How should a group with different opinions (but the same values) make decisions? In a Bayesian setting, the natural question is how to aggregate credences: how to use a single credence function to naturally represent a collection of different credence functions. An extension of the standard Dutch-book arguments that apply to individual decision-makers recommends that group credences should be updated by conditionalization. This imposes a constraint on what aggregation rules can be like. Taking conditionalization as a basic constraint, we gather (...) lessons from the established work on credence aggregation, and extend this work with two new impossibility results. We then explore contrasting features of two kinds of rules that satisfy the constraints we articulate: one kind uses fixed prior credences, and the other uses geometric averaging, as opposed to arithmetic averaging. We also prove a new characterisation result for geometric averaging. Finally we consider applications to neighboring philosophical issues, including the epistemology of disagreement. (shrink)

The present paper proposes an integrative account of social forms of practical irrationality and corresponding disruptions of individual and group-level emotion regulation. I will especially focus on disruptions in emotion regulation by means of collaborative agential and doxastic akrasia. I begin by distinguishing mutual, communal and collaborative forms of akrasia. Such a taxonomy seems all the more needed as, rather surprisingly, in the face of huge philosophical interest in analysing the possibility, structure and mechanisms of individual practical irrationality, with very (...) little exception, there are no comparable accounts of social and collaborative cases. However, I believe that, if it is true that individual akrasia is, in the long run, harmful for those who entertain it, this is even more so in social contexts. I will illustrate this point by drawing on various small group settings, and explore a number of socio-psychological mechanisms underlying collaborative irrationality, in particular groupthink. Specifically, I suggest that in collaborative cases there is what I call a spiralling of practical irrationality at play. I will argue that this is typically correlated and indeed partly due to biases in individual members' affect control and eventually the group's with whom the members identify. (shrink)

We evaluate the asymptotic performance of boundedly-rational strategies in multi-armed bandit problems, where performance is measured in terms of the tendency (in the limit) to play optimal actions in either (i) isolation or (ii) networks of other learners. We show that, for many strategies commonly employed in economics, psychology, and machine learning, performance in isolation and performance in networks are essentially unrelated. Our results suggest that the appropriateness of various, common boundedly-rational strategies depends crucially upon the social context (if any) (...) in which such strategies are to be employed. (shrink)

In this paper, I offer an analysis of the radical disagreement over the adequacy of string theory. The prominence of string theory despite its notorious lack of empirical support is sometimes explained as a troubling case of science gone awry, driven largely by sociological mechanisms such as groupthink (e.g. Smolin 2006). Others, such as Dawid (2013), explain the controversy by positing a methodological revolution of sorts, according to which string theorists have quietly turned to nonempirical methods of theory assessment (...) given the technological inability to directly test the theory. The appropriate response, according to Dawid, is to acknowledge this development and widen the canons of acceptable scientific methods. As I’ll argue, however, the current situation in fundamental physics does not require either of these responses. Rather, as I’ll suggest, much of the controversy stems from a failure to properly distinguish the “context of justification” from the “context of pursuit”. Both those who accuse string theorists of betraying the scientific method and those who advocate an enlarged conception of scientific methodology objectionably conflate epistemic justification with judgements of pursuit-worthiness. Once we get clear about this distinction and about the different norms governing the two contexts, the current situation in fundamental physics becomes much less puzzling. After defending this diagnosis of the controversy, I’ll show how the argument patterns that have been posited by Dawid as constituting an emergent methodological revolution in science are better off if reworked as arguments belonging to the context of pursuit. (shrink)

The chapter presents the prototyping of a thinking routine designed to foster good inference habits in children ages 6 to 11. The prototyping was developed at Ninho, an educational project for children from underprivileged households in Brazil. The thinking routines by Ritchhart and colleagues (2006) served as our starting point. Following a Virtue Education (VE) approach, we supposed that the repeated application would conduce to habituation. In addition, to increase peer-to-peer interactions, the teacher applying the routines worked as a facilitator (...) in a Community of Inquiry (CI). After six months of application, the results were partially successful. We identified that the repeated exposition to the magic question “What makes you say that?” made children more aware of the evidence supporting their assumptions. Furthermore, the interactions between peers made them recognize different perspectives. However, we also identified some shortcomings. Most of them seemed to arise from cognitive biases. At the individual level, belief perseverance kept children too attached to their first assumptions. They preferred self-serving rationalizations instead of accepting a counter-argument. At the group level, we identified problems of social contagion such as information cascades. The effects of a first opinion rhetorically voiced were hard to efface. Since none of the steps on the previous routines addressed these biases, we prototyped a routine to start filling this gap. Following Critical Thinking (CT) theorists, we added a step of structured instruction concerning one specific reasoning technique - the inference to the best explanation. Moreover, the recommendation of cognitive psychologists motivated the inclusion of some extra features to avoid groupthink biases. For instance, a visual table that juxtaposes contrasting arguments should facilitate comparative evaluations. Also, a star-based evaluation scale should help different individuals to discuss their assessments based on a common ground. As described in the end, the prototype with elements from VE, CI and CT presented promising results. (shrink)

