From Fulda?

Accounts of the origins of these two poems rest upon a mistake. No part of Einsiedeln 266 was ever at Fulda. This mistake arose fifteen years after Hagen discovered the text: Ernst Dümmler was editing Hrabanus Maurus’ poems for the Monumenta Germaniae Historica.Footnote ⁹ For some of the collection, the only two sources he could find were Christoph Brouwer's Reference Brouwer1617 edition and the Einsiedeln manuscript. What manuscript Brouwer used was unknown, since he explicitly declined to reveal its owner in the prologue to his edition.Footnote ¹⁰ Dümmler opted for the simplest solution: he identified the manuscript used by Brouwer as a less mutilated version of Einsiedeln 266, then he triangulated Brouwer's references to his manuscript with other mentions of a Codex Fuldensis and assumed that the two were the same.Footnote ¹¹ Hence, he concluded, Brouwer used the Einsiedeln manuscript, which had been from the time of its writing to Brouwer's own day at Fulda.

It may be true that Brouwer used a Fulda manuscript in his edition, but this manuscript was not at Fulda during Brouwer's lifetime. The whole miscellany was bound in the fourteenth century at Einsiedeln and we have the hand of the fourteenth-century librarian of the abbey, Heinrich von Ligerz, in the volume.Footnote ¹² Hence, the whole codex, including the three quires that interest us, now pp. 177–224, was at Einsiedeln in the fourteenth century, not at Fulda, and has remained there to this day.Footnote ¹³

Whatever manuscript Brouwer did use, it was not this manuscript. Brouwer explicitly states that his manuscript contained the heading: Versus Hrabani de diversis XIV. Ord. XIX. Footnote ¹⁴ If such a heading had ever been in the Einsiedeln manuscript — there is currently no such inscription — it would have been in the trimmed portion of the top of p. 208. But the folio was trimmed when the manuscript was bound, two centuries before Brouwer, and so there is no way that he can be describing the Einsiedeln manuscript. Further, on p. 212, the words we have from Brouwer's edition (p. 8), ‘sic catus et cautus attendis’, have been lost in that same trimming — there is no way Brouwer could have read these words once the manuscript was bound. Finally, Brouwer tells us explicitly that there is a loss of material of about four folios from the middle of carm. XXVIII (XXXIV Dümmler), after the line ending ‘rite canendo’ (l. 29).Footnote ¹⁵ Yet the Einsiedeln manuscript, though damaged and hard to read, continues for another twenty-five lines, as Dümmler himself prints (p. 193). Thus, there is no possibility whatsoever that Einsiedeln 266 is the manuscript used by Brouwer.

And indeed, such is the conclusion of palaeographical analysis. Even a cursory glance at the script of our three quires of MS 266 would suffice to disprove an origin at Fulda. (Hereafter, I will refer to this section simply as MS 266, with the understanding that I am only talking about these three gatherings). Bischoff assigns it to the Bodenseegebiet, that is the Bodenseeraum or the area around Lake Constance, and dates it to s. IX/X.Footnote ¹⁶ The script is peculiar, Caroline certainly, but with definite Alemannic and Rhaetian influences, both consistent with an origin somewhere in the Bodenseeraum or a centre in its intellectual/scribal sphere of influence. Bodensee characteristics include most notably the i appended below the line to letters such as m and n, as in hoɱɳ for homini; the ri ligature is also found, but the i does not extend below the line; there are also cases of ɔ for con, as in ɔtendit (p. 205) and ɔcedite (p. 206). More Caroline features include the looped g and the t without a loop on the cross-bar. Features suggesting an earlier (as in late ninth-century) date include the use of &-caudata for aet and particularly the rendering of aeterna as &a.Footnote ¹⁷ In addition, MS 266 has at least three instances of open cc style a (pp. 193, 204 and 206). In short the script is obviously related to, but not obviously that of, St Gall, or Reichenau, or Chur; had MS 266 been written in any of those places, it would hardly represent their typical script. Murbach manuscripts are hard to pinpoint, but this combination of influences is precisely what we might expect to see produced there. Further, we need an abbey in this area with the sort of library which might contain unique copies of neglected ancient texts. The obvious candidates are St Gall itself, Reichenau, and Murbach once again, the home of such exquisite rarities as the codex unicus of Velleius Paterculus.

The Contents of the Manuscript

Once we jettison the link to Fulda, the only way to begin to inquire after the origin of the manuscript is to examine the origins of its other components. The contents of the three quires of Einsiedeln 266 which include the EE have never been comprehensively described.

1. 1. 177–96: part of the Seven-Book Computus, or the Handbook of 809 (end of Book IV and all of Book V), with section numbers still attached.

2. 2. 196–202: various pieces de ponderibus et mensuris, often found along with the Computus (in, for example, the Monza and Madrid manuscripts).Footnote ¹⁸

These texts have been described as miscellaneous astronomical and philosophical fragments, which they certainly are, but they are more than that. The tell-tale sign is the numbers which precede most of the titles. These are chapter numbers from the so-called Seven-Book Computus, or Handbook of 809, a high Carolingian production put together by a team of scholars working in Aachen to provide a definitive handbook to computus, that is, the method for determining the date of Easter, for the whole Carolingian world.Footnote ¹⁹ What we have on pp. 177–96 is the end of Book IV and all of Book V of the Computus. Missing entirely is Book VII, which consists solely of Bede's De rerum natura. This is crucial, since abbreviated versions of the Computus circulated in which this book, as well as large parts of the first three books, were taken out. Ninth-century examples of abbreviated versions of the Computus include Montpellier H 334 and Bamberg HJ.IV.22 (class. 55).Footnote ²⁰ The Montpellier manuscript, just like MS 266, contains the original chapter numbers. The selection of texts found in our manuscript, including both the sections of the Computus proper and the ‘associated texts’ on weight and measures found on pp. 196–202, affiliates our text strongly with the Bamberg manuscript.

That MS 266 originally contained an abbreviation of the Computus can be proven another way. We have two gathering numbers extant in our three quires: V in the extreme gutter of p. 192 and VI in that of p. 208 (presumably there was once a VII on p. 224, but that page is far too damaged to make anything out). Now by comparing the relative lengths of identical material between the Einsiedeln manuscript and a complete manuscript of the Computus (I used Madrid 3307), we find that, were MS 266 a complete manuscript of the Computus, missing only Book VII, it should have roughly 164 pages lost at the beginning, or 82 folios. There is no way that amount of text could possibly have been contained in the four missing quires, or 32 missing folios. Therefore, MS 266 must have been an abbreviated Computus very much like the Bamberg text. Unfortunately, we do not have any sure guide to where the Bamberg manuscript was written, although it was probably somewhere in present-day France, and like the other manuscripts of the Computus in a centre with close links to the court at Aachen. The Montpellier manuscript was at Troyes in the Middle Ages, and was probably written in the vicinity.

1. 3. 202–3: Dicta philosophorum (florilegium)

This text comprises several parts: the Dicta proper, which are gleanings from thirteen authorities on the incommodities of matrimony; other unrelated material on grammar; and several mnemonic lists such as the seven punishments of Cain. The exact same material is found in St Gall MS 899, pp. 172–4, in a much clearer and more formal manuscript layout. This manuscript is probably ninth-century,Footnote ²¹ and certainly written at St Gall. On textual grounds, it is clear that the Einsiedeln manuscript was not copied from that of St Gall; they both seem to stem from a common exemplar. According to Munk Olsen, this little florilegium is only found in these two manuscripts.Footnote ²²

1. 4. 204–5: selection from Cassiodorus inst. 2.3.6–8Footnote ²³

2. 5. 206–7: Carmina einsidlensia

3. 6. 208–24: Hrabanus Maurus, Carmina

One important point about the arrangement of these texts is that all of them, after the Computus, begin on an even-numbered page, that is on a folio verso, which suggests that everything after the Computus and its associated tracts (which were certainly written all at once) was copied straight through or successively. Because of this, we must assume that the whole production was actually well planned, despite the chaotic and messy appearance of the writing and mise-en-page. For example, one might naturally assume that the lists on p. 203, including the seven punishments of Cain, the propria of the devil, etc., all crammed together haphazardly on the bottom half of the folio, are random accretions. In fact, in St Gall 899 we can find all these same texts neatly laid out in order. Hence, the scribe of MS 266 knew exactly what he needed to include, and economized as much as possible to fit everything in. It is possible, therefore, to speak of the intellectual profile of the compiler of MS 266: what seems on the surface to be a random collection of textual flotsam was deliberately assembled according to some as-yet-undetected plan, in which our carmina play an important rôle.

