I think that there is a clear relation to risk preference. Most people prefer gambles where you have 100% chance of winning $5 than a 50% chance of $0 and 50% c

I think that there is a clear relation to risk preference. Most people prefer gambles where you have 100% chance of winning $5 than a 50% chance of $0 and 50% chance of $10, even thought he expected value of each of those gambles is the same. Would you rather a world where you are born guaranteed 5 wealth units, or one where you might be born with nothing or with 10 wealth units with equal chance. But most people tolerate some risk. So I'd expect that behind the veil of ignorance, a lot of people would tolerate some inequality for the chance of the better life. So, for example, instead of 100% chance of $5, they'd take 50% chance of $4 and 50% of $6. People behave this way even in one-shot choices. I know that there are arguments about what a purely rational agent would do (though I am suspicious that the concept of purely rational is ultimately incoherent), but I am not sure how relevant a theory of justice is that relies on what such a purely rational agent would do and what no actual agent would ever do.