A group is often construed as one agent with its own probabilistic beliefs (credences), which are obtained by aggregating those of the individuals, for instance through averaging. In their celebrated “Groupthink”, Russell et al. (2015) require group credences to undergo Bayesian revision whenever new information is learnt, i.e., whenever individual credences undergo Bayesian revision based on this information. To obtain a fully Bayesian group, one should often extend this requirement to non-public or even private information (learnt by not all (...) or just one individual), or to non-representable information (not representable by any event in the domain where credences are held). I pro- pose a taxonomy of six types of ‘group Bayesianism’. They differ in the information for which Bayesian revision of group credences is required: public representable information, private representable information, public non-representable information, etc. Six corre- sponding theorems establish how individual credences must (not) be aggregated to ensure group Bayesianism of any type, respectively. Aggregating through standard averaging is never permitted; instead, different forms of geometric averaging must be used. One theorem—that for public representable information—is essentially Russell et al.’s central result (with minor corrections). Another theorem—that for public non-representable information—fills a gap in the theory of externally Bayesian opinion pooling. (shrink)

Wokeism—as an ideology—is a potent enabler of group polarization. Secondly, polarization among its upholders results from the fact that they often display those behavioral patterns typical of behavioral conformity. Thirdly, groupthink and behavioral conformity jointly lead the upholders of wokeism to disbelieve any view that question theirs.

Some religious communities argue that public policy is best decided by their own members, on the grounds that collaborating with those reasoning from secular or “worldly” perspectives will only foment error about how society should be run. But I argue that epistemology instead recommends fostering disagreement among a plurality of religious and secular worldviews. Inter-worldview disagreement over public policy can challenge our unquestioned assumptions, deliver evidence we would likely have missed, and expose us to new epistemic alternatives; when done respectfully, (...) it can also combat epistemically problematic biases and groupthink. I address two objections that members of a politically active religious community might raise: one that inter-worldview disagreement about public policy is not needed because one’s own beliefs are already true, and another that it is not needed because one’s own beliefforming processes are divinely guided. (shrink)

This chapter includes a discussion of leadership decisions and stress. Many leaders are daily exposed to stress when they must make decisions, and there are often social reasons for this. Social standards suggest that a leader must be proactive and make decisions and not flee the situation. Conflict often creates stress in decision-making situations. It is important for leaders to understand that it is not stress in itself that leads to bad decisions, rather, bad decisions may be the result of (...) time pressure in the sense that leaders have not been able to gather enough relevant information. Thus, it is worthwhile for leaders to be able to prioritize properly in order to cope with stressful situations. In some situations, a leader chooses to delegate the decisions to his/her team and then it is important to guard against «groupthink», a phenomenon where members of a team put consensus before anything else as a result of the peer pressure. A number of methods are presented that enable leaders to avoid this phenomenon. Often leaders are involved in decision-making situations where they are forced to navigate between objectives that are in strong conflict with each other. We are talking about «decision dilemmas». These are characterized by the existence of a conflict between the top leadership's desire to control the activities and their wish to give autonomy and independence to the various units. It is important for leaders to be able to strike a balance in different dilemma situations and understand how to best manage conflicts when they arise -/- . (shrink)

Nineteen Prescott Fire Department, Granite Mountain Hot Shot (GMHS) wildland firefighters (WF) perished in Arizona in June 2013 Yarnell Hill Fire, an inexplicable wildland fire disaster. In complex wildland fires, sudden, dynamic changes in human factors and fire conditions can occur, thus mistakes can be unfortunately fatal. Individual and organizational faults regarding the predictable, puzzling, human failures that will result in future WF deaths are addressed. The GMHS were individually, then collectively fixated with abandoning their Safety Zone to reengage, committing (...) themselves at the worst possible time, to relocate to another Safety Zone - a form of collective tunnel vision. Our goal is to provoke meaningful discussion toward improved wildland firefighter safety with practical solutions derived from a long-established wildland firefighter expertise/performance in a fatality-prone profession. Wildfire fatalities are unavoidable, hence these proposals, applied to ongoing training, can significantly contribute to other well-thought-out and validated measures to reduce them. (shrink)