Belief Revision: A Critique

Title: Belief Revision: A Critique

Abstract: This research critiques the methodology used in the literature to study belief change. It argues that a clear understanding of the "ontology" or scenario underlying the belief change process is missing, leading to potential issues that have been overlooked. The paper highlights two key issues: how the agent's epistemic state is modeled and the status of observations. It emphasizes the importance of considering these factors when studying belief revision, especially in the context of iterated belief revision.

Main Research Question: How can we study belief change more effectively by considering the underlying ontology or scenario?

Methodology: The research critiques the existing literature on belief change, focusing on the postulates and representations that have been proposed. It argues that these approaches lack a clear understanding of the underlying ontology, which can lead to unrealistic assumptions about the agent's beliefs and the status of new information.

Results: The research identifies two main issues that have been overlooked in the literature:

1. The modeling of the agent's epistemic state: The research argues that the existing approaches do not provide a consistent way to model the agent's beliefs, either using a set of beliefs or a richer structure like an ordering on worlds. It shows that even postulates that have been considered beyond controversy can be unreasonable when the agent's beliefs include beliefs about her own epistemic state as well as the external world. 2. The status of observations: The research highlights the importance of considering the status of new information. It argues that observations are not always known to be true, but may be believed with varying degrees of certainty. This can lead to issues when considering iterated belief revision, as the agent may need to revise her beliefs multiple times.

Implications: The research emphasizes the importance of considering the underlying ontology or scenario when studying belief change. It argues that this can help to identify potential issues and develop more effective approaches to belief revision. The research also highlights the need for a more consistent modeling of the agent's epistemic state and the status of observations.

Link to Article: https://arxiv.org/abs/0103020v1 Authors: arXiv ID: 0103020v1