2.1 Maintaining cities and empire: public cult

What kind of religious practices would a ‘gentile’ convert to Christianity have known, prior to the progressive banning and displacement of traditional religion across the fourth through sixth centuries? Neither a singular, unified ‘paganism’, nor the fractious division-within-unity of third- and fourth-century Christianity. The gods of the Mediterranean world were the objects of a vast complex of local and transregional cults, philosophical ideas, and literary, artistic, and architectural traditions. In literature and learned discourse, the Roman pantheon was equated with the Greek (Jupiter = Zeus, Juno = Hera, etc.). The same was done with the gods of other nations, by a procedure often called the interpretatio Romana or interpretatio Graeca. Thus, antiquarian writers and philosophers identified YHWH with various gods, including Dionysus, Osiris, and Saturn, or with a god not known to them (John Lydus, De mensibus [On the Months] 4.53). The historian Tacitus claimed that the Germanic peoples worshipped Mercury, Hercules, and Mars (respectively, the deities known in Old Norse as Odin, Thor [perhaps], and Týr); and that, among those peoples, some of the Suebi worshipped Isis (the Germanic deity to which Tacitus refers by the name of this Egyptian goddess is uncertain – perhaps Nehalennia or Nerthus; Germania 9, with Rives 1999a: 157–162). The local ways of worshipping the gods nonetheless remained distinct, and a reasonable case can be made that gods of the same name identified with different epithets (Zeus Philoxenos, ‘Friend of Strangers’, Zeus Olympios, ‘the Olympian’, and so on) were in fact different deities (Versnel 2011: 68–87).

Individual cities had their own gods, recognized by age-old tradition or legal enactment and worshipped according to local custom. Rome itself is an illustrative example, especially important because of its influence over the rest of the Empire (Wissowa 1912 remains an essential empirical overview; for an up-to-date survey, see Beard, North, and Price 1998 [vol. 1]). According to Roman legend, the first king, Romulus, established the auspices: the regular interpretation of divinatory signs, especially from birds, that became the duty of the augurs (augures). The public rites (publica sacra) proper were credited to the second king of Rome, Numa Pompilius. These included all those ritual acts that took place ‘with public funding on the people’s behalf, and those performed on behalf of the hills, districts, wards [i.e., various subdivisions of Rome and its environs], and shrines’ (thus the grammarian Festus, De verborum significatu [On the Meaning of Words] 284.18–20 Lindsay).

The administration of the public rites and the keeping of annals and calendar lay with the pontiffs (pontifices), whose college was associated with several other priestly officials: the priest of Jupiter (flamen dialis, an office bound about with many taboos) and his wife (the flaminica); the flamen of Mars; the flamen of Quirinus (a god eventually identified with the deified Romulus); the rex sacrorum or ‘king of sacrifices’ and his wife (the regina sacrorum); and the Vestal virgins. The ranking member of the college was the pontifex maximus. After the death of the triumvir Marcus Aemilius Lepidus, this office was ordinarily held by the emperor; as there were often multiple emperors, there could also be multiple pontifices maximi at once. The pontifex maximus was not the singular ‘head’ of a ‘religion’, nor were the pontifices the main decision-makers in matters of religion, which fell under the purview of the Senate (Beard 1990).

Some cults, upon being adopted at Rome, remained ceremonially foreign, practised either in a stereotyped ‘Greek fashion’ (Scheid 1995) or according to the custom of some other land of origin (Orlin 2010 provides a narrative overview). These so-called ‘peregrine rites’ were regulated by a college eventually made up of fifteen men, and so known as the quindecimuiri sacris faciundis (‘Board of Fifteen for Doing Rites’). The quindecimuiri also consulted prophetic, ‘Sibylline’ books at the Senate’s request. The last of the great colleges of Roman priests, the septemuiri epulones (‘Board of the Seven Banqueters’), was established in the second century BC; they had more limited duties, being responsible for the feast of Jupiter and matters to do with the games (Beard, North, and Price 1998: 100–111 [vol. 1]). A pontificate of the Sun was added much later, under the emperor Aurelian (270–275).

All of these priesthoods were held by senators in the course of an ordinary political career. They could be held simultaneously, especially in the later Empire (Rüpke 2011), and required no special training or piety – though the pontifices and augurs, at least, were expected to have religious expertise, and some holders of priestly office did distinguish themselves as writers on religion (Beard 1990: 36–40; for Latin theological works, see MacRae 2016). Lesser priestly societies – the Arval brethren, the fetial priests, and so forth – handled particular cultic duties, and various cultic assistants were involved in rituals. Divination by means of entrails and lightning strikes was exercised by the haruspices, practitioners of an Etruscan (and so ancient, but formally non-Roman) art (Orlin 2010: 95–100).

Rome’s civic institutions – of which the priesthoods were one part – served as a model for the many cities the Romans founded, or re-founded, in the Western, Latin-speaking half of the empire (Rives 1995: 28–31). North Africa enjoyed especially close ties to Rome, and its civic cults are documented by inscriptions and the works of North African Christians such as Tertullian and Augustine. As late as the 360s, when an inscription was posted that listed all local office-holders current and emeritus, the city of Timgad (Thamugadi) in Algeria was equipped with pontiffs and augurs (Chastagnol 1978). However, even a city with institutions modelled on Rome's would reverence local deities (or came to do so, at any rate, over the first centuries AD: Rives 1995: 132–169). Saturn, the Punic Baal Hammon, was of great importance in Africa, especially in the countryside where a multitude of carved stelae were erected to commemorate rites in his honour, but his cult remained in most places unofficial (Leglay 1966: 404–406; Schörner 2007). The great goddess of Carthage, however, was a divinity ordinarily known to her worshippers simply as Caelestis, the ‘Heavenly One’. Identified in literature with Juno – for example, by Vergil, in whose Aeneid Juno’s patronage over Carthage is a key plot point – she was Tinnit (or Tanit), consort of Baal and the ancient goddess of Phoenician Carthage, and was publicly worshipped both at Carthage (Rives 1995: 65-71) and in other cities (on her cult, see Lancellotti 2010, who gathers the relevant inscriptions). Like many Roman cities in North Africa, Carthage was outfitted with a Capitolium, a temple imitative of the temple of Jupiter Capitolinus, Juno, and Minerva on the greatest of the hills of Rome (Crawley Quinn and Wilson 2013). Both local and Roman gods (identified by a complex process of interpretatio Romana: Cadotte 2007) were the objects of profound devotion down to the end of formally organized pagan cult. One of Augustine’s correspondents exalted the Capitoline triad, while others asked for recompense after the death (with their approval, Augustine asserted) of sixty Christians in a riot that had followed vandalism of a statue of their city’s patron, a Hercules probably Romanized from a Punic Melqart or Milkashtart (Maximus of Madauros = Augustine, Epistle 16; Augustine, Epistle 50, to the leaders of Sufes, with Cadotte 2007: 283–288).

