Consensus between Epistemic Agents is Difficult

Abstract

We introduce an epistemic information measure between two data streams, that we term $influence$. Closely related to transfer entropy, the measure must be estimated by epistemic agents with finite memory resources via sampling accessible data streams. We show that even under ideal conditions, epistemic agents using slightly different sampling strategies might not achieve consensus in their conclusions about which data stream is influencing which. As an illustration, we examine a real world data stream where different sampling strategies result in contradictory conclusions, explaining why some politically charged topics might exist due to purely epistemic reasons irrespective of the actual ontology of the world.

5 figures

References in corpus (5)

• Local information transfer as a spatiotemporal filter for complex systems
• Updating Probabilities
• Turing patterns mediated by network topology in homogeneous active systems
• Identification of functional information subgraphs in complex networks
• Efficient sampling of constrained high-dimensional theoretical spaces with machine learning

Cited by in corpus (2)

• Semantic Information in a model of Resource Gathering Agents
• Information-theoretic description of a feedback-control Kuramoto model