Against the possibility of a formal account of rationality

Abstract

I analyze a recent exchange between Adam Elga and Julian Jonker concerning unsharp (or imprecise) credences and decision-making over them. Elga holds that unsharp credences are necessarily irrational; I agree with Jonker's reply that they can be rational as long as the agent switches to a nonlinear valuation. Through the lens of computational complexity theory, I then argue that even though nonlinear valuations can be rational, they come in general at the price of computational intractability, and that this problematizes their use in defining rationality. I conclude that the meaning of "rationality" may be philosophically vague.

Author's Profile

Shivaram Lingamneni
University of California, Berkeley

Analytics

Added to PP
2013-04-03

Downloads
329 (#46,177)

6 months
40 (#81,628)

Historical graph of downloads since first upload
This graph includes both downloads from PhilArchive and clicks on external links on PhilPapers.
How can I increase my downloads?