2. This is the usual sense of the phrase ‘ritual calendar’ in this paper. Occasionally I use ‘calendar’ in the perhaps more familiar, but narrower, sense of the written tabulated version of the cycle of festivals (see below, n. 5); the context always makes it clear where this is the case.

3. For the standard view on each of these festivals, see Scullard, H. H., Festivals and ceremonies of the Roman Republic (1981) 94-5, 108-10, 177Google Scholar. The image of primitive peasant life comes across clearly in Bailey, C., CAH VIII (1930) 435–9Google Scholar and Ogilvie, R. M., The Romans and their gods (1969) 70–99Google Scholar.

4. Perhaps the most characteristic (and strikingly successful) of Dumézil's studies of Roman ritual is his Fêtes romaines d'été et d'automne (1975), which elucidates several particularly ill-documented festivals by placing them in a broader Indo-European context. On this book, see now Scheid, J., ‘À propos de certaines fêtes d'été’, AION (archeol) 2 (1980) 41–53Google Scholar. He picks up the notion of the ‘peasant society’ underlying Dumézil's explanatory framework and asks – pointedly – ‘which society, where and at what time?’.

5. See, for example, the entry in the Praenestine Calendar (dated 6-9 B.C.) against the day of the Robigalia (25 April): ‘Feriae Robigo via Claudia ad milliarium V, ne robigo frumentis noceat’ (‘Festival to Robigus at the fifth milestone on the Via Claudia, to prevent blight (robigo) harming the crops’). For the full inscribed entry, see Degrassi, A., Inscriptiones Italiae XIII (1963) 130-1, 448–9Google Scholar.

6. Dumézil does sometimes attempt to demonstrate the continuing significance into the historical period of the ‘original meaning’ of a festival. So, for example, he claims that it was not by chance that Mark Antony chose the Lupercalia (in Dumézil's view a festival concerned with the primitive kingship of the city) to offer Caesar the royal crown; the political importance of the scene depended on the spectators knowing and understanding the original regal associations of the rite. Yet, even in this case, Dumézil admits the disjunction between his suggested religious meaning and contemporary life: ‘our information dates from an era in which we can no longer hope to obtain a complete and systematic view of rites which had long ceased to correspond to religious and social reality’. See Dumézil, G., Archaic Roman religion (1970) 346–50 (quote, p. 350)Google Scholar.

7. This particular ritual element forms part of the festival of the Parilia discussed below, pp. 3-11.

8. I refer here to the festival of the Fordicidia (15 April); for a brief account, see Scullard, , Festivals 102Google Scholar and Dumézil, , Archaic Roman religion 371–4Google Scholar.

9. Ov. Fast. 4. 725 (Parilia); 4. 905-9 (Robigalia). It is, of course, quite unknowable whether Ovid ‘really’ participated in or observed these festivals.

10. Note, for example, the bafflement of R. M. Ogilvie, who writes in relation to the Lemuria, a rituall of the dead which involved the celebrant at midnight spitting out black beans from his mouth: ‘At first sight it is difficult to imagine Livy or Horace or Agrippa getting out of bed and solemnly going through this ritual. And yet they probably did – at least in a modified form’ (Romans and their gods 85).

11. I refer almost exclusively in this paper to written aetiologies. No doubt there was also an active oral tradition; but this is now irrecoverable to us.

12. For full evidence of the surviving calendars, see Degrassi, Inscriptiones Italiae.

13. For Aeneas and the Vinalia Priora (a different festival from the Vinalia Rustica mentioned above), see Ov. Fast. 4. 879-900; Plutarch, , Roman questions 45Google Scholar (and other references collected by Degrassi, , Inscriptiones Italiae 446–7)Google Scholar. The story was (in Ovid's version) that, during Aeneas' war with Turnus, Mezentius (the Etruscan leader) promised to help Turnus, if he received a pledge in return for half of the next year's vintage; but Aeneas turned the tables on this offer by vowing the vintage, if he won, to Jupiter himself. Aeneas, of course, was victorious and from that day on the Romans offered up the first fruits of the vintage at the Vinalia to Jupiter. For the Consualia and the plot to steal the Sabine women, see Dionysius of Halicarnassus, Roman antiquities 2.31 (though there are many divergent versions, including the story told by Dionysius himself that the festival was founded by Evander: 1. 33. 2).