A Murbach Miscellany

The affiliations of these texts are revealing — on the one hand, the Dicta philosophorum shows affiliations with St Gall, while the section from the Computus suggests close associations with Lotharingia. We do not otherwise have evidence that the Computus was known in this region in the ninth century: a ninth-century codex currently in St Gall, MS 248, which contains material from it, was actually written in Laon.Footnote ²⁴

Textual arguments thus indicate the same conclusions as palaeography: that the Einsiedeln manuscript was designed by a scholar or scholars influenced by the intellectual cultures of both the Bodensee and the Middle Kingdom. Bischoff suggested that the Madrid manuscript (MS 3307) of the Computus was in fact written at Murbach — a place with precisely the intellectual affiliations I have described — although that opinion has fallen out of favour.Footnote ²⁵

On the basis of the Computus, the Dicta philosophorum and the poems of Hrabanus, we can date the formation of the miscellany to no earlier than say 850, with absolutely no reason to believe that it predates the writing of the manuscript itself, somewhat later in the century.

So where did the EE come from? They are unlikely to have been derived from an earlier miscellany, as there is no trace of them in St Gall 899, and at any rate, they are separated from the Dicta by the brief interlude of Cassiodorus. Grouping them with Hrabanus contributes nothing to solving this problem, since it merely pushes the problem back at most several decades. Further, it is very unlikely that Brouwer would not have printed them had he seen them, which implies that they were not transmitted with Hrabanus in his manuscript. It is possible that Hrabanus had seen the poems, as early as 810, but where is wholly unknown.Footnote ²⁶ The only plausible solution is that they were extracted from a complete manuscript of whatever text they come from — a text whose title may well have originally been written at the top of MS 266 p. 206 before it was trimmed. This must have been a book available to the scribe of the Einsiedeln manuscript, writing at Murbach or a related centre in Alsace or the Bodenseeraum.

Murbach offers us a catalogue from precisely the period that interests us, around 850, and among a good collection of ancient materials — for example, a copy of Lucretius, probably the one rediscovered by Poggio centuries later,Footnote ²⁷ and the earliest complete collection of the Appendix V ergiliana — it lists a Bucolicon olibrij. Footnote ²⁸ This catalogue dates from the middle of the ninth century, although it is only extant now in a copy made in 1464, by Sigismund Meisterlin, better known as the author of the Nuremberg Chronicle. The catalogue suggests that the bucolics occurred in a miscellaneous manuscript of poems from the third through the fifth centuries, roughly similar in size to the collection of the Appendix: the bucolics (number unspecified), Serenus’ De medicina praecepta, Avianus’ Fabulae, and Symphosius’ Aenigmata. Together the last three items cover about twenty-one hundred lines, compared with the roughly twenty-four hundred of the Appendix.

To retread ground covered by others, it is worth pointing out here that we do have an extant poem ascribed to an Olybrius. It is printed in the Anthologia latina, but does not come from the Codex Salmasianus. Instead its source is a florilegium put together in southern Italy, around Montecassino.Footnote ²⁹ Amid a wide variety of theological and canonical texts, this florilegium contains a section De notis which brings together a number of texts on notae — both the notae iuris and the Alexandrian critical signs — some of which are not found elsewhere. Preceding a list of eleven notae is a poetic exchange between a Campanianus, v[ir] i[llustris], and a patricius Olybrius. After that list comes a De obelis et asteriscis Platonis — a list of notes specifically put together for the reading of Plato, which is found in cognate versions in a Greek fragment on papyrus and in Diogenes Laertius.Footnote ³⁰ Following that is a list of notae which was attached to the psalm commentary of Cassiodorus. The last is the most crucial: it is very likely that the notae Platonis were transmitted through Vivarium, Cassiodorus’ monastic foundation, ultimately to Montecassino. Given the aristocratic lineage of the various Olybrii — they were closely related to the Anicii — and the fact that some of them were active into Cassiodorus’ lifetime, it makes perfect sense to assume that Cassiodorus was responsible for the preservation of the Olybrius poem and the list of notae.

All of this is relevant to the Murbach catalogue. Murbach, itself styled Vivarium, was deeply influenced as an institution by Cassiodorus’ Vivarium, and the library catalogue in particular was composed with a copy of the Institutiones at hand. This copy was itself special: there are only three early manuscripts of the Institutiones which contain both books together, and all of them are from southern Italy. But the catalogues of both Reichenau and Murbach list copies of the Institutiones in two books, a coincidence that can only be explained by direct contact between the abbeys of the Bodenseeraum and southern Italy.Footnote ³¹ In addition to listing the books that the abbey possessed, unusually the Murbach catalogue also includes books that the abbey wished to acquire, most of them taken straight from Cassiodorus’ recommendations in the Institutiones.Footnote ³²

A few entries below the Bucolicon olibrij (or, to standardize, Bucolicon Olybrii), it is no surprise then when we come across a Liber notarum.Footnote ³³ Between the two entries come several works on medicine, but the Liber notarum is separated from these by a paraph, a gamma-shaped mark indicating a new section (no. 335 Milde, p. 48). The cataloguer was proceeding topically, and it seems the Bucolicon Olybrii was contained in the same manuscript as Serenus’ poetic De medicina praecepta. Hence he inserted the few medical works that Murbach possessed between the Bucolicon and the Liber notarum, which at any rate defied easy categorization.

That the only two known references to texts of a poet named Olybrius are found in close proximity with notae cannot be accidental.Footnote ³⁴ That both of them also occur in contexts associated with Cassiodorus may not be accidental either. Rather, they suggest that the Murbach library, just like Cassiodorus’ Vivarium, contained the whole texts from which the Montecassino florilegium took excerpts. Were a new piece transmitted through Cassiodorus to turn up, the Bodenseeraum is precisely where we might expect it to be found. Indeed, the last piece of Cassiodoriana discovered, the Anecdoton Holderi, or the Ordo generis Cassiodororum, was uncovered in Karlsruhe from a Reichenau manuscript by Alfred Holder in the mid-nineteenth century, and was probably preserved in this region.Footnote ³⁵ The presence of the complete Institutiones in Reichenau and Murbach confirms this special link.

Cassiodorus is thus the link between the Einsiedeln manuscript and the Bucolicon Olybrii. As shown above, the Dicta philosophorum precede the EE in MS 266 and Hrabanus follows them. The EE accompany neither, on the basis of St Gall 899 and Brouwer's edition. They belong instead with the little fragment of Cassiodorus. Obviously, one should not push this argument too far. Selections from the second book of the Institutiones are by no means rare in Sammelhandschriften. Rather, the inclusion of Cassiodorus gives another facet of the intellectual profile of the compiler of MS 266. A late ninth-century copy of the Aachen Computus tells us that he was connected to and interested in the intellectual reforms of the Carolingian court. The Dicta localize him somewhere not too far from the Bodensee. Hrabanus — the only living author to receive a heading in the Murbach catalogue, and one of only two moderns — and Cassiodorus are two authors the compiler was particularly interested in. Where precisely MS 266 was written may never be definitively known, but this description fits precisely the intellectual, political, and cultural affiliations of the Abbey of Murbach in the ninth century.