The East, dominated since Alexander’s conquests by Greek-speakers, was more heavily urbanized; its many cities retained the gods they had worshipped before the Roman conquest – in some cases, from primordial antiquity (on the religion of Greek cities under the empire, see Lane Fox 1986: 27–261). The Acts of the Apostles bears rich testimony to the variety and intensity of Greek religion. At Athens, Paul finds a multitude of altars and cult images as well as the constant philosophical discussion of a university town (Acts 17:16–34). At Ephesus, he meets with organized hostility from artisans made rich by the worship of Artemis, the huntress of Greek mythology, but also a fertility goddess studded, on her surviving images, with the scrota of bulls (Acts 19:23–41, with Immendörfer 2017: 123–178). At Philippi, Paul runs afoul of local authorities after exorcising a girl who identifies his ministry, in apparent mockery, with the worship of the ‘Most-High God’ – in Greek, the theos hypsistos, possibly to be identified as the object of a monotheistic pagan cult (Acts 16:17–22; 2.3, below). At Lystra, he and Barnabas are greeted with divine honours after a healing (Acts 14:8–20): a reminder, like the repeated clashes between the apostles and magicians (both Jewish and gentile), of the widespread belief that the gods or lesser, intermediary spirits (daemones, hence ‘demons’) might intervene in ordinary life. In one important area at least, such manifestations of spirits were becoming more restricted. Some decades later, Plutarch (philosopher and priest of Apollo at Delphi) sought to explain why the oracles – which had played a vital role in international politics – had largely fallen silent, or devoted themselves to the petty concerns of private enquirers (De defectu oraculorum [The Obsolescence of Oracles] 7, 413a–b). Their eclipse was brief. Reviving under the Antonine emperors in the mid-to-late second century, the oracles of Apollo at Didyma (near Miletus), Delphi, and Claros attracted both official and private enquirers – including emperors – into the fourth century (Athanassiadi 1991; 1992).

The aim of public worship was to maintain the favour of the gods and so preserve the prosperity of the cities. However, even the cults of Rome were, in the first instance, the cults of a particular city. Close as their symbolic relationship was to the emperors, they were not cults of the empire and its citizens at large (which, since 212, included all free, native inhabitants). Throughout the empire, local office holders and private persons offered up sacrifices on behalf of the emperors and the imperial house, but there never was an overarching religious organization that tied together the empire’s cities, their individual citizens, or the rural peasantry. The latter group, who comprised the vast majority of the empire’s population, might make offerings in rural shrines or the temples of the cities in whose territory they lived (for a stimulating discussion of the problem of religion in, or for, the Roman empire, see Beard, North, and Price 1998: 313–363).

The imperial cult came closest to an overarching religious structure, and could be treated by an official trying Christians as the religious duty that bound all good Romans (see Passio Scillitanorum [Acts of the Scillitan Martyrs] 3, discussed further in 3.1, below; cf. Cassius Dio, 51.7; despite Ando 2000: 394–395, however, one may doubt that Tertullian, Apologeticum [Apology] 24.8 calls the emperor ‘the god of the Romans’; more likely, he means that the Christians’ God is not an official Roman one). Nevertheless, the patterns by which the goddess Roma, the living emperor, members of the imperial household, and the deified, dead emperors (divi) were worshipped vary over time and were never integrated into a single system (on the West, see Fishwick 2002–2005; on the East, Price 1984; on Rome itself, see Gradel 2002). Their cult was promulgated by central pressure (in the West, virtually always) and local initiative with imperial involvement (in the East, often), with the result that there were priests of the emperors in each of the empire’s provinces, as well as individual cities. During the last empire-wide persecution, the emperor Maximinus Daza tried to organize polytheistic worship and exert pressure on Christians through provincial and civic priests who were perhaps modelled on those of the imperial cult (Nicholson 1994; cf. 3.2, below). The initiative was brief and ultimately ineffectual. Despite overarching similarities, shared gods (or at least gods with shared names), and a common desire to ensure the prosperity of cities and empire, the public cults of the Roman world remained irreducibly plural.

2.2 Religious variety: ritual practice, social norms, and private cult

The fundamental rite of traditional cults was sacrifice, especially of animals, but also of incense and other vegetable matter; scholars routinely speak of ‘blood sacrifice’, but that idiom does not mirror Roman and Greek custom (Schultz 2016; Naiden 2012: 70–81). Sacrifices were not, however, the only rite that mattered. Public holidays were celebrated in the gods’ honour, with contests (agones; a vital institution in the Greek world), processions, chariot races, stage plays and the typically Roman institution of gladiatorial combats (see, for Rome, Latham 2016: 44–66; the ancient evidence for Greek games is vast; on their permutations over time, differences from Roman 'spectacles', and eventual decline, see especially Remijsen 2015; Rogers 1991 studies a massive inscription from early second-century Ephesus that gives instructions for a procession in honour of Artemis). These holidays were important enough at Rome to earn three of the sixteen books in Varro’s Antiquities of Divine Matters – the most extensive and learned account of Roman religion penned in the late Republic, which was a time of intensive enquiry into Greek philosophy and Roman traditions (Varro, Antiquitates rerum divinarum fr. 4 Cardauns – not a quotation, but a summary of contents in Augustine, De civitate dei [City of God] 6.3; Beard 1986). Such holidays attracted wide audiences into the late empire. However, the temples and their priests did not give the central place to morality and education about the divine assumed, from the very beginning, by Christian churches and their clergy (which did not mean that gods could not make moral demands, especially of those most devoted to their service: on the moral expectations of the cult of a popular healing god, Asclepius, for example, see Israelowich 2012: 151-152). Christian apologists underscored the difference, in an argument developed most forcefully by Augustine (Epistle 91.5, City of God 2.7, 22). There was no congregation, and no orthodoxy, in public cult.

That does not mean, however, that ancient people did not care about ‘correct’ conceptions of the gods (Morgan 2017), or that they lacked ways of seeking divine favour and expressing their spiritual aspirations. A multitude of inscriptions, objects, and literary references attest to the private offering of sacrifices, as well as (or in conjunction with) the pursuit of healing, divine counsel, and personal experience of divinity. Even Jupiter, Juno, and Minerva, who were worshipped on the Capitolium at Rome, could receive such devotion from individual worshippers, and through starkly idiosyncratic means (thus the statesman-philosopher Seneca, in De superstitione (On Superstition) fr. 69 Vottero, quoted by Augustine, City of God 6.10, with Bendlin 2000: 132; cf. Piettre 2001 for greater uncertainty about whether private worship is involved). Public and private cult – the latter ‘performed on behalf of individual people, families, and clans’ (Festus, On the Meaning of Words 284.20–1 Lindsay) – were not, in principle, opposed. Both sought to maintain the favour of the gods, and private persons as well as local authorities might perform rites ‘for the well-being of the emperor’ (on this common formula, see Fishwick 2002–2005: 352–360 [vol. 3]).

There were tensions nonetheless between individual practice and what seemed appropriate to the kind of men who held political and priestly office (see Rüpke 2016, and the burgeoning scholarship surveyed in Albrecht et al. 2018 that emphasizes the role of individual actors in all forms of ancient cult; but see also Scheid 2016 who argues – on the whole persuasively – for the primacy of public cult and its norms, at least at Rome). Some private rites were profoundly traditional, as deeply rooted in the customs of their peoples as were the rites of civic priests; thus, in the Roman cultural sphere, the worship of various familial gods, including the family lar (sometimes singular, sometimes twinned as lares), the penates or gods of the household, and the guardian spirits (for a man, genius, for a woman, iuno) of the pater familias and his wife (Flower 2017). In an increasingly cosmopolitan order, other cults could seem more threatening. Like philosophically educated Greeks, Seneca and Varro’s contemporary Cicero were suspicious of devotion that exceeded the bounds of moderation or ancestral custom and seemed to express an unbecoming fear of the gods (Seneca, On Superstition; Cicero, De natura deorum [On the Nature of the Gods] 2.70–2; cf., in the Greek world, Plutarch's treatise On Superstition, preserved unlike Seneca's in full). In his ideal ordering of the Roman republic, Cicero would have forbidden worship of any gods not received ‘publicly’ – that is, by order of the Senate – or inherited from one’s ancestors (De legibus [On the Laws] 2.19).

