A theory I've been working on: axioms are extracted from human nature, are then used to develop via reasoning a spectrum of behaviors that the people then consider to be good or bad in specific situations. The people then vote on which behaviors in specific situations are moral, and the majority rules, and rules are thus formed. These rules are absolute and can be used to deduce more specific rules. The corresponding definition of morality for this system is: "a set of rules of the form "x is right to us" or "y is wrong to us" that can be used to judge concrete behavior and represent what behavior, as theorized via reasoning using axioms extracted from human nature, is considered by the majority of humans to be good or bad for any person under specific circumstances, and from which any more specific rules, if needed, are deduced." Please excuse my sloppiness, I'm not a technical writer.

This pretty much assumes an essentialist view of human nature, otherwise we have to allow chimps to vote.