In the generation since Horsfall first submitted his commentary on Aeneid 7 as a doctoral dissertation he has been generous with publication on many specific issues and topics related to Virgil’s text: three books, Roman Myth and Mythography (B.I.C.S. Supplement 52, 1987) with J.N. Bremmer, L’Epopea in Alambicco (Naples 1991), and A Companion to the Study of Virgil (Leiden 1995); many articles in Greece and Rome, in Bulletin of the Institute of Classical Studies, and other journals, including an important study, “Camilla, o i limiti dell’invenzione” in Athenaeum 66 (1988) and countless entries on Italian names, places and things in Enciclopedia Virgiliana. But the hope that he would publish a complete commentary on Book 7 always remained and has now at last been gratified. It is an immensely learned work, often shrewdly observant about epic realities (sailing, weaving, horse-trappings, slings and boomerangs), at times witty, at times exasperating. Who else could have written a work of such depth and breadth? Not your reviewer, who has “taught” this book of the Aeneid many times without resolving a large number of its problems, and whose own commentaries have always aimed lower and probably fallen short.
But Horsfall has also given hostages to fortune in two fiercely critical reviews of the existing English language commentaries: R.D. Williams’ single volume Aeneid VII-XII, and C.J. Fordyce’s posthumous Aeneid VII-VIII. Reviewing Williams in JRS 64 (1974) 274-6, he complained that there was “nothing on the poem as allegory, nor on the massive typological links between Aeneas and Augustus, nor on Virgil’s imagery and the thematic unity it creates.” Understandably he regrets what he calls the “verbalistic approach” and Williams’ stress on “influence at the cost of antecedents.” Writing as he did for both schools and relatively unsophisticated undergraduates, Williams naturally had to bring out the full force of Virgil’s lexicon and phrasing; he could also reach more of his (often Greekless) audience with allusions to Ovid, Shakespeare or Milton than by quoting Homer, but certainly Horsfall’s new commentary does full justice to Greek “sources” from Homer to tragedy to Apollonius. Horsfall’s fully justified ideals are conveyed in his stress on “curiosity to be roused, enquiry to be stimulated, criticism to be excited.” He is less patient with Fordyce ( C.R. 29 (1979) 219-23), rightly finding him uninterested in bibliography (“not…concerned to discover or record the views of others”) and narrow in comparison with his peer Roland Austin, whose commentaries still give pleasure and understanding. He goes on to fault Fordyce for indifference to Realien, topography, Roman religion, Greek and Roman mythology and “the central legend” of Aeneas: “we can at least try whenever possible to explore fully [Virgil’s] allusions, resonances and sources. No task is more important to our evaluation of the Aeneid.”
All these are areas on which Horsfall has spent years of research and discussion with colleagues, and his publishers have rightly given him space to work at the highest level for the most demanding readership. With the commentary he provides an introduction, his own text (without apparatus), an extremely helpful translation, and three indexes, of English, Latin and Names (in, or implied by, V.’s text). The English index (sic) covers topics and authors; as a sample I give four consecutive entries: hunting, Hyginus, hymn-language and hypallage (15 references).
Horsfall’s introduction covers: 1) Structure, given in summary form. 2) Sources: epic and Kreuzung der Gattungen, which outlines Virgil’s use of Homer, all three tragedians, Callimachus (“repeatedly present”), Ennius (about whose role he has reservations), Cato, (no certain use of Origines), Sallust’s Histories, Varro (both for early Latium and the catalogue) and generic imitations of bucolic and Hellenistic epigram. 3) Language, Grammar, Syntax, Style, Metre. This section surveys V.’s lexical debts to Ennius, Lucretius, Catullus and Cicero poeta, suggesting that Cicero and Virgil were drawing their diction from the same sources: Ennius and tragedy. It has less to say about Virgilian figures (but these will be well handled in the commentary) and nothing (as he indicates on xxii) about metre. 4) Text, discusses problems such as interpolation, repeated lines and half-verses, transpositions (for which he sees no need in book 7), and the relative dating of the capital manuscripts. Significant divergences from Mynors’ Oxford Classical Text are listed on p.xxxi. In a special address ad lectorem, Horsfall explains that he has printed his text with marginal symbols O(rthography), P(unctuation) and T(ext) to mark his discussions in the commentary. Many of these provide helpful justifications of readings to which we are accustomed but where we need to be shown the grounds for rejecting well-attested variants.
