Putting in nonsensical extreme values is mot how you evaluate a claim like this. I dont think negative value is in the domain of that optimization function for

Putting in nonsensical extreme values is mot how you evaluate a claim like this. I dont think negative value is in the domain of that optimization function for any rational argumentation. Let's say it costs something to save a life, then the domain would be ]0, infinity[. He's saying that valuing 2 lives that are the same to 2 different values will not be optimal, because if one is valued more than the other, then another identical life could have been saved for the price of the lower valued one, therefore saving 2 lives for less than the inconsistently priced option. EDIT: to prove that negative values are non-sensical, assume you gain money by saving a life, then the only possible optimization is to save every life.