Corpus Augustinianum Gissense a Cornelio Mayer editum. CD-ROM, version 1.0 for DOS or Windows. Basel: Schwabe & Co. AG, 1995.

• reviewed by J.J. O'Donnell (Augustinian Studies 1997)

"Qui totum te legisse fatetur mentitur" said Isidore of Seville -- the man who says he's read all of you, Augustine, he's a liar. Such a proposition was easy enough to understand in the seventh century, when even a partial collection of Augustine's work s would have taken up a great bulk of animal hides, however trimmed and bound. Even into our own lifetimes, the reader who heard those words could look on the shelves of a good library and find the long row of tall volumes, shake his head, and agree with Isidore with no more than a visceral sense of the truth of the Spaniard's saying.

But now we know better in one way, and see more poorly in another. The surviving works of Augustine amount to 5,025,000 words. And every last one of those words may now be found on a single flimsy CD-ROM, whose outward form is more suited to serve as a coaster to protect a table top from icy glasses than to bear the weight of transmitting a culture. Those words may now be searched rapidly, quoted verbatim, and searched again. The only technological point to keep in mind is that the amount of space and expense involved in storage will drop dramatically over the next years, and the speed with which the information can be manipulated will increase equally dramatically. Augustine does not yet quite dance on the head of a pin, but he will, and before the twenty-first century is very old.

One extraordinary feature of the electronic present is that there are already three different commercially-available collections of the works of Augustine. To assess this particular version properly, it is necessary to say brief words about the others.

First, on the largest scale, there is the Patrologia Latina Database, produced by a commercial publisher who had all of PL typed into computers by inexpensive offshore labor, proofread by a combination of offshore labor and American graduate students, and made available on CD, on tape, and (apparently soon) on the Internet. The drawbacks of this version are twofold: slavish copying of the PL texts, therefore of editions going back in many cases to the sixteenth and seventeenth century; and price. Originally offered at $60,000, recent reports suggest that the "street price" to a research library is approximately $45,000. The Internet version will come instead with an annual fee greater than the price of a copy of the CAG disk under review here. [POSTPRINT note: The publisher of the Patrologia Latina Database now makes it available through research libraries for a substantial annual fee over the Internet.]

Second, the CETEDOC institute at the Catholic University of Louvain has published through Brepols for several years now the CLCLT (CETEDOC Library of Christian Latin Texts) CD product which contains a huge but somewhat idiosyncratic range of good current editions of Latin texts from earliest patristic through latest medieval times. The price of this disk is less than a tenth of the PLD product, but the coverage is much thinner: strong on patristic (complete works of Augustine and several other major figures), weaker on medieval (where the contents began by reflecting the coverage of the Corpus Christianorum Continuatio Medievalis). The underlying fact is that in both paper and electronic editions, the patristic era is much more comprehensively covered by 20th century work, while the further one goes into the middle ages, the more important works one finds where PL really is that latest and best available. CLCLT has grown in its third edition to the point where it needs two CD's, and its price has crept up accordingly; but one attractive feature is that if an institution purchases a copy, the same institution or individuals within it may purchase further copies for half price, thus bringing the product within desktop reach of at least senior scholars.

The Corpus Augustinianum Gissense began life as a collection of electronic texts of the works of Augustine prepared by C.P. Mayer and associates as a preliminary tool for work on the Augustinus-Lexikon. They gathered some e-texts from other sources, notably CETEDOC, and supplemented them with others of their own construction, based in every case on the best current editions. (N.B.: The challenge to any such e-text producer is that new current editions keep appearing. So far, no e-text producer of ancient texts that I know has succeeded in the challenge of continually refreshing corpora with the content of new editions.) As long ago as 1983, this body of texts was available in a computer in Giessen and scholars were invited to send word-search queries. I recall the mid-1980s as a time when a particular table in my study had a pile of thick printouts of such searches, shipped from Giessen: the thought today has the quaint and archaic quality one normally associates with sepia-toned pictures of the earliest automobiles.

Now the CAG texts are made available to the general public for purchase, by the same venerable publishing house that produces the Augustinus-Lexikon. The price (I hesitate to give numbers because they are labile over time and sharply affected by changes in currency conversion rates) is a bit less than the cost of the CLCLT. The contents, moreover, are distinctive.

The CAG disk contains not only the complete Latin works of Augustine, searchable in several ways, but at the same time a huge bibliography of the secondary literature on Augustine, again compiled since 1983 in the course of preparation of the Augustinus-Lexikon. It is this resource and its searchability that makes this a serious competitor for the CLCLT disks. The most serious patristic inquirer will simply have to have both. The student of Augustine who wants access to CLCLT as well may best compromise by acquiring the CAG disk personally and relying on a library copy (if possible) of CLCLT. I will return to the question of access, cost, and the creation of privileged classes of scholars below.

