A computational model of the cultural co-evolution of language and mindreading

Abstract

Several evolutionary accounts of human social cognition posit that language has co-evolved with the sophisticated mindreading abilities of modern humans. It has also been argued that these mindreading abilities are the product of cultural, rather than biological, evolution. Taken together, these claims suggest that the evolution of language has played an important role in the cultural evolution of human social cognition. Here we present a new computational model which formalises the assumptions that underlie this hypothesis, in order to explore how language and mindreading interact through cultural evolution. This model treats communicative behaviour as an interplay between the context in which communication occurs, an agent’s individual perspective on the world, and the agent’s lexicon. However, each agent’s perspective and lexicon are private mental representations, not directly observable to other agents. Learners are therefore confronted with the task of jointly inferring the lexicon and perspective of their cultural parent, based on their utterances in context. Simulation results show that given these assumptions, an informative lexicon evolves not just under a pressure to be successful at communicating, but also under a pressure for accurate perspective inference. When such a lexicon evolves, agents become better at inferring others’ perspectives; not because their innate ability to learn about perspectives changes, but because sharing a language (of the right type) with others helps them to do so.

Theory before the test: How to build high-verisimilitude explanatory theories in psychological science

“…But on the other hand they may involve additional constraints that are specific to the real-world physical implementation of the computational process under study. For instance, a learning algorithm running on the brain's wetware needs to meet physical implementation constraints specific to neuronal processes (e.g., Lillicrap, Santoro, Marris, Akerman & Hinton, 2020;Martin, 2020); evolutionary algorithms realized by Darwinian natural selection are constrained to involve biological information that can be passed on genetically across generations (Barton & Partridge, 2000); and, cultural evolution is constrained to involve the social transmission of information across generations and through bottlenecks (Henrich & Boyd, 2001;Mesoudi, 2016;Kirby, 2001;Woensdregt, Cummins & Smith, 2020). Hence, brain processes and biological and cultural evolution are all amenable to computational analyses but may have their own characteristic physical realisation constraints.…”

Human Self‐Domestication and the Evolution of Pragmatics

“…In turn, consistent with the postulated feedback loop, we expect that the emergence of early forms of language/grammar contributed to further honing and sophistication of these cognitive devices. Consistent with this perspective are studies on the relationship between language and ToM (Heyes & Frith, 2014; Moore, 2017; Woensdregt et al., 2020) as well as studies considering how the emergence of elaborated ways of expressing time through grammatical devices (e.g., verbal tenses) contributed to augment our mental travel abilities. Even if our hypothesis is that MTT is a structural precondition of language (Ferretti, 2016; Ferretti & Adornetti, 2020), we argue that the origin of language activated a coevolutive process with effects on MTT structure and functioning.…”

“…At the same time, the gradual complexification of grammars certainly provided increasingly better expressive tools for further sophistication of pragmatic principles, contributing to the mutually reinforcing feedback loop. For example, it has been suggested that important abilities for the use of language in conversational contexts, such as the ToM and perspective‐taking (more on this in Sections 8 and 9), are enhanced by the acquisition of grammatically sophisticated linguistic structures (e.g., De Villiers, 2005; Heyes & Frith, 2014; Milligan, Astington, & Dack, 2007; Moore, 2020; Rakhlin & Progovac, 2020; Woensdregt, Cummins, & Smith, 2020).…”

Coordinating social action: a primer for the cross-species investigation of communicative repair

“…We can illustrate this methodological choice using the topic of intentionality. One might argue that flexible and complex social interaction—of the type exhibited in interactive repair—requires capacities for intention attribution or mindreading that are possibly unique to humans [54,55]. This might lead us to not expect interactive repair in animals apparently lacking such capacities (but see [56]), and indeed, to date, there appear to be no reports of interactive repair in communication systems other than human language.…”