Title: "Against Risk-Aversion”
Abstract: It is widely thought that some degree of risk-aversion is rationally and ethically permissible. For example, in a choice between saving one person for sure and taking a ¼ chance of saving five people... many believe that we can be permitted to favor the guaranteed rescue over the risky gamble—even though the gamble saves more expected lives. If you attend this talk, you'll hear my best efforts to argue against this consensus. Specifically, it will be argued that non-maximizing risk-aversion conflicts with a conjunction of two attractive principles. The first principle asserts that, if you're comparing two options, and one of them gives a particular person a greater chance of rescue and affects nothing else, then you are required to favor that option. The second principle asserts that if you are required to choose X when the available options are X, Y, and some other alternatives... then you must still be required to choose X when the available options are only X and Y. Finally, it will be discussed how this argument differs from existing arguments against risk-aversion, which traditionally appeal to the fact that the risk-averse agent is likely to lose out in the long run. The argument we'll explore does not rely upon this observation and can be presented in the context of a single decision.