So we have a sequence of coincidences: the Bucolicon Olybrii are to be found at Murbach in the cultural ambit of the Bodensee, probably thanks to a copy preserved by Cassiodorus, in the middle of the ninth century. The EE are bucolics, found in a manuscript from the Bodenseeraum, possibly from Murbach itself, transmitted with a bit from Cassiodorus, written some time after the middle of the ninth century. On principle, one should not needlessly multiply bucolics: the most economical solution is that they are one and the same, or rather, that the latter is part of the former.

Medieval Catalogues and the Economy of Transmission

But there is more to this argument than a mechanical application of Ockham's razor. Medieval library catalogues are remarkably accurate as a minimal guide to the transmission of Latin literature.Footnote ³⁶ At present, there are only at most three ancient texts (more if we include patristic texts) listed in Carolingian catalogues that do not survive, including the bucolics of Olybrius. The other two are the Opuscula ruralia of Serenus Sammonicus, listed in the Bobbio catalogue, and a book of Alchimus’ declamations in the booklist of Berlin, Diez B Sant. 66.Footnote ³⁷ Given the hundreds of ancient authors and texts that do survive which are mentioned in the catalogues (indeed it is only exotic specimens like Ampelius and Julius Obsequens which are not to be found in some medieval library catalogue), these three are the genuine anomalies.Footnote ³⁸ A cautionary tale: the famous Lorsch catalogue contains an entry (no. 427) Metrum Severi episcopi in evangelia libri XII.Footnote ³⁹ Ever since this entry became known, scholars assumed the work was lost, and even suspected some corruption, since following in the catalogue are ten eclogues and four georgics by the same author (eiusdem eglogas X, eiusdem Georgicon libri IIII). In 1967, Bernhard Bischoff found in Trier some unnumbered fragments from a ninth-century codex, containing securely identifiable scraps of this work. Not until 1994 did the bishop Severus finally receive his editio princeps.Footnote ⁴⁰ The lesson of this story is that catalogues deserve a great deal of credence, and hence we should be quite sure that a text with the title Bucolicon Olybrii did in fact exist at Murbach in the ninth century, and that, given the tremendous economy of the transmission of classical literature from the Carolingians on, it is much more likely than not to be preserved somewhere. Identifying the EE with the Bucolicon Olybrii solves two difficulties at a stroke.

In terms of the history of the transmission of classical literature, linking the two is all but unavoidable. What indeed are the odds that two different, unrelated ancient collections of bucolic poems circulated in the Bodenseeraum in the ninth century, leaving scarcely a trace anywhere else? Yet according to the widely-accepted account of the EE, they could not possibly be in fact by any Olybrius since they were composed under the emperor Nero, sometime in the decade a.d. 55–65. In the second part of this study, I will examine the internal evidence for when the EE were composed.

‘Of Neronian Date’?

One thing we can examine to get a more precise sense of the dating is the language and usage. All the words contained in the EE are classical, but some of the particular usages are found primarily after the first century. In I.44, the white head of the figure whose identity must have been indicated in the line cut away by the binder ‘was shining with full honour (pleno radiabat honore)’. The sense of plenus here must be absolute: there is nothing implied or stated in the context which tells us what it is full of. Instead it means summo honore (as in Lucr., DRN 4.1155: ‘summoque in honore vigere’), or magno honore (as in Ovid, Fasti 6.658: ‘magnus et in magno semper honore fuit’), or multo honore (as in Virg., Aen. 3.474: ‘multo compellat honore’). Parallel absolute usages of plenus are rare before the fourth century (cf. TLL s.v. plenus).Footnote ⁴² Similarly, as Armstrong has already noted, the EE use totus where the meaning has to be that of omnis or summus on no fewer than three occasions (I.31 ‘toto … amore’, II.24 ‘totaque … saecula’, and II. 25 ‘tota spe’).Footnote ⁴³ It is a telling fact that Baehrens and other editors have attempted to emend all three of these passages, as if baleful corruption struck the word totus independently in three different places. In the first and last example, they emend the case of tota to give it a more regular companion (‘zonas … totas’ and ‘totas … aristas’ respectively), and in the second, they read tutaque despite the fact that tutis itself is transmitted without any trouble just a few lines later (II.36). This can be nothing but obduracy in the face of the obvious: the author of the EE has an expansive idea of the semantic range of totus. We can observe this shift in the Latin language from Apuleius on all the way to the point that most of the Romance words for universal quantity are derived from totus, not omnis (tous, todo, tutto, tot, etc., but cf. It. ogni).Footnote ⁴⁴ In fact, all three individual phrases can be paralleled by later examples (‘toto … amore’ ~ Orig., Cant. trans. Ruf. 2, p. 170 Baehrens; ‘tota … saecula’ ~ Tert., Adv. Marc. 5, p. 542 Evans; ‘tota … spe’ ~ Apul., Met. 6.5). To suggest that a Neronian author — or even worse, two Neronian authors — deploy such late features repeatedly runs counter to what we know about the development of the Latin language.

The other useful approach is to look at fontes and parallels. If there are a large number of small concurrences with post-Neronian poets, it is very likely that the EE are later, as it is far more plausible that one later author had a standard literary education which gave him familiarity with both golden and silver poets, than that the major late Neronian and Flavian authors all happened to be acquainted with this one author who has left no other trace of his existence. The evidence here is somewhat ambiguous but worth briefly presenting:

These are the all too easily obtained results of database trawling: some of them (though by no means all) a sceptic could dismiss as trivial. Nonetheless, together they suggest it is much more likely that the author of the EE wrote after the major late first- and early second-century authors. In other words, sources and parallels suggest the same thing as analysis of the language: they were probably not composed before the end of the second century.

The Bucolic Tradition

We can also situate the EE in the bucolic tradition. It is virtually certain that they postdate and respond to Calpurnius, as argued above.Footnote ⁴⁶ Since Calpurnius is either Neronian or much later, such a conclusion does nothing to clarify their dating. But there is evidence that they make use of Nemesianus (who is securely documented as third-century) as well: EE I.9: ‘fistula silvicolae munus memorabile Fauni’ is certainly connected with Nem. 1.14: ‘Iam mea ruricolae dependet fistula Fauno’, while EE II.37: ‘subeunt iuga saeva leones’ harks back to Nem. 4.54: ‘iuga … coget sua ferre leones’. The first example, where the EE gloss the fistula as a gift of wood-dwelling Faunus, is very likely a reference to Nemesianus. We also know that Nemesianus’ line about yoked lions gained a certain amount of cachet, being quoted in a poem written in 384, recently attributed to Pope Damasus, Carmen contra paganos 103: ‘vidimus argento facto iuga ferre leones’.Footnote ⁴⁷ So it is likely, but at this point not entirely certain, that the EE postdate Nemesianus, who was active in the 280s.

Besides Calpurnius and Nemesianus, another text in the bucolic tradition to which the EE relate is the De mortibus boum (DMB) of Endelechius.Footnote ⁴⁸ Endelechius taught rhetoric in Rome in the 390s; besides his authorship of this poem, he is mentioned by Paulinus of Nola (epist. 28.6) around 400, and more importantly, in the subscriptio to the ninth book of Apuleius’ Metamorphoses, written in 395 by a certain Sallustius, during the consulship of Olybrius and Probinus: ‘Ego Sallustius legi et emendavi Rome felix Olibrio et Probino v. c. conss. in foro Martis controversiam declamans oratori Endelechio.’Footnote ⁴⁹ Just like the author of the EE, Endelechius was not only a bucolic poet, but possibly a panegyrist as well; it was at his suggestion that Paulinus composed his panegyric for Theodosius. Endelechius’ only surviving literary work is the DMB. While deeply Virgilian, this poem is formally innovative, composed in asclepiads rather than hexameters; it is, moreover, explicitly Christian.

