George W. Houston’s Inside Roman Libraries: Book Collections and Their Management in Antiquity (Chapel Hill: The University of North Carolina Press, 2014) provides some information that may be relevant to the authorship of the Gospels. Scholars often claim the Gospels were originally anonymous and the names assigned to them (Matthew, Mark, Luke, John) were given at a later date. Houston’s book allows us to judge whether it is likely the Gospels were anonymous in light of how libraries operated in the Roman era.
We begin by noting Houston believes small and large libraries operated in similar ways:
I believe that there is considerable continuity between different types of collection, from personal collections of a few dozen rolls, through municipal collections such as the library of Celsus in Ephesus, to the great imperial libraries of Rome, and that all of these collections resemble one another in important respects and can be discussed within a single framework. We will find smaller collections being absorbed by larger ones, the result being that the great collections consist, to some degree, of assemblages of smaller ones and share some or many of their characteristics. Moreover, the work involved in the assembling and maintenance of small- and medium-sized collections, in a world where every manuscript is an individually handwritten product, is similar in nature if not in scale to the effort needed to establish a large one. (p. 4)
We can’t assume that if the early church’s libraries were small then they followed different rules.
Manuscripts often had titles and title tags:
Titles, authors, and whatever other identifying text there might be, such as the number of the book within the work (e.g., Iliad, Book 4), were regularly written at the conclusion of the text, either beside the last column (the usual place) or just below it, and in at least some and perhaps most cases were entered at the beginning of the text as well. These ancient equivalents of our title pages are known as end titles, final titles, or subscriptiones when they appear at the end of the text, and as initial titles, front titles, or inscriptiones when they appear at the beginning. Like modern title pages, initial titles seem generally to have been written in a script that was larger, and sometimes more elegant, than the script of the main text. Wherever they appear, titles that have survived in ancient book rolls differ in at least one crucial respect from modern title pages: they do not include information about the production of the roll. The name of the scribe, the place and date of copying, the person who commissioned the copying: none of this information is included in titles known to us. (p. 8)
In the case of both end titles and initial titles, the author’s name and the title of the work would be hidden once the papyrus was rolled up. In a few known cases, the title and author were written on the verso, that is, on the outside of the roll, so that the volume could be identified without being unrolled. More commonly, it seems, manuscripts were provided with a title tag, or sillybon. This was a strip of papyrus or parchment, roughly 3 by 8 cm, on which were written the author and title. It was glued to the edge of the roll but extended out from the edge so that, when the papyrus was rolled up, the sillybon protruded from one end of it, thus enabling the reader to identify the contents of the roll. It is not known how common sillyba were. In the absence of an external title or sillybon, and if the previous reader had rerolled the volume after use (or had a slave roll it back up), the reader would need to unroll the first twenty or thirty centimeters of the papyrus and identify the text from the initial title or the opening lines. (p. 9-10)
Despite the highly fragmentary state of the manuscripts [of a particular find at Oxyrhynchus], one of them was found with a parchment sillybon still attached, and the subscriptiones of three others survive. The sillybon, on a text of Bacchylides’s Dithyrambs (number 10), was written and glued to the edge of the papyrus a generation or more after the text was copied. It was large, 2.1 by 10.1 cm, and enough of it was left extending beyond the edge of the manuscript so that the content of the book roll would be clear. Its presence raises questions. Were all of the manuscripts provided with sillyba, either as a regular practice or as part of some systematic restoration and preservation of the collection? Or were just a few — perhaps particularly valuable manuscripts, or ones the collector wanted to be able to find easily — given sillyba? At present, we cannot answer these questions. (p. 164-165)
Book boxes of the Roman period could also be cylindrical, and they are attested in this form in both literature and in visual representations. The most helpful passage in literature is Ovid, Tr. 1.1.105-10: Ovid addresses one of his book rolls and says that it will find its mates at Rome, all neatly ordered in their cylindrical boxes (scrinia curva), with their titles clearly visible, presumably on sillyba. This must mean that Ovid thought of the rolls as being placed vertically in the scrinia, with one end — the end to which the sillybon was attached — at the top of the box. This arrangement, which would make retrieval of specific titles easy, is exactly what appears in a number of visual representations, including both paintings and sculptures. (p. 182)
There was another option for the storage of manuscripts: on shelves. In this case, the rolls were stored not vertically, but horizontally, with one of the round ends — the end to which a sillybon was attached — facing out, and on each shelf there might be a stack of book rolls two or three (or possibly more) rolls high. (p. 183)
A library would also have a list of its books:
The point is simply that when the ancients, from at least Hellenistic times on, saw a collection of books, they seem to have felt that it would be useful to draw up a list of the items it contained. (p. 40)
Typically, the lists that survive include very limited information about the manuscripts. They provide authors and titles; the authors’ names, or the titles, or both, may be listed alphabetically, but that is by no means universal. The lists often include the numbers of the books present in the collection (“Illiad, Books 1, 2, 3,” etc.), or indicate the total number of book rolls in a work (“Xenophon, Anabasis, 8 [rolls],” for example). (p. 43)
Papyrus rolls or sheets containing book lists may, then, be added to our list of a library’s standard equipment. (p. 207)
Roman book collections had no catalogs in our sense of the term, that is, a comprehensive list that included author, title, subject, and some kind of identifying catalog or call number for every roll in the collection. The sillyba, subscriptiones, and initial titles provide author, title, and book number, but in surviving examples of all of these there is never any equivalent to a call number. We must assume that those three items of identification (author, title, book number), plus the language (Greek or Latin) and probably the genre of the work, were used to sort and organize the collection and to store and retrieve the manuscripts. (p. 258)
If we imagine a first-century church with some biblical writings (OT and NT), is it probable that the Gospel manuscripts were anonymous? Is it probable that they had no title and therefore no title tag and, perhaps, no entry in the library’s book list? Given human laziness, would most of us not place a title (and, perhaps, a title tag) on each manuscript so we did not have to unroll part of the scroll and read the opening lines to determine the scroll’s contents? I think it is more probable than not that the Gospels had titles from the beginning. In light of how Roman libraries operated, it seems unwarranted to assume the Gospels were originally anonymous. This is all the more the case in light of the fact that all (as far as I know) the earliest Gospel manuscripts we do possess have titles attributing the works to Matthew, Mark, Luke, and John.