friendship. One crucial point here is that although Stroud's account is crucial to my argument, I do not need to defend it against any objections. This is because my argument might as well work as a conditional. If one agrees that Stroud's account is plausible, then friendships in echo chambers render the members less blameworthy.

5. Should we eliminate echo chambers? How?

Many start their discussion on echo chambers by characterizing the phenomenon and end it with suggested ways of elimination. As mentioned in Section 2, the trodden path leads to individualistic solutions. For instance, the members of echo chambers need to suspend their faulty beliefs and start over (Nguyen Reference Nguyen2018, Reference Nguyen2020) or learn to become more impartial (Baumgaertner Reference Baumgaertner2014). However, given that the members of echo chambers have strong reasons to be part of echo chambers I believe these suggestions are misleading. In this section, I gesture my suggestion for the problem echo chambers cause.

But before that, it should be noted that it is not obvious that echo chambers are always bad. Jamieson and Cappella (Reference Jamieson and Cappella2008: 244) claim that some echo chambers “[arm] their members effectively” and thus are conducive to active political engagement. Thus not all effects of echo chambers are negative. Having said that, the harmful effects of echo chambers, I believe, require an institutionalist solution. Consider the following discussion of Rini's on fake news:

Rini suggests that institutions need to implement clear norms that would regulate communication on social media. According to Rini, if the institutions make how the norms work clear, then more and more users would be mindful of the relevant social norms governing the platform. Consequently, the users would be more cautious about their mistakes and the tweets they retweet. So eventually clear norms would mitigate the current dissemination of content related to fake news. In the case of echo chambers, my discussion leaves us with a similar result to Rini's: Granted that the echo chambers constitute a sociological threat, the solution requires an institutional change. As mentioned briefly in the first section, both Baumgeartner (Reference Baumgaertner2014) and Nguyen (Reference Nguyen2020) suggest individualistic solutions to prevent the formation of echo chambers. But taking the argument from friendship seriously gives us the following result: focusing on individual epistemic virtues is not helpful in cases where echo chambers involve friendships. The force of this consequence, of course, depends on the frequency of friendship-infused echo chambers. Clearly, statistical evidence requires an in-depth analysis of all echo chambers, which I am not able to provide within the limits of this paper. However, based on the prevalence of the examples in the previous section, I strongly believe that most echo chambers involve friendship.

To sum up, when epistemically virtuous practices weigh more than epistemically vicious practices in a given echo chamber, the formation of an echo chamber might result in sociologically positive effects. Yet, there are still some echo chambers that are bad. If I am correct in pointing out the futility of focusing on individual epistemic vices, then the solution to the problem of those echo chambers calls for institutional change.

In this respect, Neil Levy's (Reference Levy2021) discussion on epistemically polluted environments is especially helpful. Epistemic pollution, Levy claims, occurs when the marks of true expert knowledge do not suffice to distinguish it from fraudulent experts, and so the faulty beliefs proliferate. The marks: credentials (e.g., having a PhD, publishing peer-reviewed research), track records (a record of making predictions), argumentative capacity, and agreement with the consensus (other experts recognize the truth of the research) (Levy Reference Levy2021: 117–22). Yet, Levy claims that these marks are not mimic-proof. For instance, diploma mills and predatory journals can easily generate “solid” credentials. Considering the ineffectiveness of individualist solutions (e.g., practicing critical thinking, open-mindedness, epistemic humility) against epistemic pollution, Levy (Reference Levy2021: 125–31) suggests several courses of collective action that center on restoring trust in the scientific institutions: reducing the number of predatory journals, cutting the funding for researchers publishing in illegitimate journals, reducing the media pressure on the content of publications.

Once we take the link between friendship and echo chambers into account, since the individuals are less blameworthy for their faulty beliefs, the greater extent of moral responsibility would be on the institutions. Both Levy's and Rini's suggestions would be effective for an institutional change, but I doubt that they would be sufficient. More ink needs to be spilled over the sorts of interventions required in tackling the pressing case of echo chambers. Although, I am not able to suggest policy proposals, by undermining the received view's epistemic vice narrative, the argument from friendship supplements the recent voices that raise concerns over individualistic solutions.

6. Objections

In this section, I consider three objections to my project. The first one considers the plausibility of Stroud's account; the second one points out the qualitative difference between friendships cultivated in and out of the echo chambers; and the third one raises a thought experiment to challenge the implications of my argument.

The argument from friendship runs on the plausibility of Stroud's view. One might reject Stroud's account altogether and claim that the argument does not take off the ground. This is correct. I concede that the argument needs Stroud's view to work. But I do not think this objection engages with the argument itself. The force of the argument is still there. This objection can be evaded by taking the conclusion as a conditional – as mentioned briefly in Section 4. Granted that many echo chambers involve friendship, if friendship involves epistemic partiality, then the members of echo chambers have strong reasons to choose to participate in echo chambers over epistemic demands or ideals. This move thereby implies that rejecting the antecedent only amounts to deciding not to engage with the argument.

One might object by saying that friendships formed in echo chambers are qualitatively different from the friendships cultivated outside the echo chambers. The former involves cases of friendships that we should avoid at all costs. I do not believe this is true simply because I cannot think of any qualitative difference between, say, a friendship cultivated in a CrossFit community and an academic community (assuming the former is a case of an echo chamber while the latter is not). In both cases, agents might support each other; form long-lasting bonds; exercise the practices of reciprocal love and respect; share a joint activity; and build trust and intimacy. One might still object by pointing out that Rush Limbaugh was not a “friend” to his audiences. To this, I would say yes. The parties who deceivingly benefit from the formation of echo chambers do not need to be friends with the members of the echo chambers. It is enough for my account to work if one entertains friendships with other members of the echo chambers. The only relevant point is that echo chambers might be places where individuals cultivate friendships.

One might point out the counterintuitive results of the argument from friendship as the following: Assuming that the Nazi soldiers were part of an echo chamber and that they were friends with one another, the argument suggests that these members were not blameworthy for their faulty beliefs (e.g., genocidal beliefs). Furthermore, it is individually reasonable for these members to actively participate in the echo chamber. This result challenges our intuition that the extreme beliefs of some individuals (e.g., Nazi soldiers) are blameworthy, and regardless of the friendships cultivated in echo chambers, these individuals should be held morally responsible for their faulty beliefs. Considering this objection, I think there are at least two possible routes to take. First, the Nazi soldier case might not be among the cases that are permissible in the scope of the argument from friendship. One reason for this might be that the harm of the genocidal beliefs outweighs any kind of friendship duties. So the argument remains intact. Second, even if we grant that it is permissible in the scope of the argument, the force of this objection disperses once we give proper weight to the extent of moral responsibility. Due to friendship, the individual members of the group have strong reasons to participate in the echo chamber and this implies that they are not as blameworthy as is suggested by the received view. Since the individual members bear less moral responsibility than the received view claims, the individualist solutions are misleading. The friendship aspect of echo chambers stressed in this paper adds to the importance of the social networks in which the beliefs are formed. As the growing number of dissenting voices suggests, greater care needs to be exerted on institutional change (Baumgaertner and Justwan Reference Baumgaertner and Justwan2022; Levy Reference Levy2021).