Skip to main content Accessibility help


  • Access
  • Cited by 2


      • Send article to Kindle

        To send this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the or variations. ‘’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Roman Sacrifice, Inside and Out*
        Available formats

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Roman Sacrifice, Inside and Out*
        Available formats

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Roman Sacrifice, Inside and Out*
        Available formats
Export citation


The ‘insider-outsider problem’ has had little impact on the study of religion in pre-Christian Rome. Classicists generally assume that the modern idea of sacrifice as the ritual killing of an animal applies to the Roman context. This study argues, however, that the apparent continuity is illusory in some important ways and that we have lost sight of some fine distinctions that the Romans made among the rituals they performed. Sacrificium included vegetal and inedible offerings, and it was not the only Roman ritual that had living victims. Roman sacrificium is both less and more than the typical etic notion of sacrifice.



I owe many thanks to C. P. Mann, B. Nongbri, and J. N. Dillon for their thoughtful, challenging responses to earlier drafts of this article, and to audiences at Trinity College, Baylor University, and Bryn Mawr College for comments on an oral version of it. J. B. Rives provided valuable consultation on specific points and V. C. Moses generously shared her work-in-progress on the osteoarchaeological evidence from S. Omobono. I also thank the anonymous reviewers for helpful comments, suggestions and objections that have greatly improved this piece. I have tried to respond to them all. The errors and flaws that remain are all my own.



Those working in the social sciences have, for some time, been wrestling with the ‘insider-outsider problem’: how can one person genuinely understand the thoughts, beliefs, and actions of another? Insiders can report their own subjective experience, but very often this diverges from the assessments of those observing from the outside, from what is perceived as a more objective position. The problem was cleverly illustrated in a classic article that is standard reading for introductory anthropology and religious studies courses but that is generally unfamiliar to classicists, entitled ‘Body rituals of the Nacirema’ by the anthropologist Horace Miner. The article offers a brief account of the daily habits of this North American tribe whose members place a great deal of emphasis on maintaining and improving the body, for example:

The fundamental belief underlying the whole system appears to be that the human body is ugly and that its natural tendency is to debility and disease. Incarcerated in such a body, man's only hope is to avert these characteristics through the use of the powerful influences of ritual and ceremony. Every household has one or more shrines devoted to this purpose. The more powerful individuals in the society have several shrines in their houses and, in fact, the opulence of a house is often referred to in terms of the number of such ritual centers it possesses. Most houses are of wattle and daub construction, but the shrine rooms of the more wealthy are walled with stone. Poorer families imitate the rich by applying pottery plaques to their shrine walls. 1

Although it is sixty years old, the lesson still works well. Those poor Nacirema, who despise their physical form and try to improve it through ‘ritual and ceremony’, at first seem so different from us: primitive, superstitious, unsophisticated. But then they turn out to be us. Nacirema is American spelt backwards, and Miner shows to, and interprets for, us our own bathroom habits. 2 Yet so stark is the discrepancy between his (assumed) outsider perspective and our own insider understanding of the value of a bathroom, that most readers do not recognize themselves the first time they read this piece. This is the insider-outsider problem in nuce. Neither Miner's nor his reader's understanding is right and the other wrong: they are two different views of the same set of data, and both are valuable. Working with the two of them together, we can get a more nuanced understanding of a cultural habit. Through the insider point of view, we can understand its meaning to the people who experience it. Through the outsider point of view, we can interpret it in light of comparable behaviours in other cultures.

The present study turns the insider-outsider lens on the study of Roman sacrifice: it aims to trace, through an analysis of a set of Latin religious terminology, how Romans thought about sacrifice and to highlight how this conception, which I refer to by the Latin term sacrificium, relates to two dominant aspects of modern theorizations of sacrifice as a universal human behaviour: sacrifice as violence and sacrifice as ritual meal. The argument I lay out here pertains to sacrificial practice as it was conceived by Romans living in Rome and those areas of Italy that came under their control early on, during the Republic and the early Empire. It is likely, but admittedly not certain, that the concept of sacrificium I delineate here was also at play in citizen communities throughout the Empire, at least at moments when those communities performed public rituals in the same manner as did people in the capital. The survival beyond the early Empire of most aspects of the distinction among ritual forms discussed in Section IV cannot be asserted with any confidence.

I concede that, to a certain extent, the insider-outsider lens does not show us difficulties that were previously invisible. Those studying ancient Greece and Rome in general and those focusing on Roman religion in particular have been wrestling with these issues for some time even if the terms of the discussion have not been explicit. 3 But it does bring things into sharper focus, helping the student of Roman religion to keep in view the extent to which we have interpreted the ancient sources to fit our own (rather than the Romans’) intellectual categories. Here I use it as a tool to get at one aspect of Roman religious thought; I do not offer a sustained methodological critique of contemporary approaches of Roman antiquity. The present study argues that looking at the relationship between sacrificium as it is presented in Roman sources and comparing that with modern notions of sacrifice reveals that some important, specific aspects of what has been conceived of as Roman sacrifice are not there in the ancient sources and may not be part of how the Romans perceived their ritual.

Let me be clear. Although the focus of this investigation is the recovery of some details of the Romans’ idea of sacrificium, I do not mean to imply that their concept is the ‘right’ one and that the modern idea is wrong or completely inapplicable to the Roman context. Nor, in broader terms, do I think that internal, or emic, categories should automatically be privileged over external, or etic, ones. 4 Although much work in anthropology and other social sciences has debated the relative merits of emic versus etic approaches, I find most useful recent research that has highlighted the value of the dynamic interplay that can develop between them. 5 An emic explanation is essential for understanding how people within a given system understand that system, but because it is culturally and historically bounded, its use is somewhat limited. An etic approach allows the researcher to see functions, causes, and consequences of insider behaviours and habits that may be invisible to the people who perform them, as Miner illustrated for us. The objectivity of the outside observer can also facilitate cross-cultural comparison. Our modern idea of sacrifice can, with some refinement and clarification, remain a useful concept for constructing accounts of how and why the Romans dealt with their gods in the ways they did 6 and for looking at Roman religion in the context of other religious traditions.

Yet, part of the work of a Roman historian is to try to understand how the Romans understood their world, to be aware of anachronism in our accounts thereof, and to keep in mind that the sources never truly ‘speak for themselves’. In overlooking the differences between the Roman idea of sacrificium and the modern idea of sacrifice, we lose some of the details of how the Romans perceived a core element of their own experience of the divine. Those details, once recovered, can in turn subtly reshape our own idea of what sacrifice is and what it does. Therefore, instead of privileging either the emic or etic, I argue for an increased awareness of the insider-outsider distinction and for an approach to Roman religion that makes use of both emic and etic concepts. This should prompt researchers, myself included, to greater caution when presenting a ‘native’ — in our case, Roman — point of view and to greater clarity about whether the concept under discussion at any given moment is really the Romans’ or ours, or is shared by both groups.

In what follows, I aim to clear away a few of the accretions that have arisen from more than a century of modern theorizing about the nature and meaning of sacrifice as a universal human phenomenon in order to gain a better understanding of those actions that the Romans identify by the Latin words sacrificium and sacrificare. 7 The ultimate conclusion of this investigation is that, although in many important ways this ritual comes close to aligning with the dominant modern understanding of sacrifice, Roman sacrificium is both more and less than the ritualized killing of a living being as an offering to the divine: 8 more because the Romans sacrificed things that are not animals, and less because sacrificium is not a term that encompasses every Roman ritual that involves the death of a living being. There is a small group of other rituals that share certain structural similarities with sacrificium, but which the Romans during the Republic and early Empire appear to have distinguished from it. Modern scholars sometimes group all of these rites under the rubric ‘sacrifice’. 9 Furthermore, there is reason to think that the crucial moment, or perhaps the first crucial moment, in the whole ritual process of sacrificium for the Romans was the sprinkling of mola salsa onto the victim, whereas several important modern theorizations of sacrifice place the greatest emphasis on, and see the essential meaning of sacrifice in, the moment of slaughter.


As has long been recognized, sacrificare and sacrificium are compounds of the phrase sacrum facere (‘to render sacred’), and what is sacrum is anything that belongs to the gods. 10 As in other cultures, Roman sacrifice was not a single act, but instead comprised a series of actions that gain importance in relationship to each other. 11 As Scheid has reconstructed Roman public sacrifice, 12 the ritual began with a procession that was followed by a praefatio, a preliminary offering of prayers, wine and incense. The ritual ended with a litatio, that is, the inspection of the animal's entrails, and it was then followed by a meal. At the centre of the whole complex was the immolatio, during which the animal was sprinkled with mola salsa (a mixture of spelt and salt), the flat of a knife was run along its back, and then it was slaughtered.