There can be little doubt that Endelechius modelled the beginning of the DMB on EE II.

The shared vocabulary of these two passages is immediately apparent: quid EE I.1 ~quidnam DMB 1; tacitus EE II.1, tacendi EE II.14 ~tacitum DMB 8; gravis EE II.3 ~graviter DMB 2, gravis DMB 10; altius EE II.7 ~alta DMB 5; scire EE II.11 ~scis DMB 13.Footnote ⁵⁰ But this jejune list of common words scarcely reveals the extent of the relationship between the two poems. Almost everything the poet of EE II says is said by Endelechius in different words. So ‘quid tacitus’ (EE II.1) becomes ‘quidnam solivagus … gemis’ (DMB 1–2). Glyceranus says, ‘I do not really understand’ (EE II.4); Aegon says, ‘Make your friend understand’ (DMB 4). Mystes says in response, ‘It does not please me to say everything’ (EE II.4); Buculus responds to Aegon, ‘Allow me to keep deep silence in my troubled feelings’ (DMB 5–6). After a failed suggestion, Mystes tells Glyceranus he is wrong (erras), and then that the problem is what one would least expect (EE II.7); so too Aegon tells Buculus, ‘It is the opposite of what you say, your claim is false’ (DMB 9). Both poems have an exchange of proverbs, three in each case. EE II has ‘the smooth sea does not usually become choppy without winds’, ‘satiety troubles my joys’, and ‘pleasure and sleep are often given to quarrels’ (8–10). Only the first of these is listed in Otto's Sprichwörter (no. 23), although all of them have the sententious flavour of paremiology.Footnote ⁵¹ ‘Satias gaudia vexat’, for example, is just a reworking of the old chestnut ‘la satiété engendre le dégoust’ (Montaigne's formulation), which goes all the way back to Solon (fr. 6.3 West: τίκτει γὰρ κόρος ὕβριν), already very well worn by the fourth century.Footnote ⁵² Likewise, the DMB present Buculus’ apothegm about how silence heals and then Aegon's version of Prov. 28:13, ‘a burden divided is less heavy’ (compare Sedulius, carm. Pasch. 4.76–8), along with some homespun wisdom about a boiling pot (DMB 7–11).Footnote ⁵³ None of these are listed in Otto or Walther, though they are obviously proverbial. Finally, in the EE, Mystes gives in, ‘therefore, if you are eager to know the cause of my worries’ (EE II.11), which Glyceranus interrupts with an invitation to rest under the shade of the tree, and then says, ‘you, tell what is the cause of your silence’ (EE II.14). Likewise, in the DMB, Aegon concludes his second proverb with an invitation to speech, ‘talking helps pain’ (DMB 12), and Buculus begins ‘you know, Aegon, …’ (DMB 13). From this point on, the two eclogues diverge, Mystes continuing with his song about the Golden Age, and Buculus his lament about the plague.

These two poems are structurally affiliated, and it is simply not possible that of the fifteen post-Virgilian ancient eclogues we possess two of them would display such parallels independently. Since, as it seems, EE II is modelled on Calp. 4, including its opening, the DMB must be the debtor. The DMB's dependence is also apparent in the way it simplifies the complex dialogic structure of EE II. The two dates we have for Endelechius are 395 and 400, although if we consider the DMB specifically we can obtain a broader range for his career. It has been suggested that the plague it describes took place around 386Footnote ⁵⁴ and that Paulinus alludes to it in a poem in honour of St Felix written in January of 406.Footnote ⁵⁵ Thus we have a fairly firm terminus ante quem of 405 for EE II. Further, the similarities between them — which are pronounced, striking, and immediately obvious, but, nonetheless, scarcely verbal — hint that these are two poems produced in the same milieu. This sort of imitation is more reflective of deliberate rivalry, of doing the same thing in a strikingly different way, than the imitation of past masters and school authors for literary effect. Thus, the evidence from other bucolics gives us a tentative span of about 290 to 405 in which to place the EE, with a weak preference for a date closer to the latter.

Ausonius and the EE

One significant allusion I think can locate the writing of the poems more precisely. At EE I.32–5, the poet makes a reference to Apollo, Python-slayer:

The reference here is highly specific: the invention of the paean at the victory of Apollo over the Python. This story goes back to Callimachus’ hymn to Apollo, and in Latin there are two other brief accounts of it. The first is Terentianus Maurus, where it is connected with the invention of the iambic metre (1586–95):

Terentianus’ sources are unknown, although they seem to go back to Callimachus.Footnote ⁵⁶ He also relies on Latin authors for poetic flair (compare, for example, Ovid, Met. 1.457–60 and Luc., BC 5.80–1). From Terentianus, the motif was adopted by Ausonius in a poem addressed to Iambus, the iambic foot, and sent to Paulinus, around 380 (epist. 19b.10–13 Green):

The essential connection between Ausonius and Terentianus is the iamb: Ausonius’ metapoetic enthusiasm led him to Terentianus’ de metris, where he looked up or recalled the passage on iambics, and then freely retold it in his own words, adding for example the presence of the Muses. Ausonius completely removes the people of Delphi from the story — the ones who originated the chant according to Callimachus and Terentianus.

There can be no doubt that the accounts in EE and Ausonius are linked: the lines ‘Phoebus … laetus caede draconis’ and ‘caedem in draconis … Delium’ are remarkably similar, particularly when one takes account of their different metres. Terentianus, Ausonius, and the EE (along with Claudian's preface to the In Rufinum, which shows scant similarities with the other threeFootnote ⁵⁷) present the only passages in Latin poetry that discuss the slaying of the Python in conjunction with music.Footnote ⁵⁸ Ausonius was drawing directly from Terentianus, which suggests strongly that the author of the EE was acquainted with Ausonius. Considering the contents of the story in the EE strengthens this suspicion. In the versions of this story from Callimachus through to Ausonius, the music is always vocal, not instrumental. For no particular reason, beyond the fact that Ausonius used this anecdote for an etiology (‘primus … iunxisti’), the EE combine this story with Apollo's invention of lyre-music (cf. TLL s.v. generare).Footnote ⁵⁹ The story in the EE is no old mythological variant, but a mere fusion of two separate stories, induced probably by Ausonius’ inclusion of the Muses, who indeed are mentioned two lines later in the EE.

Possible parallels with some of Ausonius’ contemporaries also deserve consideration. In Reference Knickenberg1892, Knickenberg drew up without comment a list of eight places in EE I that have parallels in later poetry, particularly Claudian.Footnote ⁶⁰ No one since has pursued this line any further (and to be fair, some of the parallels are weak). To them, I would add the following: the evocative ‘opes Heliconis’ (I.37) are found otherwise only in Claudian, carm. min. 31.19–20: ‘tunc opibus totoque Heliconis sedula regno / ornabat propriam Calliopea nurum’. The closest parallel to the Maenalides of EE II.18 is in Ausonius’ Technopaegnion: ‘nec cultor nemorum reticebere, Maenalide – Pan’ (51).Footnote ⁶¹ Nonetheless, in context, the Maenalides have nothing to do with Mt Maenalus in Arcadia; instead, they are followers of Bacchus. The only author in Greek or Latin I know of who substitutes Maenalids for Bacchae or Maenads is Nonnus.Footnote ⁶² Even more telling, II.14: ‘tu dic quae sit tibi causa tacendi’, is almost identical to the question posed to Christ by Pilate in Juvencus 4.597: ‘Pilatus quaerit quae tum sit causa tacendi’.Footnote ⁶³

Christian vocabulary, in fact, has left traces on the poems. Glyceranus asks Mystes, ‘forsitan imposuit pecori lupus?’ (II.5). Tricky wolves are not rare — the author almost certainly has Virgil, Ecl. 5.60–1 in mind: ‘nec lupus insidias pecori nec retia cervis / ulla dolum meditantur’. But imposuit specifies the nature of the trick: the wolf is an impostor (cf. TLL s.v. impono Footnote ⁶⁴) among the sheep (pecus here is a flock of sheep as in Ecl. 5.60). Proverbial as the ‘wolf in sheep's clothing’ might be to us, its origin is in the Gospels (Mt. 7:15). That we have a wolf using disguise to get access to sheep — this is why dogs are Mystes’ only sure defenceFootnote ⁶⁵ — strongly indicates an author well aware of the ‘lupi rapaces in vestimentis ovium’.

