It’s worth pausing to consider why. A mathematician looks at scientists as having a lower standard of evidence. Mathematicians prove their claims. In math, but really only in math, your earlier claim about reason being reliable when used correctly is literally true.
The difference here is that in math, one works through a rigorously logical process from postulates that are defined and accepted by all participants. Under such circumstances, the conclusions aren’t just reliable – they are
proven in a sense much more absolute than we can prove things when we talk about the real world.
If you start from given premises – and you apply logic correctly – you will absolutely get a logical conclusion.
In the real world, and in evaluating claims about the real world – which includes claims about whether beings or rules exist outside of human minds or not – the process of evidence-based inquiry can’t begin from postulates or definitions because (since we’re talking about the real world), one has to first confirm that one is beginning from premises that are true, that one’s ideas actually reflect reality. In order to confirm whether one’s premises are true, one must start from observations, from evidence.
One then tries to determine which claim is best supported by the evidence and whether the evidence is sufficient to support that claim, but
obviously this process can’t “prove” anything absolutely because – as you noted – we could always stumble across new evidence that changes everything.
So the conclusions of evidence-based inquiry are always tentative, always “the best possible answer we have now based on evidence.”
Now, the more evidence we have, the more likely a claim is to be true. Claims that are insufficiently supported by evidence can be dismissed until there is sufficient evidence for them.
Why do we insist on repeatable experiments in physics and chemistry but only historical analysis in economics and sociology? Why are we willing to rely on people’s reports in history?
Because – obviously – you can’t repeatably test something that already happened and only happened once.
For this reason, our conclusions about history are based on the best possible answers that we can come up with based on evidence…and this, by the way, doesn’t always mean blindly trusting historical reports. There are all kinds of things in “reports” of history that we don’t take at face value…for example, the history of the Roman emperors and generals contain all sorts of stories of magic and divination and other things widely thought today to not have happened. Suetonius’ account of Caesar crossing the Rubicon on the advice of a gigantic spirit is one example: it’s generally accepted that Caesar crossed the Rubicon – mostly because tons and tons of other evidence suggests that it happened – but it’s generally not accepted that spooky giant spirits talked him into it, for the reason that it’s a fantastical, supernatural account that comes from only one guy (and perhaps has its source in contemporary folklore).
But luckily for us, our discussion about morality isn’t a historical inquiry: we’re looking at whether or not we can justify the existence of a normative framework and set of rules outside of human minds. This is a claim about the world as it exists right now, for which we can gather evidence right now, so whatever the practices of historians are, they need not directly concern us for the purposes of this question.
So I still don’t know what problem you’re having exactly. Different subjects have different kinds of evidence, obviously. We can’t absolutely, positively, 100% prove the things we say about reality, which is self-evidently the case and irrelevant because we don’t need to be absolutely certain about our claims for those claims to be useful.
So what’s the problem, exactly?