Scheid's reconstruction focuses on a living victim, and this is in keeping with the ancient sources’ own emphasis on blood sacrifice. Although there is substantial evidence for other types of sacrificial offerings in the literary sources (see below, Section III), Roman authors do not discuss them at length, preferring instead to talk about grand public sacrifices of multiple animal victims. The tendency is intensified in Christian sources, which discuss pagan sacrifice exclusively in terms of blood sacrifice, distinguishing the shameful blood of animal victims from the sacred blood of Christ. 13 The Christian fathers’ equation of sacrifice with violence has shaped twentieth-century theorizations of sacrifice as a universal human phenomenon, 14 most famously those of Burkert, who identifies sacrificial slaughter as the ‘basic experience of the “sacred”’, and Girard, who begins his investigation into the origin of sacrifice by asserting its close kinship to murder and criminal violence. 15

The apparent alignment of emic (Roman) and etic (modern) perceptions of the centrality of slaughter to the Roman sacrificial process, however, is not complete. While the attention of our Roman sources is drawn most frequently to blood sacrifice, there is good reason to think that, if there was indeed a climax to the ritual, 16 the killing of the animal was not it, at least in an early period. Admittedly the Romans often used as a metonym for the whole of sacrificium the term immolatio, the stage of the ritual that includes slaughter, suggesting the special importance of that portion of the ritual sequence. 17 Yet to limit the consideration of immolatio to the moment of killing is to overlook the other actions (running a knife along the animal's back, cutting a few hairs from it) that Scheid has identified as being part of that stage of sacrificium 18 and the fact that the word immolatio itself derives from the Indo-European root *melh2− (‘to crush, to grind’): immolatio is cognate with English ‘mill’. 19 From this same root also derives the name for the mixture sprinkled on the animal before it was killed, mola salsa. 20 The Romans were aware of the link, as is made clear by Paul. ex. Fest. 97L: ‘Immolare est mola, id est farre molito et sale, hostiam perspersam sacrare’ (‘To immolate is to make sacred a victim sprinkled with mola, that is, with ground spelt and salt’), a passage which also suggests that the link between immolatio and mola salsa was active in the minds of Romans in the early imperial period when the ultimate source of Paulus’ redaction, the dictionary written by Verrius Flaccus, was compiled. 21 Studies of sacrifice have noted the etymological connection between immolare and mola salsa, but have not, for the most part, pressed its value for what it may reveal about where the Romans may have placed the emphasis. 22 Further support for the idea that the act of sprinkling mola salsa was either the single, critical moment or an especially important moment in a process that transferred the animal to the divine realm, is that mola salsa seems to be the only major element of sacrifice that is not documented explicitly by a Roman source as appearing in any other ritual or in any other area of daily life: processions, libations, prayers, slaughter, and dining all occurred in non-sacrificial contexts. 23

The importance of sprinkling mola salsa might explain a pattern in Roman public artwork from the republican through the high imperial periods. The most common images of blood sacrifice in Roman art are procession scenes of animals being led to the altar or standing before it, waiting for mola salsa to be applied to them. 24 More rare are images like those on the arches of Trajan at Benevento and of Septimius Severus at Lepcis Magna which show the moment that the axe is swung. 25 Rarest of all are images depicting the litatio, the inspection of the animal's entrails that Romans performed after ritual slaughter to determine the will of the gods. 26

In light of the importance of ritual killing in modern theoretical treatments of sacrifice, the relative paucity of slaughter scenes in Roman art requires some explanation. Art historians have debated whether ‘the choice to encapsulate the entirety of sacrificial experience in a scene of libation rather than a scene of animal slaughter (or vice versa) may tell us something about what was being emphasized as significant about sacrifice at that time or context’. 27 There is growing consensus that the answer is affirmative. While there has been tentative speculation that the reason behind a preference for procession scenes in Greek representations of sacrifice in the Archaic and Classical periods is due to a ‘growing squeamishness inside Greek culture’, 28 van Straten has offered a stronger explanation: the absence of slaughter scenes in Greek art is due to a ‘lack of interest in this particular aspect of sacrifice on the part of those Greeks whose religious beliefs are reflected in this material, shall we say, the common people of the Classical period’. 29 The basic argument transfers well to the Roman context. Huet explains the rarity of killing scenes in sacrificial reliefs from Italy by pointing out that the emphasis in these reliefs is really on the piety of the sacrificant who stands before the altar. 30 Elsner has proposed that the choice, increasingly frequent in the third century c.e., to represent the whole sacrificial ritual with libation and incense-burning scenes rather than with images involving animals is an indication of the increased emphasis on vegetarian sacrifice in that period. 31 Thinking along the same lines, it is reasonable to conclude that there are relatively few images of slaughter among Roman sacrificial scenes in public artwork of the Classical period because the emphasis in state-sponsored sacrifice lay elsewhere. The prevalence of Roman images of sacrificial victims standing before the altar, that is, of the instant before mola salsa is sprinkled on them, is due to the importance of that moment. From here, we can speculate that sacrifice was not understood by the Romans primarily as the ritual slaughter of an animal. Instead they seem to have conceived of it as the ritual consecration of an animal which was afterwards killed and eaten.


Modern theorizations of sacrifice focus on animal victims, treating the sacrifice of vegetal substances, if they are considered at all, as an afterthought or simply setting vegetal offerings as a second, lesser ritual, ‘a substitution or a pale imitation’. 32 Ubiquitous in the scholarship is the assumption that if the gods receive an animal, it is sacrifice, but if the gods receive vegetable produce and other inanimate edibles, those are something different: they are offerings. Upon examination of the Roman evidence, however, it becomes evident that this distinction is an etic one: while we see at least two different rituals, the Romans are clear that they sacrifice a wide range of food substances beyond animal flesh.

The equation of sacrifice with the offering of an animal is not completely divorced from the ancient sources. Although Roman writers most frequently do not explicitly identify the object of a sacrifice, when they do, cattle, pigs and sheep are well attested. 33 Goats and dogs are less common, and we can expand the range of species to include horses and birds if we admit animals that are identified only as the object of immolatio, if not of sacrificium itself. 34 At present, large-scale analysis of faunal remains from sacred sites in Roman Italy remains a desideratum, but analysis of deposits of animal bones from the region seems to bear out the prevalence of these species in the Roman diet and as the object of religious ritual (whether sacrificium or not it is difficult to say). 35 It is also clear from literary sources that on a handful of occasions, including instances well within the historical period, the Roman state sacrificed human victims to the gods, a topic we shall address more fully later on.

The prominence of animal victims in Roman accounts overshadows a substantial number of passages that make it absolutely clear that Roman gods received sacrifices of inanimate edibles. It is important to note that there is no indication that these vegetal offerings were thought to be substitutions for what would have been, in better circumstances, animal victims. 36 When the Romans sacrificed plant matter to the gods, it appears to be because that is what it was appropriate to do in the specific circumstance. The most famous vegetal offering occurred at the Liberalia, the festival of the god Liber, described by Varro: ‘Liberalia dicta, quod per totum oppidum eo die sedent sacerdotes Liberi anus hedera coronatae cum libis et foculo pro emptore sacrificantes’ (‘The Liberalia is so called because on that day priestesses of Liber, old women crowned with ivy, settle themselves throughout the whole town with cakes and a brazier, making sacrifices on behalf of the customer’). 37 Sacrifices of various cakes (liba, popana, pthoes) to the Ilythiae and to Apollo and Diana were part of Augustus’ celebration of the Secular Games in 17 b.c.e., a clear indication that vegetal offerings were not limited to the lower social classes. 38 There are many other non-meat sacrifices the Romans could offer. If we allow only items explicitly identified as sacrificia in Roman sources, our list includes beans, 39 wine, 40 incense, 41 milk, 42 wheat, 43 and first fruits. 44 It is unfortunate that the ancient sources on vegetal sacrifice are as exiguous as they are: it is not possible to determine what relationship its outward form bore to blood sacrifice. Were these items sprinkled with mola salsa? 45 Were they always burnt on an altar or brazier? Were they used in some form of divination? 46 Was a portion consumed later? The answers to these questions might reshape our understanding of what were the crucial elements of sacrificium.

Thus far, we have identified two points on which emic and etic ideas of what constitutes a Roman sacrifice do not align: when the critical transition from profane to sacred occurs and what kinds of things can be presented to the gods through the act of sacrificium. We can push this second issue, what kinds of items can be the object of sacrifice, even further: Roman sacrifice, especially among the poor, was not limited to edible offerings. The literary evidence for this is slender but persuasive.

The elder Pliny, in his Natural History, discusses the high regard in which ancient Romans held simple vessels made of beechwood. As proof, he recounts a story about M’. Curius Dentatus, famous for his victory over Pyrrhus in 275 b.c.e. and for his old-fashioned frugality and incorruptibility. 47 According to Pliny, Curius declared under oath that he had appropriated for himself no booty ‘praeter guttum faginum, quo sacrificaret’ (N.H. 16.185). For many readers of Latin, the most obvious translation of the Latin is ‘except a beechwood cruet with which he would offer sacrifice’, taking quo as an instrumental ablative and thereby making the vessel an instrument of sacrifice rather than the object of sacrifice itself. Sacrificare is frequently accompanied by an instrumental ablative, but in almost all cases it is clear that the ablative is the object of sacrifice, as in the phrase maioribus hostiis sacrificaverant. 48 Thus the most likely reading of the passage in Pliny is that Curius sacrificed the guttum faginum to the gods. This draws further support from the fact that the object referred to by the instrumental ablatives that accompany the verb sacrificare is almost never a knife, an axe, a hammer, or other weapon. 49 In Latin, one does not sacrifice with a knife or with an axe. One does, however, sacrifice with a cow, with a pig, or with a little cruet.