Pseudo-Octavian and the EE

Ausonius is not our author's only certain source. As Knickenberg first noted, the EE share an important half-line (I.36) with a poem in the AL (672.35): ‘huc huc Pierides’.Footnote ⁶⁶ Verdière correctly comments that this poem is late antique; he automatically assumes, however, that it is indebted to the EE. The poem in question is the famous Versus Octaviani Caesaris in laudem Maronis, beginning ‘Ergone supremis’, a rhetorical set-piece in which Augustus defends his decision to preserve the Aeneid against Virgil's dying wishes. Fairly popular in antiquity, it survives in a variety of versions, and with several ancient imitations. One of them is ascribed to the grammarian Phocas — if that ascription is correct, the poem would have been composed before the end of the fourth century.Footnote ⁶⁷ It is not only this half-line that the two poems share: the broader theme is remarkably similar, the destruction of Troy. ‘Octavian’ says (24–9):

These images of ruin and burning are meant to elide the burning of Troy with the proposed burning of the Aeneid, which he refers to as hoc opus (27). Likewise, the EE continue (I.36–41):

Commentators have already noted the connection of these lines with the story of Virgil burning the Aeneid, preparing readers for the last line of the poem in which Mantua destroys its pages.Footnote ⁶⁹ The repeated vocabulary is too similar to be coincidental.Footnote ⁷⁰

Contra Verdière, however, the relationship almost certainly goes the other way. It is the EE which allude to ‘Octavian’, and not vice versa. This can be shown on both external and internal grounds. For the first, the Versus were popular in antiquity, from the late fourth century on; from that point they began to become one of the standard pieces transmitted with the Virgilian corpus. The EE, however, do not seem to have been so widely known. On that ground alone it is far more likely that the EE copy the Versus. The contextual case is even stronger. Both passages are discussing the fall of Troy, and both passages include (implicitly at least) a Caesar. But in ‘Octavian’, Caesar intervenes so that Virgil's Aeneid is saved, while in the EE, Virgil's Aeneid ends up destroyed as Caesar's own head is crowned. This irony cannot but be intentional. A reader of the EE familiar with ‘Octavian’ would be signalled by the memorable Huc huc to call to mind Augustus’ preservation of Virgil's poetic achievement, only to be shocked (and probably amused) by a Caesar's triumph causing the destruction of the Aeneid. Getting the order wrong completely obscures the force and meaning of Thamyras’ encomium.

Interestingly, there is external evidence for the conjunction of Endelechius, who was influenced by the EE, and ‘Octavian’, who influenced them. The De mortibus boum scarcely survives: for a long time, it was thought only to be found in the printed edition of Pithou (1586), until a sixteenth-century manuscript copy turned up, at any rate closely connected with Pithou's text.Footnote ⁷¹ Besides that, the only evidence we have is the eleventh-century catalogue of the books given by Mannon to the library of St-Oyan (Besançon, Arch. Dep. 7 H 9).Footnote ⁷² There we find it in company with Claudian, Nemesianus, miscellaneous poems, Avianus’ Fabulae, the Aenigmata of Symphosius, and the Versus Octaviani (no. 89). The collection in the manuscript of Mannon (possibly from Lyon) is a context in which the EE would find themselves perfectly at home. Such also is the context of Bucolicon Olybrii, immediately followed in the Murbach catalogue by Serenus’ De medicina praecepta, then Avianus and Symphosius (327–330 Milde).

A Grammatical Education

Another feature indicating a later dating is the overwhelming influence of grammatical scholarship. For example, ‘maxime divorum caelique aeterna potestas’ (EE I.22) unquestionably comes from Virgil, Aen. 10.17: ‘o pater o hominum rerumque aeterna potestas’. But the EE are not alluding to this line as it is found in the fourth- and fifth-century manuscripts. Rather, they imitate the version found in Ti. Claudius Donatus and Servius, ‘o pater o hominum divumque aeterna potestas’. In Late Antiquity, this version is not found outside the late fourth- and early fifth-century commentary tradition. Further, the addition of caeli suggests the ‘physical’ interpretation of this line Servius ascribes to Probus, wherein Jupiter is ‘aether, qui elementorum possidet principatum’. Servius then adduces the distinction between Jupiter and Apollo (ad loc, II.385 Thilo/Hagen: ‘Aeterna autem potestas adiecit propter aliorum numinum discretionem: nam legimus et Apollinem deposuisse divinam potestatem’), just as we find in the following line of the EE, ‘seu tibi, Phoebe etc.’. Another example, even more obvious: ‘languescit senio Bacchus’ (EE II.26). The phrase is a fine Horatian tag from Odes 3.16.34–5, ‘Bacchus in amphora / languescit’. Porphyrio comments: ‘Bacchus in amfora languescit: Belle languescit, quasi senescit, ac per hoc veterescit’ (116 Holder). A lovely expression, he says, since languescit is an elegant way of saying senescit or veterescit. The pedantry of our author cannot bear to leave so subtle a phrase unexplained. Senio is patently a gloss on Horace, which indicates that our author's knowledge of the poets came straight out of the school-room.

EE II.23, where Mystes sings ‘Saturni rediere dies †redit Astraea certo†’, presents a special case. This is one of the closest imitations of Virgil in the two poems, following Ecl. 4.6: ‘iam redit et Virgo, redeunt Saturnia regna’ — whatever the second hemistich actually originally contained (Hagen rewrote it as ‘Astraeaque virgo’). Unlike Virgil, the author identifies the Virgo, calling her Astraea. Far from a use of Virgil's source,Footnote ⁷³ this lack of allusiveness, this straightforward didacticism, smacks of grammatical education. While we do not have an apposite gloss on Ecl. 4.6, see for example the ps.-Probus gloss on Georg. 1.32: ‘Erigonen dicit virginem, quae Iustitia, Astraei filia, quae cum morata esset aureo et argenteo saeculo cum hominibus, ferreo saeculo se recepit in caelum, quia ultra in terris sedem non habuerit nec sustinuerit se hominum miscere sceleribus.’ This gloss provides the key to understanding a much-discussed passage that comes just before (EE II.19–20):

Commentators agree that the sacer hircus is Virgilian, from Geo. 2.395–6:Footnote ⁷⁴

Shackleton Bailey has already pointed out that exuit ext[r]a, is almost an impossible reading, even if no one has heeded him (the TLL gives this exuit its own sub-category).Footnote ⁷⁵ He tries to solve this difficulty by emending the MS extra, but the passage in the Georgics gives us sufficient reason to go with Hagen's exta — the corruption, if it exists, must lie elsewhere. One perversity of the line (leaving aside the otherwise unattested sense of exuere) is that the goat itself must be the subject of exuit, and hence the goat divests itself of its own entrails. Under this reading, the much discussed nudatis cervicibus is not a bit of precise cultic detail, but an obvious periphrasis for the same idea as the one read into Geo. 2.395.Footnote ⁷⁶ The commentators tell us that the important word in that line is stabit, which indicates that the victim is not unwilling, since an unwilling victim was not an appropriate sacrifice (Servius, ad loc. 3.255 Thilo/Hagen): ‘tunc est enim aptum sacrificium cum dedicatum animal victimae patiens invenitur.’ The longer version, probably going back to Donatus, adds: ‘inprobant enim aruspices hostiam quae admota altaribus reluctatur.’ This is precisely the force of a bared neck from Livy to Christian texts of the fourth century and later, where it is also used in verse.Footnote ⁷⁷ Paulinus of Périgueux provides a precise parallel with ‘patet ecce innoxia ceruix / vulneribus nudata tuis’.Footnote ⁷⁸ Here we have poetic one-upsmanship: whereas Virgil's goat stood there, not resisting the sacrificing knife, the EE's goat almost performs the sacrifice itself (if the poem indeed says ‘hircus … exuit exta’), unbidden like everything else in the Golden Age.