14. Sperber, D., Rethinking symbolism (1975) 17–50 (quote, p. 48)Google Scholar.

15. See Detienne, M., ‘Rethinking mythology’ in Izard, M. and Smith, P. (eds.), Between belief and transgression (1982) 43–52 (quote p. 48)Google Scholar. Detienne is particularly concerned with intellectuals in the Greek world, but the terms of his analysis could equally well be applied to Rome.

16. For Rome's traditional tendency to incorporate religious forms from outside, see North, J. A., ‘Conservatism and change in Roman religion’, PBSR n.s. 30 (1976) 1–12Google Scholar.

17. See Ov. Fast. 4. 721-862 (lines 783-806 quoted below). In addition to the major texts reprinted in this article, see the citations on the Parilia collected by Degrassi, , Inscriptiones Italiae 443–5Google Scholar.

18. See, for example, Scullard, , Festivals 103–5Google Scholar and Wissowa, G., Religion und Kultus der Römer ed. 2 (1912) 199–201Google Scholar. On the identity of the god (or goddess) Pales, see Dumézil, G., Idées romaines (1969) 273–87Google Scholar.

19. Plutarch, , Romulus 12. 1-2Google Scholar.

20. Dionysius of Halicarnassus, Roman antiquities 1.88.3.

21. Ov. Fast. 4. 783-806. For the complex narrative of the foundation of Rome under Romulus, see 807-62.

22. Athenaeus 8. 361e-f. For the background to the Hadrianic developments reported in this passage, see Beaujeu, J., La religion romaine à l'apogée de l'Empire I (1955) 128–36Google Scholar.

23. See de Saussure, F., Course in general linguistics (tr. Harris, R., 1983, 121–5Google Scholar (quotes pp. 121 & 124)). The Roman ritual calendar is not, of course, entirely without syntagmatic meaning. There are, for example, clear relationships between festivals following one another at particular intervals in the Roman calendar: pairs of festivals, separated by one day, appear often to be complementary (the Cerealia of 19 April concerned with crops, the Parilia of 21 April concerned with herds; the Vestalia of 9 June concerned with the fire of the hearth, the Matralia of 11 June with the light of the dawn); while similarly those separated by three days sometimes suggest the same underlying preoccupation from slightly different points of view (the Consualia of 21 August and Opiconsivia of 25 August; the Consualia of 15 December and Opalia of 19 December – all concerned with the use and storage of crops). See further, Wissowa, , Religion und Kultus 437Google Scholar; Dumézil, , Fêtes 22Google Scholar.

24. This limbo time is clearly evoked by the activities of the Arcadian king Evander in Aeneid 8 – a Roman avant la lettre, who performed already ‘Roman’ cult (that of Hercules at the Ara Maxima, for example) on the site of Rome before the city's foundation by Romulus. (See also n. 27 on the Lupercalia.)

25. Tib. 2.5.28, 87-90, with Cairns, F., Tibullus (1979) 79–86Google Scholar (though Cairns' suggestion that the poem effectively incorporates two descriptions of the Parilia seems wrong; we are dealing rather with a reference to the primitive deity Pales (28), followed by a brief description of the contemporary Parilia (87-90)).

26. I am grateful to Andrew Wallace-Hadrill for raising this point.

27. Note, for example, Ovid's explanations of the Lupercalia (Fast. 2.267-452) which take the reader back to the time of the mythical king Evander, before the foundation of the city, and to the youthful exploits of Romulus and Remus (cf. Plutarch, Romulus 21.4-8; Val. Max. 2.2.9). Similarly the Consualia and Vinalia (n. 13, above).

28. Ov. Fast. 1.617-36 (refering to the Carmentalia of 15 January); see also, for a similar explanation, Plutarch, , Roman questions 56Google Scholar. For other aetiologies (including the story that the festival was founded by – perhaps – Romulus, after his capture of Fidenae), see Degrassi, , Inscriptiones Italiae 394-6, 398Google Scholar. For a brief review of the problems of the interpretation of the festival, see Scullard, , Festivals 62–4Google Scholar. A similar middle Republican aetiology is found for the Lesser Quinquatrus (13 June; Scullard, , Festivals 152–3Google Scholar) – according to which the festival was established after a strike of flute-players in 311 B.C. (see Liv. 9.30.5-10; Ov. Fast. 6. 649-710).

