This post is an attempt to answer the question: how do you define morality objectively?

Context: I used to be a moral nihilist (there is no such thing as “good” and “bad”, morality is completely arbitrary and subjective, the universe does not care). Today, I’m closer to being a moral realist, though I reserve some room for the possibility that morality is entirely a human construct. At the very least, I think arguments for objective morality should be considered seriously.

So, how do we define morality objectively?

Well, first I’d ask: how do we define anything objectively? Many people accept realism in science. They believe that science studies objective truths that are independent of the whims and convictions of any individual person. The way that science achieves this is via inter-subjective testing – we only consider a claim scientific if it can be reproduced and tested by (in principle) anyone.

Note, however, that even as we aim for objectivity via inter-subjectively testable claims, we never actually reach objectivity. We are always constrained by our fallibility and our individual subjective view into reality. We have no idea what reality is actually like. But nonetheless we consider science an objective process of knowledge-creation.1

The same could be said about morality. Morality is the domain of answers to the question: what should we do next? (Contrast this with science, which answers: how does the universe behave?). We pose answers to these questions in the form of ‘should’ statements, e.g. The US government should allocate its budget to such and such programs, or, You should not cause needless suffering upon others.

The problem is, while you can inter-subjectively test statements like “the earth orbits around the sun”, you can’t inter-subjectively test a statement like “the US government should invest more in infrastructure”. So what can you do with such a statement? How can you objectively evaluate it against another statement like “the US government should invest more in space travel”?

The key is to realize that experimental testing is not the be-all-end-all of knowledge-creation. Rather, testing is really just a special case of the more general activity of rational criticism. The reason why we care about experimental tests in science is to rule out the bad theories in favor of better ones. But there are many other ways to rule out bad theories, e.g. by finding logical inconsistencies in them or pointing out their lack of explanatory power. If I make a claim that my desk weighs 5•10²⁶ kg, you can rule out my claim before experimentally testing it (e.g. you can point out that the purported mass is a hundred times greater than the mass of the earth, which would violate our understanding of the dynamics of the earth’s motion).

Likewise, while we can’t run experiments to evaluate moral claims, we can criticize and discuss them nonetheless. We can say that because pain is inherently undesirable, one ought not to inflict pain indiscriminately on random people. If you have a better argument in defense of the view that we should randomly inflict pain on people, you can make that argument and we can evaluate it against my argument. This is not as distinct as people think from the rational evaluation (via testing) that takes place in science.2

So, the short version of the answer is: we define morality objectively in the same way that we define anything else objectively. We make inter-subjectively criticizable claims and then we critique them and discard the claims that don’t stand up to criticism. The problem is that in morality, because we are dealing with ‘should’ statements, experimental tests are not a useful method of criticism; but that does not mean we can’t make progress in our understanding of what we ought to do. Just look at all the things we used to consider acceptable that we consider morally abhorrent today. Is the difference there really just arbitrarily shifting cultural standards, or are we making objective progress?3

Ultimately, it may still be the case that statements like “you shouldn’t bully the powerless” are as much a fiction as statements like “Harry Potter is a student at Hogwarts”. But if you believe there is any domain in which objective claims can be made, you have to make a very specific argument for why the domain of ‘should’ questions is not one of them.

  1. This piece assumes the conjecture-and-criticism view of science (Karl Popper), i.e. that science is not a process of logically deriving truths about reality, but rather is a process of correcting the errors in creative, unjustified conjectures. ↩︎

  2. Also, just as we can never be sure of any of our claims in science, we can never be sure of any of our claims about morality. The same fallibility and subjectivity that hinders scientific knowledge-seeking also hinders moral knowledge-seeking: we never actually reach objective truths. ↩︎

  3. To reiterate, the position I’m (tenuously) defending—moral realism—is just the view that there are objectively better and worse answers to questions about what we should do next. In particular, I am not asserting the existence of “absolute moral axioms”, nor the existence of some cosmic judgement toward our actions, nor the existence of moral value inside physical reality. I’m only criticizing the view that that ‘should’ claims are all equally valid (and thus equally meaningless). ↩︎