This volume provides critical editions of sermons 157-183 of St Augustine, in the now canonical Maurist sequence, plus 11 additional sermons that have been discovered since the Maurists’ edition (1683), and to which Augustinian scholars have assigned numbers that slot them into the sequence (159A, 159B, 162A, 162B, 162C, 163A, 163B, 166A, 167A, 177A, 179A). The editing project is complex, because the Maurists organised the sermons thematically—all those included here are on verses from St Paul or the Catholic epistles—and only some manuscript collections of sermons are so organised. Many of the sermons here are transmitted, albeit not always exclusively, in the ‘De verbis apostoli’ collection (“DVA”), but 17 belong to other traditions.
Boodts is the overall editor, and has herself edited 28 of the individual sermons, in 6 cases reprinting her own earlier editions.1 G. Partoens reprints his recent editions of ss. 163 and 176, and C. Weidmann does likewise for 166A ( = Enarrrationes in psalmos 25.2). F. Dolbeau’s contribution stands somewhat apart: he re-edits, rather than reprints, 3 of the sermons (159A, 159B, 162C) he discovered in Mainz, Stadtbibliothek I 9, and also offers new editions of ss. 160 and 177, for which the Mainz manuscript gives valuable new evidence.2
Boodts provides a general introduction to the DVA collection. Each sermon then has an individual introduction by its editor. These discuss the sermon’s possible date, present the manuscript tradition, and set out the editor’s ratio edendi: how s/he has chosen to use the manuscript evidence. For sermons transmitted within the same collections, this involves much (acknowledged) repetition, whose purpose escapes me. A more serious frustration is the use of manuscript sigla: those employed in the general introduction are abandoned in the separate editions, and the same manuscript may then have different sigla for different sermons. This makes analysis difficult.
As a whole, the volume is a model of diligence and clarity. Where a sermon’s tradition is of manageable size, the editors have collated everything. Where it is too big for that, they have used a generous selection of manuscripts. Florilegia, including unprinted ones, are collated throughout, as are all important earlier printed editions. The editors give sure-footed and succinct guidance through this vast wealth of information: the book could serve as an excellent introduction to the medieval circulation of patristic sermons. The rationes edendi are particularly welcome: we are not left to guess at how the editors have made their decisions. The proofreading—a gargantuan task—is also exemplary.3
Less satisfactory is Boodts’s attempt to make sense of the DVA tradition. This is a very difficult task: there are “several hundred” manuscripts (xxix) and the tradition appears heavily contaminated. Building on the work of P.-P. Verbraken and of Partoens, Boodts starts from a list of 33 manuscripts, all 12th century or earlier. She eliminates 3 for practical reasons,4 and collates the rest for ss. 158 and 183 only. From these collations, and from previous scholarship, she attempts to deduce the relationships among her 30 manuscripts, and thence reduces her list to 13 witnesses, which she has collated for all the DVA sermons she had not previously edited. Some of Boodt’s proofs are convincing, and she is candid about the fragility of others. But she too often argues from readings she calls “errors” when they might be right, and, above all, she makes no attempt to test her conclusions from ss. 158 and 183 on the collations from the other sermons, either those that she is now editing for the first time or those that are re-printed here. Those collations could of course only tell us about the manuscripts that were used in each case, but it was still worth asking whether, for those manuscripts, the collations validated the ss. 158-183 conclusions. Boodts’s approach is rather to use her conclusions for ss. 158-183 to choose among readings in the other sermons. This puts too much weight on the evidence of only 2 sermons, and makes the dangerous (if tempting) assumption that the relationship between manuscripts must be the same for every sermon.5
In constituting her DVA texts, Boodts divides her manuscripts into 3 groups: those with an identifiable hyparchetype, “those which show the fewest signs of contamination” and “the remaining manuscripts”. She then seeks to build a text from agreement between groups 1 and 2, where it is backed by at least one member of group 3. When this won’t work, she prefers group 1, “as this group allows us to be more confident that the chosen reading is a not a product of contamination”.6 The reasoning here is hard to follow: groups 2 and 3 are shadowy, and there is no good cause why contaminated manuscripts cannot have the correct reading. Boodts does, very rightly, also give weight to “internal critique”, but her apparatus contains regrettably few indications of how she is doing this.
Where sermons feature not only in DVA manuscripts, there are other problems. They cannot all be discussed here, but I note in particular the case of s. 180. Boodts shows that there are two branches to the tradition (p and v), and suggests that 3 manuscripts (r1, r2, r3) may constitute a third branch. Yet, rather than balancing the 3 branches against each other, she states she will try to reconstruct only p, “with the rest of the transmission serving to filter out errors” (653). In fact, her apparatus shows that p and v have equal value. Boodts then naturally follows v when it must be right, but p and v readings needed to be given equal weight throughout, and Boodts would also have done well to consider whether the agreement of r1, r2, r3 with v against p might give the right text, as I believe it does at line 340 (see below). We find a similar problem in s. 172: a number of manuscripts are independent from DVA, but Boodts writes that because their “stemmatic positioning is based on limited evidence—though their independence from the De verbis Apostoli archetype has been adequately proved—we have elected to reconstruct the point in the stemma that is clearest to us: the archetype of the De verbis Apostoli tradition” (475). And yet the independent witnesses have “enabled us to correct the De verbis Apostoli archetype” (ibid.). The right procedure was then to reconstruct the archetype of both traditions.
