Please help me understand non-religious metaethics.

[Foucaulf]

Let me try to answer your questions in two ways. (I'm starting with metaethics in general but I'll work my way up)

In one way, metaethics asks: "what is the basis for our moral judgment?" After all, just saying that something's wrong doesn't immediately imply anything, and especially not so if you take away the concepts of sin, salvation, divine goodness and all. In response to your saying something's wrong, nothing stops me from saying that thing's right. Now we disagree. In what way could this disagreement be resolved?

At first, in order to defend their views, the individuals may appeal to some personal anecdote, data or whatever (nothing ethical in itself). Take a step deeper, and the individual may appeal to some value (like the value of life, or of freedom, or of our family). Take another step deeper, and we ask how we could justify why our values matter to someone who doesn't believe in those values.

The biggest divergence from religious metaethics is naturalism - the belief that there are right and wrong values, which best conform with natural laws. There's also anti-realism - that the values we believe are relative, and ethical argument alone cannot truly end disagreement. There's also non-cognitivism/emotivism.

You've mentioned all of this, of course, but all this setup is a way of presenting metaethics without recursion or self-reference. Starting from the simplest moral arguments, we look at facts mentioned in all those arguments or necessarily implied by them to deduce values. We then do the same for the values to deduce metaethical claims. If we agree on those metaethical claims, we are done. If not, we'll have to make more.


Now the second way. Metaethics also attempts to answer "do moral judgments mean anything?" This is the more important phrasing for secular ethics, since the easiest answer - "God says so" - is a no-no. This is also more important practically - today, we don't really get into ethical debate that often. There exist people who will respond to someone saying they're bad people not by "you are wrong", or by "you can't judge me," or "I do what I want," but by "let's not talk about it."

Against responses like these, we may have to discover a property of our world that makes moral judgments mean something, in that those judgments make reference to that property. Setting "God" aside again, we can choose from our emotions (expressivism), the laws of nature, governing our behaviour and those of others (naturalism), some unobservable but perceivable "goodness" (Moore's non-naturalistic moral realism), sense-data (empiricism), human will (Kantianism), and so forth.

Or you can say that a property of the world necessarily implies moral judgments don't mean anything, whether it is the total absence of moral facts (error theory), our inability to perceive knowledge (skepticism) and so forth.

No recursion seems to be needed here, since the process of evaluation is to propose the property and then see what kinds of sentences can be generated as a response to that proposal. This isn't self-referential as it is a possibly ill-formed algorithm.


Lot of this was cribbed from the Wiki/SEP pages, but those seem good. Caveat that I'm not a professional in any way on this.