That was wishful thinking. Roman statecraft was generally reactive and unsystematic, and only on Christianity and its offshoots – especially Manichaeism – did the Roman authorities ever conduct a systematic crackdown (for an overview of both Greek and Roman religious restriction and coercion, see Garnsey 1984). The first attempt at a more limited curtailment came in 186 BC, with the suppression of the Bacchanals (the cultic communities devoted to Dionysus, who was known in Latin as Liber Pater), following lurid accusations of sexual wrongdoing (Livy, 39.8–19; the law survives as an inscription: CIL I²581; North 1979). Roman emperors made sporadic restrictions on Egyptian and Jewish cult, but those restrictions did not gain permanency. Already worshipped in Rome in the first century BC, Isis and Serapis were promoted by multiple first- and second-century emperors, and their worship was eventually accepted (perhaps under the emperor Gaius) among the public cults of Rome (a variety of views in Salzman 1990: 170–171, note 203; Lembke 1994: 84–103; Takács 1995: 71–104). This was a change in the gods’ status at Rome, not in the empire at large or (necessarily) in ordinary persons’ estimation. Isis and her consort, Osiris, continued to attract the worship of private initiates (of whom Apuleius’ novel Metamorphoses portrays a fictional example, possibly by way of satire; Harrison 2012), while Serapis had been the great god of Alexandria, the second-greatest city in the Roman world, since the Ptolemies. Rebuilt after a fire in 181, Serapis’ temple in Alexandria was, until its destruction in 391 or 392, a sight to rival the Roman Capitolium in beauty and glory (Ammianus Marcellinus, 22.16.12; archaeological data in McKenzie, Gibson, and Reyes 2004). The limitations imposed at Rome should not, therefore, be seen as a general rule for the empire at large: the diversity of the empire’s cults was recognized by educated Romans, if sometimes with distaste. The outstanding exception was the prohibition against human sacrifice, which the Romans had practised only on occasion (for example, during the great public crisis of the Second Punic War), and ultimately forbade by senatorial decree in 97 BC (Schultz 2010). In North Africa – whose Punic great god Baal Hammon (Romanized as Saturn) was a notorious recipient of ritual infanticide – the ban was carried out in the late second century, and was helped in no small measure through transformation of local culture (the reality and extent of infanticide have been much controverted: see Shaw 2016 for a recent assessment, likening the final measures to the suppression of Hindu sati).

In the ‘mystery’ cults of Isis and other Near Eastern gods, an influential body of older scholarship found the real object of pagan devotion under the Roman Empire (Cumont 1929 is the classic statement). The cults of Isis and Osiris, Cybele and Attis, Mithras, and other ‘oriental’ gods had doctrinal content and organized priesthoods, making them more advanced (by an evolutionary view of religion) than the cults of the Greek cities and especially the statesman-like religion of Rome. They could also offer the individual devotee membership in relatively tight-knit cultic associations, often organized along family lines (a few examples in Cameron 2011: 143–144), as well as otherworldly rewards. The evolutionary model of religion is now little in favour, while ‘oriental’ seems too simplistic a category to form a useful category of analysis: marked by much later European imagination of the Middle East, it also tended to make Christianity an implicit model for cults quite different from it (Alvar 2008, a learned attempt at updating the old view, admits its weaknesses).

These cults supplemented, without replacing, more established forms of both public and private worship. These included secretive rites (so-called mysteries) long frequented in Greece, of which those of Demeter and Core/Persephone at Eleusis, near Athens, were preeminent. The first of the Near Eastern cults to be adopted at Rome, the cult of Cybele (as the ‘Great Idaean Mother of the Gods’), is illustrative: summoned to Rome from Asia Minor at the end of the third century BC, Cybele was worshipped both in the native fashion – that is, by a ‘Phrygian’ priest and priestess – and by a Roman magistrate, the praetor (Dionysius of Halicarnassus, Antiquitates Romanae [Roman Antiquities] 2.19; Orlin 2010: 101–104). The ecstatic elements of the cult, which featured ritual self-castration in imitation of Attis, were thus sealed off from the official worship of the Roman state. Devotees nonetheless continued to offer their genitals to the goddess well into the fourth century, at a time when senatorial quindecimviri (‘Board of the Fifteen’) helped private worshippers offer sacrifices to the Mother of the Gods on their own behalf – and, it appears, in the company of a ‘Phrygian’ priest (for the eunuchs, see Augustine, City of God 7.26, at Carthage; for the quindecimuiri: CIL VI 508, with McLynn 1996; Gassman 2020: 88–89). Many of these senators were holders of ancient state priesthoods. The most detailed statement of their devotion, a funerary inscription erected by a senator’s wife shortly after the defunding of the Roman state cults, records his civic and priestly offices and praises her (in his voice) for her patriotism (CIL VI 1779). Public cult and private devotion remained compatible, for tradition-minded pagans, so long as the old religious order endured.

Both public and private worship were distinguished, however, at least in theory, from illegitimate ‘magic’. Many ritual practices – the making of love-potions, casting of curses, fortune-telling, and so forth – attested in inscriptions, papyri, and other material remains lay outside the scope both of public cult and of the private cults that a statesman or philosopher, such as Cicero or Plutarch, would condone. The Roman authorities tried to prevent the subversive or harmful use of divination, including astrology (Dickie 2012: 325–329, especially on the later period). That rule still held during the fourth century, when Constantine expressly condoned private divination, so long as it was not conducted in secret, and magical rituals performed for innocent ends (Theodosian Code 9.16.1–3, 16.10.1). The tendency, however, of Christian theology and preaching, and increasingly also of imperial legislation (Sandwell 2005), was to conflate traditional religion with malicious magic, and to condemn both. Like many Christian apologetic arguments, that conflation lumped together categories that well-educated gentiles, at least, would have kept distinct. Nevertheless, it was not entirely without reason. Well before the Christian ascendancy, Platonist philosophers had linked sacrifices and the other operations of public cult, including oracles, to the daemones – capricious, if not necessarily evil, intermediaries between humans and inherently good gods who were elevated above worldly affairs (e.g. Apuleius, De deo Socratis [On the God of Socrates] 145–50, Plutarch, The Obsolescence of Oracles 13, 416e–417b). Porphyry, who opposed blood sacrifices, placed particular emphasis on the wickedness of the daemones involved therein; the intermediaries with oracles were, however, entirely beneficent (De abstinentia [On Abstinence from Killing Animals] 2.37–44; Christian influence on his thinking is not impossible: Proctor 2014). Daemones were also, as Porphyry for example complained, the interlocutors of magical specialists, although such people (who included Christians as well as pagans or Jews) may not have thought about religious differences in terms compatible with those used by Christian theologians (Boustan and Sanzo 2017). For Christians who saw demons as inherently evil and denied the authority of ancestral tradition, it was possible to reanalyse public, private, and ‘magical’ rites as instances of the same demonic system (thus, on the basis of a sophisticated semiotic and sociological theory, Augustine, De diversis quaestionibus LXXXIII [Eighty-Three Different Questions] 79, De doctrina christiana [On Christian Teaching] 2.20.30–24.37; Graf 2002; Markus 1994).

2.3 Ideas about the gods: myth and philosophy

In his Antiquities of Divine Matters, the polymath and statesman Varro distinguished three different ways of knowing about the gods: the mythical, the natural or philosophical, and the civic or political (Antiquities of Divine Matters fr. 7 Cardauns, quoted by Augustine, City of God 6.5). The ‘three kinds of theology’ are a simplification of an even more complicated reality; but Varro’s scheme, which has many echoes in other ancient authors (Lieberg 1982), may serve as a useful first approximation for thinking about the relationship between traditional cultic practice – ‘paganism’ in the narrow sense – and the wider body of ancient knowledge about the gods.