A parallel use of sacrificare is found in Apuleius’ Apologia 18, a passage which also shares Pliny's focus on poverty:

paupertas, inquam, prisca aput saecula omnium civitatium conditrix, omnium artium repertrix, omnium peccatorum inops, omnis gloriae munifica, cunctis laudibus apud omnis nationes perfuncta. eadem est enim paupertas apud Graecos in Aristide iusta, in Phocione benigna, in Epaminonda strenua, in Socrate sapiens, in Homero diserta. eadem paupertas etiam populo Romano imperium a primordio fundavit, proque eo in <h>odiernum diis immortalibus simpulo et catino fictili sacrificat.

Poverty, I say, is the ancient founder of all states throughout the ages, the discoverer of all arts, devoid of all transgressions, resplendent in every type of glory, and enjoying every praise among all the nations. For this same poverty is, among the Greeks, just in Aristides, kind in Phocion, vigorous in Epaminondas, wise in Socrates, and eloquent in Homer. Also the same poverty has established from the very beginning an empire for the Roman people and, on behalf of this, still today she sacrifices to the immortal gods a little ladle and a dish made of clay.

In both the passages from Pliny and Apuleius, the ritual implements are of diminutive size. The small size of the guttus and simpulum is assured by Varro (L. 5.124), who identifies both as vessels that pour out liquid minutatim. The catinus is a piece of everyday ware used to serve food that contains a lot of liquid (L. 5.120). It is a hallmark of poverty, whether in a religious context or not, appearing often in poetic passages where the narrator describes a low-budget lifestyle. 50

From all this, it is reasonable to conclude that the poor could substitute small vessels for more expensive, edible sacrificial offerings. While vegetal and meat offerings were on a par, inedible gifts could be sacrificed only as substitutes for edible offerings when money was a concern. These two passages from Pliny and Apuleius may provide an explanation for the hundreds of thousands of miniature fictile vessels (plates, cups, etc.) uncovered in votive deposits throughout Italy. Although they are universally referred to as votive offerings in the scholarly literature, it is possible that they are, technically, sacrifices. Furthermore, it seems reasonable to conclude that the miniature clay cows, birds, and other animals that are also commonly found in votive collections were also substitutes for live sacrificial victims. 51

There is, of course, a large leap in scale from two literary references to an explanation for a ritual practice performed in hundreds of locations over many centuries. But in reality, the relative silence of our sources about a ritual form that seems to have been available to the poor is not unique. As a comparandum, we can point to the Roman habit of creating votive deposits, collections of usually relatively inexpensive items buried in the ground: gifts to the gods that had been cleaned out of overstuffed temples and intentionally buried. Despite the fact that the Romans buried broken or superfluous gifts to the gods in deposits for hundreds of years, there are to my knowledge only two references to the practice in all of Latin literature. 52 Another example of the bias of our sources away from rituals performed by the lower classes is the dearth of references to a particular type of item found in votive deposits: anatomical votives, fictile representations of parts of the human body offered to the gods as requests for cures for physical ailments. These offerings, ubiquitous in Roman Italy through to the end of the Republic, are mentioned at most twice in extant Latin literature. 53

At first glance, the Roman habit of sacrificing items that people cannot eat (cruets and small plates) suggests that another dominant strain in modern theorizations of sacrifice might not really apply to the Roman case. Scholars frequently stress the connection between sacrifice and eating: ‘The idea of food underlies the idea of sacrifice.’ 54 But upon further reflection, in fact, the use of cruets and plates actually emphasizes the importance of the meal that concluded a Roman sacrifice. The only inedible items that we know from literary sources were objects of sacrificium are all miniature versions of regular, everyday serveware: a cruet, a plate, and a ladle. It appears that if a worshipper could not afford to sacrifice something that was itself tasty, he might fulfill his obligation by giving something that evoked the idea of it. 55

The link between consumption and sacrifice is also reinforced by a second category of sacrificial items that Romans did not eat: animals, including human animals, that were not regularly included in the Roman diet. While there is a growing body of work done on the osteoarchaeological material from other regions of the Empire, especially the north-western provinces, 56 there is a relative dearth of published studies that deal in any serious way with the collections of bones found on various sites from Roman Italy. 57 One relatively well documented example is the collection of bones dating to the seventh and sixth centuries b.c.e. from the archaic temple at the site of S. Omobono in Rome. 58 Although the remains come from a known sacred area, it should not be assumed that all of them are evidence of sacrificium or other rituals: some may be garbage, material used as fill, or even the remains of animals (like mice) that died while exploring the refuse heap. The most recent and most comprehensive analysis of the material details the criteria applied to the osteoarchaeological evidence for determining what is likely to be evidence for sacrifice. 59 Among these criteria are a clear preference for specific parts of an animal or for animals of a specific age/sex/species, unusual butchery patterns, burning or other alterations to the remains, and the association of the remains with other material (e.g., votive offerings) linked to ritual activity. The S. Omobono material shows a definite preference for certain species (sheep, goats, pigs), 60 for young animals (including foetal and neonatal specimens), 61 and for front limbs. 62

Despite the fact that the S. Omobono assemblage dates to several centuries before the Classical period, the range of faunal remains from the site are primarily what one would expect from a sanctuary based on what we know from literary texts. There is a difference, however. The vast majority of the bones come from pigs, sheep, and goats. Somewhat surprising is the considerably smaller presence of bovines, 63 a more expensive offering that dominates in literary accounts of sacrifice. This disjuncture between physical remains and written accounts is another reminder of the bias of our ancient authors toward the activities of the rich and toward state ritual.

Far less common in the S. Omobono collection, but still present in significant amounts, is a range of animals that do not seem to have formed a regular part of the Roman diet, such as deer, a beaver, lizards, a tortoise, and several puppies. 64 Of these, only dogs are attested in the written sources as victims of Roman sacrifice, albeit rarely. 65 The most famous instance occurred annually at the festival of the Robigalia in June when a red dog and a sheep were sacrificed by the Flamen Quirinalis to ward off rust from the crops. 66 Dogs had other ritual uses as well. At N.H. 29.57–8, Pliny tells us that a dog was crucified annually at a particular location in Rome, and that puppies used to be considered to be such pure eating that they were used in place of victims (hostiarum vice) to appease the divine; puppy was still on the menu at banquets for the gods in Pliny's own day. Neither of the acts that Pliny mentions is explicitly identified as sacrificium, or as any other rite in particular. The skeletal remains of dogs sometimes found interred with human remains or inside city walls are often interpreted as sacrifice by archaeologists. 67 Yet the problem remains that dogs did not form a regular or significant part of the Romans’ diet, nor did wild animals of any sort. 68 How, if these animals did not make desirable entrees, could they be considered suitable for sacrifice?

Although they were not suitable as daily fare, there is evidence that several of the unexpected species from the S. Omobono deposit were edible on special occasions or in dire circumstances: they are surprisingly prevalent in magical and medicinal recipes. In Books 29 and 30 of his Natural History, the elder Pliny includes lizards in numerous medicinal recipes to cure everything from hair loss (29.108) to lower back pain (30.53) to dysentery (30.55), and the only text we have that identifies the contents of a bulla, the amulet worn by young Roman boys, instructs the reader to put lizard eyes inside it. 69 Beavers, too, had curative properties — for example, a mixture of honey wine, anise seed, and beaver oil was thought to cure flatulence (Plin., N.H. 20.193) — and their anal scent glands (mistaken for testicles) were part of the Roman trade in luxury goods. 70 Dogs, and puppies in particular, were thought to have some medicinal and magical properties: Pliny reports that some people thought the ashes of a dog's cranium, when consumed with a beverage, could cure abdominal pain (N.H. 30.53) and, when mixed with honey wine in particular, could cure jaundice (N.H. 30.93). Dog corpses were sometimes deposited with tablets that contained curses, and dog figurines are among the required items for performing some spells. 71

The presence of bones from these species at S. Omobono should not be taken to mean that the site was what scholars call a ‘healing sanctuary’, or that it was a place where people came to cast spells on their enemies. Instead, their presence should be attributed to the status of those species as valuable and efficacious: the prevalence of dogs, lizards, and beavers in medicinal and magical recipes for potions is an indication of the exceptional value the animals were thought to have, an indication that they were somehow special, and therefore might be worthy of the gods.

Another animal sometimes sacrificed by the Romans but not regularly eaten by them is the human animal. On three occasions during the Republic (228, 72 216, 73 and again in 114 or 113 b.c.e. 74 ), the Romans followed instructions from the Sibylline Books to bury alive pairs of Gauls and Greeks, one man and one woman of each, in the Forum Boarium. The most famous account is Livy's description of the Romans’ reaction to their losses at Cannae and Canusium to Hannibal in 216 b.c.e., 75 which I quote at some length because we shall return to this passage later on:

Territi etiam super tantas clades cum ceteris prodigiis, tum quod duae Vestales eo anno, Opimia atque Floronia, stupri compertae et altera sub terra, uti mos est, ad portam Collinam necata fuerat, altera sibimet ipsa mortem consciverat; … Hoc nefas cum inter tot, ut fit, clades in prodigium versum esset, decemviri libros adire iussi sunt et Q. Fabius Pictor Delphos ad oraculum missus est sciscitatum quibus precibus suppliciisque deos possent placare et quaenam futura finis tantis cladibus foret. Interim ex fatalibus libris sacrificia aliquot extraordinaria facta, inter quae Gallus et Galla, Graecus et Graeca in foro boario sub terram vivi demissi sunt in locum saxo consaeptum, iam ante hostiis humanis, minime Romano sacro, imbutum.

