Ὁ λιβανωτὸς εὐσεβές καὶ τὸ πόπανον: the rationale of cakes and bloodless offerings in Greek sacrifice

1Sacrifice is a many-splendoured thing, and it is scarcely surprising that over the last century it has loomed so large in the study of ancient Mediterranean religion, nor that new angles should be constantly under exploration.1 Even where there is a wish to challenge the usual view of animal sacrifice as central to the religious practice of classical antiquity, attention is necessarily lavished on the whole complex of sacrifice and all its details and variants.2 For it is doubtful whether any general theory of sacrifice, or even more modest approach to the subject, can be quite satisfactory if it does not take into account the multiplicity of sacrificial forms. It is for this reason that, in attempting to contribute to one of several recent projects which mark a shift from earlier sacrificial models, I have chosen to examine an apparently minor, but still significant, aspect of Greek sacrificial ritual: the offering of bread and cakes, with a glance at bloodless sacrifice more generally. Our volume in general, rather than focusing on the sacrificial community – its response to violence, its self-definition, or its reading of its place in the universe – concentrates instead on the connexion of sacrificial actions with conceptions of the gods, the recipients of sacrifice. This piece will follow that lead, but will not avoid the sacrificing community entirely, for the relationship between deity and worshipper is of necessity part of the conceptualisation of the divine.

2In general terms, it is very noticeable that sacrificial cakes were a matter of some interest to the Greeks, and it is this interest which makes the study both possible and legitimate. Lobeck, in Aglaophamus, constructed his ‘pemmatologia sacra’ on the basis of methodical ancient treatments;³ Athenaeus, whose two discussions of bread and cakes supply much of our information, lists seven authors of treatises on these confections, and also authors on ritual and local antiquities who include material on cakes.⁴ We must ask why it was that these writers thought that cakes and bread were significant, and why they thought it important to list and distinguish different types. Some authors, no doubt, were interested primarily in explaining the multitude of food references in Middle and Old Comedy, while for the authors of ritual handbooks the important point may simply have been to ensure correct ritual performance: it is so because it is so. But these motives are not sufficient to account for all the interest; our sources include details which go well beyond these limits. Further, even in earlier sources and outside treatises of systematic classification, we can discern an interest in cakes and bread in both secular and sacred contexts, while through their numerous different names, shapes, and occasions of use, the cakes themselves supply their own implicit classification. Cakes can be perceived in terms of opposition to meat and meat dishes, as cooked food in opposition to fruits and nuts, as shaped and moulded food in opposition to pottages and gruels, and also in terms of the relationship between cake and cake – there is, so to speak, a grammar of cakes. In what follows, I shall be suggesting three contexts of significance for cakes as they are used in cult as vehicles of expression.

3But first we must consider what exactly counts as a cake, or perhaps ‘sweetmeat’ is a better word. The associations of ‘cake’ in modern languages are slightly different from the range we are discussing here. In English, we ‘bake a cake’, and in our cookery books ‘baking’ is more or less synonymous with ‘cakes and bread’. But a little further thought might suggest the more distant kinship of griddled or fried confections such as crêpes (pancakes in English) or Indian flatbreads, and Greek writers certainly list grilled breads and shallow- and deep-fried breads and cakes alongside baked ones. What all these creations have in common is the presence of a good quantity of cereal (less usually pulse) flour in their composition, the addition of liquid to permit shaping, and a cohesive final form produced by cooking or (sometimes) exposure to the sun. Within this broad category there was a fairly clear distinction between ἄρτοι, made from wheat, and μᾶζαι, from barley, and there was also a distinction corresponding in some measure to our own between ‘bread’ and ‘cake’, as we can see from the two sections in Athenaeus. His basis for the two discussions is place in the meal, with cakes appearing as part of dessert, δεύτεραι τράπεζαι. Context is quite important, then, in classification, and it is not too surprising that just as in our own culinary settings the distinction between bread and cake is far from absolute. Different writers may disagree on how to classify a particular confection. Thus the ἀμφιφῶν in the tradition deriving from Philochorus is a πλακοῦς, but the lexicographer Pollux calls it a μᾶζα.⁵ Even Pollux himself appears not to be consistent: if his text is correct, he lists the ὀβελίας among the μᾶζαι, but defines it as an ἄρτος. Although Athenaeus says that πλακοῦς is an adjective with ἄρτος understood, both he and Pollux hedge their bets with phrases like ἄρτος πλακουντώδης, ἄρτος ἐγγυτέρω πλακούντος, and μεθόριος ἄρτου καὶ πλακούντος. Generally it seems that the πλακοῦς is richer or fancier than the ἄρτος: this is clear from the story also told by Athenaeus in which Diogenes, unaccustomed to such fare, finished off a πλακοῦς and pronounced it ‘good bread’.⁶ Hence the convention of rendering πλακοῦς and similar words as ‘cake’ or ‘galette’ is not so far from the mark.

