= 27d567bc2d48d9f40e9ccc20ea370b15)
DanArmak comments on Open Thread: January 2010 - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (725)
"Calculated" based on what? What is the question that this would be the answer to?
Also, how can you define "bias" here?
As you can guess from my questions, I don't even see what an objective system of ethics could possibly mean :-)
This seems to be my biggest problem as well. I have been trying to find definitions of an objective system of ethics, yet all of the definitions seem so dogmatic and contrived. Not to mention varying from time to time depending upon the domain of the ethics (whether they apply to Christians, Muslims, Buddhists, etc.)
That does not answer my question: what is the purely objective, unbiased definition of "ethics"? We can't discuss objective systems of ethics without an objective definition of the question that ethics is supposed to answer.
P.S. my previous comment was malformatted, so you may have missed this part of it; I've fixed it now.
What do you mean by imperative? Humans have certain imperatives, whether evolved or "purely" cultural, but they are all human-specific: other creatures and minds will potentially have different ones. They can't be called "objective and unbiased between all rational thinking minds".
How can we talk about something if we can't define or at least describe it, or point to examples of it existing? Inability to define by definition means there's no concept. A concept isn't right or wrong, it just is, and it's equivalent to a definition that lets us know what we're talking about.
As for "folk concepts" of ethics, no offence intended, but aren't they roughly in the same category as religion and "sexual morals"?
Aren't you just asserting with this statement, without argument, that there is no objective ethics? Isn't it the question at hand whether or not human imperatives are specific or universal?
(Though I wouldn't exclude the higher order possibility that there could be an objective ethical system defined around imperatives in general; for any arbitrary imperatives that a subsystem defines for itself, there is an objective imperative to have them satisfied.)
Well, it's not clear to me that that's what AaronBensen meant by "objective ethics". But I do believe that human ethics are not universal, because:
Human ethics aren't even universal among humans. Plenty of humans live and have lived who would think I should rightly be killed - for not obeying some religious prescription, for instance. On the other hand some humans believe no-one should be killed and no-one has the right to kill anyone else, ever. Many more opinions exist.
I know of no reason why an AI couldn't be built with different ethics from ours, or with no ethics at all. A paperclipper AI could be very intelligent, conscious (whatever that means), but sill - unethical by our lights. If anyone believes that such unethical minds literally cannot exist, the burden of proof is on them.
Careful. We need to distinguish between ethical beliefs and 'factual' beliefs. Someone might have an ethics that says: If there is a God, do what he says. Else, do not murder. This person might want to kill Dan because he believes God wants to heathens to die. Others might have the same ethical system but not believe in God and therefore default to not murdering anyone. I'm not saying there aren't ethical disagreements but eliminating differences is factual knowledge might eliminate many apparent ethical differences.
Also, I'm not sure your second point matters. You can probably program anything. If all evolved, intelligent and social beings had very similar ethics I would consider that good enough to claim universality.
I think plenty of ethical differences remain even if we eliminate all possiblee factual disagreements.
As regards religion, (many) religious people claim that they obey god's commands because they are (ethically) good and right in themselves, and just because they come from god. It's hard to dismiss religion entirely when discussing the ethics adopted by actual people - there's not much data left.
But here's another example: some people advocate the ethics of minimal government and completely unrestrained capitalism. I, on the other hand, believe in state social welfare and support taxing to fund it. Others regard these taxes as robbery. And another: many people in slave-owning countries have thought it ethical to own slaves; I think it is not, and would free slaves by force if I had the opportunity.
I think enough examples can be found to let my point stand. There is little, if any, universal human ethics.
That is underspecified. Evolved how? If I set up evolution in a simulation, or competition with selection between outright AIs, does that count? Can I choose the seeds or do I have to start from primordial soup?
There is a confusion that results when you consider either system (objective or subjective ethics) from the viewpoint of the other.
(The objective ethical system viewpoint of human ethics.) Suppose that there is an objective ethical system defining a set of imperatives. Also, separately, we have subjectively determined human ethics. The subjective human ethics overlapping with the objective imperatives are actual imperatives; the rest are just preferences. It is possible that the objective imperatives are not known to us, in which case, we may or may not be satisfying them and we are not aware of our objective value (good or bad).
(The subjective ethical system viewpoint of human ethics.) In the case of no objective ethical system, imperatives are subjectively collectively determined. We are bad or good -- to whatever extent it is possible to be 'bad' or 'good' -- if we think we are bad or good. This is self-validation.
Now, to address your objections:
Right, human ethics do seem very inconsistent. To me, this is a challenge only to the existence of subjective ethics. In the case of objective ethics, there is no contradiction if humans disagree about what is ethical; humans do not define what is objectively ethical. In the case of a subjective ethical system, inconsistencies in human ethics is evidence that there is no well-defined notion "human ethics", only individual ethics.
Nevertheless, in defense of 'human ethics' for either system, perhaps it is the case that human ethics are actually consistent, in a way that matters, but the terminal values are so higher order we don't easily find them. All the different moral behaviors we see are different manifestations of common values.
Of course, minds could evolve or be constructed with different subjective ethical systems. Again, they may or may not be objectively ethical.
This redefinition of the word "imperative" goes counter to the existing meaning of the word (which would include all 'preferences'), so it's confusing. I suggest you come up with a new term or word-combination.
You defined objective ethics as something every rational thinking being could derive. Shouldn't it also have some meaning? Some reason why they would in fact be interested in deriving it?
If this objective ethics can be derived by everyone, but happens to run counter to almost everyone's subjective ethics, why is it even interesting? Why would we even be talking about it unless we either expected to encounter aliens with subjective ethics similar to it; or we were considering adopting it as our own subjective ethics?
That definitely requires proof. Have you got even a reason for speculating about it, any evidence for it?