Errors. Limits in Provability and Realizability Models for Distributed Rationality and Computation.

Period 01-10-2012 to 30-09-2015
Type Postdoctoral Fellowship
Fellow Dr. Giuseppe Primiero
Funding agency Research Foundation - Flanders (FWO)

Errors are the core of this research and we aim at providing both a philosophical and formal treatment of reasoning processes with errors. To do so, the constructive philosophy of logic represents a good starting point for two main reasons: 1) it strongly relies on a principle of rightness; 2) it identifies conceptually proofs and programs, thereby providing a common ground for the formal treatment of the notion of error in both human and mechanical systems. Firstly, I will give a formal treatment and taxonomy of errors. My main goal is to explain those aspects of rationality characterized by acting in the setting of multiple information sources. I shall tackle the basic issues of interactive reasoning and of revisable knowledge, studying both as processes of error reduction. Secondly, processes of error reduction will be treated as procedures to reach consensus. In this way, I will analyze the conditions in which networks of human and artificial agents encounter and resolve conflicting assignments of tasks. Finally, a full philosophical and formal study will be developed of the specific conditions under which such networks incur in error-states when additional phenomena of distrust, mistrust, doubt and risk are taken into account.