4We can now examine the relationship of breads and cakes to other food – my first area of significance. At the simpler end of the spectrum, it is clearly the case that bread of some sort was the staple ingredient of many an ordinary meal, although we should not overlook the role of grains in other forms, such as porridge or honeyed grains like the modern κόλλυβα. Solid breads have definite advantages over these alternatives, however: they are more durable, more easily portable, and they can also be used to wipe up other, wetter foods and if necessary to clean the hands after eating.7 But in either case a meal centred round bread or porridge has obviously a very different ‘meaning’ from one where meat is the main food consumed. Though the consumption of meat products may not have been preternaturally rare among the Greeks, I think it is clear that large portions of meat could never have been taken for granted as part of an everyday meal, and that most meats were valued as prestige food and a privileged part of the diet. Yet it is also clear that the relationship between meat and bread is not a simple one between foods high and low in esteem respectively. Bread is, quintessentially, what people eat, the invention of agriculture ranks among the greatest benefits given by gods to mortals,8 and apparently even in the grandest and most luxurious feasts bread in some form or other is an essential component. The meat meal will always include bread, the bread meal only on occasion meat. Of course, the grander the meal, the more elaborate the bread, or cake, as many of the creations should probably now be called. In this case, the cakes are highly valued as special treats and delicacies, as is abundantly clear from the evidence of comedy. They are not the central, most ‘important’ part of the meal, but they are indispensable to its status as a special repast.

5I return now to sacrifice. Even leaving aside theoxenia and similar practices,9 it is apparent that there is a close connexion between sacrifice and meal. The sacrifice normally results in a meal for the human participants, or at least some of them; it is also analogous to a meal, in that things commonly eaten and drunk are brought together and in some way consumed. It is not surprising that Hesiod represents sacrificial practice as deriving from an original meal shared between gods and humans.10 Whatever form the actual sacrificial meal may have taken – however much was added to the offerings for the divine or heroic recipient before human consumption – the meal analogy, in the form of the objects offered, was somewhat schematic. In sacrifice – again as opposed to theoxenia – we seldom hear of vegetables, for instance, and while fruits may form part of table-offerings they are not usually prescribed as such in sacrifice regulations. Bread and cakes, however, are quite commonly mentioned. For instance, in the well-known fourth-century Coan lex sacra LSCG 15111 we learn that the preliminary sacrificial procession in the festival of Zeus Polieus includes two animal victims, probably for different parts of the procedure, seven ‘cakes’ of the type called φθόις, honey, and a garland.12 Schematically, this seems to represent a meal, with meat as the centrepiece, honey, which will be mixed with milk as wine is mixed with water, to substitute for wine as a drink in a ritual of which the first part is wineless,13 and a garland to stand for symposiastic festivity. The cakes are an accompaniment or a follow-up to the meat – since according to the Suda, the φθόις is a cake which can accompany the offering of the entrails,14 and indeed this is the procedure followed here. At the main sacrifice on the following day, along with the ox is offered a set quantity of barley grains (a little over four litres), and two large loaves of set sizes and (in one case) form, plus libations of wine (again a set amount).15 Here again we have the basic form of the meal: meat, with accompaniments both of bread and of unmoulded cereals – perhaps the θυλήματα burned with the god’s portion in a normal sacrifice – and washed down with wine. In the one case, bread complements the meat and makes the solid food into a complete meal; in the other (the previous day) cakes appear as tasty adjuncts to a ‘meal’ already defined as special by the presence of a large quantity of good meat. The meal is in this case symbolic rather than actual, or better, it is given in its entirety to the deity, because this is a holocaust sacrifice. But whether the rite is holocaustic or participatory, the god is offered a meal of the type most humans prefer, centred around high quality meat, accompanied by bread or perhaps more often, cereal confections of a finer quality than the everyday. As Louise Bruit has shown, in sacrifice as well as in theoxenia, the gods both share and do not share the human meal, but they certainly in some manner receive a meal, not only pieces of meat.16 However important and central the meat is, the meal is incomplete without cereal accompaniments, which as I have said frequently contribute to the composition of the meal as far removed from the ordinary and add to its special quality.