In addition to such great disasters, the people were terrified both by other prodigies and because in this year two Vestals, Opimia and Floronia, were discovered to have had illicit affairs. One was killed at the Colline Gate, under the earth as is the custom and the other took her own life … Since this horrible event which occurred in the midst of so many terrible things, as is wont to happen, was turned into a prodigy, the Board of Ten Men was ordered to consult the Books. Q. Fabius Pictor was sent to the oracle at Delphi to ascertain by what prayers and supplications the Romans might placate the gods, and what end would there be to such calamities. Meanwhile, from the Sibylline Books some unusual sacrifices were ordered, among which was one where a Gallic man and woman and a Greek man and woman were sent down alive into an underground room walled with rock, a place that had already been tainted before by human victims — hardly a Roman rite.

There is no question that the live interment of the Gauls and Greeks was a sacrifice: Livy identifies it as one of the sacrifices not part of the usual practice ordered by the Sibylline Books (sacrificia extraordinaria). Livy also uses the language of sacrifice when he describes the underground room as a place that had already seen human victims. 76

Now, the Romans did not eat people, so how does their performance of human sacrifice reinforce the link between sacrifice and dining? The answer is that human sacrifice, which the Romans are quick to dismiss as something other people do (note that, although Livy is clear that the burial of Gauls and Greeks is a sacrifice, he also says that it was ‘hardly a Roman rite’), is closely linked in the Roman mind with cannibalism. For illustration, we can turn once again to the elder Pliny, who writes about the habits of the Gallic tribes north of the Alps: ‘et nuperrime trans Alpis hominem immolari gentium earum more solitum, quod paulum a mandendo abest’ (‘And very recently, on the other side of the Alps, in accordance with the custom of those peoples, individuals were habitually sacrificed, which is not all that far from eating them’ N.H. 7.9).


The preceding discussion has, I hope, made clear that the Romans’ own notion of sacrifice is broader and more complex than is generally perceived. The modern assumption that sacrifice requires an animal victim obfuscates the full range of sacrificium among the Romans. In this section, I make the case that the related and equally widespread notion that all Roman rituals that required the death of an animal were sacrifices obfuscates the variety of rituals that Romans had available to them, effacing some of the fine distinctions Romans made about the ways they approached their gods. Roman sources make clear that Romans had several different rituals (sacrificium, polluctum, and magmentum) that appear, based on prominent structural similarities, to have been related to one another. Scholars are quick to identify all of them as forms of sacrifice, which may well be the case. It is important to remember, however, that no ancient source articulates any sort of relationship among these rituals. Furthermore, although all of these rites were performed on foodstuffs at altars or at least in sanctuaries, there are some critical differences among them and the ways they are discussed by the Romans. Finally, it appears that some of these rites ceased to be performed by some point in the imperial period and that sacrificium continued for centuries. Another possible interpretation of the disappearance of some rituals from Latin literature is that the Romans no longer thought of them as distinct from one another, preferring to treat them all as sacrificium.

It is understandable that, from the etic viewpoint, two rituals performed in roughly the same way should appear to be identical to each other, even if emic accounts distinguish between them. As illustration, let us return to Livy and the human sacrifice in 216 b.c.e. Our author makes clear that the sacrifice of two Gauls and two Greeks happened alongside another ritual: the punishment of an unchaste Vestal Virgin. The same two interment rituals would be performed alongside one another again just about a century later, in 113 b.c.e. This repeated coincidence of ritual performances suggests that the two forms of ritual killing 77 were linked. 78 Indeed these two rituals appear at first glance to be identical — live interment in underground chambers, though admittedly in different locations within the city and with different victims. Livy, however, treats each burial in a distinct way. He stresses the traditional nature of the burial of the one Vestal with the phrase ‘as is the custom (uti mos est)’ and describes her death in neutral terms (necare). 79 He does not use the language of sacrifice, that is, he does not call the ritual a sacrificium nor does he identify the Vestal as a victim. 80 In contrast, as I have pointed out, Livy uses the language of sacrifice to describe the second interment and in the next breath expressly distances Roman tradition from it, calling it ‘a rite scarcely Roman (minime Romano sacro)’. Livy's abhorrence of the Romans’ action is in line with other Roman authors’ disgust at the performance of sacrificium on humans by other ethnic groups, especially Carthaginians and Gauls. 81

Here we have two rituals that look, to an outsider, almost identical, but Livy takes pains to distinguish between them. The burial of Gauls and Greeks was a sacrifice, but one that Romans ought not to have performed. We do not know what name the Romans gave the ritual burial of an unchaste Vestal Virgin, but we know it was not sacrifice. Nor was it secular, capital punishment; the punishment of criminals usually took a more direct and swift form: strangulation, beating, crucifixion, or precipitation (i.e., throwing someone off a cliff). 82

Live interment was only performed by the Romans as ritual killing, but live interment was not the only form of ritual killing (whether human sacrifice or not) that the Romans had available to them. There are at least two other rituals that the Romans performed that also required the death of a person. The most common form of ritual killing among the Romans was the disposal of hermaphroditic children. 83 On fourteen occasions between 209 and 92 b.c.e., androgyne infants and children were included among the prodigies reported to the Roman Senate. The children were drowned by the haruspices, usually in the sea. The Romans then observed a regular set of expiatory rituals, most importantly offerings made to the goddesses Ceres and Proserpina by matrons of the city and the procession of a chorus of twenty-seven virgins.

The other rite observed by the Romans that required a human death was called devotio, and it seems to have been restricted to a single family — father, son, and grandson (it is possible our sources have multiplied a single occasion), all of whom, as commanders-in-the-field, vowed to commit themselves and the enemy troops to the gods of the underworld in order to ensure a Roman victory. Devotio is frequently called ‘self-sacrifice’ by modern scholars, 84 and indeed it certainly fits the modern notion of an act by which one suffers great loss for the benefit of others. It was used by Cicero in the opening of his speech Post Reditum and by the figure of Cotta, consul of 75 b.c.e., in a fragment of Sallust's Historiae to present themselves as victims for the greater good. 85 But while Roman devotio aligns well with our idea of self-sacrifice, it appears that the Romans did not draw a similar connection between devotio and sacrificium. It appears that no Roman source ever uses the language of sacrificium to describe devotio, 86 nor does any Roman author ever express any sort of discomfort with this rite akin to Livy's shrinking back from the sacrifice of Gauls and Greeks. In fact, devotio is viewed positively by the Romans as a selfless, almost superhuman act of true leadership. 87 Of the various forms of ritual killing that were part of their religious experience, the Romans only reacted with disgust to that form they identified as human sacrifice, a distinction in value sometimes lost when all these ritual forms are grouped together under the rubric ‘sacrifice’. 88

The Romans performed at least four forms of ritual killing, only one of which was sacrifice. There is also evidence that the Romans had a variety of rites, only one of which was sacrificium, that involved presenting foodstuffs to the gods. While the evidence does not allow us to recover precise distinctions made among these rites (sacrificium, magmentum, and polluctum), it does strongly suggest that the Romans — at least through the period of the Republic — conceived of these rituals as somehow different from one another. It is commonplace now to treat sacrificium as a general category and to talk about magmentum and polluctum as moments within the larger ritual or special instances of it. 89 It is possible that this genus-species relationship in fact existed in the Roman mind, as is perhaps suggested by the fact that sacrificare means ‘to make sacred’, and these other rituals seem to be different ways of doing the same work, namely transferring items from human to divine ownership. It is important to note, however, that we cannot determine conclusively from the extant sources what relationship, if any, existed among them in the Roman mind.

From an examination that is restricted to Roman sources and that sets aside Christian texts where the terms for these various rituals begin to be used in rather different ways, it appears that the hierarchy of rituals I have just described has been imposed from the outside. Although Roman sources identify some specific types of sacrifice (e.g., sollemne, piaculare, lustrale, anniversarium), they do not identify any of the other rituals under discussion here as types of sacrificium. 90 Of this class of rituals, sacrificium does seem to have been somehow different from the others. It is the only one of these terms that does not come to be used outside the realm of the divine. It is also noteworthy that sacrificium appears to be the only member of this class to require mola salsa. Finally, while other rituals seem to have fallen into desuetude, or at least to have fallen out of the literature, by the late Republic or early Empire, sacrificium remained a vital part of Roman religious life for centuries.

