Johnson et al. have taken a radical step in relation to “radical uncertainty.” They have created a model for understanding the wide body of research demonstrating that individuals tend to overestimate their own rationality and decision-making capacities, including, according to the authors, academics, politicians and economists, who collectively tend to have overestimated human rationality. Their case is profound in arguing that the network of norms and institutions society gathers into a system of rules for identifying truth – what Jonathan Rauch, a senior fellow at the Brookings Institution, recently termed a constitution of knowledge (Rauch, Reference Rauch2018) – is mistaken in relation to understanding how decisions are made.

Here we seek to extend the implications of Conviction Narrative Theory in relation to the role of imagination and its relationship with the social context. Johnson et al. argue that both factors are insufficiently incorporated into decision-making theory: We additionally propose that they are dynamically related to one another, and that it is the nature of this relationship that determines individuals' capacity to respond adaptively to new information. One defining difficulty with conviction narratives is that, notwithstanding evidence that challenges their position, individuals are often unable to adjust their narratives. One way of understanding this is irrationality; another way is through petrification of the narrative in the light of epistemic mistrust: New information that challenges the narrative is understood but not adopted into what Polanyi (Reference Polanyi1959/2014) calls “personal knowledge,” the sum total of what any individual knows, because it is not regarded as relevant (Fonagy, Luyten, & Allison, Reference Fonagy, Luyten and Allison2015). Inflexible, dysfunctional conviction narratives are indications of a lack of trust in incompatible but pertinent information. We understand the possible reasons for this epistemic closure and unwarranted certainty as inter-relational in nature (Fonagy et al., Reference Fonagy, Campbell, Constantinou, Higgitt, Allison and Luyten2022). In a developmental psychopathology context, we described the factors that drive this entrenchment as early relational insecurity (Fonagy, Luyten, Allison, & Campbell, Reference Fonagy, Luyten, Allison and Campbell2017a, Reference Fonagy, Luyten, Allison and Campbell2017b). But we have also more broadly come to understand this epistemic shut-down as an outcome of functioning within a social environment in which the individual does not feel recognised in the communications they receive. We have suggested that such lack of experience of recognition in others' narratives in relation to the self may constitute something of the “missing link” in understanding the relationship between socio-economic deprivation and/or alienation and psychopathology (Fonagy et al., Reference Fonagy, Campbell, Constantinou, Higgitt, Allison and Luyten2022).

Epistemic dysfunction does not exclusively manifest as outright mistrust, it also presents as lack of discrimination in relation to the communication of knowledge, rendering the individual vulnerable to exploitation or misinformation, which we have termed epistemic credulity (Campbell et al., Reference Campbell, Tanzer, Saunders, Booker, Allison, Li and Fonagy2021). In a state of epistemic credulity, individuals misread or are misled by the narratives presented by others, imagining that they are being recognised and understood, and thus reducing adaptive epistemic vigilance towards the information they receive (Sperber et al., Reference Sperber, Clement, Heintz, Mascaro, Mercier, Origgi and Wilson2010). We might think here of a populist politician who, despite their blatant self-interest, creates a powerful narrative of being able to represent the sense of betrayal and injustice felt by the ordinary person, to the extent that obvious misinformation about election results is believed.

It is easy to pathologise these processes as they manifest in their most egregious forms, perhaps partially driven by a social media-escalated breakdown in epistemic consensus. In fact, such ways of thinking together hold important functions in maintaining social stability (Mercier & Sperber, Reference Mercier and Sperber2017) in an expectable environment (Cicchetti & Lynch, Reference Cicchetti, Lynch, Cicchetti and Cohen1995). Community and collaboration are supported by a set of mental processes reserved for shared cognition or relational mentalising, also described as the “we-mode.” When we intend to accomplish an outcome jointly with others, we adopt a “first-person plural perspective” – the we-mode (Gallotti & Frith, Reference Gallotti and Frith2013, p. 160). The we-mode may be organised around cognitive and neural structures that are intrinsic to individual make-up and are the product of a distinct developmental and evolutionary history (Tomasello, Reference Tomasello2019). Building on joint intentionality, the joint agent emerges where mental states are aligned to achieve a common goal, grounded in respect born of each having a role in the collaborative activity (Tomasello, Reference Tomasello2016). The route by which the we-mode is triggered may be understood thus: (1) The learner's imagined sense of self (their personal narrative) (2) is imagined by the instructor, establishing a prospect for the we-mode and (3) this image is perceived by the learner, reinforcing the potential we-mode and (4) compared with the learner's personal narrative and (5) in case of a match, co-representation is created: The we-mode removes the I-mode's protection from change and the channel for rapid, efficient modification of personal knowledge is opened. Relational mentalising is thus key to establishing epistemic trust, it is dependent on both instructor and learner taking an imaginative leap in relation to the other's mind, but using that imagination in a manner that is contained within a reasonably accurate assessment of both one's own and the other's mental state. Mistrust might be understood as a failure of imagination, credulity as manipulated or untethered imagination.

Johnson et al. describe narratives as representations in individual minds; shared narratives as being formed by the common elements fragmentarily shared across a social group. We suggest that it is harder to separate individual from shared narratives and knowledge than this conceptualisation suggests. Conviction narratives are particularly relevant for social contexts where strong boundaries around social groups exist and information incompatible with current beliefs are discarded if it comes from individuals who are members of an out-group to whom epistemic trust is not extended (Tong, Wang, & Danovitch, Reference Tong, Wang and Danovitch2020). As information incompatible with beliefs accumulates, paradoxically mistrust increases as beliefs come to serve to identify the social group who hold them. This is an outcome of epistemic vigilance (sometimes hypervigilance) precluding the integration of socially conveyed information into knowledge schemata. We have recently argued that models for understanding social processes and social cognition have failed to learn from developmental psychopathology (Campbell & Fonagy, Reference Campbell and Fonagy2022); here we perhaps make a bolder point that the psychopathology of everyday life – the lapses in communication, ruptures and the outcomes of inhibited or disorderly imagination – is the engine of culture, because these processes create the narratives that hold (or fail to hold) groups together.