I

A review of the evidence for legal restrictions on temple cult prior to the 390s must begin with Constantine’s ban on sacrifice, which is controversial ground and requires lengthier treatment than subsequent legislation. Issued at some point after his victory over Licinius in September of 324, Constantine’s prohibition of sacrifice is not extant but is established primarily on the basis of two pieces of contemporary evidence. Eusebius of Caesarea, writing in the 330s, claims that Constantine issued a law ‘preventing the pollutions of idolatry which of old were performed in every city and country region, that no one should dare to erect cult statues, or engage in divination and their other superstitious vanities, or indeed sacrifice at all’ (ὁ [scil. νόμος] μὲν εἴργων τὰ μυσαρὰ τῆς κατὰ πόλεις καὶ χώρας τὸ παλαιὸν συντελουμένης εἰδωλολατρίας, ὡς μήτ’ ἐγέρσεις ξοάνων ποιεῖσθαι τολμᾶν, μήτε μαντείαις καὶ ταῖς ἄλλαις περιεργίαις ἐπιχειρεῖν, μήτε μὴν θύειν καθόλου μηδένα).Footnote ⁴ The final element in the tricolon is unambiguous: a total ban on θυσίαι (sacrificia), that is, not merely animals but also wine, votives, incense and every other type of bloodless offering.Footnote ⁵ Ostensibly, this would also encompass sacrifice both at public temples and on private property — a scope that Eusebius may reinforce by postponing the grammatical subject of the infinitives (μηδένα) to the position of emphasis at the end: ‘that absolutely no one should dare’, et cetera.Footnote ⁶ However, since the first reported element in the law — the ban on erecting cult statues — appears to pertain to temples, it may be that temple cult was especially in view. The middle element is less clear but generally has been understood to refer to private magical practices.Footnote ⁷ While not impossible, this interpretation results in a rather heterogeneous law, and Eusebius’ syntax may indicate something else. The collocation μήτε μὴν καθόλου (‘nor indeed at all’) in the last colon suggests a genus, in this case τὸ θύειν, of which the preceding is a species.Footnote ⁸ If so, then the penultimate clause is a species of sacrifice and a very different interpretation suggests itself: the form of ‘divination’ in question may be haruspicy (or more precisely extispicy), which presupposes blood offerings, and ‘the other superstitious vanities’ may be the many other ritual elements in the grand spectacle of public animal sacrifice.Footnote ⁹ This would result in a more coherent law: a ban on (1) erecting cult statues at temples, (2) public animal sacrifice Romano ritu and (3) every other species of sacrifice, including all bloodless offerings. But at any rate, whatever one makes of the penultimate clause, the ultimate one is clear: according to Eusebius, Constantine banned sacrifice altogether.

There is a common belief that Eusebius’ account of Constantine’s reign is uniquely dishonest.Footnote ¹⁰ However, not only is the assumption unproven, it is also prima facie unlikely. Eusebius was an intellectual with a very public profile who was writing in the first instance for an audience of eyewitnesses. Indeed, in the case of this law, he was reporting not only a recent political fact but also a current legal reality. Since the unambiguous claim of a total ban on sacrifice could have been instantly verified or falsified by everyone, a bare-faced lie would have accomplished nothing other than the immolation of his own credibility and standing.

Of course, while public intellectuals were inhibited from inventing the current or recent political facts that were known to all, they could and did subtly manipulate them through rhetorical framing: conjectural or specious imputation of meaning or motive, calculated omission, deliberate ambiguity, misleading juxtaposition and many other techniques. Eusebius is no different from any other in this regard. In fact, there may be an example of this sort of subtle rhetorical manipulation in his report of Constantine’s law. Eusebius does not say that it was issued immediately after the conclusion of civil war in late 324 — that is a claim that could have been instantly verified or falsified by his original audience — but his narrative order gives the impression that this may have been the case. Late 324 is therefore the conventional date in modern scholarship.Footnote ¹¹ However, despite Eusebius’ solicitous care for chronology elsewhere in his œuvre, his Vita is frustratingly vague on the timing and sequence of events, heightening suspicion that he may have been using a studied vagueness at points to his rhetorical advantage. In this case, one should be alert to the possibility that Constantine may not have issued his ban immediately.