This section also explains the reference system, as is essential in a commentary so comprehensive in bibliography. Over a hundred major works are listed in the bibliography which precedes the text. Other material is either cited at the end of or (as Horsfall would say, “alas!”) during a comment, or printed with a reference in bold type to the lemma where the full reference can be found. These different modes can be a source of confusion and delay. Apart from reviewers, few will read through a commentary continuously, and even as I did I would stumble over an author’s bare name for which the reference itself was up to seven pages earlier. Finally (p. xxxiii) he cites Encyclopedia Virgiliana with a Michelin system: a bare reference does not recommend (but the modestly anonymous author can be Horsfall himself), a surname alone (he chooses as his example the vox et locus nihili Della Morte V* 962) tolerates, but a reference with initials or first name actually commends the article.
First to review the textual changes: a period and new paragraph separate 58 sed variis portenta deum from 57; 351 vipeream spirans animam,(not inspirans because of Virgil’s preference for unprefixed verbs and because the infection is still to be described: but Horsfall allows reasons for choosing the compound); 430 laetus in arma para (not arva : the repetition from armari 429 is no problem, and arva are out of place here); 571 levavit (instantaneous relief, not the generally transmitted levabat, probably a secondary error resulting from a scribe writing b for v); 587 pelago (only in one major manuscript, V) emphatically repeating the dative pelago of 586; and in 588 circumlatrantibus with compound verb instead of separate adverb—suggested by Seneca’s use of the same compound and Virgil’s own regular anastrophe of the adverb (as immediately after in 589-90 circum /…fremunt). At 684 Horsfall prints Heinsius’ pascis, arguing that anaphora of quos requires the second person to share the syntax of the vocative Amasene pater. He is surely right in printing ex agmine at 703 rather than examine, which would evoke a swarm of bees without contextual support for the image. At 727 he follows Delz in construing Sidicina…aequora as accusative parallel to acc. Cales (775 poenigenam I reserve for later comment). Twice he offers a rare toponym: Hülsen’s conjecture Amitina for Atina at 630, on the grounds that Atina is too far away, and Sturae (801) for Saturae, in view of the nearby river Storas and the place Astura. Perhaps it would be appropriate, too, to mention here two passages of similar form where he has suggested a felicitous new interpretation without changing a letter: at 343 Laurentis tecta tyranni / celsa petit and 807 sed proelia virgo/ dura pati, he reads the epithet as f. sing., describing Allecto and Camilla, instead of the more obvious acc. pl. with tecta, proelia.
Still focusing specifically on the text I noted helpful discussions on unfinished sections (248), duplications…”the result of haste or incompleteness” (259-73), the possibly unfinished and certainly inconsistent 664-9 (how can Aventinus be both mounted and on foot, both swirl his lionskin hood and cloak and wear it?), and the two similes echoing Homer’s catalogue at 699-702 and 703-5. Horsfall suggests that since, unlike the Homeric similes, they seem to duplicate each other, this is an obvious case of alternative drafts, and in view of the incomplete 702, 699-702 (already used at G. 1.388) will be the later version.
It would be fatuous for me to single out individual comments for admiration, but let me offer a sample of valuable or stimulating discussions. 9: candida…luna, and 25-36: the need for Aeneas and his men to avoid Circe and the economy of the narrative which is hastening them to their destination explain why Virgil must provide a full moon and following wind to carry the men past Monte Circeo; it also motivates their hunger on arrival, for they had no overnight pause and evening meal. 37-45: the antecedents of the proem are analysed; Iliad, Odyssey, Apollonius (for Erato but also for the landing at the Tiber mouth recalling the landing that ends AR book 2), Aeneid 1 (7.41 evoking 1.8) and historiography (for the language of rerum/…status (his punctuation) and of causality). There is an admirable note at 107-47 on the omen of the table-eating, with a new take on the inconsistency with Book 3 (“V. did not narrate the episode in which Anchises unravelled the latent ambiguity in Celaeno’s prophecy”) backed up by 123 on the tradition of Anchises’ prophetic powers and 134 on Aeneas’ gratitude as reason for including Anchises in the prayers: the setting of the whole episode is made vivid by recognising the boyish speech of Ascanius and the time (dusk) evoked at 138 by the prayer to night and the rising stars. And Horsfall is alert to the inconsistency that supplies fresh wine for libation (134) and festive food (146), where before they had had only wild fruits and the mysterious mensae —(which he diagnoses as pitta).