At their root, myth, cult, and philosophy were distinct in antiquity. Although the epic tradition told stories about the gods worshipped in the Empire’s cities, their cults gave a doctrinal role neither to Homer and Hesiod nor to later poets such as Vergil, author of the Roman national epic, the Aeneid; Ovid, who wrote the often scurrilous mythological epic Metamorphoses and a massive poem on the Roman cultic calendar, the Fasti; or Horace, who wrote what is surely the most famous imperial-era poem meant for performance at a public ritual, the Carmen Saeculare (sung at the ludi saeculares, the so-called ‘Secular Games’ which were conducted, at Augustus’ orders, in 17 BC). In turn, philosophers as early as Plato (Republic 2, 377a–80d) expressed deep antipathy for mythological tales of the gods’ foibles, strife, and crimes, preferring to believe the gods inherently good and harmonious. The disjunction still troubled Varro’s Roman contemporaries when they began to engage in serious thinking about the shape of their tradition and its relationship to Greek philosophy. Thus Cicero, in his dialogue On the Nature of the Gods, has his Stoic interlocutor, Balbus, reject the gods of mythology as ‘fictitious and made-up’ (On the Nature of the Gods 2.70), in favour of an enlightened conception of divine providence, of the stars as gods, and of the traditional pantheon as divine personifications of natural phenomena. Stoicism was not, however, the only philosophical option in the first century BC. Seeking a way to bring together philosophical learning and traditional cult, Cicero’s interlocutors scout Epicureanism, which rejected both divine providence and the fear of the gods on which much traditional cultic practice was based (cf. On the Nature of the Gods 1.54–6, Lucretius, 1.62–150). The pontifex Cotta, like Cicero himself a Skeptic, has the final word in the dialogue; having rejected Epicureanism, he also challenges Balbus’ Stoic account of divinity, which he thinks impossible to harmonize with the Roman cultic tradition in which he invests his confidence (On the Nature of the Gods 3.5).

Long afterward, Cicero’s Christian readers would take Cotta’s arguments as an exposé of the intellectual and spiritual vapidity of public cult – which Cicero had, they asserted, retained only for public order (Lactantius, Divine Institutes 2.3.2; cf. Arnobius, Adversus Nationes [Against the Nations] 3.7; Gassman 2020: 31–37). Cicero himself knew that an atheistic conclusion was possible (De divinatione [On Divination] 1.8), but insisted that the intent of Cotta’s arguments – his own literary creation – was to reveal the problems in Stoic ideas concerning the gods (our interpretation of his meaning has not been helped by the loss of some material from the Skeptical third book). Cicero was writing in a world that was polytheistic by default, in which such arguments did not yet bring the cultural stakes that they would after the rise of Christianity (North 2005: 135–136). Cotta’s objections nonetheless show the logical tensions inherent within Graeco-Roman religion. A ‘gentile’ might not simply worship different gods for different purposes or at different times. He might hold that the gods were inherently good and governed the world by providence; laugh at a comedy that lampooned the adulteries of Jupiter; enjoy reading Vergil; and, as a priest of his city, preside over a rite of a ‘Heavenly’ goddess identical (if he lived in North Africa) with the consort of an ancient Phoenician Baal. Furthermore, this priestly task would in no way preclude him from also acting as a priest in the imperial cult, seeking divine blessing by sacrificing a bull to the Mother of the Gods, or consulting an astrologer or diviner after disturbing dreams.

Such an individual might hold that all the gods he and the many peoples of the world worshipped were, in some sense, one. Often called pagan ‘monotheism’ (Athanassiadi and Frede 1999), this constellation of views was fully compatible, for most of its ancient proponents, with vigorous practical polytheism; ‘henotheism’ might be a better term, though that has also been used to denote the unique exaltation of one god within a pantheon (cf. Versnel 2011: 239–244). To apologists, at any rate, apparently monotheistic statements in the philosophers and poets were useful proofs for their own doctrine of the unicity of God (e.g. Lactantius, Divine Institutes 1.5). Nonetheless, belief in Christ and rejection of other mediators still remained the non-negotiable dividing line (Edwards 2006: 223). Actual pagan monotheism may be found, according to one interpretation of the evidence, in the cult of the theos hypsistos or Zeus hypsistos, the ‘Most-High God’ attested in numerous inscriptions (Mitchell 1999; 2010; otherwise, Belayche 2005; Parker 2018: 124–130, summing up the debate, leans against the idea of ‘pagan monotheism’). The father of Gregory of Nazianzus was an adherent of a group of ‘hypsistarians’, whose worship his son perceived as a mixture of pagan and Jewish elements (Oration 18.5). Whether there was a single cult of a single theos hypsistos or not, a connection of such worshippers of the Most High with the God-fearers of the biblical book of Acts is plausible. Additionally, there is a possible parallel in a little-understood, apparently Jewish-Christian sect called the ‘worshippers of heaven’ (caelicolae) attested in late Roman Africa (Augustine, Epistle 44.13, Theodosian Code 16.5.43, 16.8.19, with Simon 1978; Shaw 2011: 277–279). As a rule, however, the acceptance of henotheistic ideas could undergird, and did not undermine, worship of many gods.

Though philosophy, poetry, and cult did not form a unified doctrinal and ritual system (a single ‘religion’), they were not without influence upon one another. Poetry had never been sealed off from cultic practice and could shape, or be shaped by, the ideas of the worshippers (for all that these were not necessarily consistent; cf. Feeney 1998: 14–16). The comedies and tragedies that Greek-speakers continued to read in the course of their education had been composed in Athens in the fifth and fourth centuries BC, to be staged at festivals in honour of Dionysus. Under the Roman Empire, plays were still performed at traditional holidays associated with the gods, a fact that Christian apologists were determined not to let pagans ashamed of divine immoralities forget (Augustine, City of God 2, brings this argument to a peak of rhetorical effect). Such dramas did not constitute any kind of official ‘doctrine’ about the gods, but they were part of the ritual apparatus that maintained divine favour and, perhaps even more important, they communicated ideas about the gods (thus, on ancient Athens, Parker 2005: 136–152). The images of the gods, furthermore, showed many of the same attributes and physical features for which the gods were celebrated in poetry; philosophical criticism of the one could thus implicate the other (a point evident, though not developed, in Cicero, On the Nature of the Gods 1.81, 2.70).

Varro’s three-fold division (between the philosophical, poetical, and civic) is too neat to fit the Neoplatonist thought current in late antiquity. More sanguine than Plato himself about the Homeric epics, later Platonists debated the philosopher’s proper stance toward traditional ritual. Some, such as Iamblichus, sought to draw nearer to the transcendent divinity by means of theurgy, ritual performances that their critics treated as magic (perhaps not entirely unfairly: Dillon 2007). The many works of Porphyry of Tyre – who was the most influential Neoplatonist after Plotinus – are largely fragmentary, and both their chronology and their import are much controverted (Magny 2014, for example, offers cautions against over-confident reconstructions). What survives shows Porphyry criticizing the offering of blood sacrifices (On Abstinence from Killing Animals 2.38–50); exegeting a scene from Homer in part by comparison with the cave shrines of Mithras (De antro nympharum [The Cave of Nymphs]); explaining the nature and roles of the particular gods with references to their images (De imaginibus [On Statues]); and attempting to construct a philosophical theology out of oracular utterances (De philosophia ex oraculis haurienda [On Philosophy from Oracles]). All these remained endeavours of the individual philosopher, without direct bearing on public cult, except during Julian’s brief attempt to rejuvenate traditional religion (3.3, below). Philosophers might draw on oracular utterances – which offered substantive theological insights and helped to foment imperial hostility toward Christians (Robert 1971; Lactantius, De mortibus persecutorum [The Deaths of the Persecutors] 11.7) – just as they could draw upon Homer and on esoteric writings. Such works included the Chaldaean Oracles, the books of the Etruscan haruspices, and works attributed to ancient sages such as Hermes Trismegistus (the Egyptian Thoth) and Orpheus (see, e.g. Fowden 1986). Public cult nonetheless remained a distinct enterprise, conducted according to norms laid down in antiquity and adapted over time.

