It’s a sticky business, and different ethicists will frame the words different ways. On one view, objective includes “It’s true even if you disagree” and subjective includes “You can make up whatever you want”. On another, objective includes “It’s the same for everybody” and subjective includes “It’s different for different people”. The first distinction better matches the usual meaning of ‘objective’, and the second distinction better matches the usual meaning of ‘subjective’, so I think the terms were just poorly-chosen as different sides of a distinction.
Because of this, my intuition these days is to say that ethics is both subjective and objective, or “subjectively objective” as Eliezer has said about probability. Though I’d like it if we switched to using “subject-sensitive” rather than “subjective”, as is now commonly used in Epistemology.
So, this isn’t the first time I’ve seen this distinction made here, and I have to admit I don’t get it.
Suppose I’m studying ballistics in a vacuum, and I’m trying to come up with some rules that describe how projectiles travel, and I discover that the trajectory of a projectile depends on its mass.
I suppose I could conclude that ballistics is “subjectively objective” or “subject-sensitive,” since after all the trajectory is different for different projectiles. But this is not at all a normal way of speaking or thinking about ballistics. What we normally say is that ballistics is “objective” and it just so happens that the proper formulation of objective ballistics takes projectile mass as a parameter. Trajectory is, in part, a function of mass.
When we say that ethics is “subject-sensitive”—that is, that what I ought to do depends on various properties of me—are we saying it’s different from the ballistics example? Or is this just a way of saying that we haven’t yet worked out how to parametrize our ethics to take into account differences among individuals?
Similarly, while we acknowledge that the same projectile will follow a different trajectory in different environments, and that different projectiles of the same mass will follow different trajectories in different environments, we nevertheless say that ballistics is “universal”, because the equations that predict a trajectory can take additional properties of the environment and the projectile as parameters. Trajectory is, in part, a function of environment.
When we say that ethics is not universal, are we saying it’s different from the ballistics example? Or is this just a way of saying that we haven’t yet worked out how to parametrize our ethics to take into account differences among environments?
I think it’s an artifact of how we think about ethics. It doesn’t FEEL like a bullet should fly the same exact way as an arrow or as a rock, but when you feel your moral intuitions they seem like they should obviously apply to everyone. Maybe because we learn about throwing things and motion through infinitely iterated trial and error, but we learn about morality from simple commands from our parents/teachers, we think about them in different ways.
So, I’m not quite sure I understood you, but you seem to be explaining how someone might come to believe that ethics are universal/objective in the sense of right action not depending on the actor or the situation at all, even at relatively low levels of specification like “eat more vegetables” or whatever.
Did I get that right?
If so… sure, I can see where someone whose moral intuitions primarily derive from obeying the commands of others might end up with ethics that work like that.
“the proper formulation of objective ballistics takes projectile mass as a parameter”
I think the best analogy here is to say something like, the proper formulation of decision theory takes terminal values as a parameter. Decision theory defines a “universal” optimum (that is, universal “for all minds”… presumably anyway), but each person is individually running a decision theory process as a function of their own terminal values—there is no “universal” terminal value, for example if I could build an AI then I could theoretically put in any utility function I wanted. Ethics is “universal” in the sense of optimal decision theory, but “person dependent” in the sense of plugging in one’s own particular terminal values—but terminal values and ethics are not necessarily “mind-dependent”, as explained here.
I would certainly agree that there is no terminal value shared by all minds (come to that, I’m not convinced there are any terminal values shared by all of any given mind).
Also, I would agree that when figuring out how I should best apply a value-neutral decision theory to my environment I have to “plug in” some subset of information about my own values and about my environment.
I would also say that a sufficiently powerful value-neutral decision theory instructs me on how to optimize any environment towards any value, given sufficiently comprehensive data about the environment and the value. Which seems like another way of saying that decision theory is objective and universal, in the same sense that ballistics is.
How that relates to statements about ethics being universal,objective, person-dependent, and/or mind-dependent is not clear to me, though, even after following your link.
