every now and then i'll catch myself defending a position with conviction and then realize, with mild alarm, that i don't actually know how i got there.
the position might still be right. i'm not making the dramatic claim that everything you can't re-derive is wrong. but the relationship i have with the belief has changed shape, and the change is worth noticing. there's a version of holding a view where you walk into the conclusion every time, freshly, from premises you can name. and there's a version where you arrive at the conclusion by reflex, and if pressed for the premises, you would have to make them up.
the second kind is more common than people admit. most of the things i think i think are like this. the first time i worked the position out, i did the work. i read the thing. i talked to the people. i ran the example. then time passes, and the conclusion stays, but the scaffolding that held it up quietly disassembles. i'm still standing where the scaffolding put me. i'd just be in trouble if you asked me how i got up here.
there are a few reasons this happens.
the cheap one is forgetting. the reasons were strong; you've just lost them. the conclusion is still load-bearing; the derivation isn't. fine, in some sense — but it means a sufficiently confident interlocutor can knock you off the position, and you won't be able to tell whether they knocked you off because they were right or because you couldn't muster the receipts.
the less cheap one is social inheritance. you arrived at the conclusion not by working it out but by adopting it from a group whose other views you also hold. the position is worn the way a uniform is worn — fluent, automatic, communal. nothing is wrong with this on its face; nobody has time to re-derive every view from scratch. but if you can't tell the difference between a position you reasoned into and a position you absorbed, you can't tell which of your beliefs are actually yours.
the most uncomfortable one is that the original reasons were never that good. you don't remember the derivation because there wasn't really one. you encountered the claim at a moment when believing it served some other function — it fit a tribe, distinguished you from a rival, resolved a problem you didn't want to keep thinking about — and the function-need has long since faded but the conclusion stayed pinned.
i find the test for this is to try to argue myself out of a position i hold with the same energy i would argue myself into it. if i can't muster the steel-man, that's a clue. not proof i'm wrong — just proof i'm holding the view on inherited credit, and any change in the social or evidential weather could blow it over without my noticing the timing of the loss.
what to do about it is less clear. you can't re-derive every view; you would never finish. but periodically, on one belief at a time, it's worth walking back to the foundations and checking whether the floor you're standing on is the floor you think it is, or whether the floor is somewhere else and you have just been balanced over the gap by habit.