Defeaters and Bayesianism

1. Bayesian reasoning is a type of defeasible reasoning.
2. One can understand Bayesian reasoning without having the concept of a defeater.
3. So, not all types of defeasible reasoning require the concept of a defeater to understand.

Contextualism: An argument in search of a conclusion

"You know you are a moderately competent art history graduate student. You are visiting a 15th century Italian church Professor Takayama, the world's foremost expert in Fra Angelico. You see a painting that you think is quite definitely not by Fra Angelico. Professor Takayama points to it and says: 'I am quite certain this is by Fra Angelico.' You conclude, on the authority of the great Takayama, that the painting must indeed be by Fra Angelico. Then Professor Takayama turns to you and asks: 'Do you agree?' You respond, quite honestly: 'Yes, quite certainly.'"

Evolution and moral knowledge

Consider this argument for moral scepticism (this formulation is based on a comment by Christian Lee). The existence of our moral beliefs can be given an evolutionary explanation which makes no reference to the truth of these beliefs, so that:
(*) If there were no moral truths, we'd still believe in moral truths.
This, the argument continues, is an undercutting defeater for our moral beliefs even if in fact there are moral truths.

I'm going to argue that at least on four metaethical views, (*) is false. But first note a complication. On many theories of morality, moral truths are necessary truths. But then (*) is a counterfactual with necessarily false antecedent and hence on Lewis semantics trivially true. However, this can't be how the evolutionary moral sceptic understands (*), since then (*) is going to be equally trivially true if one replaces moral claims with mathematical ones, and I take it that the moral sceptic isn't trying to argue for scepticism simpliciter. Rather, in the case of those theories of morality which make morality a necessary truth, the sceptic understands (*) to be a true per impossibile counterfactual. Now on to considering four moral theories.

1. Kantianism. On Kantian ethics, morality is very closely tied to practical rationality. If, per impossibile, there were no moral truths, there would be no practical rationality. If there were no such thing as practical rationality, there would be no agents or even potential agents. It is not clear whether on a Kantian view there could be any beliefs if there were no agents. If not, (*) is a false per impossibile counterfactual. Suppose that non-agents could have beliefs. Well, still, they wouldn't be our beliefs, because we really are agents, and being the sort of entity that has at least a potential for agency is essential to who we are.

2. Natural law. On natural law ethics, morality is grounded in the teleological features of our nature. If those features were other than they are, or if they were absent altogether, the nature would be significantly different, and the beings with that nature would be essentially different from us--they wouldn't be human. Thus, on natural law ethics, were there no moral truths, or even were the moral truths different from what they are, we wouldn't exist, and hence (*) is false.

3. Divine command metaethics of the divine-will variety. On this view, the right thing is what God wills us to (arbitrarily on some versions, or out of the goodness of his nature; the will in question is the antecedent will). Now, presumably God's deliberation as to what kinds of a universe to create was tightly intertwined with what actions, if any, he willed his creatures to do. It is very likely true that if there were no moral truths, i.e., if God had not willed any actions, then God would have created a significantly different universe (perhaps one in which finite agents would not have arisen), and similarly it is likely that if he had willed different moral truths, he would have set up evolutionary processes or intervened (or whatever is the right story about God's cooperation with evolution) to produce different creatures from the ones he had. In particular, it is unlikely that we would exist then, and hence (*) is probably false.

4. Divine nature metaethics. On this family of theories, the right is grounded in God's nature. For instance, the right may be imitation of God, or it may be a certain kind of participation in the Good which might be identified with God. But now consider the counterfactual. Were there no moral truths or were moral truths different, then God's nature would be different from what it is (e.g., it wouldn't be good). But God created us out of the goodness of his nature. Thus, likely, we wouldn't exist were God's nature different in moral respects from what it is. Hence, (*) is probably false, once again.

Note that on theories 2-4, I've also argued that the following variant of (*) is probably false:
(**) If moral truths were significantly different from what they are, our moral beliefs would be no different.