The Smaller Libraries of Roman Egypt

George W. Houston

University of North Carolina at Chapel Hill

When they have looked at Roman Egypt, library historians have invariably focussed on the great library at Alexandria, ignoring all others.  Papyrologists, meanwhile, have noticed evidence of several sorts for book collections and possible libraries in Egypt, but they have not tried to assess this evidence as a whole, nor put it within the broader context of library history generally.  In this paper, I will survey the types of evidence that have emerged, consider some of the problems involved in working with this evidence, and present a first attempt at assessing it.

Some seventeen papyri of Roman date contain lists of books (R. Otranto, Antiche liste di libri su papiro, 2000).  It can be difficult to tell what these lists represent:  a list of books a teacher wants his students to read?  A list of desiderata, things to buy?  Books in a bookstall, or in a library?  For our purposes, we need only establish that the list represents a real, existing, collection of books.  If the list contains duplicates, for example, then it is probably not an assignment, nor a list of desiderata.  On the basis of such criteria, at least nine papyri can be taken as lists of real book collections (Otranto nos. 3, 6, 8, and 14-19), or small libraries.  There are, of course, problems.  The papyrus fragments, for example, are very small, and give us only partial lists, not complete catalogues.  Still, they give us some hint of what might be contained in a book collection, and that is precious information indeed.  We will look at one such list (Otranto no. 16 = PSILaur. inv. 19662 verso) as an example of the problems and potential payoff.

On several occasions, dozens or hundreds of literary papyri have been found together, or can be shown to have belonged together.  Grenfell and Hunt made three distinct "large literary finds" in January of 1906, and the contents of the first two of these groups can be worked out in some detail:  we can name sixteen texts in the first collection (W.H. Cockle, Hypsipyle 22, n. 14), and about eighty from the second (Funghi and Savorelli., Tyche 7.1992.77).  Other groups, varying in number from just three books to fourteen, can be established as well.  Each such group can be taken as being some part of a particular book collection.  Needless to say, there are problems here too.  Find spots of individual papyri were not carefully recorded; the owners of individual literary works are almost never known; we cannot tell for sure whether each collection came from a private or a public library.  Despite this, these concentrations of papyri, like the lists discussed above, give us unparalleled insight into the contents and nature of Roman-era book collections.  We will look at one such collection as an example. 

Finally, we have indirect evidence on Egyptian book collections from the use of the word θυρις to describe the cabinets or wall niches in which books were kept (G. Husson, Journal of Juristic Papyri 19.1983.155-62).   Combined with archaeological evidence, this indicates that books were generally stored in cabinets of modest size, not in large stack rooms.

What emerges from all of this?  Some tentative conclusions:  we have specific evidence of some fifteen different book collections in Roman Egypt, a surprisingly large number.  They contained very few books in Latin, and the division of libraries into Greek and Latin sections, so familiar from the City of Rome, clearly did not hold in Egypt.  In the lists, books are regularly divided into genres, lending support to the general assumption that Roman libraries were organized by genre.  With two or three noteworthy exceptions, the collections contain standard authors and few rare works.  That, plus the use of (single) cabinets as containers, suggests that most book collections were small, and when we say "Roman library" we might want to think in the first instance of a collection of from five to several dozen volumes.  Alexandria, it appears, was exceptional beyond all imagining.