There may also be some positive evidence that Constantine did not issue the law until years later. In Jerome’s translation and continuation of Eusebius’ Chronicon, an entry under the year 331 states: ‘The pagan temples were overthrown by Constantine’s edict.’Footnote ¹² This could have something to do with his confiscation of their cult statues, other possessions and sources of revenue.Footnote ¹³ However, an entry from the preceding year states: ‘Constantinople was dedicated after the spoliation of almost all the cities.’Footnote ¹⁴ Everyone has understood this, no doubt correctly, to refer to the plundering of the temples, which therefore had occurred already prior to 330 (even if the building also continued after this date).Footnote ¹⁵ What, then, could the edict of 331 be — the edict that undermined the temples — other than the ban on central aspects of temple cult? In fact, this timing would also explain the part of the law, as reported by Eusebius, that is anomalous and has not garnered much attention: the ban on erecting cult statues. In the period prior to 330, Constantine had deprived the temples of their statues (not every temple in the world, but apparently many of them).Footnote ¹⁶ In his edict, he then forbade anyone from erecting new statues and from performing sacrifice. It appears that Constantine may not have issued his ban until 331.Footnote ¹⁷

If this is correct, the great rhetorical advantage gained by Eusebius through his narrative sequence is that the reader is encouraged to interpret Constantine’s Letter to the Eastern Provincials, issued in late 324, in light of the edict banning sacrifice, which had not yet been issued. The Letter, quoted in full by Eusebius (Vit. Const. 2.48–60), concludes with a statement of religious toleration — of a sort, but the framing is vastly different from the neutral statement of this principle in the ‘Edict of Milan’ (313). It makes perfectly clear that Constantine holds temple cult in extreme contempt and would be happy to see it eradicated, but it also admits that this is not yet possible and it calls on Christians to refrain from applying force to their non-Christian neighbours.Footnote ¹⁸ Eusebius’ narrative placement of the ban prior to the Letter was a clever way to prejudice the reader’s perspective on this document, or more charitably to signal what the emperor had intended all along.Footnote ¹⁹ However, for many modern readers, in the absence of firsthand knowledge of the political realities, Eusebius’ narrative order has had a rather different impact. Some have used the Letter, which expresses grudging toleration, as evidence that the ban preceding it could not have existed.Footnote ²⁰ Malcolm Errington has attempted to rescue the situation by proposing that Constantine did issue the law described by Eusebius, in late 324, and that the Letter then rescinded it — an attempt at restriction followed by a quick return to toleration.Footnote ²¹ This is implausible, not least because Constantine’s law is cited as valid in a constitution of 341 (discussed in the next paragraph).Footnote ²² But if the argument presented here is correct, the perceived problem evaporates: the legal restriction of temple cult did not begin until 331, roughly seven years after the Letter to the Eastern Provincials. Constantine’s views and policies, which had shifted noticeably between 313 and 324, evidently continued to harden during the final thirteen years of his reign.

The matter of its date aside, Eusebius’ report of the edict’s contents is independently confirmed by a surviving law of 341, issued in the names of Constantius and Constans (the latter name dropped out in transmission):Footnote ²³

This is a clear ban on sacrifice, without qualification, and by its own admission it merely reiterates a law that had been issued by Constantine. The recipient was L. Crepereius Madalianus, vicar of the diocese of Italy, indicating that this law emanated from the western court of Constans.Footnote ²⁴ T. D. Barnes has suggested that Constantine’s initial ban was limited to the east — perhaps in its promulgation but certainly in its enforcement — and that this law represents an extension of his policy to the west.Footnote ²⁵ But this cannot be quite right. If Constans had wished to promulgate a general western law, edictal letters would have been issued to the two western praetorian prefects. These officials were responsible for disseminating general legislation, but the imperial address was not altered for each lower official who received a copy.Footnote ²⁶ Therefore, the letter from which this law was excerpted was issued by Constans directly to the vicar of Italy. This fact, along with the fact that it admits to instituting nothing new, indicates that it was responding to a consultatio from Madalianus, who perhaps simply wanted another clear imperial condemnation of the practice.Footnote ²⁷ In other words, Constantine’s original prohibition was already the law of the land in the west, as well as the east, and Constans’ response to Madalianus merely confirms it. Not only is Eusebius’ claim of a ban on sacrifice vindicated; so too is the impression that he gives of its universality.Footnote ²⁸

The convergence of two independent contemporary witnesses — one literary and one documentary — supported by some ancillary evidence that is less decisive is a powerful evidentiary basis.Footnote ²⁹ The only thing that could overturn them would be contradictory evidence of even greater quality, although such an occurrence in this case is scarcely imaginable. As it happens, there is only one directly contradictory piece of evidence that sceptics have been able to adduce, Libanius’ Pro Templis, and in particular the following claim:Footnote ³⁰