3.1 The beginnings to the 240s

The first holders of imperial power to encounter Christians seem to have met their message with a mixture of interest, incomprehension, and indifference. In Acts, the proconsul of Cyprus, Sergius Paulus, is converted after the blinding of his associate, the Jewish magician Elymas (Acts 14:6–12). As proconsul of Achaea, L. Junius Gallio Annaeanus (brother of the philosopher Seneca) refuses either to intervene in the intra-Jewish dispute brought to him by Paul’s opponents or to prevent the beating of the synagogue leader by the ashamed Jews (Acts 18:12–17; Sherwin-White 1963: 99–107 lays out possible interpretations of the legal situation). Paul, upon returning to Jerusalem, uses his Roman citizenship to avoid being beaten by a low-ranking army officer (Acts 22:25–9; cf. Acts 16:38), then deals repeatedly with the local governors Felix and Festus. Neither displays hostility, and Paul’s eventual passage to Rome to be tried by Caesar (now, Nero) is the result of his own request, though doubtless it is to Festus’ liking (Acts 23–7; for the legal procedure, see Sherwin-White 1963: 48–70). The charge that the Christians were trying to subvert Caesar’s laws had already been raised to the discomfiture of local authorities at Thessalonica (Acts 17:6–8); but if that charge gained more than formal acknowledgement by Roman officials, Acts gives little sign of it. That the account closes with Paul under house arrest must have rung ominous to later Christians, who knew the sequel, but there is no evident allusion to Paul’s death in Acts 28.

Each of the officials appears to be encountering nascent Christianity without prior knowledge of its tenets, and none is acting on the basis of an established legal precedent: a fact that invalidates speculations that a decree of the senate was issued against Christians in 35, following a report of the events in Palestine to Tiberius (thus rightly Barnes 1968: 32–33). In fact, what Tertullian asserts, some 160 years later, is that Tiberius wished to acknowledge Christ as a god, and, after the Senate’s refusal, issued threats against any opponents of Christianity (Apol. 5.2). If there is any truth to the story (which is corroborated by no independent source, despite Sordi and Ramelli 2004; see now Gassman 2021), it suggests only that Christianity remained legal, a subject of intra-Jewish disputation of little, or only personal, interest to Roman authorities.

Imperially authorized violence against Christians began later in Nero’s reign, when the emperor made them scapegoats for the fire that destroyed much of Rome in 64. That Christians were specially targeted – and thus whether one may really speak of a ‘Neronian persecution’ – has been doubted, as has the traditional connection with the deaths of Peter and Paul (thus Shaw 2015, Shaw 2018, disputed by Jones 2017 and Cook 2020). However, both later Christians and later pagans certainly perceived Nero to have had Christians in his sights. Nero’s cruelty is noted by two pagan writers of the second century: the historian Tacitus (Annals 15.44), the main source for the events in 64; and the imperial biographer Suetonius (Nero 16.2). Both still denigrate Christians as adherents of a ‘superstition’. (In the Latin of their era, superstitio denoted a despised ‘religion of others’ as well as the ‘excessive practice’ of Roman religion condemned by Cicero and Seneca; Grodzynski 1974: 44–50). Later Christians would see Nero as responsible for the first of several bouts of persecution, and report that a second took place under Domitian (Eusebius, Historia ecclesiastica [Church History] 3.17; Tertullian, Apology 5.4). Lactantius, writing in the 310s to celebrate the ending of persecution, implied that the Church had peace between Domitian’s reign and that of Decius, in the 240s; others enumerated, more carefully, successive periods of repression under later emperors (Lactantius, On the Deaths of the Persecutors 3; Augustine, City of God 18.52).

These accounts tend to put too much emphasis on the emperors’ attitudes. What followed the events under Nero was in fact the same kind of haphazard, arbitrary judgment on display in Acts, though in the earlier period with greater friendliness toward the new movement. Whether there was ever a single decree of the Senate or emperor on which such prosecutions rested is doubtful (a careful article by Barnes 1968, surveying the many and contradictory references, concludes that there was none). All of the basic considerations – local animosities, proconsular discretion, indirect and reactive leadership from the emperor, shaky accusations of grand crimes, and the persecution ‘for the Christian name’ – are already in evidence in the two extant official documents emanating from a contemporary governor’s dealings with Christians (for what follows, see especially, de Ste. Croix 1963. Streeter 2006 offers a retrospective on this classic article, while Corke-Webster 2017 rightly emphasises that Pliny’s own overarching concern is not to understand or even control Christians per se, but to justify himself before the emperor). In a letter written to Trajan in early 110s, while he was governor of Bithynia, the younger Pliny expresses great uncertainty about proper punishment of the Christians who have been reported to him, and even about the charge: is it the ‘name’ or the ‘crimes associated with the name?’ His investigations found no crimes, but he takes for granted that Christians are to be punished, and not simply for the stubbornness that irks him (Epistle 10.96). He required Christians to recant and prove their sincerity by cursing Christ and venerating ‘your [i.e., Trajan’s] image and the statues of the gods’ with wine and incense. In his reply (Epistle 10.97), Trajan approves of the procedure, only forbidding the admission of anonymous accusations.

Although Trajan’s letter was technically binding only on him, Pliny’s report describes the basic procedure followed up to the middle of the third century. Eusebius reports various outbreaks of persecution across the Roman empire. For example, he documents an incident at Vienne and Lyons in France in 177 by quoting a lengthy contemporary letter. According to the letter, local hatreds, inflamed for whatever reason (the letter’s writers blame the Devil), led to a violent riot against Christians. They were imprisoned by municipal authorities, pending the governor’s arrival. When the governor arrived, a prominent member of the congregation, Vettius Epagathus, tried to offer a defence but was allowed only to say whether or not he was a Christian. The result was a division between the steadfast and those who renounced their faith. Fearing torture (a Roman judicial routine), the Christians’ slaves, who happened to be pagans, accused the Christians of cannibalism and incest, and the crowd was enflamed to anger. Under torture, two particularly heroic martyrs, Blandina and Sanctus, insisted over and over again, ‘I am Christian’ (Church History 5.1).

Individual officials clearly differed in their approach. In an open letter to the proconsul Scapula, written in 212 (Barnes 1985: 38), Tertullian described the behaviour of Scapula’s peers and predecessors. Some were harsh and met a suitably unpleasant end (Ad Scapulam [To Scapula] 3.4–5): Vigellius Saturninus, who had presided over the condemnation of several Christians from Scillium in Numidia, went blind (the trial, though not the blindness, is described in the Acts of the Scillitan Martyrs, for which see now Rebillard 2017: 351–359); Lucius Claudius Herminianus, a governor in Cappadocia whose wife had converted to Christianity, was eaten by worms, then, repenting, died ‘almost a Christian’; Caecilius Capella, who had condemned an African Christian to be mauled by animals, died some years later in a siege at Byzantium (so Birley 1991 on a difficult passage). Others were gentler (To Scapula 4.3): Cincius Severus coached Christians on the responses needed for their acquittal; Vespronius Candidus punished a Christian as a mere rabble-rouser; Asper, who had expressed his regret at having had to take on such a case, pardoned a Christian who had given in after ‘moderate discomfiture’ – torture, presumably, but not to the full ability of its practitioners – without making him sacrifice to the gods; Pudens tore up an anonymous accusation, releasing the accused Christian without charge (To Scapula 3.4-5, 4.3). Tertullian also appeals to the behaviour of Roman emperors, but focuses on their direct interactions (real or rumoured) with Christians in the palace, Senate, or army, and not with ordinary provincials (To Scapula 4.5–6). The day-to-day direction of justice was in the hands of provincial officials, and so the discretion, in the case of North Africa, was left to Scapula.