6So far I have been using the word ‘special’ to indicate simply a degree of richness and niceness which contrasts with everyday fare. But there are other senses in which cakes can be special: they may either be formed in a particular way that suggests sacred use, or they may be specific to a particular deity, cult or festival. This is my second context of significance, and the one which comes closest perhaps to the main theme of this volume. An appropriateness for sacred as opposed to secular use is indicated, for instance, by Polybius’ reference to ‘the πόπανα with ὀμφαλοί which are placed on sacrifices’;17 here we have a round, flattish cake with one or more circular protrusions on its surface, which is defined for Polybius’ readers by its use in sacrifice, and more particularly by being offered and placed in addition to or after the main event.18 Again, Pollux divides his list of μᾶζαι (but not ἄρτοι or πλακοῦντες) into ‘sacred’ and ‘those which humans use’,19 so it is clear that, even allowing for regional differences in nomenclature, there are many different types of cakes which are used in cult. This in turn suggests at least a partial differentiation of cakes as appropriate in different cult contexts. The scholars of later antiquity were aware of several cakes used in specific cults, some mentioned in sources such as comedy, but also treated and explained in ritual handbooks, local histories, and specialist works on food. Thus for instance we find the βασυνίας, offered to Iris by the Delians, while we are told (though epigraphical evidence changes the picture slightly) that the βοῦς is offered to Apollo, Artemis, Hekate and Selene.20 In an earlier paper, I suggested that in perhaps the majority of cases the important thing may be not so much a particular appropriateness in the cake, a perceived match between the qualities of the festival or deity and the confection offered, as the simple fact that one particular ritual circumstance calls for the use of one particular cake.21 The ritual is identified and given specificity by – no doubt among other things – the kind of cake offered. Cakes are able to perform this function because of the very great number of variables involved in their making. In contrast with the few liquids available for libations, and the relatively limited number of animal species used in sacrifice, even allowing for differences in age, colour, and so on, cakes, being a human creation, can be made in a far greater variety of shapes and sizes, so that it is easy for a particular form to be confined to a particular ritual context. Since, as we have seen, cakes were also often – perhaps normally – carried in the sacrificial procession, this form was also presented very visibly, so that for both spectators and participants it would necessarily come to characterise a particular ritual in distinction to another. But can we get further than this?

7However much the Greek mind loved explanations, it still seems unlikely that every confection had an explicable correlation with the festival in which it was employed or the deity to whom it was offered. But some cases may seem more promising than others. Let us begin with the bous. I suggest that here we can find three types of meaning. Cakes representing animals, like this one, are a not uncommon phenomenon: we know also of a deer-cake (ἔλαφος), and the Delian ἀχαΐνη, whose name also means a deer, while we hear more generally that animal-shaped cakes were offered when animals could not be.22 On this level, the message is an ambiguous one. If the cake-βοῦς alludes to a sacrificial animal βοῦς, it might be saying that the sacrifice of bovines is appropriate for the deity – that this is a substitute for the real thing, which could be given – or, on the contrary, that the deity receives the cake because the real thing is unacceptable, either to the deity or to the sacrificer. However, it is clear that this last rationale cannot be a necessary implication of the βοῦς, as it can be offered in conjunction with animal sacrifice; it could however be taken to indicate a theoretical willingness to offer a more expensive cow rather than merely a sheep or goat. On another level, or rather another type of meaning, if Pollux is right that the βοῦς is felt to be particularly appropriate in the cults of Apollo, Artemis, Hekate and Selene (he is not right that it is used exclusively so), it follows that the cake serves to reinforce a perception that this group of deities has something in common. Here, it is not surprising to find Artemis linked to Hekate, as certain forms of Artemis are very close in character to this goddess, and the connexion with the Moon also belongs in this area, from at least the fifth century; it is rather more of a surprise to find this aspect of Artemis thus linked to Apollo, although a connexion with Apollo was probably felt in the sevenfold nature of the cake, the seventh being the day of the month particularly associated with that deity.23 One effect of the βοῦς cake, then, might be to affirm the unity between the more independent, Hekate-like forms of Artemis and her Apolline aspect – one wonders if the cake was offered also to the third member of the Delian triad, Leto.24

