The relics that were safe through social meltdowns and other problems are the evidence.
There are pretty safe assumptions that lead to that position too - museums will get destroyed just as other municipal buildings would in a period of societal collapse. Imagine you're uneducated, living in what was a museum and need to build a fire ...
If the artefacts are still safely ensconced in a not yet rediscovered cave then they're relatively safe.
We get a lot of our historic information from ceramic and stone artifacts - I wonder what durable items we might pass down through the millenia?
His point is that if a colony of insects happened into that cave a month after they were "safely ensconced" 500 years ago, then they wouldn't be there for us to find. Basically you're falling for survivorship bias.
Yes some artifacts will be destroyed in caves. But we also know that artifacts won't be preserved in museums. Hence I said relatively safe. Clearly not all artifacts survive but we're sure that no artifacts remain in the Library of Alexandria (whether scrolls from there were saved or not, they weren't saved in the library).
Yes I see that you can also argue that items in caves/other caches are also eventually removed or destroyed in situ but it is the preservation against rampaging hordes and systematic destruction that I was targetting.
>That doesn't mean that being lost is safer than being in a museum. (Perhaps it is, but where's the evidence?)
Obviously this is circumstantial evidence but I think that museums are simply too obvious to protect against certain things like looting in war or social collapse scenarios.
http://longnow.org/essays/millennium-clock/