1. In the proem to the Iliad the only explicit reference to divine forces is Δι⋯ςδ'⋯τε⋯ετο βουλ⋯(5), in the Odyssey βο⋯ ς ’ Yπερ⋯ονος‘Hελ⋯οιο (8). At the start of Argonautica there is no hint whatever that the human events in Apollonius's narrative are closely and continually linked to divine actions and initiatives and the first reference to a deity, to Athena's role in the building of the Argo (18–19), is almost casual.

2. In the Latin tradition there was ample precedent in Naevius(BP 13, 16, 21 Marmorale) and Ennius (A 22, 175, 291, 457 Vahlen). See Perret, J., Virgile (Paris, 1965), p. 132Google Scholar. That the status-conferring role of the gods was essential even to epic poems on recent historical subjects is clear from Cicero's practice of introducing into his autobiographical epic pieces a concilium deorum and a long address by Urania to the poet (ad Q.fr. 3.1.24, Div. 2.17). Serious doctrine thus lies behind the satiric context of Eumolpus's assertions in Petron. Sat. 118: ‘non enim resgestae uersibus comprehendendae sunt, quod longe melius historici faciunt, sed per ambages deorumque ministeria et fabulosum sententiarum tormentum praecipitandus est liber spiritus ut potius furentis animi uaticinatio appareat quam religiosae orationis sub testibus fides’.

3. On the divine ingredients in Livy's history see Liebeschuetz, W.JRS 57 (1967), 45 ffGoogle Scholar.

4. As in our reaction to the ghost of Hamlet's father or to the statue of the Commendatore in the final act of Don Giovanni. There is a clear contrast with the incidental use of divine activity as a metaphor of physical events, e.g. Bibaculus's, ‘interea Oceani linquens Aurora cubile’ and ‘Iuppiter hibernas cana niue conspuit Alpes’ (MacrobGoogle Scholar. Sat. 6.1.31, Quint. 8.6.17). Both the dramatic effect and the status conferred by such images are independent of belief in the gods.

5. Otherwise the divine machinery merely ensures that the story is kept at a distance from reality, within the region of fantasy. See Quinn, K. F.Virgil's Aeneid: a critical description(London, 1968), p. 305Google Scholar for precisely this view of the gods here.

6. After a discussion of the heavenly bodies (54) and the deification of natural products – Ceres, Liber (60), functional abstractions – Fides, Mens etc. (61), and human benefactors – Hercules, Castor etc. (62). The Fides group has Greek parallels of course, but it was characteristic of primitive Roman religion to proliferate such deities, denned precisely by their prouinciae (a practice ridiculed in Aug. Civ. Dei 4.21) but otherwise quite vaguely conceived (cf. Varro ap. Serv. in Georg. 3.1 on whether Pales was male or female). They were therefore unsuited to pictorial or poetic elaboration, unless like Cupido or Fortuna they were personified in the Greek fashion.

7. Jupiter is the universal father who helps (iuuat) all his children; Juno also the ‘helper’, the atmosphere between earth and heaven; Venus the procreative desire that comes (uenit) to all creatures. The Penates dwell in the penetralia of the home, their name derived from the food store (penus), etc. Etymologies were as important for the Stoics as for Plato and usually just as fanciful.

8. Cf. Lucretius's, explicit use of Bacchus and Ceres (2.655 ff.)Google Scholar and the introduction of Venus into the proem, inspired by Empedocles's use of Ἀφροδ⋯τη for his creative principle φιδ⋯της (Simplic. Phys. 158.24), as the personification of the creative process by which concilia of atoms are formed.

9. Protests to this effect were as old as Xenophanes (Sext. Math. 9.193), whose antianthropomorphism (Clem. Strom. 5.109.3) was, however, not total (Sext. ibid. 144, Simplic. Phys. 23.20).

10. Cicero elsewhere says explicitly ‘at hoc quidem commune est omnium philosophorum … numquam nee irasci deum nee nocere’ (Off. 3.102).

11. ‘pietate ac religione atque hac sapientia quod deorum numine omnia regi gubernarique perspeximus omnis gentis nationesque superauimus’, Cicero boasted on a public occasion (Harusp. Resp. 19); ‘… nostrae ciuitatis quae numquam profecto sine summa placatione deorum immortalium tanta esse potuisset’, declares even the sceptical pontifex maximus Cotta in N.D. 3.5.

12. Especially topical for the Augustan reader were the ceremony of the temple gates of Janus (7.601 ff.) and the ara maxima of Hercules (8.185 ff.).

