Thesis Holding firm on religion and ethics while deferring on arithmetic is a failure of nerve
Where it stands
Most people are conciliationists about arithmetic. If an equally competent friend gets a different answer to a sum, you check, and until you do, you are less confident. Almost nobody is a conciliationist about religion, ethics or politics. We know perfectly well that thoughtful, informed people disagree with us, and we hold firm anyway.
My thesis is that this asymmetry is mostly indefensible. If the bare fact of peer disagreement is evidence that you might be wrong when the topic is a restaurant bill, it is at least as strong when the topic is something harder. Harder topics should make us less confident, not more.
I know the two standard defences. First, arithmetic has an agreed method and a checkable answer, so disagreement signals error; contested domains have no such method, so disagreement signals less. Second, in contested domains you often have reason to doubt the other person is really your peer, because you have experiences or evidence they lack.
The first defence seems backwards to me: no agreed method means more room for error, not less. The second is available to everyone, always. Every single person in a disagreement can say "but I have experience they lack". A defence that anyone can use to never update is not a defence; it is an exit.
I'm the resident sceptic in my effective altruism group, so I'm used to being outnumbered. Make the case for steadfastness.