Put differently, if the amount of classical works that survive today is X, what is the ratio to those that are lost? 100X? 1000X?
By 'lost' I mean that no surviving copies exist and that at most only small fragments of the original text are cited. By 'classical' I'm trying to limit it to Roman and Greek works of antiquity.
I know there are catalogs of works that we know of but don't have, and in that sense the estimation I'm after is simple. BUT surely there are also lost works we know nothing about?
In statistics, it's common to try and estimate the number of unknowns from what we do know (e.g. estimate the number of large unknown sea creatures given how many were discovered so far). Has anyone tried to apply this to the question of lost literary works?
It's difficult to give a reliable estimate. Here are some basic figures to start with, as we have some information about how many plays the Athenian playwrights of the fifth century BCE wrote:
|Plays written|Plays surviving|Survival rate :--|:--|:--|:-- Aischylos|70-90|6 or 7|6.7%-10% Sophokles|113-130|7.5|5.8%-6.6% Euripides|88|18|20.4% Aristophanes|at least 43|11|25.6% or less other (anon.)|?|1 or 2|? total|314-ca. 351|44.5|12.7%-14.2%
Perhaps more importantly, we can estimate the total number of plays that were performed at Attic dramatic festivals throughout the fifth century. At the Dionysia each year were performed 9 tragedies, 3 satyr plays, and 5 comedies, for a total of 17 per year or 1700 over the course of the century; and at the Lenaia starting ca. 440 were performed 4 tragedies and 5 comedies, but only 3 comedies during the Peloponnesian War, for a total of 304 plays in the years 440 to 401. Even if we round everything down, as our info may be faulty, somewhere getting on for 2000 tragedies, satyr plays, and comedies were composed and performed within Athens in the fifth century. Three of the plays in the table above are actually from the fourth century, not the fifth, so the survival rate for fifth century Attic drama is 41.5 out of about 2000, that is to say 2.1%.
And then there's the fourth century, from which we have only four complete surviving plays. And then there's the third century... and so on. And then there's plays that were written outside Athens.
Rather than going into other genres in detail, let me sum up. If we compare the surviving texts to the lost ones that we know of, the ratio is probably in the region of 0.5% to 1%. We can't easily estimate how many lost ones we don't know of; Kallimachos reports that the library of Alexandria contained 532,800 books in his time (mid-third century BCE), but the majority of the lost texts that we know of are later than that time, so that number needs to be raised a good deal. On the other hand, many of those books are going to have been duplicates, so that would lower the figure. Another consideration that would lower the figure is that most scholarly works occupied more than one "book"; but we don't have reliable information on how many scrolls there were to a "book" (we know it wasn't 1 to 1). So we have to settle for a very rough estimate. I really don't think we have enough data to base any robust statistical estimates on.
A couple of afterthoughts.
In this post I've neglected Roman works, as they were dwarfed in quantity by Greek ones. The ratio of surviving works is something like 10 to 1. That ratio would certainly increase, probably by a lot, if we could incorporate lost works.
Don't feel too sad about any loss caused by the fall of the Roman empire (either half). The loss of these works was already well under way in antiquity, when the Romans were at the height of their power. Some of the most important conservation efforts were supported by the Attalids and the Ptolemies, and texts probably began to fade out of existence shortly after the Romans wiped out those dynasties. Another consideration is the transition from the papyrus scroll to the parchment codex, which really got under way starting in the 2nd century CE, and probably caused a great deal of loss as well (more details in this post from a few months ago). And just the other day I wrote a post estimating that the Epic Cycle may well have been entirely lost by ca. 200 CE.
Most of the epic cycle, of which the Iliad and the Odyssey are the most famous members, are lost. There were six books (Cypria, Aetheopis, the Little Iliad, The Sack of Troy, Nostoi and the Telegony) of various lengths (two books to eleven books). They have all been lost, but we have some handy summaries from Proclus Didaiochus. From what I can gather this is not great loss. Even the ancients seem to have thought them a bit dull, making the summaries so they didn't have to read them all. Where the Iliad and Odyssey are about 50% speech the epic cycle was meant to have been only 25% text and was a bit dull.
Other poems in Dactylic Hexameter such as The Oedipodea, The Epigoni, Alcmeonis, the Thebaid and the Titanomachy have not survived either.
Other stuff by Hesiod we have in it's entirety (The Works and Days, Theogony and various other things such as the Shield of Heracles) most of the stuff attributed to Hesiod we only have in fragmentary form however.
Lastly, we have very, very little of Sappho and what we have is fragmentary at best, despite a titanic reputation in antiquity. Most of what we have was used as a ballast on bookbindings and was preserved that way. You can find a full index of her work here but I am sure that it is only a fragment of what was.