I will describe the search functions and effectiveness now in some detail. For those who weary of the detail, let me say at the outset that the product is excellent in every respect, but has some limitations the purchase or user needs to be aware of.

The product runs under DOS or Windows 3.1 or higher; I have run it quite successfully on Windows NT. The screens are intuitively enough organized that recourse to the manual after installation is very rare indeed. Pull-down menus are in German but still largely intuitive, and even the utterly German-less user would, with the help of a cheap dictionary, become independent of dictionary and manual in a matter of hours at most, more likely minutes. On my system, accented and umlauted vowels from European languages do not display the diacritical marks; I cannot tell whether that is a problem of interface with an American system or some deeper problem in the disk.

The presentation of the texts is transparent and easy to read and follow. Like the CLCLT disks, CAG discourages downloading of full texts. One understands in principle the reluctance to see the product "given away" by illicit copying, but at the same ti me wonders at the gravity of the problem in comparison to me inconvenience. It is possible to capture one paragraph at a time, and one could build up a complete e-copy of even an extensive work if one were very patient (or could afford a very patient student's time), but the nuisance seems unnecessary.

A more serious caution needs to be raised about contents. This is a disk containing the works of Augustine, that is to say, words that he wrote or shared in writing. This means that the corpus does not exactly match the printed editions we are used to using, where, e.g., the letters are published in a corpus that comprises letters to and from Augustine. In this version, letter 88 is present (written in the name of the clergy of the church of Hippo), presumably because Augustine had a hand in writing it, but letter 135 from Marcellinus to Augustine is not present. Both choices are possibly misleading: it was the omission of 135 and others like it that surprised and inconvenienced me, but it is also at least slightly misleading to offer texts with multiple authorship -- what if some word shows up only in such a text? Should we assume that it is a word Augustine wrote himself? I would have preferred a more catholic inclusiveness of texts and then a software system that marked texts more clearly with their authorship when they came up in searches.

A particularly lovely feature of the work (going well beyond CLCLT) is the tagging of biblical quotations in the texts. Passages marked by modern editors as quotations show up in boldface in the texts, and as you move your mouse's arrow over them, the biblical reference pops up in a margin. Given that the tagging is present, it is frustrating that in the present version it is not possible to search (as near as I can figure out) by biblical reference, so one is left to the old e-text searcher's device of hunting biblical passages by looking for two distinctive words in proximity. And it is also necessary to remember that only explicit "quotations" are tagged this way. The allusions and echoes that fill the apparatus scripturisticus of so many of A.'s works are not registered, and so even if it were possible to search the tags, there would be further work necessary to track down even those other allusions and echoes already identified by editors.

The simplest task for a user, beckoning as you boot up the program, is to do a word search. Such searches may seek individual words, stems of words with "wildcards" (hence "pagan*" gets all words beginning pagan-, and a dialogue box lets you see the list of words that will be searched and edit it down -- if, e.g., you really want only paganus and its declined forms, but not paginitas and paganismus), and proximity searches of two words within a user-specified number of words of each other. The searches are rapid and generate a list of paragraphs of Augustine's works within which the search target occurs.

Here it is not quite obvious to the user and necessary to understand that the traditional paragraph of Augustine's printed works is the unit of measure on which all else relies. Works are cited by book and paragraph (but not by chapter, thus e.g. Conf. 1 .12 rather than 1.6.12 -- given the co-existence of "chapter" and "paragraph" numbers in our editions and the non-Augustinian origin of both, I prefer the fuller citation as unambiguous and those who share this preference will have some momentary confusions in using this system), and when a "hit" comes up on a search, you get the whole paragraph with the target word highlighted.

Hence when you search for pagan*, you are told that there are 345 occurrences in A.'s works and given a list of passages. But in fact, that means there are 345 *passages*, i.e., paragraphs, but some of those contain multiple occurrences of the same word. From the list of passages offered it is not possible to tell which paragraphs have multiple occurrences, and so your first impression of a word's use is always understated. A separate "frequency search" will give you a list of A.'s works with a count of the number of times any give words occurs in it, and from that we learn that the string pagan* actually occurs 517 times in all of A.'s works. (The frequency search, just to complete that thought, would be more helpful if it listed the works in which there are no occurrences. Scanning the list of works and the frequency with which "pagan*" occurs in them, it is hard to force oneself to think about which of Augustine's works are missing from that list, but the information is part of what one wants to know.) The context searching for two words within so-and-so many words of each other is very quick, responsive, and is a real boon to the user. (I have had scattered concerns that some proximity searches were not finding every occurrence, but I have not been able to test that concern rigorously.) Similarly, one may concentrate on a single work of Augustine's and search within only that work very easily: another great tool.

