How much medieval (and ancient) literature is lost?

STATISTICAL RECOVERY OF GHOST MANUSCRIPTS: Study finds 90 percent of medieval chivalric and heroic manuscripts have been lost. Researchers used ecological "unseen species" model to estimate size of medieval European lit (Jennifer Ouellette, Ars Technica).

But can an ecological model really be applied so readily to such a markedly different scholarly domain? "Intuitively, it's a weird thing to say that a literary work behaves like a species," Kestemont acknowledged. "In fact, this method isn't even specific to ecology." Chao1 is so general that it's been used in lots of other fields, with "species" standing in for classes of stone tools (archaeology); types of die for ancient coins (numismatics); different causes for a given disease (epidemiology); genes or alleles (genetics); and distinct vocabulary words (linguistics), to name a few.

"It's even been used to estimate the number of stars in a galaxy or the number of bugs in a piece of software that haven't been discovered yet," said Kestemont. "The real question is 'in which conditions wouldn't we be able to apply it?'" As long as you can distinguish something akin to "species," the model works very well.

In adapting the model, Kestemont and his co-authors treated literary works as species and manuscript copies as sightings of a species. They counted works that only appeared sparsely in the historical record and then used those counts to calculate F0—in this case, the number of works that once existed but scholars have never observed. A work was considered "lost" when none of the documents that once preserved it still survived.

I would say that is a fair standard for "lost."

Seriously, this looks like a useful statistical method. I don't know how well it would work for manuscripts of ancient literature. The sample is smaller and the chronological gap between many of our surviving manuscripts and the originals is much wider. But it might be worth trying.

Intuitively, I would guess that the survival rate for ancient (i.e., let's say, pre-Islamic) literary manuscripts from the Middle East and the Mediterranean regions would be on the order of 1% or less, and shrinking the farther back you go.

By "manuscripts" I mean works written on papyrus and parchment. Mesopotamian literature inscribed in cuneiform on clay tablets is a different problem.

The underlying article in Science is behind a subscription wall. Cross-file under Lost Books.

Visit PaleoJudaica daily for the latest news on ancient Judaism and the biblical world.