8Finally, for a third type of meaning, we might consider the form of the cake itself. While Pollux deals with this cake just after the σελήνη, saying that both are named from their shape, he describes it simply as a cake with moulded horns. The Suda, on the other hand, in two places describes a βοῦς ἕβδομος – a form which is well attested epigraphically – as a cake placed on six σελήναι cakes, which we know to be round discs in form, and itself shaped with projecting horns – κέρατα ἔχοντα – which, we are told, are ‘in imitation of the new moon’ (κατὰ μίμησιν τῆς πρωτοφανοῦς σελήνης).25 This full assembly of cakes, then, according to the lexicographical tradition has a double allusion to the moon, which seems to fit very well with the earlier testimony connecting the βοῦς cake with Selene and other goddesses with lunar aspects. But this connexion is perhaps unlikely to be very old. If a cake is called βοῦς, after all, it does not need any explanation for its horns beyond that fact; and there is indeed also listed in the tradition another kind of cake or maza sometimes said to have horns, the βήρηξ, for which no lunar connexion is suggested.26 Further, the description of the tips of the moon’s crescent as ‘horns’, though it may seem familiar and obvious to us, seems to have been less so in archaic and classical Greece; I have not been able to find an instance earlier than Aratus, who does, however, use the word several times.27 Even the depiction of the Moon with a crescent on her head, which might (but need not) suggest horns, is far from canonical before the hellenistic period. Earlier images show just as frequently the crescent beside Selene’s face, or else the goddess with a disc rather than a crescent.28 It is possible, then, that the lunar meaning comes to be ascribed some time after the name and form of the βοῦς cake had gained currency; possible also, therefore, that the offering to Selene, who in any case is not well attested in cult in the earlier period,29 is also a later development. In this case the name of the round cakes forming the base for the horned cake on top would have referred originally merely to their disc like shape without necessarily evoking a very strong association with the actual moon.30 The potential for a strong association is activated later, as both Selene herself and the lunar associations of Artemis/Hekate come further to the forefront, and as the idea of the moon’s ‘horns’ becomes better established in both vocabulary and iconography. The ‘meaning’ of a sacrificial offering may change and develop over time.

9To sum up what we have learned from this example, the ἕβδομος βοῦς shows that one admittedly rather elaborate and multiform cake can have several simultaneous meanings. Working in reverse order this time: if we accept the lunar connexion, the complete cake assembly, including the ‘full-moon’ round selenai and the crescent shaped horns on the ox itself, draws attention to the moon’s phases and strongly suggests a lunar association for Artemis and Hekate. Where there is from other angles less plausibility for such a connexion, as with Apollo, or the offerings to Hermes, Kourotrophos and Hestia which are attested epigraphically, this aspect will naturally be less conspicuous, even in the later period. It is there to be ‘activated’ when other factors suggest that this is appropriate. In this area, the cake is at least capable of saying something about the deity to whom it is offered, approximating Artemis and Hekate to the moon, and in the case of the Moon herself, drawing attention to her regular changes. It is in this area, too (let us now call it area 1), that the Apolline connexion with the number seven is operating. In another area (area 2), the shared use of the cake emphasizes a dynamic between deities, while thirdly (area 3) the name and schematic animal form necessarily suggest some interplay between cereal offering and animal sacrifice, even if the nature of this link is ambiguous, and this says something about the relationship between deity and worshipper.

