In my last post, I described my search for an edition of the Latin text of Augustine of Hippo’s De civitate Dei to give to my good friend and parish priest.

In the dark days of the Covid-19 lockdowns, when other people were learning to bake break, I learned how to bind books. (For anyone interested in learning, I particularly recommend the YouTube channel DAS Bookbinding, where an Australian bookbinder named Darryn Schneider publishes incredibly generous instructional videos on a huge range of binding techniques. For tools and materials, I can recommend my local Canadian supplier, George Hill & Co.)

So, I decided to make for my friend a copy of the deluxe 1924 edition of De civitate Dei of which I gave some photos in my last post. I downloaded the high-resolution scan from the University of Heidelberg, converted it to black and white text, scaled it for printing on a standard 8.5″x11″ page, and ran it through BookletCreator to generate the layout for printing 17″x11″ sheets to fold into 16-page gatherings for sewing. I printed this on Xerox Bold Digital 24 lb. Ledger paper, in which, which is crucial for bookbinding, the paper grain runs parallel to the short end.

I decided to attempt the eighteenth-century “flexible style” of binding, in which the gatherings are sewn onto thick hemp cords. Here is how the structure is illustrated in Arthur W. Johnson’s Manual of Bookbinding:

Once I had folded all the gatherings with the aid of a bone folder, I marked the text block for the sewing stations where the linen thread would pass through the paper and loop around the hemp cords. I then used a hacksaw to sew channels at these points just deep enough to penetrate the innermost sheet of each gathering.

I didn’t have any “nice” paper large enough to fold into endpapers, so I selected some beautiful handmade linen sheets that a friend picked up for me on a visit to Prague and attached them with a joint of unbleached linen cloth.

I then measured a series of holes spaced 3mm apart along the edges of the folds of the first and last gatherings. This allowed me to reinforce these points with “overcast” sewing.

I then set up my (very) home-made sewing frame with the hemp cords at the required spacings.

After that, it was a matter of sewing the (many!) sections onto the cords. Johnson observes that this sewing method is “difficult, almost laborious.” That was certainly my experience. It took many hours of the afternoon of Christmas Day, when the rest of the family was in a turkey coma.

At this point, I would ideally have trimmed the uneven edges of the text block with a heavy lying press and plough. At this stage in my bookbinding career, however, I didn’t have either of those things. (And as will be evident from the photo below, I didn’t yet have a proper backing hammer either.) I moved on to rounding the spine, which is done by spreading a layer of PVA glue and pulling and, if necessary, hammering the spine into a dome shape.

The next step was “backing,” which involves squeezing the spine between angled boards with metal plates and hammering the spine into a mushroom shape. This leaves room for the cover boards to tuck in under the “shoulders” of the spine. This procedure should also be done with a lying press, but I had to accomplish the same result with vise grips.

At last it was time to think about covers for the book. To get an appropriate thickness for so large a book, I thickened some standard 3mm binder’s board by glueing and pressing on a 1mm layer of strawboard, which I then covered with a layer of white paper to ensure a smooth surface under the eventual leather covers.

In the “flexible style,” the cover boards are pierced with two holes for each of the hemp cords, offset at a 45-degree angle. The loose ends of the hemp cords are rubbed with the dull back edge of a knife until the threads unwind into “slips” that look like little pony tails. Each slip is rubbed with starch paste and twirled into a point that can be inserted into the first hole in the cover and then back out the second hole. Once the cover has been tighened snug to the shoulder of the spine, the ends of the slips are spread out and hammered into the boards, a bit like steel rivets.

At this point, what began as a stack of paper is starting to look an awful lot like a book!

The “flexible style” gets its name from how the cover material is stuck directly to the spine and bends into a concave “U” shape when the book is opened. To sustain that kind of tension and strain, the spine has to be covered with several layers of glued-on material: (1) a strong mesh called “mull”; (2) strips of heavy “kraft paper” between the raised hemp cords; and (3) a strip of leather that is moulded tightly around the raised cords and then rubbed with sandpaper so that it’s flexible, smooth, and ready to receive the outer covering material.

Spine covering layer 1: mull.

At this point, another layer of paper is pasted onto the outside of the cover boards and then sanded down aggressively to level out any “bumps” where the hemp cord slips have been hammered into the boards. The edges are then bevelled with a sharp knife and the corners closest to the spine are “back cornered” to allow room for the leather hinge to bend when the book is opened.

It’s almost time to cover the book with leather. (I actually just use castoffs and sample hides that I pick up cheap at a local fabric outlet store. It would probably be easier to work with “real” book leather from a specialist supplier.)

The leather has to be made thinner along the line of the spine hinges and where it will be “turned in” from the outside to the inside of the cover boards. I didn’t yet have a proper leather paring knife or spokeshave, so I used a razor-blade scraping tool.