This would carry more conviction if it had not been claimed six decades after the fact, not to mention in a rhetorical context in which Libanius was attempting to convince Theodosius that he ought to tolerate temple cult.Footnote ³¹ The corollary of the general rule proffered above — that public intellectuals sometimes subtly manipulated their presentation of current or recent political facts but did not invent them — is that they are far less trustworthy on the distant past. This is both because they are less likely to be well informed themselves and, perhaps more importantly, because their original audience was in no position to verify or falsify the claims; as a result, they were free to indulge in cruder forms of distortion.

Moreover, it may be that Libanius is doing something much more interesting here than either demonstrating his ignorance or telling a bare-faced lie.Footnote ³² His claim is frequently interpreted as though it denies that Constantine issued a law banning sacrifice, but this is not in fact what he says. What he says is that Constantine disturbed no aspect of worship that was κατὰ νόμους. One could take this to mean worship that was ‘in accordance with custom’, but one could take it equally to mean worship that was ‘in accordance with law’. On the second reading, the statement is certainly true. After all, Constantine’s legal ban on sacrifice left ‘everything else’ (ἅπαντα τἄλλα) — priests, processions, spectacles, prayers, hymns — intact and present for all to see.Footnote ³³ The ambiguity further degrades the quality of this evidence. But it is also scarcely credible that the ambiguity was unintentional. Libanius constructed a statement that could be read in two incompatible ways — one true and one false — and left it up to his audience to decide. It is a studied ambiguity that absolves him of the charge of having told an open lie.

Libanius arguably employs the same rhetorical trick later in the Pro Templis, too, when he embarks on an account of the punishments suffered by Constantine for his transgressions and in an aside says that he will pass over a fact that is best left untranslated: οὐκ ἐπὶ τὰς θυσίας προῆλθε.Footnote ³⁴ Everyone has taken the preposition in a hostile sense and interpreted this as another denial of Constantine’s ban: ‘he did not proceed [i.e. legislate] against sacrifice’. However, the clause is strikingly vague and is susceptible of other plausible readings. For example, it could be construed to mean that Constantine ‘did not come forward to sacrifice’.Footnote ³⁵ This certainly would have been true at any point after about 312. Indeed, there was a famous story circulating in the east in the late fourth century according to which Constantine was expected to come forward to perform public sacrifice in Rome but did not, causing great offence.Footnote ³⁶ Libanius himself refers to this episode in a different rhetorical context in two other roughly contemporary orations for Theodosius, claiming that Constantine defused the situation by issuing a law to the effect that the people of Rome were permitted to express discontent with their emperor — an act of magnanimity that won him great acclaim.Footnote ³⁷ But to return to the ambiguous clause in the Pro Templis, if Libanius had wished to claim that there was no Constantinian law prohibiting sacrifice, why would he not have said simply οὐ τὰς θυσίας ἀνεῖλε/ἐκώλυσε (‘he did not ban sacrifice’)? These are unambiguous transitive verbs that he uses regularly of legal bans when he is not being evasive.Footnote ³⁸ The answer is presumably that he did not wish to tell an outright lie, but for rhetorical reasons, he did wish to present the opportunity for his audience to form a false impression.

At the age of about 75 in 390, Libanius was one of the few survivors with a living memory of the political realities during the final years of Constantine’s reign, and he undoubtedly remembered the truth even six decades later. However, he coyly left the door open for his juniors, who had no firsthand knowledge, to form a false impression. Not only is the solitary and late testimony of Libanius insufficient to dislodge the clear, independent and convergent contemporary testimonies, it is also less explicit and rhetorically trickier than has been hitherto observed.

The present author is compelled to draw a confident conclusion, based on the available evidence, that Constantine did indeed ban all sacrifice in his domain. How do proponents of the ‘minimalist’ position deal with the powerful evidentiary basis for this conclusion? There is a long scholarly tradition of simply dismissing Eusebius, but this convenience becomes considerably more difficult when his claim converges with the archival legal evidence of Cod. Theod. 16.10.2, which would appear to be unassailable. The ‘minimalist’ stratagem devised in this case is critical because it is deployed repeatedly in the case of subsequent evidence, too, to argue that all aspects of public cult remained legal down to the 390s. The claim is that Cod. Theod. 16.10.2, contrary to all appearances, did not ban sacrifice, but merely secret haruspicy.Footnote ³⁹