I should also mention Horsfall’s cumulative analysis of V.’s complex picture of the Latins (and Italians too) as both peaceful and warlike, “innocent and not innocent” (46, cf. 183); “Though it is Trojan, not Latin religion that will prevail in the eventual fusion…V. represents the Latins as godly in their own right” (153). Introducing the catalogue he argues (p.421) that far from being sentimental about the Italians V. is blunt about their violent and lawless nature: in 681 Horsfall alliteratively describes the Praenestines as “brave but backward, simple but savage, hardy heroic and hopeless”—all this developed from the single word agrestis ! But he rightly refers the reader to his detailed study of V.’s Italians in the self-praises of Numanus Remulus ( Latomus 30 (1971) 1108f, now in Oxford Readings 305ff.)
One characteristic of all Horsfall’s work on Virgil is his brave recognition that the poet was not much interested in consistency. It is something to consider in more detail when we come to the Catalogue. But Horsfall does not limit himself to Italian lore: compare his valuable notes drawing attention to V.’s knowledge and use of Greek tragedy, e.g. 323-40, on Allecto and the debt to Euripides and Aeschylus, and 373-405 on Amata’s oreibasia (a mosaic reflecting a dozen phrases and half lines of Eur. Bacchae among shards from other poets).
Perhaps it will be most helpful to illustrate Horsfall’s skills (and problems) as a commentator from the sequence beginning with the threefold assault on Latinus (583-600) before the opening of the Gates of War (601-20). Horsfall pinpoints the moral keys to events: infandum contra omina bellum,/contra fata deum and alerts us to the problem in perverso numine, which can only make sense only if this numen is Allecto. Horsfall has already drawn attention (46) to Latinus’ strength in passive opposition, underestimated by modern readers. At 586-90 resistit, he notes “the whole simile is to be read in the same spirit [of resistance]…, while resistance is excluded from the narrative.” Here his text of 587 with epanalepsis of (dative) pelago, strengthens the element of resistance within the simile, reinforced by sese mole tenet, and refunditur. At 591, the apparent collapse of Latinus’ refusal, Horsfall notes that “the careful reader will discover that more is to follow.” Yet I would like to have seen some comment on frangimur heu fatis beyond the shrewd recognition of frangi here as shipwreck. Latinus seems to contradict his own awareness of the actual will of fate while he still keeps his sense of the wrong being done (595-7 ipsi has pendetis sacrilego sanguine poenas…te, Turne, nefas, te triste manebit/supplicium). Surely this speech and his foreshadowing of their punishment by the gods after his own death must be, rather than follow, his invocation of the gods ( testatus, 593 with present sense?). There is the same tension between Latinus’ flagging strength and continued sense of divine will in rerumque reliquit habenas, as when he reenters the action at 11.231 deficit ingenti luctu but immediately summons and advises a council of war, until it is disrupted by news of Aeneas’ approach.
Now for the Gates of War. With his knowledge of Varro’s res divinae and related antiquarian material Horsfall is at his best on ritual (compare 81-106 on the disparate ingredients of Latinus’ incubation, 154 on the velamenta of the Trojan envoys, and 170-91 on Latinus’ strangely Roman palace/temple and the dual nature of Picus, his statue garbed as a Roman priest-king but mythically and magically a victim of Circe). The twenty lines on the geminae Belli portae foreshadow the scene of Octavian’s triumph on the shield and come as close to the epic Augustus wanted as V. would ever venture. The ritual of opening was itself in abeyance until Augustus revived the practice of closure for peace (Horsfall 611 and 612). Here the Senate’s decision is implemented by the consul, clad in the same ritual garb ( Quirinali trabea cinctuque Gabino) that V.’s readers have just visualized on the statue of Picus at 187-8 (Horsfall ad loc.), I would differ from his reading only in stressing the evocation of the republic itself (rather than Augustus’ toy republic), colored as Augustan by the choice of enemies and just warfare (those Parthians, those standards) in 606-7. It takes only six lines from 616 hoc et tum for Juno to usurp Latinus’ ritual function and for preparations to flare into war. (I wish Horsfall had followed the elements of sound and sense that mirror 615 aereaque adsensu conspirant cornua rauco in the sound of 623 and sense of 628.)
