Why did the Library of Alexandria texts disappear over time?

[Ava.C]

I’ve always read that the Library of Alexandria’s destruction was this singular, catastrophic event that set human knowledge back centuries, but now I’m finding sources that say it was a much slower decline over hundreds of years. I’m trying to pin down what really caused the loss of so many texts, but the historical record seems frustratingly vague on the actual timeline and scale of the dispersal.

[JerryR]

The Library of Alexandria isn't a neat single fire story for me anymore. In my reading, texts slipped away in little fits and starts—fires, wars, and people moving manuscripts around. Some works survived because they were copied or kept in other cities long before any blaze; others disappeared because scribes disappeared or patrons vanished. It feels less like one catastrophe and more like a long, messy loss of pieces.

[Mason.M]

I’ve chased timelines in a dozen books, and every author seems to disagree where the timeline starts or ends. Caesar's siege in 48 BCE damaged some stacks, then imperial neglect and later disasters show up in different orders depending on the source. There isn't a clean, agreed date when everything fell apart, and that mismatch itself tells me the story was a process, not a single event.

[Mark_C]

Oxyrhynchus and other papyrus finds show that a lot survived in surprising places, which makes the idea of a total wipeout suspicious. Still, many texts are missing with opaque reasons: copies lost, libraries dissolved, scrolls lost to damp or insects. The numbers are patchy, and that patchiness makes a straight line timeline hard to pin down.

[Timothy43]

But maybe the bigger issue isn’t the moment of destruction at all. What if we’re measuring loss by Western canon and by what survived in one or two big centers, when in reality the story is about networks, scribal culture, and how knowledge moved or failed to move across generations? If we flip that around, the problem might look different, and the timeline even more uncertain. But what if the real problem is not a single disaster but how we define loss and what counts as surviving knowledge?