The distinction between sacrificare and mactare was lost by Late Antiquity, but it was still active in the Republic and early Empire. 91 Mactare is another ritual performed on animals (referred to as hostiae and victimae) at an altar, but also on porridge (Nonius 539L). The ancients derived the term from magis auctus and understood it to mean ‘to increase’ and by extension ‘to honour with’. 92 The corresponding substantive is magmentum, a type of offering laid out only at certain temples. 93 Unlike sacrificare, which remained solely in the divine realm, mactare did not need to involve the gods: mactare is something that one Roman could do to another, both literally (one can mactare someone else with a golden cup, for example) and metaphorically (with misfortune or expense). Another way that mactare is different is that gods can mactare mortals — at least in comedy, where characters sometimes wish that the gods would honour their enemies with trouble. 94

Another example of a ritual that looks a lot like sacrificium but is not identical to it is polluctum. Pollucere is an old word, appearing mostly in literature of the second century b.c.e., 95 but in later texts as well. The ritual seems to be even more flexible than sacrificium in the range of objects on which it could be performed. The elder Cato instructs his reader to pollucere a cup of wine and a daps (ritual meal) to Jupiter Dapalis (Agr. 132.1–2). One can also pollucere grain, wine, oil, cheese, meat, fish with scales, a host of other food items, and even unidentified (and presumably inedible) goods. 96 The ritual is so closely tied to the notion of dining that polluctum could be used for everyday meals (e.g., Plaut., Rud. 1419). In the sacred realm, Romans could also pollucere a tithe to the god Hercules. 97 Other than the range of items that can be polluctum, the only other thing we know about the ritual is that it involved an altar, which is, of course, the proper locus of sacrifice. Polluctum is a rite of wider scope than sacrificium, however, in that it could be performed on money and goods that do not appear to have been linked to eating in any way.

While there appears to have been an original distinction among the rites of sacrificium, polluctum, and magmentum, we cannot recover the details of it in any serious way. The hypothesis that only sacrificium required mola salsa is strongly supported by the sources, but because that is an argument ex silentio, it cannot be proved beyond all doubt. Some more support for the notion that these were not interchangeable can be drawn from material evidence, visual representations of the moment of ritual slaughter. Aldrete's survey of images commonly identified as sacrifice scenes makes clear that Roman art depicts different procedures (hitting with a hammer, chopping with an axe) and implements (hammers, axes, knives), and that the preference of implement changes over time. Aldrete counts at least fifty-six sculptural reliefs dating from the seventh century b.c.e. to the fourth century c.e. that contain scenes of ritual slaughter where the implements can be clearly discerned. 98 Of the fifty-six reliefs, forty-one show officials carrying axes. Hammers appear in only fifteen scenes, two-thirds of which date between the first century b.c.e. and the second century c.e. Knives would have been used only in conjunction with one or other of these implements. To explain the decision to sometimes portray one weapon instead of the other, Aldrete posits that ‘various gods, cults, and rituals may have dictated certain procedures or tools’. 99 Perhaps these reliefs preserve the performance of one or more of the rituals that seem to have faded in popularity by the high imperial period: magmentum and polluctum.

That we cannot fully recover what were the critical differences among these rites is frustrating, but the situation is certainly not unique in the study of Roman religion. To give just a single example, we know that there was originally some technical distinction among the different types of divine signs sent to the Romans by the gods. But we can no longer recover — indeed it appears that Romans of the early Empire could no longer recover — what was the difference between a monstrum, a prodigium, a portentum, and an ostentum. 100 The limited sources we have are imprecise in their use of the terms — even Cicero, who was an augur and was surely aware of the distinction. 101


Looking at Roman sacrifice through the insider-outsider lens lets us see more clearly that, for the Romans, sacrifice was both more and less than it is for many scholars writing about it today. Once we have recognized that there are two notions of sacrifice at play, we can set aside our etic, outsider ideas for the moment and look at the Roman sources anew. What we find is that for the Romans, to sacrifice was not simply to kill in a ritual fashion. Sacrificium is the performance of a complex of actions that presents the gods with an edible gift by the sprinkling of mola salsa and the ultimate goal of which seems to be the feeding of both gods and men. We also find that the gods were open to receiving sacrifices of vegetables, grains, liquids, and, when those were not available, miniature versions of the serveware that would normally have contained them. In this way, the native, or emic, Roman view of sacrifice is more expansive than ours. The expanded range of sacrificium suggests that meat and vegetal produce were both welcomed by the gods, and that we should not assume that meat offerings were necessarily privileged over other gifts in every circumstance. In addition, the acceptability of miniature serveware as objects of sacrificium shows the ability of the ritual to accommodate the varying social status of those performing it. All of this indicates a certain flexibility and elasticity in the ritual of sacrificium that suggests, especially if a similar flexibility could be demonstrated in other ritual forms, a need to moderate the emphasis — both ancient and modern — on the orthopractic nature of Roman religion. On a wider scale, the arguments made here about the nature of Roman sacrificium further undermine the increasingly discredited idea that sacrifice as a universal human behaviour is primarily, if not exclusively, about the violence of killing an animal victim.

Sacrifice was just one of several rites (alongside polluctum and magmentum) that the Romans had available to them that look to us, standing outside their religious system, as if they were all identical — or nearly so. By placing this variety of rites that the Romans had under the single rubric of ‘sacrifice’, we have lost sight of some of the complexity and nuance of Roman ritual life. This has repercussions for our understanding of some elements of Roman religious thought. For example, the apparent contradiction between Roman abhorrence of ritual killing and the frequency with which Romans performed various forms of it is, to a large extent, explicable once it is recognized that the Romans objected only to the performance (by themselves as much as by others) of sacrificium on human victims. Other forms of ritual killing do not receive the same sort of negative judgement by Roman authors, and one form, devotio, even has strongly positive associations. Furthermore, because there were multiple rituals — not just sacrificium — through which the Romans could share food and other goods with their gods, we can see that the Romans had a wider range of ritual tools available to them for communicating with the divine. By looking at Roman sacrificium through the insider-outsider lens, by keeping in sight what is there in the sources, what we add to it, and where our modern notion of sacrifice does and does not align with the Romans’ own idea, we have a sharper, more detailed picture of one aspect of Roman antiquity.



Cornell, T. J. (ed.) 2013: The Fragments of the Roman Historians, 3 vols, Oxford


Hornblower, S., and Spawforth, A. 1996: The Oxford Classical Dictionary (3rd edn), Oxford

1 Miner 1956: 503. The article is reprinted in McCutcheon 1999, a volume that offers in its introductory chapter a very good overview of the insider-outsider problem and that includes a selection of some of the most important scholarly contributions to the debate within the study of religion. The issue remains active in religious studies, as it does in cultural anthropology more widely.

2 I presume that Miner's observations apply also to bathroom habits elsewhere in North America and Europe.

3 See, for example, Feeney 2004, an excellent discussion of the application of theoretical models of sacrifice to the poetry of Vergil and Ovid. For a treatment of this methodological issue on a broader scale, see the rather pointed critique in Hopkins 1978: 180–8.

4 Emic and etic, terms drawn originally from the field of linguistics (Pike 1967: 37–44; reprinted in McCutcheon 1999: 28–36), are one of several pairs of words used to present the insider-outsider distinction. Others include first-order vs. second-order categories, particular vs. universal, descriptive vs. redescriptive, and local vs. global.

5 See, for example, Morris et al. 1999 and Berry 1990.

6 See, however, C. Ando's concluding essay in Faraone and Naiden 2012 along with A. Hollman's review of that same volume in BMCR 2013.04.44 and, in the same vein but with reference to ancient Egypt, Frankfurter 2011.

7 For this discussion, the metaphorical extension of the English word ‘sacrifice’, by which one can sacrifice for one's family or hit a sacrifice fly in baseball, is not relevant: this meaning is completely unknown to the Romans of the Classical period. Also unfamiliar to the Romans would be another use of ‘sacrifice’ now current in the life sciences, as a term for euthanasia of research animals with no real religious significance The plea of an editorial in the Canadian Journal of Comparative Medicine and Veterinary Science from 1967 (p. 241) that researchers abandon the term because there is no deity involved in the act of euthanizing laboratory animals, fell on deaf ears: ‘sacrifice’ remains common in animal management literature.

8 OED, s.v. ‘sacrifice, n.’.

9 e.g., J. Scheid, s.v. ‘sacrifice, Roman’ in OCD 3, 1345–6; Prescendi 2007: 122–5; Rüpke 2007: 137–8. Cf., n. 89 below.

10 Macr., Sat. 3.3.2, citing the late republican jurist Trebatius; Prescendi 2007: 25–6.

11 McClymond treats sacrificial events as ‘clusters of different types of activities’, including prayer, killing, cooking, and consumption, which are not in and of themselves sacrificial (they are frequently performed in other contexts), but which become sacrificial in the aggregate (McClymond 2008: 25–34).

12 Scheid 2005: 44–57; 2007: 263–9. Scheid's reconstruction and interpretation is followed by Prescendi 2007: 31–48. Rüpke 2005 offers a different interpretation of the meal that follows the sacrifice.

13 On the early Christian appropriation and transformation of Roman sacrificial imagery and discourse, see Castelli 2004: 50–9.

14 ‘Arguably, then, it is the Christians who bequeathed to future generations the metonymic equivalence of sacrifice and violence’, Knust and Várhelyi 2011: 17. Elsner 2012 emphasizes the heavy influence of early Christian writers on modern theorizations of sacrifice.