Indeed, the connection between this couplet and the Georgics is even stronger when one examines the lines just above in Virgil (2.388–9): ‘et te, Bacche, vocant per carmina laeta, tibique / oscilla ex alta suspendunt mollia pinu.’Footnote ⁷⁹ The laeta carmina go along with the tibia laeta that canit. (Is it going too far to suggest that the sound of tibique played a rôle?) The next line also gives something hanging from a tree (suspendunt ~pendunt); in Virgil, the tree is a pine, in the EE, an elm. This is no symbolic deviation but a learned innovation, since the commentators are clear that the elm as used for training vines is particularly sacred to Bacchus/Liber (Servius on the line-ending ab ulmo in Ecl. 1.58: ‘ulmus lignum est quod sub vinea fit’; cf. Manilius 3.662: ‘tunc liber gravida descendit plenus ab ulmo’; and Petronius, carm. 33.2: ‘uvaque plena mero fecunda pendet ab ulmo’).

Further in Virgil, it is not a bare-necked goat that is hung from the tree, but rather oscilla.Footnote ⁸⁰ The Virgil commentators, such as Servius and ps.-Probus, were at pains to identify these oscilla: one explanation tied them to the story of Icarius, the father of Erigone, who was killed for distributing Liber's wine by peasants who could not distinguish between intoxicants and toxins. In despair, Erigone hung herself from a tree. A plague arose afflicting the young women of Attica, who went mad and likewise hung themselves from trees; ultimately, the Athenians discovered from the oracle that Icarius’ death had to be avenged before the plague of suicidal madness would end. Hence, little dolls called oscilla were hung from trees in memory of the dead women. In one version of the story, Icarius, Erigone and his dog were placed among the stars as Boötes, Virgo and Canis minor (sometimes maior), respectively. This is a minor alternative to the usual identification of Virgo with Iustitia, and the resulting confusion is compounded by the fact that the names Astraea and Erigone could be used both for the daughter of Icarius and Iustitia.Footnote ⁸¹

EE II capitalizes on this confusion. The common thread running through the whole first part of Mystes’ song is Bacchus/Liber: from II.17 ‘spirant templa mero’ to 26 ‘languescit senio Bacchus’, we get a collection of Virgilian tags and allusions, all loosely organized around the same theme, completely recombined in a virtuoso display of grammatical education.Footnote ⁸² Astraea, with her double significance, is thus the connection between the Georgics-inspired rites of Bacchus in this half and the Golden Age of Eclogues 4 in the second half. A literate audience surely would have appreciated this subtle nod to the grammaticus’ art. Far from displaying any genuine cultic knowledge, the author of the EE gives us a romp through the commentary tradition to Virgil, combining different elements as he sees fit: hanging the goat from the tree instead of the oscilla, reserving Erigone or Astraea for rewriting Ecl. 4.6, changing the tree to an elm to reflect its associations with vines. Virgil provides the raw material; grammatica provides the glue which makes the passage hang together. It is not possible that this poem was written without recourse to commentary traditions cognate to those which survive today, and which began to take their mature form in the fourth century.

Grammatical training also guided the poet in naming his characters. In fact, the EE are the only ancient Latin bucolics which do not use a single Theocritean character.Footnote ⁸³ The names they do use are suspect: Glyceranus has already been singled out by Courtney as particularly problematic, not just on literary, but also linguistic, grounds.Footnote ⁸⁴ All the others are chosen from the Latin tradition, and usually for facile reasons. Midas the judge comes from Ovid; Thamyras and Ladas are both associated with competitions, although only the former with poetry.Footnote ⁸⁵ It becomes clear that the author chose the name Ladas on account of Juvenal's mention of the runner (sat. 13.97), when one examines the old scholion, composed probably in the fourth century: ‘Ladas fuit inter nobilissimos cursores Olympico certamine, cui aemulus Talaris eandem gloriae palmam tulit, sed apud Elidem coronatus est.’Footnote ⁸⁶ With that biography, read again the exchange in EE I.13–16:

The conceit of two rivals vying for the gloria of the palma is lifted from the reading of Juvenal under the grammaticus’ rod. Sometimes the author can be quite clever in this game of literary onomastics: in the same exchange, Ladas says that [Apollo] ‘laudatam … chelyn iussit variare canendo’ (I.18). The force of this boast to Thamyras only makes sense in the context of the chapter on the history of music in Pliny the Elder, ‘cithara sine voce cecinit Thamyris primus’ (7.204). Unlike his rival's namesake, Ladas makes music with both lyre and voice. Mystes has nothing to do with the slave-boy in Horace, and everything to do with mysteria. We cannot chalk it up to accident that it is Mystes who says ‘nec me iuvat omnia fari’ (II.4).Footnote ⁸⁷

The question of the names in Latin bucolic is complex: by using Theocritean names, Virgil signals that he is writing the kind of poetry Theocritus wrote, and using non-Theocritean names carves out an independent space for his own poetry. Likewise, Calpurnius uses both Virgilian and Theocritean names to establish his place in the pastoral tradition. The school tradition held otherwise. Commentators on the Eclogues such as Servius explain in their prefaces the principles behind bucolic onomastics:

Servius’ comments tell us all we need to know — it is not only that bucolic names should have something to do with pastoral or bucolic activity, but that nomina ficta in general should be created on the basis of their meaning. In fact, this passage of Servius (or that of his source) is very likely where ‘Glyceranus’ came from.Footnote ⁸⁹ The easiest way to turn a sweet woman into a sweet man is to add a masculine suffix, here -anus, and Glycerium becomes Glyceranus. Precisely the same tendency can be observed in Endelechius, who introduces one non-Virgilian character, one with a large (and ailing) herd of cows, called Buculus.Footnote ⁹⁰ We might sneer at Endelechius’ naiveté in naming his character, but to the mind of the pre-eminent teachers of Virgil in his own day, he was simply following in Virgil's footsteps.

The Centonists

Another late feature of the eclogues is the overwhelming influence of Virgil. Obviously, from his own lifetime, Virgil exercised a potent sorcery over his successors. Nonetheless, it was only from the shadowy third century on that that sorcery began to demand constant, unremitting verbal echoes, as part of a larger shift in Latin poetic practice. The cento is only the most extreme form of a poetic tendency observable in poetry of almost every genre.Footnote ⁹¹ That is also the way that the EE use Virgil. The last line of EE II is directly lifted from the fourth eclogue: ‘casta fave Lucina: tuus iam regnat Apollo!’ (=ecl. 4.10). If the EE were Neronian, they would offer the earliest example by a century at least of a whole line of Virgil being incorporated into an independent piece of poetry.Footnote ⁹² But that is not the only example of centonizing. We have already seen that EE I.22 and II.23 are close reworkings of Virgil (above, preceding section). Similarly, the hemistich ‘quod minime reris’ (EE II.9) is taken directly from Virgil (Aen. 6.97). It is found in precisely the same way in the bucolic cento of the otherwise unknown Pomponius (AL 719a, l. 50; c. 400). Just like the centonists — the best examples are found in Ausonius’ racy Cento nuptialis — the poet of the EE occasionally rips phrases out of their context to repurpose them in bizarre ways. There can be no doubt that I.28 ‘stetit ostro clarus et auro’ comes from Virgil, Aen. 4.134–5 ‘ostroque insignis et auro / stat’; the poet evidently does not mind applying Virgil's description of a horse to a god. Some lines are almost perfect centos, such as EE II.16: ‘annua vota ferat sollemnesque incohet aras’, which combines Aen. 5.53: ‘annua vota tamen sollemnisque ordine pompas’ and 6.252: ‘tum Stygio regi nocturnas incohat aras’.Footnote ⁹³ Other fourth-century poets employed Virgil in this way; compare, for example, the opening of Paulinus of Nola's miniature epic on John the Baptist, ‘summe pater rerum caelique aeterna potestas’, with EE I.22. The same line of Virgil is found entire in Proba's cento (29, cf. 463).Footnote ⁹⁴ I am not positing a direct relationship between Proba or Pomponius or any of the centonists and the EE; nonetheless, the use of Virgil in the EE is centonizing in an analogous sense. It fits well placed after Proba and Ausonius; but placing it earlier, particularly all the way back in the Neronian age, would make its use of Virgil highly anomalous.