29. For Catulus' vow, see Plutarch, , Marius 26Google Scholar. For a similar example, note the foundation of the Temple of Bellona in the Circus Flaminius, vowed in 296 B.C. by Appius Claudius Caecus (Liv.10.19.17).

30. See Dio, Histories 43.42.3 (although note that these games appear quickly to have fallen out of use; cf. Histories 45.6.4). Beaujeu, , La religion romaine 131Google Scholar suggests that Caesar, like Hadrian later, was rather attempting to turn the festival into an ‘official’ festival of the anniversary of Rome.

31. For the Liberalia marked in calendars as the anniversary of Munda, see Degrassi, , Inscriptiones Italiae 426Google Scholar with Appian, Civil war 2.106.442 and Dio, Histories 43.44.6. The traditional associations of the Liberalia were as a festival of Liber Pater (who became equated with the Greek Dionysus); it was also the day on which Roman boys traditionally ‘came of age’. See, briefly, Scullard, , Festivals 91–2Google Scholar.

32. For ‘tension’ between different styles of Roman religious discourse, see Beard, M., ‘Cicero and divination: the formation of a Latin discourse’, JRS 76 (1986) 33–46Google Scholar. There is, of course, no reason to assume (as has often been done) that Ovid and his sophisticated contemporaries would have ‘really believed’ the ‘scientific’, ‘rationalising’ explanations, and would have effectively dismissed the ‘mythological’ exegesis as primitive mumbo-jumbo. The point is that the two forms could, if awkwardly, exist side by side.

33. I must stress here the fortuitous coincidence between Plutarch's symbolic narrative and my ‘historical’ reconstruction. I am not in any sense going back on my primary contention that these exegeses are an integral part of Roman religious experience, not external explanations of it.

34. In attempting to reconcile the account of Athenaeus with the archaeological evidence for the new temple, I follow here the narrative of the Hadrianic developments suggested by Beaujeu, (La religion romaine 128–33)Google Scholar: that the first formal celebration of Rome's birthday on 21 April (a celebration of a Romaia that is, rather than just a Parilia) probably took place in 121 A.D.; that the decision was taken at that time to found a temple of Venus and Rome (or Fortune of Rome or just Temple of the City, as it is variously called), whose foundation day thus fell on 21 April (although it was not completed till some years later); that in later years the celebrations of 21 April combined the commemoration of the birthday of the city and (as a further symbolic marker) the commemoration of the foundation of the temple. One must remember, of course, that the specific association of the day of the Parilia with the birthday of Rome long predated Hadrian; the earliest surviving (Republican) calendar of the city marks the day as ‘Parilia: Roma condita’ (Parilia: Rome founded); see Degrassi, , Inscriptiones Italiae 443Google Scholar.

35. The Protestant church has, of course, other ways of incorporating changes into its festal year. A different political ideology, for example, can be assimilated by providing prominent positions for political leaders in the enactment of ritual or by carrying out rituals in locations normally associated with the political establishment.

36. The incorporation of ‘imperial’ festivals into the traditional calendar is not necessarily as ‘innocent’ a process as my analysis may suggest. In part, we are no doubt dealing with a seamless web, ‘naturally’ assimilating the new political order; in part, we may imagine a consciously ideological use of existing tradition by Caesar, Augustus and their followers.

37. For a full (but dull) account of this exciting document, see Dilke, O. A. W., Greek and Roman maps (1985) 41–53Google Scholar. The map is there treated in entirely utilitarian terms; only at the end of the account (p. 53) is there a hint of anything more: ‘Also the full extent of the Roman Empire could be seen at a glance’.

38. I say this image was principally synchronic, for of course it also allowed speculation on the (diachronic) narrative history of the growth of the Roman world.

39. For other discussions of the importance of calendars in cognitive systems, see Bourdieu, P., Outline of a theory of practice (1977) 96–109CrossRefGoogle Scholar and the various papers collected in Systèmes de pensée en Afrique noire 7 (1984)Google Scholar (special issue on ‘Calendriers d'Afrique’).