15 Burkert 1983: 3; Girard 1977: 1. Concise surveys of the major modern theories of sacrifice in the ancient world can be found in Knust and Várhelyi 2011: 4–18, Lincoln 2012, and Graf 2012. A wider range of scholarly approaches is presented by McClymond 2008: 1–24.

16 It is entirely possible that the search for a single, critical moment where a change from profane to sacred occurs is, in fact, a modern preoccupation.

17 This is made clear in numerous passages from several Roman authors. As an example, I offer Var., R. 1.2.19: ‘Itaque propterea institutum diversa de causa ut ex caprino genere ad alii dei aram hostia adduceretur, ad alii non sacrificaretur, cum ab eodem odio alter videre nollet, alter etiam videre pereuntem vellet. Sic factum ut Libero patri, repertori vitis, hirci immolarentur, proinde ut capite darent poenas; contra ut Minervae caprini generis nihil immolarent propter oleam, quod eam quam laeserit fieri dicunt sterilem’ (‘And so therefore, it has been established by opposing justifications that victims of the caprine sort are brought to the altar of one deity, but they are not sacrificed at the altar of another, since on account of the same hatred, one does not want to see a goat and the other desires to see one perish. Thus it happens that goats are immolated to Liber Pater, who discovered the vine, so that they pay him a penalty and, by a contrary logic, caprine victims are never immolated to Minerva on account of the olive: they say that whatever olive plant a goat bites becomes sterile’).

18 Scheid 2005: 50–5.

19 Ernout and Meillet 1979: 411 s.v. molo; Walde and Hofmann 1954: 2.104–6 s.v. molo; de Vaan 2008: 386–7 s.v. molo.

20 Fest. 124L, s.v. mola.

21 cf. Serv., A. 4.57 and 10.541.

22 e.g., Faraone and Naiden 2012: 4; Prescendi 2007: 36 and 108–9. An exception is Scheid 2005: 52.

23 Plin., N.H. 31.89 is usually taken to refer to sacrifice (so Prescendi 2007: 105) but the text mentions only sacra, not sacrificia.

24 Two famous examples are found on the altar of Domitius Ahenobarbus (Ryberg 1955: fig. 17a–c) and the Cancellaria relief (Ryberg 1955: fig. 37a–b). I follow Elsner 2012: 121 in setting aside the plethora of images of the tauroctony of Mithras and the taurobolium of Cybele and Attis.

25 Ryberg 1955: figs 83 and 89b. Another famous instance of this scene is on the Boscoreale cup (Aldrete 2014: 33, fig. 1).

26 As in a relief from the Forum of Trajan now in the Louvre (Ryberg 1955: fig. 69a).

27 Elsner 2012: 123.

28 Graf 2012: 46–7.

29 Van Straten 1995: 188.

30 Huet 2005; 2007.

31 Elsner 2012: 160–1.

32 The lack of interest in vegetal sacrifice is widespread in the field of religious studies (McClymond 2008: 65). Significant exceptions to this rule in the study of Roman sacrifice are the treatment of the sacrifice of wheat and wine in Scheid 2012 and the argument for the increased popularity of vegetal sacrifice in Late Antiquity advanced by Elsner 2012. The quotation comes from Frankfurter 2011: 75.

33 This assertion is based on a search of ‘sacrific*’ on the Brepolis Library of Latin Texts A.

34 Goats: Var., R. 1.2.19; Liv. 45.16.6. Dogs: Fest. 358L. Horses: Plin., N.H. 28.146; Fest. 190L s.v. October equus. Birds: Suet., Calig. 22.

35 MacKinnon 2004: 59–74.

36 Scheid 2005: 100–2; 2012: 84. Contra Prescendi 2007: 22–3.

37 Var., L. 6.3.14. Compare Var., R. 2.8.1.

38 CIL 6.32323.139–40 = ILS 5050.139–40 = Pighi 1965: 117 (from Rome).

39 Paul. ex Fest. 77L, s.v. fabam and Fest. 344L, s.v. refriva faba.

40 Var., L 5.122. See also Scheid 2012: 90–1.

41 Sacrifices of wine and incense are common in the Commentarii Fratrum Arvalium, e.g. Scheid 1998: nn. 58.47, 64.1.46–7, and 68.1.49.

42 Var., R. 2.11.5.

43 Paul. ex Fest. 423L, s.v. sacrima.

44 Fest. and Paul. ex Fest. 286L and 287L, s.v. pecunia sacrificium; Paul. ex Fest. 423L s.v. sacrima.

45 This is suggested by Ov., F. 1.127–8.

46 There is a small amount of evidence for a form of auspicium performed with beans: Fest. and Paul. ex Fest. 344L and 345L, s.v. refriva faba; Plin., N.H. 18.119.

47 Analyses of the traditions about Curius and his contemporary Fabricius, both famous for prudentia and paupertas, are found in Berrendonner 2001 and Vigourt 2001.

48 e.g., Liv. 22.1.19; 45.16.6; Plin., N.H. 36.39; Tac., Ann. 6.34. Cf. Macr., Sat. 3.2.16.

49 To my knowledge, the sole exception is a phrase preserved twice in the Commentarii Fratrum Arvalium (Scheid 1998: nn. 55.1.20 and 58.13) where the presence of an accusative object of immolare necessitates that cultro be instrumental in the traditional sense: ‘ture et vino in igne in foculo fecit immolavitque vino mola cultroque Iovi o(ptimo) m(aximo) b(ovem) m(arem), Iunoni reginae b(ovem) f(eminam), Minervae b(ovem) f(eminam), Saluti publicae populi Romani Quiritium b(ovem) f(eminam).’

50 Lucil. frag. 450 Krenkel; Hor., Sat. 1.3.90 and 1.6.115; Juv. 6.343 and 11.108.

51 As suggested by Bouma 1996: 1.238–41.

52 Aul. Gel. 2.10.3–4, quoting a letter of Varro, and Paul. ex Fest. 78L, s.v. favisae.

53 Augustine, Civ. D. 6.9 (which probably draws on Varro) and possibly Paul. ex Fest. 93L, s.v. ipsilles with 398L, s.v. subsilles.

54 Douglas 1982: 117. This line of interpretation has enjoyed a wider influence in the study of Classical Antiquity than work along the lines of Burkert 1983 and Girard 1977, and the bibliography is enormous. Foundational is the collection of essays on Greek sacrifice in Detienne and Vernant 1989.

55 It is entirely possible that miniature ceramics were not, in reality, less expensive offerings than actual foodstuffs. Pliny and Apuleius may reflect an élite misconception about the religious praxis of lower class worshippers, offering an incorrect, emic interpretation of an observable phenomenon. Even if this is the case, the argument still stands that these passages underscore how essential was consumption to the ritual of sacrificium.

56 e.g., Martens 2004 and Lentacker, Ervynck and Van Neer 2004 on a mithraeum at Tienen in Belgium, King 2005 on Roman Britain, and the various contributions to Lepetz and Van Andringa 2008 on Roman Gaul.

57 A brief survey of the bone assemblages from sites in west-central Italy is offered by Bouma 1996: 1.228–41.

58 Ioppolo 1972; Tagliacozzo 1989.

59 Moses, forthcoming.

60 Moses (forthcoming, table 2) reports that these species account for 89.9 per cent of the total number of individual animal specimens recovered.

61 Tagliacozzo 1989: 66.

62 Moses, forthcoming, table 8.

63 8.9 per cent of the total, according to Moses, forthcoming, table 2.

64 There is some limited zooarchaeological evidence for the consumption of dogs at some Roman sites, such as the inclusion of dog bones bearing marks of butchery among bone deposits that comprise primarily bovine and ovine remains, but it is not widespread. As the most extensive survey of meat production in Roman Italy has concluded, ‘Dogs were variously trained as guards, protectors, companions, and pets, but they were not raised to be eaten’ (MacKinnon 2004: 74).

65 The only Roman reference to the sacrifice of a deer pertains to a Greek context: Ov., F. 1.387–8 where the deer is sacrificed to Diana as a substitute for Iphigenia.

66 Ov., F. 4.901–42 with Fest. 358L, s.v. rutilae canes; Var., L. 6.16. The offering of a dog to Robigus may be the same ritual as the augurium canarium referred to by Plin., N.H. 18.14. Columella 2.21.4 might also refer to dog sacrifice, but the verb (feceris) leaves it ambiguous as to which ritual was being performed. Pliny reports a ritual, possibly sacrifice (‘res divina fit’, 29.58), involving a dog in honour of the little-known goddess, Genita Mana (cf. Plut., RQ 52 = Mor. 277A–C). Plutarch is the only source for dog sacrifice at the Lupercalia (RQ 68 and 111 = Mor. 280 B–C and 290D; Rom. 21.5).

67 See, for example, Wilkens 2006 and De Grossi Mazzorin and Minniti 2006.

68 On the general absence of wild meat from the Roman diet, see MacKinnon 2004: 190–2. Although there is some evidence for Roman consumption of dog in the form of canine skeletons with butchery marks (e.g., De Grossi Mazzorin and Tagliacozzo 1997: 437–8), there is no evidence that dogs were raised for meat production (MacKinnon 2004: 74). For an argument that wild animals are more common in ancient Mediterranean, and specifically in Etruscan, sacrifice than is generally acknowledged, see Rask 2014.