10It may well be that this degree of complexity was not to be found in every sacrificial cake. But equally, it is clear that it is a theoretical possibility; the boxes are there waiting to be filled in, and given the fragmentary state of our knowledge, it is only too likely that there were more specific meanings commonly ascribed to individual cakes than we can now recover. Let us look at a few more examples. Interesting because of its relatively early date is a case in the Derveni Papyrus.31 The writer refers to πόπανα ἀνάριθμα καὶ πολυόμφαλα ‘countless cakes with many navels’, which are offered by mystai to the Eumenides. Of course, many things are unclear about this text, but one thing that is certain is that the author explicitly identifies the Eumenides with souls, and explains the cakes ‘because the souls, too, are countless’, ὅτι καὶ αἱ ψυχαὶ ἀνάριθμοί εἰσι: doubtless the many omphaloi are felt to be subsumed in this explanation of the appropriateness of large numbers. It is a relatively simple explanation, which works in only one area of significance (our first area), applying to the deity alone; a quality of the cakes correlates with a quality of the recipient. Whatever the ritual these mystai are engaged in (this is an issue we need not explore here),32 the generally polemical attitude of the Derveni author makes it unlikely that his explanation is the ‘official’ line, the viewpoint of hierophant, mystagogue or initiator; but it does show the kind of explanation that could be offered and, presumably, in some quarters accepted. One is tempted to say that these cakes were polysemous as well as polyomphalous. But what must inevitably have been expressed in this offering of πόπανα, regardless of our author’s interpretation, comes out in the word προθύουσι: the cake offering to the Eumenides is a preliminary sacrifice before the main event, and thus the mode of its offering indicates something in our second and third areas, those of the relationship between deity and deity, and between deity and worshipper. The Eumenides are not the main object of cult, but they are associated with that object (whether this is an Eleusinian or a Dionysiac milieu) in a subordinate capacity (area 2), and it is absolutely necessary for the worshipper to approach them before there can be access to the rites connected with the main deity (area 3). Given the nature of the Eumenides as usually conceived, we might guess that in this case the πρόθυμα was thought of as propitiatory, designed to placate entities with the power to obstruct the good things which the mysteries promised. More generally, we should note that the preliminary sacrifice of cakes is quite a frequent procedure, particularly well attested for the subordinate deities in the entourage of Asklepios,33 and by the evaluation system we have already touched on, in which cakes are seen as good things but less valuable than meat, this prothyma inevitably indicates a position of lesser importance for the deity who receives it – not necessarily in general, but in the context of that particular ritual complex.

11The cakes for the Eumenides indicate that, besides those cases where several meanings may be simultaneously present to the same observer, different observers and probably different participants may find different and varied significances in the same ritual objects. Returning to the main pemmatological sources, a further example will briefly show more possibilities. This is the μυλλός, made in Syracuse and other Sicilian cities for the Thesmophoria. It is a cake in the shape of the female genitals, and though we are not offered an explanation in Athenaeus,34 it is apparent that several associations are possible. First, of course, the cake could operate as something to giggle at, something to draw the women celebrants together in a naughty conspiracy. Something like this may be suggested by the analogous case of the breast-shaped κριβάναι, displayed and eaten at women’s dinners in Sparta, where a divine recipient appears to be absent.35 But given the complexity of the Thesmophoria and its centrality in women’s religious observances, it is unlikely that this would exhaust the meanings of the cake. Its form could refer to sex, or to fertility, or to femaleness, all of which are generally thought to have some part to play in the Thesmophoria, so any one of these references might make it an appropriate offering to the Thesmophorian Goddesses (thus an area 1 significance). If femaleness is felt to be emphasized, an area 3 meaning will also be apparent, marking out an affinity between the Goddesses and their (on this occasion) exclusively female worshippers. It is also quite possible that a more formal kind of explanation was offered to the participants; we can easily imagine an observer in the Herodotean style remarking that ‘there is a ἱερὸς λόγος for the offering of the cake’. But this, of course, is speculation.

12We could speculate further that certain features in the preparation and presentation of sacrificial cakes are likely to have attracted explanation on some level or other, whether by actors or observers. Number symbolism, for instance, would seem irresistible; we have noted the numerological link between Apollo and the hebdomos bous. We may not now know why precisely one dried fig and three walnuts were used in the basynias cake, but we can be fairly confident that someone did, or thought they did.36 And we recall that in the Coan procession on the day before the sacrifice to Zeus Polieus, exactly seven cakes were displayed, though only two deities at most seem to be involved; surely someone knew of a reason for this.