13. No doubt a feature of Hellenistic κτ⋯σις epics; cf. Callim. Aet. 2.75–83 Pf on the religious legends about the origin of Zancle. For Apollonius see Powell, J. U.Coll. Alex. 4–8Google Scholar. But the blend of secular and sacred was so fundamental to Roman political institutions that the combination of Homeric divine machinery and Roman cult must have been prominent already in Naevius and Ennius.

14. Cic. N.D. 2.6; cf. Liv. 2.20.12.

15. The quasi-scientific rules governing divination, which enabled them to be assimilated to more rationalized theologies like Stoicism, distinguished the two groups very sharply: portents were interpreted by the highly systematized diuinatio artificiosa; dreams, which were significantly grouped with furor uaticinantium and thus more closely related to the condition of ⋯νθουσιασμμ⋯ς, were the subject of diuinatio naturalis (Cic. Div. 1.11, 34 etc., with A. S. Pease's commentary).

16. Lines 987–8; cf. Nisus's words to Euryalus (Aen. 9.184–5): ‘dine hunc ardorem mentibus addunt, / Euryale, an sua quoique deus fit dira cupido?’, followed a couple of lines later by the forthright ‘mens agitat mihi’. This is not a clue to Vergil's concept of the gods’ role in events but part of his characterization of Nisus, to be set beside, for instance, the agnostic tone of ‘Iuppiter aut quicumque oculis haec aspicit aequis’ (ibid. 209).

17. Equated in classical times with Ἑςτ⋯α and the μεγ⋯λοι θεο⋯ of Aegean religion (see nn. 21, 22), they are more likely to be indigenous cognates than derivatives of these.

18. The incident illustrates the fragile boundary between internal and external phenomena. Hector appears in a dream to Aeneas but the sacra that he passes to him belong to the external world and are thereafter carried by the hero wherever he goes. By contrast the theophany of Venus, ‘aetherios inter … nimbos / dona ferens’ (8.608 ff.), though private to Aeneas, is an external event, the visit of a mother to her son. Hence the reality of the arms given to him poses less difficulty; but the incident is correspondingly harder to demythologize than either the Hector-dream or even the appearance of Venus to Aeneas at 1.314 ff.

19. Dido also has her Penates (1.704); so has Evander (8.123), and the early morning sacrifice made jointly by Evander and Aeneas to the local Lar and Penates (8.543) underlines their association with the intimate hospitality to strangers in the home. That their precise function was uncertain is clear from Cicero's brief account of them (N.D. 2.68).

20. The difficulty of reconciling in somnis (151) with nee sopor (173) already troubled ancient readers, as Servius's note shows. Perhaps it is an indication that Aeneas himself, who is telling the story, was uncertain.

21. That their identification was debated in classical times is clear from Macrobius (Sat. 3.4.6–13), who cites (1) Nigidius's opinion that they were Trojan versions of Apollo and Neptune, both of whom appear independently in the Aeneid; (2) Varro's opinion that they were originally from Samothrace and that ‘qui diligentius eruunt ueritatem Penates esse dixerunt per quos penitus spiramus, per quos … rationem animi possidemus’. (3) Hyginus's view that they were θεο⋯πατρῷοι ‘gods of our ancestors’ or ‘of our fatherland’ (cf. patriosque Penatis at 2.717). In the aedes deum penatium on the Velia, which was restored by Augustus (R.G. 19, 36), they were depicted as two seated youths in military dress, according to Dionysius of Halicarnassus (A.R. 1.68), who like Varro identifies them with the Samothracian μεγ⋯λο⋯ i.e. the Καβἲροι Moreover, there is evidence in coinage of the 1st century B.C. to suggest that this iconography owed something at least to the Dioscuri; see Lloyd, R. B., AJPh 77 (1956), 38ffGoogle Scholar.

22. The line-ending magnis dis is Ennian (A 201 V). It is not clear whether the magni di are here distinct from the Penates (cf. magnos… penatis at 9.258). Macrobius (loc. cit. n. 21) cites the view of Cassius Hemina that the di magni were the Samothracian θεο⋯ μεγ⋯λοι and then adds ‘eosdemque Romanorum penates proprie dici θεοὺς μεγους', citing Aen. 3.12, which he presumably understood as ‘and with the Penates, the great gods’. It has been conjectured that dis magnis was actually inscribed upon the statue of the Penates in the Velia temple. See Williams, R. D.Aeneidos Liber Tertius (Oxford, 1962), pp. 54–5Google Scholar.

23. The indigenous Neptunus was defined solely by his prouincia (cf. n. 5), as was Portunus– ‘deus portuum portarumque praeses’ (Varro ap. Schol Vet. Aen. 5.241; cf. Cic. N.D. 2.66) – who is also assigned to a Trojan context (5.241).

24. Nautes owes his skill as a counsellor explicitly to Tritonia Pallas (5.704).

25. This Italian cult is, however, carefully linked to the Greek Artemis by the strange tale of Hippolytus's resurrection.

