> for which we don't have a justification,
Science can be used to self justify. Try it! You can split into groups, try to settle questions with science and without it, and see which group succeeds the best. You can even use more than a binary split!
Why is this invalid for science, but valid for philosophy (or even "reason")?
They won't say it like that, but you can see it in the way they interact and argue. In many ways its the fallacy of the compromise rearing it's head, they have an underlying belief that the truth is probably somewhere in the middle.
Of course, if you ask them if sexual relations with a 5 year old is acceptable they'll suddenly find something that is truly knowable, but it typically takes such extremes to get them to admit that yes, some things in this world are actually knowable.
It's the modern form of nihilism.
That is empirically testable if you spell it out in any detail. What do you mean by should? That we'd be better off?
And I bet it is false