It’s a sticky business, and different ethicists will frame the words different ways. On one view, objective includes “It’s true even if you disagree” and subjective includes “You can make up whatever you want”. On another, objective includes “It’s the same for everybody” and subjective includes “It’s different for different people”. The first distinction better matches the usual meaning of ‘objective’, and the second distinction better matches the usual meaning of ‘subjective’, so I think the terms were just poorly-chosen as different sides of a distinction.
Because of this, my intuition these days is to say that ethics is both subjective and objective, or “subjectively objective” as Eliezer has said about probability. Though I’d like it if we switched to using “subject-sensitive” rather than “subjective”, as is now commonly used in Epistemology.
So, this isn’t the first time I’ve seen this distinction made here, and I have to admit I don’t get it.
Suppose I’m studying ballistics in a vacuum, and I’m trying to come up with some rules that describe how projectiles travel, and I discover that the trajectory of a projectile depends on its mass.
I suppose I could conclude that ballistics is “subjectively objective” or “subject-sensitive,” since after all the trajectory is different for different projectiles. But this is not at all a normal way of speaking or thinking about ballistics. What we normally say is that ballistics is “objective” and it just so happens that the proper formulation of objective ballistics takes projectile mass as a parameter. Trajectory is, in part, a function of mass.
When we say that ethics is “subject-sensitive”—that is, that what I ought to do depends on various properties of me—are we saying it’s different from the ballistics example? Or is this just a way of saying that we haven’t yet worked out how to parametrize our ethics to take into account differences among individuals?
Similarly, while we acknowledge that the same projectile will follow a different trajectory in different environments, and that different projectiles of the same mass will follow different trajectories in different environments, we nevertheless say that ballistics is “universal”, because the equations that predict a trajectory can take additional properties of the environment and the projectile as parameters. Trajectory is, in part, a function of environment.
When we say that ethics is not universal, are we saying it’s different from the ballistics example? Or is this just a way of saying that we haven’t yet worked out how to parametrize our ethics to take into account differences among environments?
I think it’s an artifact of how we think about ethics. It doesn’t FEEL like a bullet should fly the same exact way as an arrow or as a rock, but when you feel your moral intuitions they seem like they should obviously apply to everyone. Maybe because we learn about throwing things and motion through infinitely iterated trial and error, but we learn about morality from simple commands from our parents/teachers, we think about them in different ways.
So, I’m not quite sure I understood you, but you seem to be explaining how someone might come to believe that ethics are universal/objective in the sense of right action not depending on the actor or the situation at all, even at relatively low levels of specification like “eat more vegetables” or whatever.
Did I get that right?
If so… sure, I can see where someone whose moral intuitions primarily derive from obeying the commands of others might end up with ethics that work like that.
“the proper formulation of objective ballistics takes projectile mass as a parameter”
I think the best analogy here is to say something like, the proper formulation of decision theory takes terminal values as a parameter. Decision theory defines a “universal” optimum (that is, universal “for all minds”… presumably anyway), but each person is individually running a decision theory process as a function of their own terminal values—there is no “universal” terminal value, for example if I could build an AI then I could theoretically put in any utility function I wanted. Ethics is “universal” in the sense of optimal decision theory, but “person dependent” in the sense of plugging in one’s own particular terminal values—but terminal values and ethics are not necessarily “mind-dependent”, as explained here.
I would certainly agree that there is no terminal value shared by all minds (come to that, I’m not convinced there are any terminal values shared by all of any given mind).
Also, I would agree that when figuring out how I should best apply a value-neutral decision theory to my environment I have to “plug in” some subset of information about my own values and about my environment.
I would also say that a sufficiently powerful value-neutral decision theory instructs me on how to optimize any environment towards any value, given sufficiently comprehensive data about the environment and the value. Which seems like another way of saying that decision theory is objective and universal, in the same sense that ballistics is.
How that relates to statements about ethics being universal,objective, person-dependent, and/or mind-dependent is not clear to me, though, even after following your link.