If agents cannot reliably distinguish self-deception from genuine moral reasoning, the category of '...