26. Hence her brother Apollo's rejection of the second part of Arruns's vow at 794–8.

27. Cf. Poseidon's hostility to the Greeks during and after the Trojan War, aggravated by Odysseus's treatment of Polyphemus (Od. 1.68 ff.).

28. The Olympian action at a distance in both instances is typical, and it is an oversimplification to interpret the two agents Allecto and Iris as ‘scarcely more than a personification of Juno's suffering’ (Putnam, Michael C. J., The Poetry of the Aeneid(Cambridge, Mass., 1965), p. 88)Google Scholar.

29. Perret, J., op. cit. p. 131Google Scholar, plausibly suggests that in the traditional version of the story Turnus was indeed motivated by plain jealousy, as the Latins were by resentment towards the alien usurpers.

30. An Italian Achilles, also born of a goddess (6.89–90; cf. 9.742,10.76), with a distinguished human pedigree (7.56), including an Argive connection (371–2). Subsequently we learn of his outstanding physical prowess (783 ff.) and his piety (9.23–4, 10.620; cf. 7.438–9). Every detail combines towards a re-creation of the Homeric ideal hero, to set against the Vergilian Roman ideal embodied in Aeneas.

31. The line describing the actual intervention, ‘Irim de caelo misit Saturnia Iuno’ (9.2) echoes – again significantly – the descent of Iris to the Trojan women at 5.606, from a context where the goddess's machinations were frustrated.

32. It would be interesting to apply the external/internal interpretation of divine interventions used in the present study to these and other passages in Homer.

33. Who later provides explicitly the content of the simile describing Aeneas's first glimpse of Dido (498 ff.).

34. There is a precedent in the Odyssey, where Telemachus's despondency is abruptly ended by his encounter with the disguised Athena (1.113 ff.) and he assumes a confidence and maturity that amaze even his own mother (360–1) as well as the suitors (381–2).

35. 567–88 contain enough unusual features of style and language to justify the suspicion that they are, as they stand, not Vergilian but rather the work of an early interpolator (presumably prior to Lucan 10.59), one of the many (Suet. Vita = Donatus Praef. Buc. 41) who were unable to resist the challenge of the uersus imperfectos in the poem. However, the content of the passage is surely an astonishing choice for an interpolator to have made, unless he had some evidence of Vergil's intentions. The whole of this section of the poem still awaited its final revision at Vergil's death, as witness the half-lines at 468, 614, and 623, and it may reasonably be conjectured that Vergil had left enough lines and fragments (there are many phrases that certainly sound authentic scattered through the suspect passage) to leave his intentions clear. The disconcerting picture they gave of the hero would have made Varius grateful to have the excuse of their fragmentary character for excluding them from his edition. The same considerations would explain why the passage as we have it, cobbled together from the poet's disiecta membra, was not admitted into the authorized version that is reflected in the manuscript tradition of the poem itself and in the abundant citations of the Aeneid in the grammarians, but surfaced only in Servius's preface to his commentary and in the Servius Auctus note on 566. For a detailed discussion of the passage and a very different conclusion see Goold, G. P., HSCP 74 (1970), 130–68.Google Scholar. Austin's, R. G. view of the passage, Aeneidos Liber Secundus (Oxford, 1964), pp. 217 ffGoogle Scholar., is close to the one proposed here, though not all Austin's interpretations of detail are acceptable; e.g. sceleratas (576) is indeed ‘a piece of selfcondemnation’, the judgement of Aeneas the narrator on Aeneas furens (595).

36. infelix in 712, as elsewhere, probably has both the senses ‘barren’ and ‘ill-fated’. Dido had borne Sychaeus no children. At 4.327–30 she refers poignantly to the fact that, though she and Aeneas have lived together as husband and wife, she is not pregnant. This is significantly rare, if not unique, in classical myth and legend, where it is normal for even a single act of sexual intercourse to produce offspring.

37. Iulus-Ascanius thus plays a crucial part in both the beginning and the end of the affair.

38. Otis, B., Virgil: a Study in Civilized Poetry (Oxford, 1964), p. 226Google Scholar, sees Jupiter's interference as an external parallel to the stirring of Aeneas's conscience; but this reduction of Vergil's divinities to something like their Apollonian role ignores the subtle detail of the Vergilian narrative.

39. Otis, op. cit. p. 94Google Scholar rightly sees amor andfuror as threatening pietas andfatum, but in imposing this antithesis on Dido and Aeneas respectively he reduces them to Morality personae and diminishes the essential tragedy of their relationship.