At the root of this interpretation are two laws issued by Constantine in 319, while he was Augustus of the west only.Footnote ⁴⁰ These prohibit diviners (haruspices) and pagan priests (sacerdotes) from entering private homes to practice their craft, obviously due to the potential political dimension of sacrifice with haruspicy occurring behind closed doors. Moreover, both laws explicitly, if also contemptuously, allow pagans to sacrifice in public: ‘But if you think it profitable, go to the public altars and shrines and celebrate your customary rituals; for We do not prohibit the offices of an obsolete rite to be performed in the open.’Footnote ⁴¹ In the comparable portion of the parallel law, public sacrifice is classified as superstitio.Footnote ⁴² Both laws clearly permit, but also disparage, traditional cult. The ‘minimalist’ position is that this remained Constantine’s policy to the end of his life (almost two decades later).Footnote ⁴³ And if the law of 341 was citing Constantine as precedent, the thinking goes, then it, too, must have been proscribing only secret haruspicy. However, if this was the legislators’ intent in 341, it is difficult to see how anyone could have discerned it. Christian legislators were perfectly capable of specifying their target when it was secret or nocturnal sacrifices, private divination or any of the dark arts.Footnote ⁴⁴ In this case, the target was sacrificia without qualification, and they were merely confirming Constantine’s law.Footnote ⁴⁵

This is perhaps the place to address two rescripts from the 330s. The Orcistus epigraphic dossier is not especially relevant, but it has been claimed in support of the ‘minimalist’ view.Footnote ⁴⁶ It contains responses from Constantine precipitated by two civic petitions, separated by a number of years.Footnote ⁴⁷ The first, from late 324–326, restored Orcistus to the dignity of a civitas — a status that it had lost at some earlier point when it was subordinated to the neighbouring city of Nacoleia. We learn from this rescript the interesting detail that the Orcistans had claimed in their petition to be entirely Christian — presumably a confession designed to incline the imperial ear to their plea.Footnote ⁴⁸ The second rescript, again favourable, was issued on 30 June 331. This one responds to a second petition in which the Orcistans evidently had complained that, despite the earlier ruling, the Nacoleians were still compelling them to contribute financially to their civic cults.Footnote ⁴⁹ The timing is a point of interest. It was suggested above, on the evidence of Jerome’s Chronicon, that Constantine’s ban on aspects of temple cult may have been issued in 331. One wonders, therefore, if the Orcistan petition of the same year was inspired precisely by the publication of this edict. The new legal situation may have offered an opening for the Christian occupants of this city to press their claim against their pagan neighbours, especially if the elements of cult being funded (unspecified in the surviving text) included sacrifice.

The Rescript of Hispellum is more obviously relevant. Issued in 333–335 or 337 (summer) in response to a petition from the Umbrian city of Hispellum, this imperial response grants the request to erect a temple to the gens Flavia, with a high priest and spectacles, but with the stipulation ‘that the temple dedicated to Our name not be polluted by the deceits of any contagious superstition’.Footnote ⁵⁰ The ‘minimalist’ interpretation is that this forbids only clandestine sacrifices with a divinatory purpose.Footnote ⁵¹ However, it would be an odd stipulation that secret haruspicy not occur at a public temple. Moreover, as noted above, Constantine had been classifying public sacrifice as superstitio since at least 319.Footnote ⁵² One presumes that this same equation was also made in his general ban, for it recurs in Constans’ confirmation of the law in 341.Footnote ⁵³ It is difficult to avoid the conclusion drawn by Noel Lenski: ‘The restriction against contagiosa superstitio is thus a blanket ban on sacrifice and perhaps also idolatry at this new imperial cult site in Umbria.’Footnote ⁵⁴ One may add that this was merely a reminder of the blanket ban that Constantine had already implemented in his entire domain.

The ‘minimalist’ view becomes even more strained when it is confronted with Constantius’ legislation. In the case of the first of two laws from 356, this statement is preserved:Footnote ⁵⁵

The purpose and context of this fragment are uncertain. The second law from 356 is more expansive:Footnote ⁵⁶

This confirms Constantine’s ban on sacrifice, but goes further by closing the temples and prohibiting access to them. This was a ban on all temple cult, and it certainly appears to be universal in scope. Not only is it addressed to a praetorian prefect (copies were presumably sent to his colleagues), it also clearly states that it applied to temples ‘in all rural locations and cities everywhere’ (omnibus locis adque urbibus universis). This phrase is difficult to square with a supposition of merely local applicability.Footnote ⁵⁷ (For an argument that it applied even to Rome and Alexandria — an annulment of their exemptions, but a short-lived one — see below, Section III.)