Accusations of conspiracy and gross vices occur (see esp. Minucius Felix, Octavius 9, discussed in section 4). The emperor Hadrian, writing to another governor in 124 or 125, may have allowed prosecution of Christians only if they had acted against the law (Justin, Apology 1.68; for doubts about the rescript’s authenticity, see Corke-Webster 2017: 400). Still, such charges were generally subordinated to religious concerns. A recent study has speculated that a third-century Egyptian papyrus that mentions a conspiracy may be our earliest verbatim record of a trial of Christians (Huebner 2019), but in fact its arguments tend to confirm the primacy of religious over non-religious concerns (and indeed the case probably does not concern Christians at all: Dolganov and Rebillard 2021). When, in 258, Cyprian of Carthage was accused of leading a ‘nefarious conspiracy’, the charge was bracketed by the claim that he had ‘long lived with a sacrilegious mind’ – i.e., as a Christian, rejecting the worship of many gods – and had shown himself to be ‘an enemy to the Roman gods and sacred cults’ whom the emperors ‘were unable to recall’ to their forms of worship (Acta Cypriani [Acts of Cyprian] 3.4, of which there are two recensions and much confusion in the manuscripts: Bastiaensen 1990: 216; 224). The conspiracy was, in other words, a conspiracy to reject the worship of the gods. Though Pliny, for one, clearly bore a suspicion (and likely heard accusations) of Christian involvement in nefarious crimes (Epistle 10.96.7, Corke-Webster 2017: 381), his own goals were primarily religious: as he declared in the conclusion of his letter to Trajan, the quick-spreading ‘superstition’ had been checked, and so ‘the temples now almost deserted have begun to be thronged, the sacred festivals long interrupted are conducted again, and the flesh of sacrificial victims, which until now hardly found a buyer, is for sale everywhere’ (Epistle 10.96.10). The emperor’s image and oaths by his name were involved, since the fundamental aim was that of public cult throughout the Roman world: ensuring the prosperity of the city, the entire empire, and the persons of the emperor and his house (Tertullian, Apology 32–6). By refusing to sacrifice, Christians were endangering the religious order on which the empire rested.

3.2 The empire-wide persecutions: Decius to the Tetrarchy

Around the time that Tertullian wrote to Scapula, the jurist Ulpian, collecting legal precedents for the use of proconsuls, gathered the instructions various emperors had given to enquiries on the Christian problem (Lactantius, Divine Institutes 5.11.19; Rives 1996: 21). Trajan’s letter surely appeared among them, but this part of Ulpian’s collection is lost. As late as the 210s, it appears, no emperor had issued a generally binding edict pertaining to Christians. Late in the year 249, that changed. After overthrowing Philip the Arab (according to later reports, a secret Christian; Eusebius, Church History 6.34), Trajanus Decius wished to restore the favour of the gods for a Roman empire now gripped by foreign and civil wars, and so commanded every inhabitant of the empire to offer sacrifices (Rives 1999b). Repurposing the bureaucratic machinery that enabled reasonably efficient census-making, Decius required each individual to obtain certification that he or she had, in fact, sacrificed or deputized another to do so. Numerous receipts survive from Egypt, including one issued to another person on behalf of a Christian; another, issued to a priestess of a local deity, confirms the universality of the decree (most of the inscriptions are collected in Knipfing 1923). Decius meant to make everyone sacrifice, not just Christians, though their impiety might well have been the new policy’s trigger.

The edict, whose exact terms are not reported anywhere, was a new step in Roman law: for the first time an emperor had commanded a universal display of religious adherence. The result was turmoil in the empire’s churches, as most Christians sought some compromise with the command – if not by sacrificing outright, then by obtaining a certificate by proxy or bribe; or, less controversially, by running away (Clarke 1984–1989: 21–39 [vol. 1] offers a fine account). Surviving bishops were forced to confront mass apostasy and, later, to decide what to do with the ‘lapsed’; the result, in Carthage and Rome, was the Novatian schism and momentous debates over the sacraments and the theology of martyrdom and penance.

The religious-political consequences were also vast. Decius’ policy worked with, not against, local cult. At Carthage and Rome, sacrifices were made on the capitolium, presumably to Jupiter and his consorts (Cyprian, De lapsis [The Lapsed] 8, Ep. 21.3.2, from Celerinus). Sacrifices were also made in Smyrna at the temples of the chief local goddesses, the Nemeseis (though here we cannot be sure: the key source, Martyrium Pionii [The Martyrdom of Pionius] 18.13–14, is held to be especially early and authoritative by an eminent historian - Robert 1994: 1–9 - but judged a redaction c. 400, on primarily philological grounds, by Zwierlein 2014: esp. 2: 39–40; for an overview and speculations about the text's redaction, see, Seeliger and Wischmeyer 2015: 172–177. The incorporation of inscribed documents into a fictitious life of a [historical] second-century churchman from Asia Minor, Abercius — Thonemann 2012 — simultaneously undermines a historical case for the authenticity of The Martyrdom of Pionius and allows for the reliability of contextualizing details, such as the role of the Nemeseion). Whatever the exact situation at Smyrna, the result in the empire at large was a de facto split between Christians – those faithful to the death, at least – and the worshippers of traditional gods (Jews seem to have been exempted from the command). The division between ‘Christians’ and ‘gentiles’ that devout Christians had so long taken for granted amongst themselves was now becoming a social reality (cf. Beard, North, and Price 1998: 241 [vol. 1]).

Decius died in battle with the Goths in 251. Despite the death of Cornelius of Rome (exiled by Trebonianus Gallus in 253), the persecution that Cyprian feared at Carthage did not break out again until 257, when the emperors, now Valerian and his son Gallienus, began a new empire-wide persecution (Clarke 1984–1989: 4–17 [vol. 3]; 8–14 [vol. 4]). The emperors' motive appears again to have been religious. The complaints voiced by a pagan associate of Cyprian a few years before reveal the contemporary outlook: all the empire’s many sufferings, on whose severity Cyprian expatiated in his reply Ad Demetrianum (To Demetrianus), were the fault of Christians. At first, Christian clergy were ordered to sacrifice and, upon refusal, were exiled with orders not to hold meetings on pain of execution (thus Dionysius of Alexandria, ap. Eusebius, Church History 7.11.6-11, and the proconsul’s dialogue with Cyprian in Acts of Cyprian 1). A year later, Valerian, answering an enquiry from the Senate, ordered clergy to be executed forthwith; high-ranking Christian men to be stripped of rank and property and, if they persevered, to be executed; Christian ladies to be stripped of property and exiled; and members of the imperial household to be shipped in chains to the imperial estates (Cyprian, Epistle 80.1.2–3). Cyprian himself would be beheaded on 14 September 258, several weeks after Xystus (Sixtus) II was executed at Rome (see 3.1, above; Cyprian, Epistle 80.1.4).

Though the proconsuls responsible for Cyprian’s exile and execution referred to ‘Roman’ cults and gods (Acts of Cyprian 1.1, 3.4), the emperors clearly wanted to uphold traditional cult in all its variety, and not just some set of cults especially associated with Rome, the empire, or the emperors’ persons (whose veneration rarely appears in accounts of the empire-wide persecutions; Millar 1973). Thus, at Alexandria, the bishop Dionysius reported being commanded to sacrifice to the ‘natural gods’ on the emperors' behalf (Eusebius, Church History 7.11.9). The same basic conviction that the worship of the gods — all the ancestral gods — was necessary for the empire’s prosperity undergirded the anti-Christian edicts of Diocletian (284–305) and his colleagues in the Tetrarchy. Now, however, the inconsistency of Roman policy would lead to markedly different results in different parts of the empire. Valerian had met a bitter end in captivity to the much-victorious Shah of Persia, Shapur I, and Valerian’s son Gallienus issued a general amnesty (Eusebius, Church History 7.13). In the decades following there was little danger in being a Christian, except for a few men caught between pacifism and forced military service (if accounts such as the Acts of Maximilian can be trusted; see Seeliger and Wischmeyer 2015: 400–405). The emperor Aurelian, from whom persecution was later feared, was regarded highly enough by Christians that they called upon him, successfully, to enforce the deposition of Paul of Samosata from the see of Antioch (Eusebius, Church History 7.30.19–21). The first move against Christians in some forty years occurred in 299, when the emperor Galerius purged Christians from his court and army after his haruspices – practitioners of the ancient divinatory system of Etruria in Northern Italy, long in use by Romans – blamed the failure of sacrificial divination on watching Christians who had made the sign of the cross (Lactantius, The Deaths of the Persecutors 10, with Borgeaud 2010; Barnes 1981: 18–19). His senior colleague Diocletian, at Galerius’ instigation (or so Lactantius, formerly a professor in Diocletian’s court, claims), consulted with his officials and sent a haruspex to ask the advice of Apollo at Miletus. Receiving the desired oracle, he began a general crackdown on Christians, beginning with his court at Nicomedia, on 23 February 303, the festival of the god Terminus (Mort. 11–12).