69 Marcellus, de Medicamentis 8.50; Palmer 1996: 23–4.

70 Devecka 2013.

71 Ankarloo and Clark 1999: 75–6; Wilburn 2012: 87–90.

72 The numerous sources for this event are collected and analysed in Engels 2007: 416–18, 443–8. For the possible link between this instance and the revelation of an unchaste Vestal, see Schultz 2012: 126 n. 18

73 Liv. 22.57.2–6; Cass. Hemina fr. 32 Peter = FRH F33.

74 Plu., RQ 83 = Mor. 283F–284C; Liv., Per. 63.

75 Liv. 22.57.2–6, discussed also in Schultz 2012: 126–7.

76 On the Latin terminology for living sacrificial victims, see Prescendi 2009.

77 I use ‘ritual killing’ as a blanket term for any rite, including but not limited to sacrifice, that involves the death of a human being. For the difference in Roman attitudes toward human sacrifice and other forms of ritual killing, see Schultz 2010.

78 The exact nature of the connection between the two rituals is not clear, but I agree with Eckstein 1982 that we should not see the sacrifice of Gauls and Greeks as some sort of atonement for the unchastity of the Vestals.

79 Adams 1973: 280–90.

80 There is no evidence, contra Parker 2004 and Wildfang 2006: 58–9, that the Romans ever perceived the punishment of a Vestal as sacrifice. For a more extended analysis of the distinction between the punishment of unchaste Vestals and, on the one hand, sacrifice and, on the other, secular capital punishment, see Schultz 2012.

81 For example, Cic., Rep. 3.15 and Font. 31; Plin., N.H. 36.39; Tac., Ann. 14.30; Sil. 3.763–829.

82 Prescendi 2007: 224–41 and, arriving at the same conclusion by a different path, Schultz 2012: 132–3.

83 MacBain 1982: 127–35; Schultz 2010: 529–30; 2012: 129–30.

84 e.g., O'Gorman 2010: 121–7 and Versnel 1976.

85 Cic., Red. pop. 1; Sall., Hist. 2.47.10 (M) = 2.44.10 McGushin. See Rosenblitt 2011 for the connection between these two passages.

86 The closest any Roman source comes to linking devotio and sacrifice is Cic., Off. 3.95: ‘Quid <quod> Agamemnon, cum devovisset Dianae quod in suo regno pulcherrimum natum esset illo anno, immolavit Iphigeniam, qua nihil erat eo quidem anno natum pulchrius?’ Because the context is Greek, it is safe to assume that Cicero is using, as he often does elsewhere when addressing a general audience, technical terms in a very general way. Devotio is primarily a form of vow that is, ideally, followed by a death (‘si is homo qui devotus est moritur, probe factum videri’ (Liv. 8.10.)). See Oakley 1998: 481 and Sacco 2004: 316. In Livy's account of the first devotio in 340 b.c.e. at the battle of the Veseris between Rome and the Latins (8.9.1–14), the ritual consists of the recitation of the dedicatory formula by the consul P. Decius Mus while in the midst of battle. The description of Decius’ ensuing death is very spare and devoid of any sacrificial imagery or terminology. At 8.10.11–12, Livy notes that a commander could devote one of his soldiers rather than himself. If the devotio was not successful (i.e., the devotus somehow survived), expiatory steps had to be taken: the burial of a larger-than-life-sized statue and piaculum hostia<m> caedi. If the commander who devoted himself did not die in battle, he was interdicted from performing any ritual on behalf of the state (publicum divinum).

87 As illustrated by Livy's description of the first Decius to perform the ritual as he rode out to meet the enemy: ‘aliquanto augustior humano visu, sicut caelo missus’ (8.9.10).

88 Schultz 2010: 520–2.

89 As is implied in all the relevant entries in the OLD. See also n. 9 above.

90 This statement and much of what follows is based on a series of searches in the Brepolis on-line database of Latin literature, Libraries A and B ( conducted throughout the summer of 2015. The database is a very useful, but not infallible tool.

91 The distinction is preserved by Suet., Prat. 176 and Serv., A. 4.57. Nonius 539L identifies mactare with immolare, but the texts he cites do not really support his claim. Plaut., Amph. 1034 seems to draw an equivalence between sacrificare and mactare (cf. Val. Max. The relationship between magmentum and augmentum (Paul. ex Fest. 113L, s.v. magmentum; Serv., A. 4.57) is not clear.

92 Paul. ex Fest. 112–13L, s.v. mactus; Serv., A. 9.641. Modern etymologists disagree on the origin of the term. Ernout and Meillet 1979: 376 s.v. mactus; Walde and Hofmann 1954: 2.4 s.v. mactus; de Vaan 2008: 357 s.v. mactus.

93 Var., L. 5.112; see also Cic., Har. Resp. 31. Magmentum also appears in two imperial leges sacrae pertaining to the observance of the Imperial cult preserved in inscriptions found in the Roman colonies of Salona in Dalmatia (CIL 3.1933, dated to 137 c.e.) and Narbo in Gallia Narbonensis (CIL 12.4333, dated to 11 c.e.). The two texts are nearly identical and perhaps go back to the original lex sacra of the altar of Diana on the Aventine hill in Rome, to which the inscriptions explicitly appeal.

94 See, for example, citations from Pomponius and Afranius in Non. Mar. 540–1L.

95 Plaut., Stich 233; Cato, Agr. 132; Cass. Hemina fr. 13 Peter = FRH F17.

96 Fest. 298L, s.v. pollucere.

97 Plaut., Stich. 233; CIL 12.1531 = ILLRP 136 = ILS 3411 (from Sora). Also Var., Men. 413 = Macr., Sat. 3.12.2. It is probable, but not certain, that this is the same as the polluctum of ex mercibus libamenta mentioned by Varro at L. 6.54. Cato's instruction to pollucere to Jupiter an assaria pecunia refers to produce valued at one as (Agr. 132.2; Scheid 2005: 136–9). Paul. ex Fest. 287L, s.v. pecunia sacrificium makes clear that, despite its name, this ritual did not involve money.

98 Aldrete 2014: 32. Of these, three-fourths come from the first and second centuries c.e.

99 Aldrete 2014: 48.

100 The problem is widely acknowledged, but see specifically Moussy 1977; 1990; Engels 2007: 259–82. Similar difficulties beset efforts, both ancient and modern, to reconstruct the technical differences among the concepts of sacer, sanctus, and religiosus: see Rives 2011.

101 Throughout his corpus Cicero uses a range of technical divinatory terms, including augur, ostentum, and portentum, in rather general ways, even in De Divinatione where one might reasonably expect him to be more precise.