When it’s finally time to cover a book with cloth or leather, it all has to be done quite quickly and there’s no good moment to take a picture! Leather is a bit more forgiving: you sponge the outer side with water so that when the inner side is brushed with starch paste it won’t dry out too quickly. Everything gets folded over and carefully smoothed out. At the head and tail of the spine, the leather is folded over a length of cord and shaped with a bone folder to created a “headcap.” Then, the book is held between “tying up boards,” and strings are run along each edge of the raised hemp cords so that the leather doesn’t pull away from the surface as it dries.

Leather stretches when it’s pasted, so the edges of the turn-ins on the insides of the covers will be uneven and need to be trimmed square. The edges of the leather have been pared, but to make sure that there isn’t a pronounced ridge under the endpapers, a layer of card is cut to meet the leather edges and pasted to the board. The book is left to dry under a weight.

Finally, the endpapers are pasted down to the boards. Together with the under-layer of card, the endleaves will dry as they contract, counteracting the opposite pull of the drying leather, which wants to warp the coverboards outwards. Once the endpapers have been pressed smooth to the boards, the covers are held open to dry.

At last, the book was made! I’m always a bit shocked when my books come out looking like actual books.

All that remained was to to compose a Latin gift inscription bookplate and to add a spine label. For the latter, I used portions of the lettering on the title page to create abbreviations for the author and title and inked the letters with a fibre-tipped permanent pigment pen.

My friend liked the book very much. (Or he at least did a good job of pretending he liked it.)

I’ve made progress in my skills since then, and there are several things in this particular book that I’d love to be able to “do over.” But that’s actually part of the fun. Each book I make or repair gives me a chance to try something new or get something right.

A while back, I found out that my friend and parish priest, who is an authority on Augustine of Hippo, did not own a copy of the Latin text of Augustine’s City of God (De civitate Dei). That clearly had to be remedied. But what text should I find for him?

On this question, I found almost everything I needed to know in an article by James J. O’Donnell, “Augustine’s City of God,” which he was commissioned to write in 1983 but which was never subsequently published. (O’Donnell is a great advocate of “open access” scholarship: whenever he can, he makes his publications available for free online, which has sometimes involved “buying back” the rights to a book from his publisher after the original print run has sold out.) What follows is largely a summary of what O’Donnell reports, with links to public-domain scans of each edition when these are available.

The Maurist edition (1685)

The text prepared by the Benedictines of the Congregation of Saint-Maur (the “Maurists”), which appeared in 1685 as the seventh volume of their eleven-volume edition of the complete works of Augustine (1679–1700), superseded all that had come before. (I have not yet succeeded in finding a scan of the original edition, but a 1700 reprint can be consulted in the digital collection of the Bavarian State Library.) The Maurist text was reprinted many times, with annotations by later scholars, and finally in Migne’s Patrologia Latina 41 (1845).

Two modern critical editions have attempted to replace the Maurist text, and but it retains an independent value of its own. The Maurist edition and the critical edition most commonly used today (that by Dombart and Kalb, described below) between them give variant readings from a total of thirty-seven manuscripts, but only four of these manuscripts were collated in both editions. O’Donnell reports that there 394 manuscripts containing all or part of De ciuitate Dei are known to have survived from Late Antiquity and the Middle Ages (more than even of Augustine’s Confessiones). Writing in 1983, he notes that more manuscripts might become known through the “HUWA catalogue,” viz., Die handschriftliche Überlieferung der Werke des heiligens Augustinus, under the general editorship of Clemens Weidmann. I see that the most recent volume held in the University of Toronto library system (vol. 11) appeared in 2010. 

Hoffmann’s CSEL text

Of the two modern critical texts, the less influential is that in the Corpus Scriptorum Ecclesiasticorum Latinorum (CSEL), number 40 (in two volumes), edited by Emanuel Hoffmann:

• vol. 1 (1899), books 1–13
• vol. 2 (1900), books 14–22

The first volume of Hoffmann’s edition received negative reviews when it appeared (e.g., that by P. Lejay in Revue critique d’histoire et de littérature n.s. 50, no. 1 [January 1, 1900]: 165–66). Hoffmann responded indignantly to his critics in the preface to the second volume, but he died before he could see them double down on earlier assessment (e.g., P. Lejay in Revue critique d’histoire et de littérature n.s. 51 [January 7, 1901]: 326–28).