There is no space here to trace out all the twists of the so-called Great Persecution. Originating in the East, the first anti-Christian measures were enforced by the senior Western emperor, Maximian; his subordinate, Constantius Chlorus, who reigned in Gaul and Britain, commanded only the demolition of church buildings (Mort. 15.7). Neither emperor enforced the edict of universal sacrifice issued by Diocletian in 304, though some especially zealous governors commanded sacrificing on their own initiative (de Ste. Croix 1954; Barnes 1981: 23). After the retirement of Diocletian and Maximian in 305, the persecution lapsed throughout the West; though Eusebius, The Martyrs of Palestine (Mart. Pal.) 13.12, puts a round number of two years on the persecution in the West, toleration was only declared, in Italy and Africa, on the accession of Maxentius, Constantine’s future opponent, in 306 (Optatus, 1.18, Eusebius, Church History 8.14.1–6). The consequences of the persecution were still vast, as accusations that some African clergy had collaborated by handing over copies of the scriptures for burning (an innovation by Diocletian) led to the Donatist schism (for the theological background, see Coogan 2018). In the East, imperial hostility was much more intense, and was further encouraged by appeals from cities seeking favour by showing their zeal against Christians (Mitchell 1988). There, the emperors issued occasional and inconsistent decrees of toleration (most famously, a deathbed recantation by Galerius in 311: Lactantius, Mort. 34; for a useful survey, see Van Dam 2007: 163-169), but persecution was only definitively ended on the defeat of Maximinus Daza by Constantine’s colleague Licinius in 313. Laws from both Galerius and Maximinus (Lactantius, The Deaths of the Persecutors 34; Eusebius, Church History 9.7.3–14) describe the emperors’ motivations: by recalling the Christians to their ancestor’s religion, they intended to maintain not just the public order but also the favour of the gods for the emperors and their subjects.

3.3 The Christian empire: Constantine to Justinian

Constantine fought against Maxentius, self-declared emperor in Rome and himself tolerant toward Christians, under the sign of the cross (thus Lactantius, The Deaths of the Persecutors 44.5, Barnes 1981: 48, doubts in Harris 2015). Constantine was the first Roman emperor to take Christ as his own patron. With his Eastern colleague Licinius, he formulated a policy of general toleration that acknowledged all cults, including Christianity, as routes to divine favour: the so-called ‘edict of Milan’, preserved in imperial letters sent East in 313 (Lactantius, The Deaths of the Persecutors 48.2–12; Eusebius, Church History 10.5.2–14; see now Lenski 2017). Later, Constantine became increasingly opposed to traditional cults and Licinius turned (or so Eusebius claims, Vita Constantini [Life of Constantine] 1.49–41, Church History 1.8.8–19) against Christians. In 319 and 320, Constantine issued laws that allowed haruspicy to continue to take place in the city, but he defined it as a private ritual of some Romans, not of himself or the state (Theodosian Code 9.16.1–2). On appeal, he authorized divinatory reports after lightning strikes on imperial monuments (Theodosian Code 16.10.1); but, after defeating Licinius in 324, he took a hard line against pagan cult. According to Eusebius’ report, he shut down oracles, forbade erection of new idols, and banned sacrifices (Life of Constantine 2.45.1). The law has not been preserved and many scholars have doubted Eusebius’ construal of its contents, but the ban on sacrifices is confirmed by a law of the emperor Constans issued in 341 (Theodosian Code 16.10.2; on Constantine’s laws, Bradbury 1994 is especially lucid and accurate). Constantine had turned the policy inaugurated by Decius and pursued by Diocletian and Maximinus Daza on its head: what had been a mandatory act of worship was now forbidden.

To some Christians, the long-awaited victory over the Devil and the cults by which he deceived humanity was at hand: so Constantine himself reports in his ‘Letter to the Eastern Provincials’ (Eusebius, Life of Constantine 2.60.2). The fact that many Christians lived outside the Roman Empire and continued to be liable to persecution impinged only occasionally on Graeco-Roman consciousness (see, e.g. Augustine, City of God 18.52, for an acknowledgment of persecution in Persia and among the Goths). Constantine, aware of his subjects’ devotion, refused in the same letter to destroy the temples and other apparatus of traditional cult. His hesitation came despite his antipathy toward traditional cult of all kinds, which he expresses in the letter as well as in an extant speech. Though he showed preferences for pro-Christian cities (Lenski 2016: 162–164), he did not mandate any act of Christian worship from non-Christians, a precedent that held good for over two centuries. Eusebius names only a few isolated incidents in which temples especially offensive for prostitution or oracular agitation were destroyed (Life of Constantine 3.26, 53–8). In the mid-330s, Constantine granted a town in northern Italy permission to erect a temple of the imperial household, with a priest and games; it was, however, to be devoid of ‘superstition’ – in Christian parlance, sacrifices and all the other trappings of idolatry (CIL XI 5265, with Lenski 2016: 114–130; the law might also date from the interregnum after Constantine’s death in 337: Barnes 2014: 20–23).

The sons of Constantine followed their father’s policy, reiterating the ban on sacrifices and superstition; when Constantius II reigned as sole emperor in the 350s, the penalty for these offences was death (Theodosian Code 16.10.2, 4). Maintaining their father’s flexibility, they allowed processions connected with temples – which were not to be demolished – so long as ‘superstition’ could be omitted, and appointed new postholders to the Roman priesthoods (Theodosian Code 16.10.3). That action was taken during Constantius II’s visit to Rome in 357, only a year after a reiteration of the capital penalty for sacrificing (Theodosian Code 16.10.6). Constantius went on a tour of the city’s greatest sights, including its temples (Ammianus Marcellinus, 16.10.13–17; Symmachus, Relatio [State Paper] 3.8); the only concession to his Christianity, possibly made to please senatorial Christians, was the temporary removal of the Altar of Victory from the Senate-house (see now Moser 2018: 292–298; Gassman 2020: 53–54). Little if any evidence suggests that the laws of Constantine and his sons had (or, perhaps, were meant to have) more than a chilling effect (Bradbury 1994). Years later, the pagan orator Libanius would commemorate an older acquaintance for his constancy in worshipping the gods, despite the death penalty the emperor had threatened (Oration 1.27). A central rite of pagan worship was forbidden – but it could, with care, continue.

