One of the oldest and deepest questions across the cognitive sciences concerns the architecture of the mind. Which cognitive capacities are supported by domain-general computations that apply to a broad set of inputs, and which are supported by domain-specific computations, that are specialized for a particular mental function? This question has motivated a broad set of research programs, ranging from the evolution and development of human knowledge (Tomasello, Melis, Tennie, Wyman, & Herrmann, 2012), to the specialization of neural functions (Kanwisher, 2010), to building machines to think and learn in the same ways that we can (Lake, Ullman, Tenenbaum, & Gershman, 2017). 

Work from the last few decades has led to the proposal that in addition to our capacities for domain-general processing, the human mind and brain have ‘domains of cognition’ (Gelman & Noles, 2011): that is, computations that operate on a limited set of inputs, that represent those input in terms of causal and abstract content, and that guide learning and thinking in virtue of specific functional goals. For example, researchers have proposed a range of candidate cognitive domains, including objects, agents, number, space, social relationships, and language (Spelke, 2022; Carey, 2009).

The workshop raises an understudied question: Granting that human minds have some domain-specific systems of cognition, how do these domain-specific systems work together? For example, how do human adults bring together information about minds, actions, spatial relations, and objects to make sense of the sentence “Alice saw three keys on top of the table, and knew that one of them must open the door”? How do humans and other animals represent and make decisions in a spatially extended world with enumerable entities like physical objects and conspecifics? One simple answer to this question is that while domain-specific systems operate in parallel and relatively independently of each other, domain-general cognitive systems (on some accounts ‘central cognition’, Fodor, 1983; on other accounts, natural language; Spelke, 2022) flexibly combine the outputs of each of these systems, delivering representations that integrate information cross domains. Yet, it is also plausible that the functions of domain-specific systems depend on each other in more direct ways, for example to represent an agent as an object with a mind (Liu, Outa, & Akbiyik, 2024), or to represent an agent as the cause of physical objects’ features (Jara-Ettinger & Schachner, 2024). In this example, where does that leave the status of and relationship between these two cognitive domains: agents and objects?

In this workshop, we will bring together researchers who work at the intersection of two or more cognitive domains, and to consider the following questions:

• What are ways that domains of cognition could in principle interact with each other?

• How does evidence for specialization of a cognitive system bear on the question of its independence from other domains?

• Are these domains (objects, agents, number, space, social relationships, and language) the right ones? What’s missing and how could we tell?

• What do alternative hypotheses about mental architecture predict about the development, neural basis, and evolution of the human mind? What sorts of studies on young children, the human brain, and non-human animals would help us to adjudicate between these hypotheses?

• What facilitates the integration of information across domain-specific systems? Do the representations within and/or outputs of domain-specific systems share a common format? How do such representations interface with language?

You can find the full proposal here.