Horsfall has meditated long and often over the Catalogue (especially in his investigations of Virgilian invention in Alambicco and Mythos in Mythenloser Gesellschaft [ed. F. Graf, Tübingen 1993] 131-41). At 641-817 he offers eight pages of introduction subdivided into nine sections: on the Homeric and post-Homeric inheritance (relating the warriors listed to those that feature in action); on geographical distribution (arbitrary); on the component elements selectively used for each contingent; on the adapation of the Homeric similes at 699-702 and 703-5; on V.’s antiquarian reading and specifically his use of Varro Res Humanae 11 (sections v and vi); on foundation legends and the ontological status of V.’s leaders (discussed below); on other Varronian material such as Varro de Gente Populi Romani, suggested by many references in de Lingua Latina; and finally on the evocative or emotional aspect of V’s local names, on which Horsfall is regretfully skeptical—”a lot…are obscure to the point of unidentifiability.”
Let us return to the names of the leaders. Horsfall analyses them into three groups: those deriving from the Aeneas-legend, Turnus and Mezentius and Clausus from Roman gentile tradition; those (names, not person-alities) attested independently of Virgil—the founders of Tibur and Praeneste (see Horsfall / Bremmer), Halaesus and Messapus and finally the many names unattested elsewhere: Aventinus son of Hercules, Oebalus the Campanian, Ufens, Umbro, Virbius filius and Camilla. As with Horsfall’s separate investigation of phrases like ferunt fama (765), his careful investigations end in (his and our) frustration. Facing the traditions about Halaesus, son of Neptune associated with the Faliscan area, and Messapus, the South Italian from Ennius’ homeland, Horsfall asks why V. switches their land and parentage and can only conclude (p.422-52) that V. wrote out of “geographical indifference rather than Reine Willkür.” There is another kind of question to ask here, I feel. If V. is happy to use legendary figures (with modifications to taste) and to invent his Camilla, his Marsic snake-charmer Umbro, and a son for Virbius, why does he take over a category with an existence in tradition (and we may add Oebalus, whom he has moved from Tarentum to Capri, and the Alban Aventinus, become a proto-Roman) but refuse to keep their traditional associations? This is wilfulness, but it must have a motive. Might I suggest that in the case of Aventinus V., who knew the variant tradition (D.H.1.32 and Servius on Aen 8.51) that Pallas, our son of Evander, was the child of Hercules and hinted at it in 10.464f, was creating a substitute Roman son for Hercules, and naming his mother as a substitute for Ilia?
Now for Virbius, the one figure with a clearly attested non-Varronian source, where we have Servius’ confirmation (on Aen. 7.778) that Callimachus told his posthumous Italian life as an aition for the taboo on horses at Aricia. But “the story may have been taken over by Varro” (p.495). V.’s 22 lines devote 765-76 to Hippolytus’ tragic death and resurrection by Asclepius, the punishment of Asclepius and Diana/Trivia’s rescue of her devotee to spend an inglorious life ( aevum) in her grove under Egeria’s protection with a new identity as Virbius; only 778-80 retail the horse-taboo. But the section opens and closes with the Italian leader, the entirely unexpected son: the first four lines (” Hippolyti proles rather suggests a personal sense for mater too,” says Horsfall, with my sympathy) make him son of a nymph Aricia, reared by Egeria in the grove of Diana; the last two lines return to him as he hastens ( ruebat modifies the surprisingly deliberate sounding impf. exercebat) to join the warfare in his chariot, undeterred ( haud setius) by his father’s fate. Here for once ferunt fama (765) refers to a specific source, known to us by the chance of survival in Servius. How many others were there unknown to Servius?
Let us return to Oebalus. At 735 Horsfall speaks of Virgil “improvising a background for Oebalus out of skimpy and diverse materials…V. simultaneously distances himself.” (cf. Alambicco p.127 [my translation]) ” Fertur neither guarantees nor denies what we have just said; rather it indicates the lack of commitment of seriousness or interest on V.’s part in arguments of this type.”) But isn’t it likely that he transferred to Oebalus (as to Halaesus or Messapus) a fama about the parents of another mythical founding hero?