13But these are only guesses; we can see that there is some tendency to attach a special significance to the form of the cake offering, whether explicitly or implicitly, but it does not at all follow that such significances were found in all or even the majority of cases. It seems plausible that significances of the second and third types, relating the deity to other deities or to the worshippers, were more widespread, as these associations tend to be looser, less tangible and less explicit than those which relate the offering directly to an aspect of the deity. It is also likely that connexions and explanations of a more conscious sort are more plentiful in the fourth century than the fifth, and in the hellenistic period than in the fourth century. Curiously, perhaps, there is little evidence for a kind of catch-all explanation which is commonplace in a similar area in the Hindu ritual context, and which might seem an easily adopted approach: that is to explain a particular offering as a deity’s favourite. That ‘Ganesha likes laddoos’ is frequently stated, and though laddoos may be offered to any deity, they are felt to be particularly appropriate for Ganesha, who is indeed often pictured holding a round sweetmeat or a bowl of these sweets – although South Indians interpret the iconography as showing a different kind of cake, the modakam, and say ‘Ganesha likes modakams’.37 This type of explanation by no means precludes explanations on other levels within the system, and it might seem that it ought to have appealed to the Greeks, who seem to have had no problems with attributing human-style preferences for things such as trees to their deities. True, there is the apparently isolated case of Apollon Opsophagos, ‘lover of dainties’, worshipped in Elis, who no doubt received unusual offerings, perhaps of fish, perhaps specially elaborate trapezomata.38 And there is the explanation of the third/second-century BCE historian Mnaseas, in his Περὶ Ἀσίας, for the sacredness of fish to the Syrian Goddess and the taboo on their consumption by lay people: Atargatis had been a queen who had ordered all fish to be given to her ‘because she liked the food’ (διὰ τὸ ἀρέσαι αὐτῇ τὸ βρῶμα).39 But this is set outside Greek religion proper, and more tellingly, it is in the context of a euhemeristic account, so that the preference is essentially attributed to a human being rather than a god. Perhaps this apparent reluctance to imagine the gods’ favourite foods is then connected with the whole problematic of divine consumption of the products of sacrifice; where Hinduism has available a sophisticated thought system able to explain how the divine and human can both partake of food offered to the gods, the Greeks by contrast found the whole question troubling, and produced no satisfactory resolution to their conflicting concepts on the matter.

14But one further suggestion of a deity’s preference in foodstuffs may serve to introduce the final part of this essay, in which we examine the position of cakes and bloodless offerings more generally as an alternative to blood sacrifice. This is a late and whimsical epigram from the Palatine Anthology, in which Ares speaks: ‘Who set down rich cakes, grapes and rose petals for me, Ares the sacker of cities? Let people offer these to the Nymphs; I, Ares of bold designs, do not accept sacrifice without blood’.40 Here clearly we have moved away from the apparently arbitrary preference for one form of cake over another to a semi-humorous statement about the general form of sacrifice that is appropriate for a particular deity. Leonidas of Alexandria, the imperial period author of this isopsephon verse, is here reversing the conceit of Aristophanes in Peace 1019-20: ‘Peace doesn’t enjoy slaughter, and her altar isn’t made wet with blood’.41 Of course, Aristophanes isn’t reflecting the actual cult of Peace in Athens, which was almost certainly established later than the play, and which certainly did include animal sacrifice, as we know from the dermatikon accounts.42 It seems that the idea of a deity ‘not liking’ either kind of offering is conceivable, but primarily in a joky context, and further, it also seems difficult to maintain that the offering of cakes or bloodless sacrifice as such was frequently related in some more general way to the perceived character of the deity who receives it.

15Detienne indeed has asserted that in one of the most celebrated cults which permitted only bloodless sacrifice, that centred round the ‘pure altar’ or ἁγνὸς βωμός on Delos, the vegetal offerings (wheat, barley, popana) suited the character of Apollo Genetor as giver of fruits, καρπῶν δοτήρ – but when we look at his evidence for this interpretation of the epithet in Plutarch, it does not refer to the Delian cult at all, rather to offerings at Delphi made by the Magnesians and Eretrians, with a strong colouring of Plutarch’s own universalising conception of Apollo.43 It is certainly not legitimate to assume from this that the character of the Delian Genetor was similar. Much more to the fore in the cult of the ἁγνὸς βωμός would appear to be its marked contrast with the nearby ‘horn altar’, the keratinos, where the slaughter of animals is emphasized by the use of their body parts to build up the altar itself.44 Other examples bear out the idea of an opposition between different styles of offering within the same cult area. The offering of cakes to Zeus Hypatos in Athens, for instance, is contrasted by Pausanias in a rather freewheeling way with Lykaon’s supposed institution of human sacrifice in Arcadia, but on the Athenian acropolis itself it contrasts with another cult of Zeus, that of Zeus Polieus, in which the act and repercussions of the killing of the victim receive a peculiar emphasis.45 In the case of the fireless, and bloodless, sacrifices to Athena in Rhodes, it has been convincingly shown that we have both a contrast between a ritual of cattle sacrifice and one of bloodless offering, and very likely an instance of conscious archaising practice beginning in the mid-fifth century.46