Philosophy

One of the more enigmatic passages in the EE is the song of Ladas in EE I in praise of Apollo, and in particular its cosmogony (29–31):

Much can be and has been said about these lines. The passage as a whole is inspired by Appius’ visit to the Sibyl in Lucan 5.86–120, although the influence does not extend to verbal echo. Here I would point out the implicit triad of divina potestas, artifex, and amor. This is at once a vulgarization of the Platonic triad of demiurge, ideas, and matter or world soul — connected to the Plotinian hypostases of Good, Mind, and Soul, and Power, Wisdom, and Goodness — and the Christian trinity. It is hard to imagine a cosmogony with an implicit divine triad that is independent of Neo-Platonic or Christian influence. That the details are not precise is itself the point: our poet is no philosopher, and it was exactly this sort of vulgar philosophizing that the grammatici considered their particular purview. The same Platonism can be observed in Servius’ comments on Aen. 6.724–9, the most famous philosophical passage in Virgil, and one which the EE directly echoes (6.727 ‘magno se corpore miscet’ ~ EE I.31 ‘toto miscet amore’); these lines are also cited in one scholion on Lucan 5.95–6.Footnote ⁹⁵ An extremely similar notion can also be found elsewhere in the scholia to Lucan: ‘hunc spiritum summum deum Plato vocat “artificem” permixtum mundo omnibusque quae in eo sint.’Footnote ⁹⁶ The only element in the EE missing in Virgil, Servius, or the Lucan scholiast is amor, but this could well have come from another Platonic source. Calcidius, for example, in his commentary on the Timaeus, equates divina providentia with caelestium amor (cap. 254). Hence it should come as no surprise when we find the Christian centonists producing passages much to the same effect as the EE. Compare Pomponius (10–13):

With this in mind, we can return to the opening of Ladas’ song, where the cosmogony begins (EE I.22–9):

The first two lines present a near equivalence between the supreme God and Apollo. This is a Middle Platonic notion, defended at length by Plutarch in the De E apud Delphos (393b–394c). The further equivalence of this god with the mens caeli and the solis imago comes ultimately from Cicero's Dream of Scipio 17: ‘deinde subter mediam fere regionem Sol obtinet, dux et princeps et moderator luminum reliquorum, mens mundi et temperatio, tanta magnitudine ut cuncta sua luce lustret et compleat.’ But one can hardly doubt the influence of the philosophical commentators on this text, such as Macrobius (comm. 1.20.6): ‘[sol] mens mundi ita appellatur ut physici eum cor caeli vocaverunt.’ Macrobius is drawing on an earlier source; compare cognate passages in Firmicus Maternus (mat. 1.10.14): ‘Sol optime maxime, qui mediam caeli possides partem, mens mundi atque temperies, dux omnium atque princeps, qui ceterarum stellarum ignes flammifera luminis tui moderatione perpetuas’, and especially Ammianus (21.1.11):

Ammianus provides us with the direct link, since he suggests a connection between the sun, the mens mundi, the Sibyls, and thunder, just like the EE's solis imago, mens caeli, raving virgo, and thundering hand (I.29). Ammianus’ passage is in a digression on divination, and the close verbal correspondence between him and Macrobius suggests that they are both drawing on a common source.

It is theoretically possible that Ladas’ speech in EE I is a complex, syncretistic amalgam of Stoic and Platonic philosophies with Roman, Greek, and Eastern theologies. But it might just as well be representative of the pop philosophy of the later fourth century we find everywhere in the grammarians — in Macrobius, Servius, Favonius Eulogius, and the scholars whose work is preserved in the scholia to Horace, Lucan, and Statius. In either case, however, it is hardly Neronian.

Ennodius and the EE

In 1986, David Armstrong noted contrary to received opinion on the EE: ‘Nothing in them gives any indication that the emperor praised in them must be Nero; he might be any pagan emperor to Julian or Eugenius for all we know.’Footnote ⁹⁸ This would seem to rule out a connection with any Olybrius — since the family, from the first bearer of the name Q. Clodius Hermogenianus Olybrius on, was known as staunchly Christian.Footnote ⁹⁹ Thanks to the work of Alan Cameron, however, we now know that there ought to be no strict line between ‘pagan’ and ‘Christian’ literature in the fourth century, nor between pagan and Christian authors and audiences.Footnote ¹⁰⁰ Christians often had literary tastes just as classicizing as their pagan colleagues, and in many cases were more devoted to ancient literature. Anne-Marie Turcan-Verkerk has brilliantly shown us how a supposedly pagan panegyrist — Pacatus Drepanius — was himself a Christian poet, who exercised his skill on so pious a theme as the Paschal Candle.Footnote ¹⁰¹ Hence, we have no reason to rule out a priori the possibility that the author of the EE or the Caesar they mention was a Christian. Indeed, as shown above, we have good evidence that the author was well acquainted with Christian texts and idioms.

The later history of the EE is obscure. There is only one secure touchstone: the late fifth- and early sixth-century bishop and poet Ennodius. It is virtually certain that Ennodius knew these poems. I have already quoted the passage on Apollo Python-Slayer above (EE I.32–5):

Compare Ennodius, Carm. 1.3.22–4: ‘eorum Pindareus flumina uicit auus, / docta Camenali cecinit qui carmina plectro / cuius Apollinei nil tacuere chori’; as well as Carm. 1.2.5–6: ‘docta Camenarum coeat pia turba sororum, / offerat arguto pollice quod loquitur’. These two passages are obviously related; the fact that they are each closely modelled on two lines of the EE occurring in the same passage close together cannot be coincidental. The latter is so similar to EE I.35, it should be considered a direct rewriting (doctarum turba sororum ~docta … turba sororum, venerat ~coeat, Camenarum as a gloss); in addition the line following in Ennodius has the same last word as the line preceding in the EE (loquitur and loquuntur).

Likewise, EE I.23 contains a striking phrase ‘to pluck the speaking strings’: ‘seu tibi, Phoebe, placet temptare loquentia fila, / Et citharae modulis primordia iungere mundi.’ Loquor used in the context of making music is not common (cf. TLL s.v. loquor Footnote ¹⁰²), but the collocation with filum is only found in three other authors, all fifth- and sixth-century: Paulinus of Petricordia (Vita S. Mart. 6.105), Venantius Fortunatus (carm. 6.10.3: fila loquacia), and twice in Ennodius (27, preface to carm. 1.8: loquacia fila; and 208, dict. 24, carm. 2.90). Paulinus offers no clear evidence either way, and Venantius probably got the idea from Ennodius, but both the instances in Ennodius have clear connections with the larger context in the EE.

