![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
People suffer from a variety of cognitive shortcomings when forming and updating their political beliefs. Three pervasive shortcomings are confirmation bias, disconfirmation bias, and motivated reasoning. The emotional state of anxiety can help us overcome these biases given the open-minded, information-rich, reflective deliberation with diverse people it may promote—although mass and social media may hinder this type of deliberation.
Notes
1 Indeed, Gilbert Harman (Citation1999 and Citation2000), based on social-psychological evidence, argues against the existence of character traits. John M. Doris (Citation2002) more moderately argues that “narrow” but not global character traits can be empirically supported. Such traits have limited explanatory power, since agents do not manifest them in most or all circumstances.
2 See Alston Citation2005 for truth promotion and further features that are desirable from an epistemic point of view.
3 Belonging to a certain group and identifying with it can be, and often is, shaped by contingent matters of the individual’s history without any regard for the truth of one’s beliefs.
4 Consider this: if one takes a group of people to be one’s epistemic peers (i.e., they are as likely as oneself is to reach the truth on some matter) and they all disagree with oneself (as in Asch’s famous experiment), it might be epistemically adequate to change one’s mind and believe as they do.
5 Many factors, as mentioned, can in principle play an etiological role in the formation and updating of political beliefs. Although here we focus on some truth-deflecting sub-personal shortcomings, satisfying psychological explanations are certainly likely to concern a series of factors of many kinds (e.g., personal, social, and environmental). So our claim regarding the counteracting power of anxiety needs to be restricted given that other factors might interfere with this power. Moreover, even if such sub-personal shortcomings sometimes play a more significant role than other factors (sub-personal and otherwise), one would need to experience anxiety at the right times and to the right degree in order for open-minded, information-rich, reflective political deliberation to be promoted (we come back to this issue in section IV). This certainly might not often happen, which might in turn explain why we don’t often see such political deliberation among individuals (leaving aside the fact that other factors could also preclude it).
6 Of course, this is not to deny that political belief-formation and updating might not be, even quite often, the result of automated heuristics: namely, low-information, “peripheral” mental shortcuts (Kahneman Citation2011). Peripheral processing, which relies on quick, implicit inferences, may exploit informational cues: say, regarding candidate traits (Popkin Citation1991), party identification (Campbell et al. Citation1960), or public mood (Rahn Citation2000). But just as heuristics may badly mislead people (Kuklinski and Quirk Citation2000), they may also allow “low-information rationality” (Popkin Citation1991). Heuristic processing per se needn’t be epistemically suspicious, unlike processing that is the result of confirmation and disconfirmation bias and motivated reasoning.
7 But see Friedman Citation2019, ch.5, which proposes a non-motivational explanation of the phenomena that the theory of motivated reasoning is meant to explain, which he calls dogmatism and attributes to a “spiral of conviction” that rationally and unavoidably screens in confirmatory evidence for whatever interpretation allows us to make sense of evidence, while screening out disconfirming evidence.
8 The relevant groups can be, and often are, political parties (Campbell et al. Citation1960; Cohen Citation2003): they are the most salient groups in democratic politics and most people identify with some political party. But ethnic groups, religious groups, genders, and more local groups can be relevant too.
9 After all, people normally have pre-existing political beliefs when collecting reasons (in which case confirmation bias is very likely to kick in); they often reflectively form and update such beliefs (in which case disconfirmation bias is likely to kick in); and they tend to have group attachments too, which may motivate the protection of their social identities when forming and updating political beliefs (in which case motivated reasoning is likely to kick in).
10 For any emotion one can have a corresponding trait (i.e., the tendency to have the emotion).
11 Pathological anxiety can manifest in the form of phobias, panic disorder, generalized anxiety disorder, obsessive-compulsive disorder, and post-traumatic stress disorder.
12 Cf. LeDoux Citation2015 and Rachman Citation1998, arguing that anxiety is exclusively consciously experienced uneasiness or nervousness. For reviews of different accounts of emotions, see Deigh Citation2009 and Price Citation2005.
13 We can distinguish between two kinds of cognitive accounts: those that maintain that specific propositional attitudes are both necessary and sufficient (pure cognitive theories) and those that maintain that they are only necessary (hybrid theories in which emotions are complexes of propositional attitudes and feelings). The pure cognitive alternative can be ruled out in our case, given that it neglects the feelings associated with emotions (see also Goldie Citation2000).
14 The work of several research teams supports and replicates the relevant results (to be introduced below). These teams exploit diverse paradigms and methodologies, including experimental designs of simulated election campaigns, political websites and news pieces that induce anxiety, and survey research.
15 For example, one party may advance reason R1 in favor of some claim and another one may respond by introducing a counter-reason or defeater D that speaks against the claim; then the first party may introduce a defeater of the defeater DD or concede R1 that has been defeated—or weakened to certain extent—and perhaps introduce some new reason R2, and so on. Eventually they weight the reasons for and against to see how strong is the case for the claim. Here, of course, we mean the epistemic weighting of reasons. Each party attempts to rationally persuade the other parties by them conveying the quality of the reasons (not by, say, manipulating or bargaining with them).
16 Interestingly, as Pfau Citation2007 shows, both Aristotle and Demosthenes argued for the need for some such emotional state to move the citizenry to effective political deliberation.