16In searching for meaning applied to bloodless sacrifice, then, it may be pertinent to approach the question in more ‘traditional’ style through first examining the modalities of the practice itself and perceptions about it. It may be that within the general category, cakes and other cereals occupy a position of primacy, coming to typify a whole style of offering symbolically opposed to the sacrifice of animals. Ares mentions the cakes before the other girly things he rejects – in fact they are the first word of the poem. Again, while we have learned to be cautious about reading too much into the aition of the Bouphonia, what is interesting from the pemmatological point of view is that the story does not simply oppose blood to bloodless sacrifice, but specifically the killing of the ox to the offering of cakes or grains – depending on the account – which were the intended gift to the deity and which the offending animal ate.47 Plutarch, in discussing the taboos applying to the Flamen Dialis in Rome, draws a parallel between flour (being intermediate between seed, which can still germinate, and bread or cake) and raw meat, which is intermediate between the living animal and cooked meat, ready for consumption.48 Much can be said about this interesting perception, as has been done by Vernant, but my point here is that Plutarch thinks the comparison between the two series a reasonable one – animal food can parallel cereals because cereals, rather than pulses, vegetables or fruit, are the paradigmatic plant-derived nourishment.

17But what are we to say more generally about cults which prescribe or encourage bloodless offerings, usually headed by the ‘living’ grains and/or the ‘complete’ cakes? As is well known, such rituals were often associated with extreme antiquity – perhaps a supposed primitive vegetarian state – or sometimes exoticism – the Delian ἁγνὸς βωμός having surely some connexion with the mysterious offerings of the Hyperboreans. But far more commonly bloodless offerings are given to deities who in other circumstances or at other times would accept animal sacrifice; very few deities, we can suppose, when presented with cakes, grapes and roses would react as indignantly as the poetic Ares. On the other hand, as we have seen, it is normally the case that just as meat is valued more highly than bread, so animal sacrifice is more costly, sumptuous and emphatic than offerings of cereal origin, suitable for public festivals and major private occasions, rather than an everyday event. The significance of a bloodless offering for a worshipper was then often lack of funds or the fact that the occasion was not a particularly unusual one. By the fourth century, however, this rather negative evaluation of the choice for a bloodless offering could be turned inside out. While a straightforward assumption might be that the gods were best pleased with large, high-quality and frequent sacrifices, such as those that endear Hector to Zeus in the Iliad,49 it was also easy to observe that large sacrifices could have as much to do with ostentation as with piety. If, then, the gods are held to be interested in their worshippers’ intentions, as they are increasingly through the fifth century, a modest and sincere offering might seem preferable to a large one designed as much to impress one’s fellow citizens as to please the deity. According to Xenophon, Socrates would frequently quote the line of Hesiod ‘According to your ability make sacrifice to the gods’, with stress on ‘according to your ability’. It would be intolerable, he said, if the rich and unjust were more favoured by the gods than the poor and upright.50 When Knemon in Menander’s Dyskolos complains about constant, over-the-top sacrifices, he is showing himself up as a comically exaggerated misanthrope, but he is also tapping into a well-known and respectable way of thought. ‘They sacrifice not for the gods, but for themselves. Incense is pious, and cakes; those are placed on the fire complete, and the god receives them; but these people put out the end of the tail and the gall bladder, the inedible bits, for the gods and gobble up the rest themselves.’51 Similar points are made more than once in fourth-century comedy, undoubtedly forming a commonplace and being linked with the idea that the modest – and therefore usually bloodless – offering is morally superior. A bloodless offering cannot be designed to impress one’s fellow humans, and neither can its real purpose be to give oneself a nice meal. However agreeable cakes are, as Knemon says, they are typically burnt whole on the altar fire; they may also be deposited on an offering-table, and become priestly perquisites, but they are almost never consumed by the worshipper.

18Increasingly, then, the offering of cakes without animal sacrifice comes to be bound up with a certain self-image for the offerer, involving a more thoughtful piety than that perceived as the default position. But beyond this, it could be seen as setting up an improved dynamic of reciprocity, a better relationship, between the human and divine partners in sacrifice. The humans surrender their offerings completely to the gods and do not directly benefit from it themselves; the gods are wise and reasonable, looking at the intention and, for Socrates, the general moral status of the humans. We have come a long way from the cake as the distinctive accompaniment of a particular deity or cult, but I think we can see the potential here too, in this more generally applicable style of offering, for cakes to indicate important things about the gods and about their relationship with their worshippers.