Curve-fitting on an individual level can be selfishly driven, as you say, by a desire to become more coherent in order to avoid money-pumps and develop a better “active posture” guide for action / decision theory.
But one reason that people care so much about curve-fitting (when, as you note, those individual selfish motives seem kinda weak) is that curve-fitting is a very important societal project. In order to create societies where many people (with conflicting desires, values, etc) can live together harmoniously/prosperously, we need to have some objective-ish standards (morality, criminal law, norms, etc) that are simple and universal enough that people can actually use them to make decisions and assign status and enforce punishments for violations.
So, there is a middle ground between trying to create the simplest possible curve (like total hedonic utilitarianism) in an attempt to capture realist moral truth with scientific elegance, versus drawing a ridiculously overfitted curve because “whatever”. In practice, the degree of curve-fitting / “overfitting” to do is determined by:
Society is curve-fitting to the “data points” of the moral intuitions of many people in society, not a single individual. Fortunately, humans are sorta similar, so most people can be mostly satisfied by enforcing that everyone follow the “average” of everyone’s different intutions. This prohibits the most severe kinds of overfitting.
Societal rules also have to be somewhat coherent and universal in order to work—everyone might simultaneously believe that “stealing other people’s stuff is fun” and “it sucks when people steal my stuff”, but those desires are in conflict so society will have to come up with some unsatisfying compromise like “stealing is never okay for anyone” or “stealing is always okay for everyone” or “only the king can steal other people’s stuff, and nobody else can steal” or etc, otherwise people will constantly get into fights. This universality / coherence requirement also prohibits extreme, self-serving types of overfitting.
Otherwise, the rules should be overfit as much as possible to try to capture as many details of the average human moral intuition as possible, so that people can live as comfortably and happily as possible with the rules.
But! The rules can’t be too complex, because then people couldn’t remember them or apply them in normal life. This also constrains the amount of overfitting that can be done.
Of course, as an ethical anti-realist, you don’t have to participate in the societal project of trying to create the moral theory which will result in the most coherent, satisfied, prosperous society or whatever! You can just say “screw society’s rules” and decide to be total hedonic utilitarian if you want, because you like scientific-style parsimony! But in practice, human psychology and society has been deeply shaped by this project of creating societally optimal curves for morality / laws / norms. So, you probably were raised in a culture that imbued you with a desire to have a moral system that is some intermediate level of overfitted. So that’s where the feeling that we should quest for a moral law of intermediate overfittedness comes from.
Hopefully this was helpful or at least thought-provoking. I’d be happy to talk more about this if you are interested: - how this intersects with the examples of slavery, factory farming, and other moral circle expansions - how we might be able to overfit harder and thus better satisfy human preferences if we got better at remembering the rules by writing them down in law or applying them automatically or something - how the situation might change if the diversity of values/desires/etc in society went up or down, making the average of everyone’s intuition more or less satisfying to the typical person - analagous situations (under capitalism, we use money as an imperfect curve-fit-style representation of who owes favors to whom—but the requirements for universality and coherency force some weird deviations from the ordinary/intuitive way that favors are reckoned) - why individual people might develop preferences for relatively more or less overfit ethics
Anyways, I really loved your post and it helped clarify some important stuff for me! Thank you for writing it.
Curve-fitting on an individual level can be selfishly driven, as you say, by a desire to become more coherent in order to avoid money-pumps and develop a better “active posture” guide for action / decision theory.
But one reason that people care so much about curve-fitting (when, as you note, those individual selfish motives seem kinda weak) is that curve-fitting is a very important societal project. In order to create societies where many people (with conflicting desires, values, etc) can live together harmoniously/prosperously, we need to have some objective-ish standards (morality, criminal law, norms, etc) that are simple and universal enough that people can actually use them to make decisions and assign status and enforce punishments for violations.
So, there is a middle ground between trying to create the simplest possible curve (like total hedonic utilitarianism) in an attempt to capture realist moral truth with scientific elegance, versus drawing a ridiculously overfitted curve because “whatever”. In practice, the degree of curve-fitting / “overfitting” to do is determined by:
Society is curve-fitting to the “data points” of the moral intuitions of many people in society, not a single individual. Fortunately, humans are sorta similar, so most people can be mostly satisfied by enforcing that everyone follow the “average” of everyone’s different intutions. This prohibits the most severe kinds of overfitting.
Societal rules also have to be somewhat coherent and universal in order to work—everyone might simultaneously believe that “stealing other people’s stuff is fun” and “it sucks when people steal my stuff”, but those desires are in conflict so society will have to come up with some unsatisfying compromise like “stealing is never okay for anyone” or “stealing is always okay for everyone” or “only the king can steal other people’s stuff, and nobody else can steal” or etc, otherwise people will constantly get into fights. This universality / coherence requirement also prohibits extreme, self-serving types of overfitting.
Otherwise, the rules should be overfit as much as possible to try to capture as many details of the average human moral intuition as possible, so that people can live as comfortably and happily as possible with the rules.
But! The rules can’t be too complex, because then people couldn’t remember them or apply them in normal life. This also constrains the amount of overfitting that can be done.
Of course, as an ethical anti-realist, you don’t have to participate in the societal project of trying to create the moral theory which will result in the most coherent, satisfied, prosperous society or whatever! You can just say “screw society’s rules” and decide to be total hedonic utilitarian if you want, because you like scientific-style parsimony! But in practice, human psychology and society has been deeply shaped by this project of creating societally optimal curves for morality / laws / norms. So, you probably were raised in a culture that imbued you with a desire to have a moral system that is some intermediate level of overfitted. So that’s where the feeling that we should quest for a moral law of intermediate overfittedness comes from.
Hopefully this was helpful or at least thought-provoking. I’d be happy to talk more about this if you are interested:
- how this intersects with the examples of slavery, factory farming, and other moral circle expansions
- how we might be able to overfit harder and thus better satisfy human preferences if we got better at remembering the rules by writing them down in law or applying them automatically or something
- how the situation might change if the diversity of values/desires/etc in society went up or down, making the average of everyone’s intuition more or less satisfying to the typical person
- analagous situations (under capitalism, we use money as an imperfect curve-fit-style representation of who owes favors to whom—but the requirements for universality and coherency force some weird deviations from the ordinary/intuitive way that favors are reckoned)
- why individual people might develop preferences for relatively more or less overfit ethics
Anyways, I really loved your post and it helped clarify some important stuff for me! Thank you for writing it.