17 For example, according to a recent survey (Newman et al. Citation2020), people from the United Kingdom, United States, and Chile, where the Internet penetration is 95 percent, 89 percent, and 77 percent, respectively, normally get the news online (including through social media: 77 percent, 72 percent, and 86 percent, respectively). Not only are online media the main source of news in these and many other countries, but in the last few years there has been a constant and significant global drop in the number of people whose sources of news are television and print newspapers (for example, in the United Kingdom in 2013, 79 percent relied on TV and 59 percent relied on newspapers, whereas in 2020 only 55 percent and 22 percent did, respectively); see Newman et al. Citation2020.
18 In the third quarter of 2020, 1.82 billion users were checking Facebook daily (and 2.74 billion monthly), which represents a 12 percent increase over 2019. This is a tendency that continues to repeat year after year; see Facebook Reports Third Quarter 2020 Results.
19 Moreover, it is of lower quality than other kinds of platforms such as news forums. In particular, the reasoning and argumentative engagement in deliberation on social media is significantly lower (Esau et al. Citation2017). This does not suggest that effective online deliberation is impossible. In fact, it has been argued that platforms with a not-uncommon online design that favors anonymity and asynchronicity can promote better high-quality deliberation than the especially designed offline meetings of Deliberative Polls (De Brasi and Gutierrez Citation2021).
20 This is due to the fact that, first, one chooses whom to interact with, and often these people are family, friends, and acquaintances who share many relevant views. Second, the algorithms present one with the type of posts that, given one’s clicking history, one seems to prefer. Now, according to research conducted by Facebook employees on 10.1 million U.S. users (Bakshy et al. Citation2015), although the social algorithm limits viewpoint diversity, how much cross-cutting content individuals encounter depends primarily on who their friends are. However, this research was done before a 2016 change of algorithm that aimed to further personalized the experience.
21 Google enjoys more than 90 percent of the global search market (Oberlo Search Engine Market Share 2021). In December 2009, personalization was adopted for all users by Google (Personalized Search for Everyone), based upon previous search activity linked to an anonymous cookie in the browser and 56 other signals (Pariser Citation2012).
22 We suffer from an overconfidence bias due to overestimation that makes this outcome likely (Hoffrage Citation2017).
23 Nor does our focus on social media suggest that legacy media are a panacea. It has been shown that one’s use of mass media often results in selective exposure both for strongly ideologized people and for those who maintain more neutral positions (Bou-Hamad and Yehya Citation2016), which is related to greater polarization of political beliefs (Stroud Citation2010).
24 See Lynch Citation2019, 147-49, and Redlawsk et al. Citation2010, 590, for the possibility of change even when holding extreme views very confidently. Having said that, the well-known phenomenon of belief perseverance (e.g., Ross et al. Citation1975) seems to speak against our capacity to change our minds in response to belief-incongruent reasons. Indeed, there may even be a “backfire effect” (Nyhan and Reifler Citation2010) whereby we increase our confidence in our beliefs when facing belief-incongruent information, updating our confidence in the wrong direction. However, the settings for the experiments where these phenomena are meant to show up are not deliberative; hence the studied phenomena are not incompatible with the ability to change one’s mind in response to reasons offered in deliberation. Moreover, a recent meta-analysis regarding the psychological efficacy of disconfirming information suggests that the debunking effect is more effective when people are provided with new detailed reasons, which is likely to be the case in deliberation with relevantly diverse individuals, which enables them to adjust the mental model justifying their beliefs (Chan et al. Citation2017). Furthermore, as our anxiety increases because more counterevidence is confronted, we pay more attention to it and start making the right adjustments to our beliefs (Redlawsk et al. Citation2010). So it is not surprising that the backfire effect has not been replicated (Wood and Porter Citation2019), especially since changes of mind due to deliberation, as Barabas Citation2004 shows, depend on factors such as deliberators’ increase in knowledge, their diversity of opinions, and their willingness to keep an open mind.
25 Given that argumentation is a social enterprise, we do not merely rely on each other for information but also for critical pushback. For more on this and a plausible evolutionary story supporting it, see Mercier and Sperber Citation2017.
26 Discussion with like-minded people exposes us, as we have seen, to a high concentration of reasons that confirm our antecedent views. Yet another process is social comparison: group members attempt to maintain their reputation and self-conception by emphasizing the attitudes they perceive to be normative within the group. But this again is likely to happen only in a group of like-minded people, so group heterogeneity can also remedy group polarization due to social comparison. One might wonder whether, if a group is heterogeneous, minorities within it might silence themselves for fear of being ridiculed or fear of some other social punishment (Sunstein Citation2006, 68), but a group norm that welcomes dissent can nudge the minority to contribute (Paluck and Green Citation2009). Of course, whether someone is then heard when speaking is another issue; people might not listen to others because of differences in social identity, gender, etc., but as we have seen, anxiety predisposes one to such listening.
27 This is not to suggest however that motivated reasoning is consciously switched off but rather that anxiety seems to sub-personally trigger a certain type of cognitive processing. For a worry concerning the metaphysics of sub-personal motivation, see Friedman Citation2019, 246.