I also do not think it was wise of Boodts (and Partoens) to provide an “exhaustive” (lxxix) apparatus, recording all non-orthographical variants. Where no stemma can be built, it is perhaps justifiable to record everything that could be right. But do we truly need every unique error, including those the scribes have themselves corrected (let alone the misprints of previous editions)? The result is an enormous apparatus, and I wonder who will read it. The procedure is especially puzzling in s. 170, where (we are told) the archetype is extant, but we are still given the readings of 13 other manuscripts, and in s. 176, where Partoens gives the readings of 31 DVA manuscripts, although he wishes to discard the DVA text in favour of the alternative ‘De paenitentia’ tradition. Moreover, for all their detail, the collations are not always accurate.7
Despite these problems, the text adopted by all the editors is generally very good: free from non-Latin, carefully punctuated, and boldly faithful to the loose grammar and hammering repetitions and questions of Augustine’s preaching. Given how few predecessors they have, it is no discredit to the editors that their text cannot always be accepted. I suggest corrections, as follows:8
Better reading in apparatus:9 s. 158, 175: credemus; s. 159, 123: qui; s. 160, 25 aliquo; s. 163, 284: et ex; s. 163B, 14: accept Morin’s addition (see 27-30); s. 164, 302: nunc; s. 165, 122-3: qui autem … quod] quis autem … cui; s. 166A, 202: sed; s. 169, 207: iustitia1: iustitia tua; s. 174, 69 om. et1; 215: qui non; s. 176, 80-1: sed ex … superare] sed quia ex …. superavit; s. 178, 100: ergo] totum; s. 180, 329 sanctum] falsum, 340: clament; s. 182, 80: si non.
Repunctuate: s. 161, 204: est, sed] est sed; s. 162A, 503-4: gentibus.] gentibus?; s. 165.122 habet cui] habet, cui; s. 167A, 6 illum et] illum, et; s. 181, 163: habebat, ut] habebat ut; 178-9: hic, exhibet] hic exhibet.
Conjecture: s. 159, 180-1: timorem … poenarum] timorem doloris timoremque poenarum; s. 159B, 684: quis] is; s. 160, 152 ire] irent; s. 162C, 242: del. et; s. 176, 113: se] me; s. 180, 132 διὰ] μὰ.
Deeper corruption: s. 162, 201-2 (alienum … esse 2): I can neither construe nor make sense of this. s. 172A, 240-5 (denique … praecisum): There seems to be a lacuna before this passage, as the image of the eye is not followed through. The worm image is also hard to grasp, and perhaps needs fixing.
On balance, despite its imperfections, this volume is a very valuable contribution to Augustinian scholarship. Where the editors reprise their earlier work, one welcomes the gathering of their scattered editions in a single volume. Where they re-edit sermons last edited 100-300 years ago, they offer huge improvements on the earlier text, and much precious new evidence. This will be the edition of reference henceforth, and rightly so. But it is to be hoped that its qualities will not lead to another 3-century gap in editorial work on these sermons, but will rather serve as a spur to detailed, sermon by sermon, studies of what Augustine said and how he said it. The sermons are well worth the trouble.
1. s. 168, 169 (with G. Partoens and M. Torfs), 170, 180, 182, 183.
2. Dolbeau writes in French, whereas all the other editors use English.
3. A slip at p. lix: iam3 should read iam2; p. 542: S1 cannot be right. Corrections to apparatus: s. 161, 32 inv. is ambiguous; s. 166, 78: a is the archetype; s. 175.30: which ‘ridere’?; s. 176: ‘164’ should read ‘165’; s. 178, 189: ‘etiam’ is a false lemma, 201: si1 or si2?; s. 180, 348: v reading given twice. The various layers of editing have also caused some slips. Boodts uses v throughout for DVA in her apparatus, but doesn’t tell us what it means till s. 172. d for the ‘De verbis domini’ collection at s. 171 is also unexplained. In s. 169 and 174, we are not told that s of the apparatus is σ1 of the stemma. We are also not told that Dolbeau and Weidmann use the asterisk in their apparatus to mean “this reading could be right”.
4. But Paris Lat. 14292, “no longer consultable due to damage to the binding” (xxxv-vi), has been online Gallica BnF since 2011.
5. For instance, Boodts has Valenciennes 157 and Vat. Lat. 476 as closely related. They do share characteristic errors in s. 157, 159, 161, 178, 181, 183, but not elsewhere.
6. I quote from the rationes edendi , formulated in near-identical terms in the introductions to most of the DVA sermons.
7. As a sample, I have re-collated Vat. Lat. 474 DigiVatLib.474 for s. 159 (errors: 6 hac] ac a.c.; 10 secundum] cundum a.c.; 12 ut om. p. c.; 16 tantum] tamen; 19 nonnulla] nulla; 89 infidelitate] infideli te a.c.; 100: gustate et videte] gusta et vide p.c.; 143 delectatione] delectatatione a.c. ; 155-6 gratius minus] gratus munus (?) a.c.; 203 minantia] minacia p.c.; 225 ait bis a.c.); British Library Add. 17292 for s. 168 (errors: folios should be 22r-24r; 79 ubi vestimenta inv. a.c.; 93 om. et3; 129 et] ut; 153 sic] si a.c.; 160 dignus vocari inv. a.c.; 164 ipsa vacat inv. a.c.; 177 non orat inv. a.c.); Lyon 604 Bibliothèque municipale de Lyon, Ms. 604 for s. 170 (errors: 121 quaere] quare; 193 resurrectionem] resurrectione a.c.; 214 om. sumus; 219 creditis] credetis a.c.; 235 qui2] quia).
8. I hope to justify these suggestions elsewhere.
9. A number of these readings return to the Maurists’ text, which is never to be dismissed lightly.