In no. 6, we have an interesting variation on the genre/author organ-
ization: the basic distinction seems to be between prose, arranged by
author (Plato, then Xenophon), and poetry, beginning with Homer.
Plato’s works, though, were apparently subdivided by genre: theSympo-
sium(line 1) is not classed as a dialogue, but separately, and perhaps some
other genre had been specified earlier.^39 Thus the basic and traditional
system of classification here was applied flexibly, in a way that took
account of the contents of the individual collection and, perhaps, the
particular interest of the owner.
Finally, no. 3 (writers of comedy), as we have seen, may be all that
survives of a comprehensive booklist from a library. If so, we have just
part of the library’s holdings in comedy—primarily old comedy at that, so
we must assume that new comedy had its own section—and we may well
have a list, our only such, that actually follows with some care the rule of
organization by genre and authors arranged alphabetically. If this is cor-
rect, and if the library provided coverage of other genres (tragedy, epic,
lyric; history, oratory, etc.) and miscellaneous works (among them works
of reference) at the same level of completeness as it did comedy, then we
might easily posit a collection of several thousands of volumes, impres-
sively large for antiquity.^40
In sum, these lists provide information that allows us to probe several
aspects of book collections in Roman Egypt. In many particulars what we
see is very similar to what we know of contemporary collections in Rome
and Italy. We can see how the collectors conceptualized their collections
by genre, subject, or both, in some cases the subcategories they used, and
whether they were rigorous or casual in compiling the inventories. The
lists we have often imply only the most rudimentary sorting and organ-
ization, and the goal in most cases seems to have been simply to count the
have been left disorganized or even chaotic to our eyes. There are examples of such from the
libraries of monasteries in late antiquity. See, for example, Coquin 1975, a catalogue in
which there are some logical groupings (New Testament texts, say), but the same groups
reappear at more than one place in the list, and unrelated works are mixed in with them. So,
too, Crum 1893, 60 2, no. 44, in which, as Crum noted, the books ‘‘are not arranged
according to their contents.’’ Both of these lists are in Coptic.
- Puglia 1996, 52. The symposium itself may well have been thought of as a genre. Or
perhaps Plato’sSymposium, dominated as it is by long speeches, was considered as rhetoric. - For some comments on the capacities of Roman libraries, and the many uncertainties
involved, see Dix and Houston 2006. Also, van Minnen 1998, 100, provides comparative
numbers from medieval and Renaissance libraries. They are invariably small by modern
standards. Similarly, booklists from Anglo Saxon libraries in England range in size from a few
books to a maximum of 65 (Lapidge 1985). These were codices, and thus the largest libraries
would have contained the equivalent of about 250 to 300 papyrus rolls. What I would
emphasize here is that we have in our lists from Egypt reasonably clear evidence for what
common sense would suggest: that Roman era libraries varied in size along a wide range from
just a few volumes to several thousands, and that collections of even a few dozen volumes
might well be considered, and in fact were, impressive.
246 Institutions and Communities