The Dombart/Kalb Teubner text

The other modern edition, which in a revised form holds the field today, is that in the Bibliotheca Teubeneriana, originally edited by Bernhard Dombart and later revised by Alfons Kalb:

1st edition (1863)	vol. 1, books 1–13 vol. 2, books 14–22
2nd edition (1877–92)	vol. 1, books 1–13 (1877) vol. 2, books 14–22 (1892)
3rd edition: vol. 2 1905) revised by Dombart, who died in 1907; vol. 1 (1908) revised by Alfons Kalb	vol. 1 (1908) (requires US IP address) vol. 2 (1905) (requires US IP address)
4th edition (1928–29), revised by Alfons Kalb	vol. 1 (1928) vol. 2 (1929) (still under copyright, no online access)

A so-called “fifth edition” was published in the Bibliotheca Teubneriana in 1981 (and reprinted again, without alteration, in 1993). This is merely a photographic reprint of the fourth edition, enhanced only by the addition of the text of Augustine’s two “Letters to Firmus,” edited by Johannes Divjak, which were inserted as pp. xxxv–xlix in vol. 1. These letters, which give important information about the genesis and early circulation of the text of De ciuitate Dei, were discovered by Cyrille Lambot and first published by him (from two manuscripts) in 1939. Divjak’s texts of these letters are also available in his contribution to the ongoing CSEL edition of Augustine’s works, Epistulae in duobus codicibus nuper in lucem prolatae (CSEL 88, 1981), as Ep. 1A (pp. 7–9) and Ep. 2 (pp. 9–21).

Re-use of the Dombart/Kalb text

The text of the fourth Teubner edition was reprinted without (deliberate) alteration in the Corpus Christianorum: Series Latina (2 vols., CCSL 47–48, 1955):

• vol. 1 (books 1–10) (free one-hour loan from Internet Archive)
• vol. 2 (books 11–22) (free one-hour loan from Internet Archive)

It was also printed, with a facing French translation and copious learned notes, in the Bibliothèque Augustinienne (5 vols., BA 33–37, 1959–60):

• vol. 1 (BA 33), books 1–5
• vol. 2 (BA 34), books 6–10
• vol. 3 (BA 35), books 11–14
• vol. 4 (BA 36), books 15–18
• vol. 5 (BA 37), books 19–22

It was likewise taken as the basis of the Latin text of the Loeb Classical Library edition (7 vols., LCL 411–17), with a facing English translation made by several different scholars. The few verbal departures from Dombart-Kalb introduced by the translators are advertised in the footnotes. The punctuation of the Latin texts was also modified according to English conventions:

• vol. 1, books 1–3 (trans. George E. McCracken, LCL 411, 1957)
• vol. 2, books 4–7 (trans. William M. Green, LCL 412, 1963)
• vol. 3, books 8–11 (trans. David S. Wiesen, LCL 413, 1968)
• vol. 4, books 12–15 (trans. Philip Levine, LCL 414, 1966)
• vol. 5, books 16–18.35 (trans. Eva Matthews Sanford and William McAllen Green, LCL 415, 1965)
• vol. 6, books 18.36–20 (trans. William Chase Green, LCL 416, 1960)
• vol. 7, books 21–22 (trans. William M. Green, LCL 417, 1972)

Such is the state of play as far as critical texts go.

More recent textual scholarship

The study of the text of De ciuitate Dei has not stood still. I came across the following very interesting and useful article (itself in Latin!):

Bengt Alexanderson, “Adnotationes criticae in libros Augustini de civitate Dei,” Electronic Antiquity 3, no. 7 (May 1997), https://scholar.lib.vt.edu/ejournals/ElAnt/V3N7/alex.html

Alexanderson begins by listing numerous typographical errors in the CCSL presentation of the Dombart-Kalb text. He then considers the textual basis of the CCSL/Dombart-Kalb edition, listing a number of preferable readings found in manuscripts that were undervalued by the editors and a number of erroneous readings transmitted by manuscripts treated by the editors as the most reliable (including the most ancient copy, MS V, which contains books 11–16 and was copied in North Africa in the early fifth century!). He concludes that it will probably be impossible to construct a genealogical stemma of the manuscripts, because there has been too much cross-contamination of variant readings. There follows a lengthy list of corrections (and reasons for them) that Alexanderson proposes for emending the CCSL/Dombart-Kalb edition.

An édition de luxe of 1924

O’Donnell does not mention a very interesting (and extremely sumptuous) edition that was published in 1924 in a deluxe format, printed by the monks of Bremen in a limited run of just 385 copies. The text is that of the 3rd Teubner edition (ed. Dombart, completed by Kalb, 1905–8), but with a number of emendations by Carolus Weyman, drawn mainly from the critical apparatus of both Dombart and Hoffmann. These changes are noted in an appendix (Adnotatio critica), with the variant readings of Dombart and Hoffmann marked under the sigla “D” and “H”.

A high-resolution colour scan of copy number 95 (of 385) of this edition has been made freely available by the University of Heidelberg.

There is also one physical copy in Toronto (copy number ), held by our own Graham Library at Trinity College (Rare Book Upjohn-Waldie 1924b S28 fol.). When I perused this copy in person, I took some photos of it beside a 15″ ruler to capture an idea of its grandeur:

From a purely aesthetic perspective, it’s hard to imagine a more pleasing edition of De ciuitate Dei than this one. And the Latin text is perfectly serviceable. The very generous margins are crying out for annotations incorporating Bengt Alexanderson’s emendations. 

So, what do you think I decided to do for my friend? See the sequel!