After Constantine’s reign, only one emperor espoused the traditional gods. Julian (361–363), whose family had been massacred in the power struggle after Constantine’s death, was a convinced Neoplatonist of a mystical stripe. He wrote theological treatises, among many other works, including a polemic against those he termed ‘Galilaeans’ (incomplete fragments survive, most in the extant books of Cyril of Alexandria’s Against Julian). Julian appears to have aimed not just at reviving local cult in all its pomp (an intention which did not always meet with enthusiasm from pagans themselves; Bradbury 1995) but also at reforming pagan religion. The goal was to transform it into something more closely resembling – and better able to compete with – the organized, episcopally dominated, humanitarian Christian churches (see Greenwood 2017, which surveys and builds upon a vast scholarship). Like his attempted rebuilding of the Jewish temple at Jerusalem and his ban on Christian teachers of Classical (that is, the ordinary) education, these efforts came to naught with his death during the disastrous Persian campaign of 363. (The historian Ammianus Marcellinus reports Julian’s death, offers a retrospective on his reign, and describes the succession of Jovian at 25.3–5; though an admirer of Julian, he criticizes the law on teachers at 25.4.20; for further reading, see Watts 2006: 48–78). His successors – the short-lived Jovian, and Valentinian I and Valens – pursued a tolerant, even pluralistic policy (on Jovian, see Themistius, Oration 5). Although attempts to restrict nocturnal worship (long associated with magic) could have forbidden even respectably ancient mysteries, if a well-connected pagan had not intervened on their behalf, Valentinian I condoned haruspicy and all ancestral rites as proper religion (Zosimus, 4.3.2–3, Theodosian Code 9.16.9).

That policy was to change under his sons. Appointed emperor as a child, Gratian was reigning in his own right in Milan in 382, when he issued measures against the cults of Rome, which had hitherto been left alone. A much later account has him refuse a pontifical robe, from which older scholarship inferred a renunciation of the title pontifex maximus in 382 (Zosimus, 4.36; full discussion in Cameron 2007). Mistaken – at best – in its details, the report captures the spirit of Gratian’s moves: confiscation of the property of the priestly colleges, a ban on inheritance by the Vestals, and removal of the altar of Victory from the Senate-house. This last was a profound symbolic statement, since senators had sworn allegiance to the emperors at the altar since Octavian’s triumph in 29 BC (Symmachus, State Paper 3.5; Herodian, 5.5.7, Suetonius, Divus Augustus [Life of Augustus] 35.3, with Gassman 2020: 108, note 7; the broad scope of Gratian’s measures has been successfully defended by Cameron 2011: 33–59, against Lizzi Testa 2007). Gratian did not force the great senators to leave their priesthoods, but he was going (it seems) to allow the priesthoods to die out for lack of support and presumably also of new appointees. A senatorial appeal was refused an audience, following transmission of a petition by Christian senators from Damasus of Rome, by the hand of Ambrose of Milan, to the court. The sequel – by divine wrath, as it seemed to pagan senators (Symmachus, Ep. 2.7.3) – was Gratian’s death in a revolt and famine in Italy. In 384, the eminent senator and pontifex Quintus Aurelius Symmachus sent a renewed appeal for restoration of priestly privileges (Relatio [State Paper] 3). Matched by open letters from Ambrose (Epistle 72[17]–73[18]), the appeal failed to win over the thirteen-year-old Valentinian II and his advisors. Despite future appeals to multiple emperors (Ambrose, Epistula extra collectionem [Epistle outside the main collection] 10[57]), the Roman priesthoods would remain permanently defunded. The ancient Roman cults had officially ceased to be public.

To bring idolatry to an end was an avowed purpose of Valentinian II’s colleague and successor, Theodosius I. Repeated laws issued by Theodosius and his descendants banned traditional rites, sometimes with minute specificity (Theodosian Code 16.10.7–9, 16.10.11–25). Theodosius, however, was reasonably lenient in enforcement of religious legislation (Errington 1997), and both emperors and their subordinates continued to promulgate and enforce laws unevenly. Thus, a particular crisis, the destruction of the great temple of Serapis at Alexandria in 391 or 392 – which followed on violent clashes between local Christians and pagans led by Neoplatonist philosophers – probably resulted from over-zealous construal of a recent law by the bishop Theophilus (Hahn 2008; for the date, Burgess and Dijkstra 2013). Afterward, pagans said that the destruction had been prophesied by a clairvoyant philosopher; and in the hearing of students at Constantinople, professors who had taken part in the fighting boasted of having killed multiple Christian opponents (Eunapius, Vitae Sophistarum [Lives of the Sophists] 6.96; Socrates Scholasticus, Historia ecclesiastica [Church History] 5.16.9–14).

The church historian Rufinus claims that demolition of temples throughout Egypt followed the destruction of the Serapeum (Historia ecclesiastica [Church History] 11.28). To both contemporary Christians and many modern scholars, Theodosius’ defeat of a revolt by the pagan general Arbogast and his hand-picked emperor, Eugenius, in Italy in 394 was a similarly fatal blow. It represented the crushing of a final rally by senatorial pagans, whose defeat occasioned widespread conversion (see Harris 2015 against the forceful but overstated arguments of Cameron 2011: 93–131). A plausible case can be made for significant senatorial conversion under Constantine (Barnes 1995), but the limited data do indeed suggest large-scale conversion in the last decades of the fourth century (Salzman 2002). The process was slower in North Africa, for example, where imperial commissioners ceremoniously demolished idols and shut down temples in March 399. Christians in Carthage in June 401, who had just witnessed (or taken part in) ritual vandalism of a statue of Hercules, called for their city to imitate Rome in eliminating its idols (Augustine, City of God 18.54, Sermon 24, with Magalhães de Magalhães de Oliveira 2006; Kelly 2015).

The diffuseness of polytheistic cult and the unevenness of laws’ promulgation and enforcement make it impossible to locate a definitive ‘end’ of all traditional cult, public and private, at Rome or in the empire at large (cf. Brown 1998; Trombley 1993–1994 is expansive, but not always accurate). The decades following the flamboyant measures of the 390s saw continued outbreaks of hostility between pagans and Christians – most famously the lynching of the philosopher Hypatia at Alexandria in 415 (on Hypatia, see Watts 2017). In the largest outpouring of anti-pagan literature since the Tetrarchy and the reign of Constantine, large-scale apologies were produced by bishops in both East and West: Cyril of Alexandria’s Against Julian; Theodoret of Cyrus’s Cure of Hellenic Maladies; and, most influential, Augustine’s City of God, to which the Spanish priest Orosius added seven books of Histories against the Pagans.

Though Eastern emperors forbade pagans from holding office, they were indulging in wishful thinking when they wondered, in 423, whether there were any pagans left (Theodosian Code 16.10.21–22). Even if the ‘paganism’ of the ancient Lupercal festival, celebrated at Rome in the 490s to the chagrin of Pope Gelasius, cannot be taken for granted (McLynn 2008), the Neoplatonists who troubled Cyril were still to be found in the Alexandrian philosophical schools as late as the 480s, where they were alleged, at least, to participate in idolatrous rites in temples (Zacharias Rhetor, Vita Severi [Life of Severus of Antioch] 17–18). Alexandrian Neoplatonists eventually found a way of coexisting with Christianity (indeed, many were Christians). Meanwhile at Athens, where the Academy was led by the prolific and deeply religious Proclus, a resolutely pagan Platonism endured decades past the closure of the cities’ temples, removal of their idols, and cessation of festivals, likely in the 460s (Marinus, Vita Procli [Life of Proclus] 28–29; Watts 2006: 79–256). It would not be until the reign of Justinian, in around 531, that an emperor would require his subjects to be baptized (Codex Iustinianus [Code of Justinian] 1.11.10); the contemporary miaphysite churchman John of Ephesus allegedly converted tens of thousands of pagans in the hinterland of Asia Minor (Menze 2008: 56–65). In 529, Justinian had issued a law forbidding divination, which led to a cessation of mystical Neoplatonist teaching at Athens; Damascius, who was leading the Academy once directed by Proclus, left for Persia (Agathias, Histories 2.30.3–4). An ‘end of paganism’ may not be placed even here: that would require ‘paganism’ to have a definite content and require certainty regarding the sincerity of converts who might, even if they really were baptized in such great numbers, have known essentially nothing of Christianity. Still, the succession of philosophical teachers who had energized both non-Christian and Christian thinking in late antiquity had ceased, as had the age-old worship practised by the cities of the Graeco-Roman world. The Eastern Roman Empire and the Germanic kingdoms in the West were now well on their way to becoming Christendom.