By itself, the fact that Ennodius possibly had access to the EE would hardly be of any interest. But the reference in 27 is interesting for another reason: it is addressed to a figure named Olybrius. The whole of the preface to 1.8 is full of bucolic imagery, recently analysed by G. Vandone, wherein Ennodius presents himself as an agrestis pastor playing the pan-pipes and Olybrius as an inrisor urbanus with an Apolline lyre.Footnote ¹⁰³ Similarly, carm. 1.2, which contains the most obvious imitation of the EE, is addressed to Eugenes, probably the brother of the same Olybrius.Footnote ¹⁰⁴ Since there is solid evidence for thinking that the EE were transmitted as the Bucolicon Olybrii, Ennodius’ use of this material begins to make a lot more sense.

The other factor pointing positively to an identification of the author of the EE with a poet Olybrius is provided by the Montecassino florilegium mentioned above. One of the correspondents is identified as Olybrius in the manuscript, and seems to have quite a bit in common with the author of the EE.

Stemma has a number of meanings, including the generic meaning of ancestry, but it can be used specifically for the maternal line.Footnote ¹⁰⁵ Could this be a reference to Proba? That would provide a nice contrast with maiorum similis beginning the previous line, probably referring to Petronius Probus and Olybrius’ ancestors on the Petronian side. At any rate, just as we have already seen with the EE, there can be no doubt that this exchange is deeply dependent on Ausonius. See his dedication of Ludus septem sapientum to Pacatus Drepanius (XXVI.1-18 Green):Footnote ¹⁰⁶

The links between this preface of Ausonius and Olybrius extend beyond the clear resonances in the poems. In line 14, Ausonius says he will consider the obeli Pacatus applies to his text palmae, not culpae. Fault, or culpa, is what critical marks like the obelus are supposed to indicate.Footnote ¹⁰⁷ This is a sophisticated bit of wordplay: the Greek lēmniskos (L. lemniscus) can be used both for a crown for poetic achievement (cf. Ausonius, epist. 20.6) and as a critical mark. Ausonius is playing here with its Latin analogue palma, as both poetic reward and a critical mark. The use of palma for a critical mark is found in only one place: the collection De notis antiquorum, which follows the exchange between Olybrius and Campanianus, uses palma as a translation of the Greek obelos or lēmniskos. Such a meaning of palma is otherwise not found in Latin literature (it is not in the TLL).Footnote ¹⁰⁸ The fact that they both use this meaning of palma suggests very strongly that Ausonius and Olybrius are working in the same milieu.

Just as this Olybrius was already a well-known poet (maiorum similis), so was the author of the EE, following the standard interpretation of ‘laudatam … chelyn’ of I.18. Olybrius, despite his vertiginously high standing as a patricius, was actively working in the grammatical tradition (‘dux bonus … doce’), just as the author of the EE was. Olybrius may have made much of his descent from Proba (‘stemma poetarum’), and the author of the EE is working in the same vein as the centonist. Both Olybrius and the author of the EE were influenced by Ausonius, who stood for the later fourth century as a model of how to combine scholarship, poetry, political power, social influence, and imperial service. Indeed, Ausonius was himself consul in 379 along with the first Olybrius, Q. Clodius Hermogenianus Olybrius, the son of Proba.

Anicius Hermogenianus Olybrius, Author of the EE?

I have been able to narrow down the date of the two poems to a span of about fifteen years, from 390 to 405. This span coincides with the youthful prime of Anicius Hermogenianus Olybrius, the grandson of Q. Clodius Hermogenianus Olybrius; he was probably born around 375, held the consulship in 395, and was dead by 410.Footnote ¹⁰⁹ According to Claudian, both Olybrius and his brother co-consul, Anicius Probinus, were accomplished poets; ‘Pieriis pollent studiis’, he says in his panegyric for them.Footnote ¹¹⁰ We have at least one epigram by Probinus extant in the Bobbio collection (no. 65), which demonstrates that Claudian's praise had some basis in reality.Footnote ¹¹¹ We also have a poem-letter of Claudian addressed only to Olybrius (carm. min. 40), beginning:

The conclusion of the poem is particularly interesting (23–4):

Here we get the same Caesar-Virgil conceit that marks both ‘Octavian’ and EE I. A coincidence perhaps, but this is not a common topos in late antique poetry. Claudian audaciously casts Olybrius himself in the rôle of Caesar, and himself in the rôle of Virgil.

Almost every other contemporary reference to Olybrius and Probinus mentions their excellence in the liberal arts. Their studia liberalia are prominently featured in the letters of Symmachus addressed to them.Footnote ¹¹² A rhetor named Arusianus Messius dedicated a book to Olybrius and Probinus on appropriate usage illustrated with lines from Virgil, Cicero, and Sallust.Footnote ¹¹³ Given the scholastic features and occasional pedantry of the EE, it would be no surprise if they were composed with the aid of such handbooks.Footnote ¹¹⁴ One possible borrowing is particularly interesting: EE I.13: ‘quid iuvat insanis lucem consumere verbis’ is definitely reminiscent of Virg., Aen. 2.776: ‘quid tantum insano iuvat indulgere dolori’; but Arusianus alone presents a variant of this line much closer to the EE, ‘quid iuvat insano tantum indulgere labori’ (no. 260 Della Casa).

One might hope that the Montecassino epigrams would help in identifying our Olybrius. Unfortunately, this Olybrius cannot be identified, while his correspondent Campanianus can be identified but not dated.Footnote ¹¹⁵ Iulius Felix Campanianus was successively it seems comes ordinis primi et formarum as v[ir] c[larissimus] et spectabilis, urban prefect of Rome as v. c., and finally styled v. i. in the Montecassino poem.Footnote ¹¹⁶ His successor in the first office was one Tarpeius Anneius Faustus, v. c. et spectabilis, also otherwise unknown.Footnote ¹¹⁷ We do, however, have attested in 384 a procurator of Q. Clodius Hermogenianus Olybrius called Tarpeius, also v. c., and these two are the only references to a Tarpeius, v. c., in PLRE. Footnote ¹¹⁸ Accounts which seek to place Campanianus in the middle of the fifth century assume that this Tarpeius must have been a descendant of the older.Footnote ¹¹⁹ Since, however, both Campanianus and Tarpeius are independently connected to an Olybrius, it seems to better fit the evidence to put them both at the end of the fourth or the beginning of the fifth century, in which case we need no recourse to hypothetical descendants. The extraordinarily close links between Campanianus’ poem and Ausonius only strengthen this identification.Footnote ¹²⁰

Connecting these eclogues with the family of the Olybrii fits well with their origin, reception, and transmission. The line begins from Faltonia Betitia Proba, the mother of Q. Clodius Hermogenianus Olybrius, and (probably) the foremost Christian centonist. The family may have been keen to advertise their relationship; see, for example, the note from the lost codex Mutinensis copied by Montfaucon:

Similarly, we have the incipit of the codex Palatinus:

The problem of Proba has been wrangled over for decades now; it seems that the traditional identification is still intact.Footnote ¹²³ All I will point out is that these two inscriptions, which certainly go back to some ancient information, both identify Proba as uxor and mater. I wonder, however, if the purpose of this biography is not to advertise her illustrious connections (a fact superfluous to the poem), but rather for her family to advertise their connection to her.

The name Olybrius persisted through the fifth century, with an emperor Anicius Olybrius reigning briefly in the West in 472.Footnote ¹²⁴ Olybrii maintained their prominence under the Ostrogothic kings, with the consul of 491, the emperor's grandson, and the correspondent of Ennodius (whatever their relationship to each other may have been),Footnote ¹²⁵ and into Cassiodorus’ lifetime with Olybrius, cos. 526.Footnote ¹²⁶ Such a history perfectly matches the fortuna of the EE: composed under the influence of Ausonius, imitated and alluded to by Ennodius, transmitted with Cassiodorus.