The bibliographical search function is powerful and intuitive in most regards, certainly easier to learn than the similar function of the Database of Classical Bibliography that is producing an e-Marouzeau. One may search by the conventional hooks of author, title, and keyword, but there is a non-intuitive feature to the last. By "keyword" this database means a pre-defined keyword coded into the data. Hence the entry for a given scholarly article will end with a list of the official keywords. These may be searched. If a word is not thus approved, then you are thrown back to searching for the word if it occurs in a title and, especially given the multilingual nature of the bibliography, you must remember to curtail the word for searching as much as possible. Hence, if you wish to find what moderns have said about Augustine and abortion, you may search abort* or abortion in titles, and you get only one reference. If you search "abortion" in keywords you get nothing. If you see that the approved keyword is the Latin "abortus" (I have the impression that all keywords are Latin words, but I could be missing some), then you get a longer list of articles. The bibliography is limited to work on Augustine, with some omissions that may be rectified in later versions, omissions at the margin: hence a small surprise at discovering that Peter Brown's Body and Society is not included, for all that it is one of the most widely-read and influential works on Augustine of the last decade.

On overall evaluation, CAG compares very favorably with the other products available. There is obvious appeal to me in the PLD product's completeness, but the age of its texts and the huge price are real negatives for every user. CLCLT contains more text, but it must be admitted that CAG has software that is easier to use, more intuitive to learn, and much faster in response time on the same CD drive. The most significant concern an owner of one of these disks should have is with the life-expectancy of data in is form; the most significant concern a scholar should have is with access.

Life-expectancy: For all that they are miracles of near-instantaneous access today, CD's will indeed be drink coasters within our lifetimes. They are at best a transitional instrument for distribution of information. Within a very few years, what seems near-instantaneous today will seem painfully slow; and we already see with CLCLT's two-disk edition that there is room for improving the amount that can be stored on one disk. But once you have sold physical objects to readers, there are real costs on all sides, to say nothing of nuisance, in upgrading to a later version. That CAG, like CLCLT, comes with its own dedicated software is also an inconvenience. It must run in a window of its own, on a CD of its own, and the reader must adjust mentally to t he new program every time it is loaded -- an adjustment greater than we ever make, e.g., changing from one printed book to another. It is reasonable to expect and hope that these products may find a way to transfer their riches to the Internet in a standard format that allows maximum access and interoperability.

Access: But access is the real problem. Though in principle, electronic information technology allows near universal access nearly instantaneously, the economics of publishing make for artifacts like this which can be made available now only to a privileged few. CAG, to be sure, comes out of European academic environment where Internet access is far less common than in the US, and so to the creators the Internet may not yet seem like a cure-all. But CD-management is a problematic business for libraries to a much greater extent than book management ever was. How to provide enough terminals, how to secure the product from theft, how to assure users access to the documentation they need to figure out yet another brand of software -- these are real problems carrying real additional costs with them. CAG does not restrict access once the product is purchased, but note that PLD, for example, requires purchasing institutions to limit access to their own faculty, students, and staff. The faculty member from a small college who visits a great library now finds that some of its treasures are off-limits to her, not by decision of the library, but by decision of the publishers -- a decision the librarians keenly regret. The "implicit reader" of these products remains for the moment the relatively wealthy and independent private individual -- the senior faculty member with a research fund. The great loss here is not only that Augustine scholars in less privileged institutions are shut out, but that students and scholars of subjects other than Augustine are subtly deterred from making use of a tool which could, in principle, open Augustine to them as never before.

But the publishers will counter that they have real costs in making these texts available, costs they must recover, and for the moment they are right. One could dream of a form of subsidy for such publication that would allow publishers to distribute material more freely, and there are those who do dream and practice an Internet circle-of-gifts culture, producing resources on the cheap and distributing them freely. The present writer has already done a bit of that for Augustine, but is acutely aware of the limitations of such possibilities. Alternately, we await a moment when "cybercash" or the like will allow us to pay very small amounts (a few pennies) per hit for visits to given sites. At the outset, that technology (which will arrive quite soon) w ill create all kinds of new anomalies and new costs as things we had been used to getting for free now start costing money; but in the long run, supply and demand will determine which resources we can afford and which we can't, and subsidy can be reserved for cases where it is truly necessary. The sobering lesson is that technology has gotten well ahead, for the moment, of our economic and social structures. We are left to contemplate, and occasionally use, tools of great power, while knowing that they cannot yet fulfill their real potential.

In sum, then, the Corpus Augustinianum Gissense is a marvel. The serious student of Augustine will want to have it and use it regularly. Put another way, no serious scholarship on Augustine will now be possible that does not take advantage of such tools . We are in for a bumpy time as we accommodate our institutions to their use, but in the long run our studies will be much enriched.