Minimal Virtues: From Mindreading to Ethics via Joint Action (Munich)

Date given: Wednesday, 8th February 2023

Handout for all of Minimal Virtues: From Mindreading to Ethics via Joint Action (Munich) as pdf. (This has the same content as the web version you are reading here.)

Notes

These are notes and slides for a talk based on some work-in-progress.

Abstract

How do agents ever perform optimally when time is pressing and cognitive resources such as working memory are scarce? One strategy is to rely on minimal models of a domain. This is well established for goal selection, where habitual processes rely on stimulus–action links in contrast to slower processes which compute expected utilities (Dickinson, 2016); and for physical cognition where broadly perceptual processes rely on impetus mechanics in contrast to slower processes which, in experts, can apply a Newtonian model (Hubbard, 2022).

Are minimal models similarly critical for explaining resource-limited performance in mindreading, joint action and ethical cognition? This talk will review puzzling patterns of discoveries which motivate constructing minimal models and testing predictions about their use.

In the mindreading domain I will invite discussion of a recent challenge for the minimal approach. Two labs have explored connections between fast mindreading processes and motor cognition. Mixed results hint that new ideas are needed.

On joint action, the puzzle is developmental. The leading view has it that infants’ earliest joint actions already involve shared intention, which in turn requires sophisticated forms of mindreading; yet there is also a body of evidence for the view that those sophisticated forms of mindreading are a consequence of cultural learning, which characteristically involves joint action. To escape the circle we can construct a minimal model of joint action using the notions of collective goal, dyadic action plan and task co-representation.

In ethics there is a puzzle about whether fast ethical processes are characteristically deontological (Greene, 2014) or characteristically utilitarian (Kurzban, DeScioli, & Fein, 2012). There is also a puzzling pattern of discoveries about whether time pressure reduces or increases characteristically utilitarian responses (Suter & Hertwig, 2011; Rosas & Aguilar-Pardo, 2020) or neither (Bago & De Neys, 2019). This is a clear invitation to construct minimal models of the ethical.

In conclusion, minimal models and the associated predictions they generate have already proven necessary for explaining resource-limited performance in nonsocial domains: they are likely to be no less essential in mindreading, joint action and ethics.