40. See Williams, R. D., op. cit. p. 134Google Scholar.

41. Prompted by the continuing strife between the goddesses, it is a far cry from the serenity he showed (1.254 ff.) in comforting the despairing Pöschl, Venus. V., Die Dichtkunst Virgils: Bild und Symbol in der Aeneis (Innsbruck, 1977), pp. 16 ff.CrossRefGoogle Scholar, makes altogether too much of that serenitas. For it is only one facet of Jupiter's character, as it is of the Roman ideal that Poschl seems to regard the god as symbolizing allegorically. Not for nothing did Jupiter wield the least serene of punitive weapons (1.230); on occasion he could exhibit as much saeuitia as Juno (cf. 1.4 with 11.901).

42. By contrast in tuorum fata (1.257–8) and fatis contraria nostris/fata Phrygum (7.293–4) the genitives and nostrisare all ‘objective’, viz. ‘the fate that has been assigned to your people’ etc.

43. The construction of the sentence is odd. Helenus can hardly be forbidden by Juno to divulge what he has been barred from knowing. So we must either understand something like etiam sisciret with fari (‘Juno forbids him to speak, even if he did know’) or take -que twice over, first as coordinating prohibent and uetat and then scire and fari (‘The Fates prevent Helenus and Juno prohibits him from knowing and speaking the rest.’)

44. Cf. the Stoic definition of τ⋯χη as ⋯δηλος αἰτ⋯α ⋯νθρωπ⋯νψ λογισμῷ (Stob. Ecl. 1.7.9). A good illustration is Palinurus's use of forte (6.349) in describing his own death, which, he does not know but we do, has been contrived by a god.

45. The classic Stoic posture of resignation; cf. Sen. Ben. 2.18.8: ‘ducunt uolentem fata, nolentem trahunt’.

46. In fact Fortuna represents the impact of Fate, for prosperity or the reverse, upon those who are affected by it but cannot comprehend the broader providential plan.

47. The phrase is from Tac. Ann. 6.22, where a view of Fate (probably Stoic) is cited that leaves us with electionem uitae but, once our choice is made, commits us to certum imminentium ordinem. We may choose our role in the drama but the part is already written and we cannot conjecture much of its content in advance.

48. Cf. the doctrine that magna di curant, parua negligunt (Cic. N.D. 2.167), where di is virtually synonymous with Fatum. Unlike the more austere concepts of Fate this doctrine did leave some freedom of action to the individual, provided of course that he was left unmolested by divine forces.

49. It is not clear whether Aeneas's attribution of the prophecy to Anchises (7.122–3) is a mark of the unfinished state of the poem or a subtle indication from the poet that the hero's memory of his early wanderings is still dominated by his father.

50. The phrase again is from Williams, R. D.; The Aeneid of Virgil Books 7–12 (London, 1972), p. 432Google Scholar.

51. More typical is the god's motivation of another Etruscan captain to attack Camilla's forces: ‘Tarchonem in proelia saeua/suscitat et stimulis haud mollibus inicit iras’ (11.727–8). For here there is no suggestion that if left to himself he would have joined the battle at this moment.

52. The ambiguity is very potent, for instance, in Venus's irate question to Jupiter (10.34–5): ‘cur nunc tua quisquam/uertere iussa potest aut cur noua condere fata’?

53. Perret, , op. cit. p. 130Google Scholar, follows Boissier in seeing ira Iunonis as causing all the important incidents in the poem, but this ignores the love affair at Carthage. Pöschl, , op. cit., regards Juno first (p. 14)Google Scholar as the mythological personification of the historical power of Carthage. This could certainly apply to Ennius's Juno (see A. 291 V, cited by Servius on 1.281). But it accords ill with Vergil's account of Juno's reconciliation to Rome's illustrious future (12.827), which would leave the other addressees of Dido's famous prayer (4.607–10) with the task of implementing its demands (622–9). Nor does Jupiter's reference to the future wars with Carthage (10.11–15) necessarily imply Juno's involvement. Later (p. 17) Pöschl converts the goddess into a divine symbol of the demonic forces of violence and destruction. But the violence she is responsible for comes from her opposition to Fate and is far from being her monopoly in the poem. For Otis, , op. cit., p. 309Google Scholar, Juno is the external stimulus corresponding to and exciting the inner response; this seems once again to reduce the deity to Apollonian redundancy.

54. Likewise Vergil's contemporaries the elegists continued to fill the gap in the aetiology of love with the traditional gods of erotic literature: the lover who acts unpredictably and out of character is afflicted with dementia, insania, the result of being vanquished and possessed by Venus or Amor.

55. The view that the gods of the Aeneid were allegorical figures of psychological phenomena, symbols that the reader decodes as he goes along, was first and most influentially elaborated by Heinze, R., Virgils epische Technik (Stuttgart, 1914 ³), pp. 291–318Google Scholar.