I do not want to underestimate the value of Horsfall’s extended comment on Hippolytus/Virbius and his son, but I would like now to illustrate from this sample some of the difficulties of using Horsfall’s commentary. It is here that he adopts Servius’ variant reading poenigenam (born of Apollo’s punishment of Coronis), a good choice, but it seems to have generated one of the two displacements in Horsfall’s text, for fulmine is printed as the last word on 772, as, oddly, is litora in 763 instead of opening 764—accidents unparalleled in the text as far as I can tell. More to the point is his note on 778, the third reference to these luci in this passage (cf. 763, and secretis sedibus…nemorique…in silvis, 774-75). At 763 he refers us back to 95, which simply refers back to 82, calling Albunea a nemus, near synonym of a lucus. The first note on lucus is at 29, where Horsfall refers forward to almost twenty others including 778,780. But only at 778, on the archaeologically identified precinct of Aricia, does he note that “clearing” would often be a better translation than “grove” and that Castagnoli sees lucus as a technical term for an extra-urban sanctuary.
Another significant and harmful postponement is the note clarifying the application of Laurens, Laurentes which comes unprompted by a textual verbal cue at 151, followed by the caution against the non-existent *Laurentum, provoked by ante urbem 162, and more comment at 171 Laurentis…Pici. But the name has already occurred at 47 and 63, where the reader is simply referred to Horsfall’s article in E.V. Incidentally it is a relief to see that after his three previous treatments (most recently in RMM) Horsfall has settled (see 209 Corythi) on a realistic acknowledgement that this form could just as well denote a place or an ancient king and cannot be pinned down as Cortona or Tarquinia: there is “no reason to believe that Virgil had any clear ideas about where that was.” The important note at 778 exemplifies another problem of consultation arising from Horsfall’s substantial bibliography and his tendency to insert references and comments in mid-sentence instead of at the end of the sentence or comment. Apart from simple cross-reference within book 7, it contains six parentheses including one of more than 2 lines. (The very first note of all, on tu quoque, contains thirteen of these parentheses, six of them extended.)
Intrusive bibliography and delayed references are recurring problems, as can be some of Horsfall’s single letter abbreviations either to words in the lemma or to manuscripts, but I will mention below only a relatively few minor confusions or difficulties. The other crying need was for a local map to help us with Horsfall’s intimate command of modern regional place names in Latium. Having once tried to place the Italian leaders onto a map of Italy, I realize this is conceptually impossible. But those of us who cannot wander through the Campagna need to locate Tor Tignosa, Monte Sant’ Angelo and their like. Even the great forthcoming Barrington Classical Atlas has a detailed map of Rome, but no separate map of Latium.
Bibliography, p.xli. Nelis’ Virgil and Apollonius, consulted by Horsfall in typescript, still awaits publication.
55. Aeneas ante alios pulcherrimus omnes (4.141) should be quoted instead of the bare ref. under ante alios followed by an unreferenced allusion under pulcherrimus.
223. The Hardie reference should be to CI (see bibliography).
342-3. Clearly Horsfall means “noun precedes adjective” (but he goes on to make the bold suggestion that celsa refers to Allecto).
419. On female priests (what Horsfall and I call priestesses) note that Ov. F. 5.573, far from being female, is Julius Caesar. And we need a cross-reference to 659 Rhea Sacerdos.
462. Something has happened to the text, probably the loss of a line (p.312 line 5: “Tu.’s orders group is instant and raging”).
612. There is an unexplained reference to “a timeless Roman magistrate for Pomathios.” If this is a modern reference I have not been able to find it by working back through the commentary. Could it be a misprint for the historian Promathion cited at 679?
The triviality of these queries only shows the predictably immense learning and the impressive understanding of V. offered by Horsfall’s commentary. However great the intermittent difficulty of threading through the maze of parentheses to follow his argument, it is enormously worthwhile. This is a commentator who loves his author for better or for worse, even when the poet is at his most enigmatic. I am deeply grateful that the somewhat tantalizing alembic has matured into this rich and rewarding commentary.