Adams, J. N. 1973: ‘Two Latin words for “kill”’, Glotta 51, 280–92
Aldrete, G. S. 2014: ‘Hammers, axes, bulls, and blood: some practical aspects of Roman animal sacrifice’, Journal of Roman Studies 104, 2850
Ankarloo, B., and Clark, S. (eds) 1999: Witchcraft and Magic in Europe: Ancient Greece and Rome, Philadelphia
Berrendonner, C. 2001: ‘La formation de la tradition sur M’. Curius Dentatus et C. Fabricius Luscinus: un homme nouveau peut-il être un grand homme?’, in Coudry and Späth 2001, 97–116
Berry, J. W. 1990: ‘Imposed etics, emics, and derived etics: their conceptual and operational status in cross-cultural psychology’, in Headland, T. N., Pike, K. L. and Harris, M. (eds), Emics and Etics: The Insider/Outsider Debate, Newbury Park, 8499
Bouma, J. W. 1996: Religio Votiva: The Archaeology of Latial Votive Religion, 3 vols, Drachten/Groningen
Bradley, M., and Stow, K. (eds) 2012: Rome, Pollution and Propriety: Dirt, Disease and Hygiene in the Eternal City from Antiquity to Modernity, Cambridge
Burkert, W. 1983: Homo Necans: The Anthropology of Ancient Greek Sacrificial Ritual and Myth, trans. Bing, P., Berkeley
Castelli, E. A. 2004: Martyrdom and Memory: Early Christian Culture Making, New York
Coudry, M., and Späth, T. (eds) 2001: L'Invention des grands hommes de la Rome antique, Paris
De Grossi Mazzorin, J., and Minniti, C. 2006: ‘Dog sacrifice in the ancient world: a ritual passage?’, in Snyder and Moore 2006, 62–6
De Grossi Mazzorin, J., and Tagliacozzo, A. 1997: ‘Dog remains in Italy from the Neolithic to the Roman period’, Anthropozoologica 25/26, 429–40
Detienne, M., and Vernant, J. P. 1989: The Cuisine of Sacrifice among the Greeks, trans. Wissing, P., Chicago
De Vaan, M. 2008: Etymological Dictionary of Latin and the Other Italic Languages, Leiden
Devecka, M. 2013: ‘The traffic in glands’, Journal of Roman Studies 103, 8895
Douglas, M. 1982. ‘Food is not feed’, in Douglas, M., In the Active Voice, London, 117–24 (originally published in J. Kuper (ed.) 1977: The Anthropologist's Cookbook, New York, 1–7)
Eckstein, A. M. 1982: ‘Human sacrifice and fear of military disaster in Republican Rome’, American Journal of Ancient History 7, 6995
Elsner, J. 2012: ‘Sacrifice in late Roman art’, in Faraone and Naiden 2012, 120–63
Engels, D. 2007: Das römische Vorzeichenwesen (753–27 v. Chr.), Potsdamer Altertumswissenschaftliche Beiträge 22, Stuttgart
Ernout, A., and Meillet, A. 1979: Dictionnaire étymologique de la langue latine (4th edn), Paris
Faraone, C. A., and Naiden, F. S. (eds) 2012: Greek and Roman Animal Sacrifice, Cambridge
Feeney, D. 2004: ‘Interpreting sacrificial ritual in Roman poetry: disciplines and their models’, in Barchiesi, A., Rüpke, J. and Stephens, S. (eds), Rituals in Ink: A Conference on Religion and Literary Production in Ancient Rome, Potsdamer Altertumswissenschaftliche Beiträge 10, Stuttgart, 121
Frankfurter, D. 2011: ‘Egyptian religion and the problem of the category “sacrifice”’, in Knust and Várhelyi 2011, 75–93
Girard, R. 1977: Violence and the Sacred, trans. Gregory, P., Baltimore
Graf, F. 2012: ‘One generation after Burkert and Girard: where are the great theories?’, in Faraone and Naiden 2012, 32–51
Hopkins, K. 1978: ‘Rules of evidence’, Journal of Roman Studies 68, 178–86
Huet, V. 2005: ‘La mise à mort sacrificielle sur les reliefs romains’, in Bertrand, J.-M. (ed.), La Violence dans les mondes grec et romain, Paris, 91119
Huet, V. 2007: ‘Le sacrifice disparu: les reliefs de boucherie’, in van Andringa, W. (ed.), Sacrifices, marché de la viande et pratiques alimentaires dans les cités du monde romain, Food and History 5, Turnhout, 197223
Ioppolo, G. 1972: ‘I reperti ossei animali nell'area archeologica di S. Omobono (1962–1964)’, Rendiconti della Pontificia Accademia Romana di Archeologia 44, 339
Johnston, S. I. (ed.) 2004: Religions of the Ancient World, Cambridge, MA
King, A. 2005: ‘Animal remains from temples in Roman Britain’, Britannia 36, 329–69
Knust, J. W., and Várhelyi, Z. (eds) 2011: Ancient Mediterranean Sacrifice, Oxford
Krenkel, W. 1970: Lucilius Satiren (2 vols), Leiden
Lentacker, A., Ervynck, A., and Van Neer, W. 2004: ‘The symbolic meaning of the cock: the animal remains from the Mithraeum at Tienen (Belgium)’, in Martens, M. and De Boe, G. (eds), Roman Mithraism: the Evidence of the Small Finds, Brussels, 5780
Lepetz, S., and van Andringa, W. (eds) 2008: Archéologie du sacrifice animal en Gaule romaine, Montagnac
Lincoln, B. 2012: ‘From Bergaigne to Meuli: how animal sacrifice became a hot topic’, in Faraone and Naiden 2012, 13–31
MacBain, B. 1982: Prodigy and Expiation: A Study in Religion and Politics in Republican Rome, Collection Latomus 177, Brussels
MacKinnon, M. 2004: Production and Consumption of Animals in Roman Italy, Journal of Roman Archaeology Supplementary Series 54, Portsmouth, RI
Martens, M. 2004: ‘Re-thinking sacred “rubbish”: the ritual deposits of the temple of Mithras at Tienen (Belgium)’, Journal of Roman Archaeology 17, 333–53
McClymond, K. 2008: Beyond Sacred Violence: A Comparative Study of Sacrifice, Baltimore
McCutcheon, R. T. (ed.) 1999: The Insider/Outsider Problem in the Study of Religion: A Reader, London
Miner, H. 1956: ‘Body ritual among the Nacirema’, American Anthropologist 58, 503–7
Morris, M. W., Leung, K., Ames, D., and Lickel, B. 1999: ‘Views from inside and outside: integrating emic and etic insights about culture and justice judgment’, The Academy of Management Review 24, 781–96
Moses, V. C. forthcoming: ‘Zooarchaeological report on the Area Sacra di Sant'Omobono ’, in Brocato, P. and Terrenato, N. (eds), Ricerche nell'area dei templi di Fortuna e Mater Matuta
Moussy, C. 1977: ‘Esquisse de l'histoire de monstrum ’, Revue des études latines 55, 345–69
Moussy, C. 1990: ‘Un problème de synonymie: ostentum et portentum ’, Revue de philologie, de littérature et d'histoire anciennes 64, 4760
Oakley, S. P. 1998: A Commentary on Livy, Books VI–X, vol. 2, Oxford
O'Gorman, E. 2010: ‘History as group fantasy’, Cultural Critique 74, 117–30
Palmer, R. E. A. 1996: ‘Locket gold, lizard green’, in Hall, J. F. (ed.), Etruscan Italy: Etruscan Influences on the Civilizations of Italy from Antiquity to the Modern Era, Provo, 1727
Parker, H. 2004: ‘Why were the Vestals virgins? Or the chastity of women and the safety of the state’, American Journal of Philology 125, 563601
Peter, H. 1914: Historicorum Romanorum Reliquiae (2nd edn), 2 vols, Leipzig
Pighi, I. B. 1965: De Ludis Saecularibus (2nd edn), Amsterdam
Pike, K. L. 1967: Language in Relation to a Unified Theory of the Structure of Human Behavior (2nd edn), The Hague
Prescendi, F. 2007: Décrire et comprendre le sacrifice, Potsdamer Altertumswissenschaftliche Beiträge 19, Stuttgart
Prescendi, F. 2009: ‘La vittima non è un'ostia: Riflessioni storiche e linguistiche su un termine di uso corrente’, Mythos 3, 145–56
Rask, K. A. 2014: ‘Etruscan animal bones and their implications for sacrificial studies’, History of Religions 53, 269312
Rives, J. 2011: ‘Control of the sacred in Roman law’, in Tellegen-Couperus, O. (ed.), Law and Religion in the Roman Republic, Mnemosyne Supplement 336, Leiden, 165–80
Rosenblitt, J. A. 2011: ‘The “devotio” of Sallust's Cotta’, American Journal of Philology 132, 397427
Rüpke, J. 2005: ‘Gäste der Götter – Götter als Gäste: zur Konstrucktion des römischen Opferbanketts’, in Georgoudi, S., Piettre, R. Koch and Schmidt, F. (eds), La Cuisine et l'autel: les sacrifices en questions dans les sociétés de la Méditerranée ancienne, Turnhout, 227–39
Rüpke, J. 2007: Religion of the Romans, trans. Gordon, R., Cambridge
Ryberg, I. 1955: Rites of the State Religion in Roman Art, American Academy in Rome Memoirs 22, Rome
Sacco, L. 2004: ‘Devotio’, Studi Romani 52, 312–52
Scheid, J. 1998: Commentarii Fratrum Arvalium Qui Supersunt: les copies épigraphiques des protocoles annuels de la confrérie Arvale (21 av.–304 ap. J. C.), Roma Antica 4, Rome
Scheid, J. 2005: Quand faire, c'est croire: les rites sacrificiels des romains, Paris
Scheid, J. 2007: ‘Sacrifices for gods and ancestors’, in Rüpke, J. (ed.), A Companion to Roman Religion, Malden, 263–71
Scheid, J. 2012: ‘Roman animal sacrifice and the system of being’, in Faraone and Naiden 2012, 84–95
Schultz, C. E. 2010: ‘The Romans and ritual murder’, Journal of the American Academy of Religion 78, 516–41
Schultz, C. E. 2012: ‘On the burial of unchaste Vestal Virgins’, in Bradley and Stow 2012, 122–35
Snyder, L. M., and Moore, E. A. (eds) 2006: Dogs and People in Social, Working, Economic or Symbolic Interaction, Proceedings of the 9th ICAZ Conference, Durham 2002, Oxford
Tagliacozzo, A. 1989: ‘Annalisi dei resi faunistici dell'area sacra di S. Omobono’, in Il Viver quotidiano in Roma arcaica: materiali dagli scavi del tempio arcaico nell'area sacra di S. Omobono, Rome, 65–9
Van Straten, F. T. 1995: Hiera Kala: Images of Animal Sacrifice in Archaic and Classical Greece, Leiden
Versnel, H. 1976: ‘Two types of devotio ’, Mnemosyne 29, 365410
Vigourt, A. 2001: ‘M’. Curius Dentatus et C. Fabricius Luscinus: les grands hommes ne sont pas exceptionnels’, in Coudry and Späth 2001, 117–29
Walde, A., and Hofmann, J. B. 1954: Lateinisches Etymologisches Wörterbuch (3rd edn), 3 vols, Heidelberg
Wilburn, A. 2012: Materia Magica: The Archaeology of Magic in Roman Egypt, Cyprus, and Spain, Ann Arbor
Wildfang, R. L. 2006: Rome's Vestal Virgins: A Study of Rome's Vestal Priestesses in the Late Republic and Early Empire, London
Wilkens, B. 2006: ‘The sacrifice of dogs in ancient Italy’, in Snyder and Moore 2006, 132–7