Robust, efficient, dynamic Theory of Mind

Roth, Marion (2024) Robust, efficient, dynamic Theory of Mind. PhD thesis, University of Glasgow.

Abstract

Theory of Mind is a psychological term that describes the ability to reason about others’ mental states. This includes their thoughts, feelings, beliefs, goals, or intentions of future actions. In reasoning about another’s mental state, an individual can generate a vast number of possible theories about what is on their mind. In a rich and complex social situation, the options could be infinite. In theory, therefore, mindreading is a very sophisticated and energy-consuming ability. Computational simulations of detailed and sophisticated Theory of Mind require lengthy computations and elaborate processes. Yet, humans are able to reason about others’ mental states with ease. Research has suggested that humans are not perfect at handling the contents of another’s mind. Rather, human Theory of Mind shows errors and biases. This work argues for a conceptual approach to Theory of Mind as a responsive, selective, and incomplete cognitive ability. From a perspective of limited cognitive energy, it proposes that mindreading processes are shaped by individual and situational demands and resources.
This thesis is supported by four studies which explore what heuristics may be involved in reducing the cognitive energy an individual spends on Theory of Mind. The first study suggests ad hoc representations as a structural component reducing cognitive costs and identifies various inferential patterns that may narrow down the number of options to be considered in Theory of Mind reasoning. The second study explores stereotypes as a mindreading heuristic and considers the balancing act of navigating the own want and needs, others’ reactions, social belonging, and experiencing certainty. The third study investigates the phenomenon of egocentrism in more detail and the fourth study proposes that flexibility and gaps in mindreading can shape its efficient use.
The four studies are based on quantitative, qualitative, and mixed methods, and deliver insights into the complex and heuristic processes underlying Theory of Mind. Results are discussed with both psychological and computational considerations. Throughout this work, an overarching conceptual framework guides the exploration of different elements, within which various heuristics are identified and discussed. The thesis demonstrates a richness in Theory of Mind strategies that reflects large individual differences and imperfect yet robust abilities to reason about others. This work highlights the value of interdisciplinary work to generate novel perspectives and considerations and facilitate innovative research in both fields. Implications for future research and artificial intelligence applications are discussed.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: T Technology > TA Engineering (General). Civil engineering (General)
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Supervisor's Name: Marsella, Professor Stacy and Barsalou, Professor Lawrence
• Date of Award: 2024
• Depositing User: Theses Team
• Unique ID: glathesis:2024-84820
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 16 Jan 2025 12:32
• Last Modified: 17 Jan 2025 09:25
• Thesis DOI: 10.5525/gla.thesis.84820
• URI: https://theses.gla.ac.uk/id/eprint/84820

Related URLs

• Conference proceeding