Despite what is evidently a ban on all temple cult in 356, Roland Delmaire has argued that sacrificia here are once again to be interpreted as merely secret acts of haruspicy. (One may note that, as in the case of Constans’ law of 341, haruspicy, secret or otherwise, is not even mentioned.) Furthermore, Delmaire argues that the closure of the temples was merely a step taken to prevent such clandestine acts, and that traditional blood sacrifice performed at open-air altars in front of the shuttered temples was still allowed under these laws.Footnote ⁵⁸ This is implausible. If one had wanted to conceal a sacrifice conducted for divinatory purposes, surely one would have selected a private property shielded from prying eyes — as reflected in Constantine’s two laws of 319, discussed above — and not the interior of a public temple. Moreover, Delmaire’s hypothesis is flatly contradicted by the evidence. Libanius describes Constantius’ policy on sacrifice simply as μηκέτ᾿ εἶναι θυσίας (‘sacrifice was no longer allowed’) without qualification.Footnote ⁵⁹ There is plenty of other evidence, too, that this law amounted to a ban on temple cult tout court and that the restrictive legislation of a generation or more had been broadly effective.Footnote ⁶⁰

When Julian succeeded his cousin as sole Augustus, he faced a situation in which the eastern temples, at least, were in a ruinous condition. He immediately rescinded the religious legislation of his predecessors and issued his own, reopening the temples, legalising sacrifice and generally resurrecting the worship of the gods.Footnote ⁶¹ There is no indication in the sources that these legislative acts were limited to reinstating the clandestine procedures that some believe were the only rites to have been prohibited. Julian personally oversaw the restoration of old temples or the creation of new ones, because some ancient temples were now physical ruins (not by natural means).Footnote ⁶² According to Libanius, in the period leading up to Julian’s ascension to the position of sole Augustus, all of Hellas was praying silently for this outcome but without resorting to altars, because there were none; sacrifices had ceased altogether but were reinstated by Julian in person at the Parthenon in Athens and elsewhere.Footnote ⁶³ Moreover, the disappearance of temple cult was not a recent phenomenon. Libanius recalls that when Julian revived sacrifices and festivals, there were only a few who could remember their performance in earlier times.Footnote ⁶⁴ Julian himself lamented that no one in Cappadocia was sacrificing during his reign and those who wanted to worship in this fashion did not know how to do it.Footnote ⁶⁵

Some would object that these conditions were due not to imperial legislation, but rather to the organic process of Christianisation and a general malaise surrounding aspects of traditional cult, especially animal sacrifice.Footnote ⁶⁶ While these may have been factors, the attempt to deny the role of imperial legislation encounters serious obstacles in the evidence. For instance, in his appeal to Julian on behalf of Aristophanes of Corinth, Libanius claims that during Julian’s time in Gaul (355–361), Aristophanes came to ‘what was left of the temples’ (τὰ λείψανα τῶν ἱερῶν) and brought offerings of tears but ‘no incense, no animal victim, no sacrificial fire, no libation, because it was not possible’ (οὐ λιβανωτόν, οὐχ ἱερεῖον, οὐ πῦρ, οὐ σπονδήν, οὐ γὰρ ἐξῆν).Footnote ⁶⁷ In his Epitaphios for Julian, Libanius claims that the future Augustus was deeply affected as he observed not only temples in ruins and altars overturned, but also ‘religious rites suspended’ (τελετὰς πεπαυμένας) and ‘sacrifices banned’ (θυσίας ἀνῃρημένας).Footnote ⁶⁸ Reflecting after Julian’s death on the period preceding his reign, Libanius says the following:Footnote ⁶⁹

This combines the legal restrictions, on the natural reading of them, with paralegal acts of property destruction and profanation. And for Libanius, although he did not want to admit it, this ‘war’ against the temples had been initiated by Constantine.Footnote ⁷⁰ Even allowing for the possibility of exaggeration, one cannot in any way square these contemporary assessments with the view that traditional public cult was allowed to continue unperturbed or that imperial legislation had no appreciable impact.

Evidence for the west outside of Rome during the same period is almost non-existent, but the epigraphic record from North Africa is interesting.Footnote ⁷¹ After a Carthaginian inscription attesting the restoration of a temple of Magna Mater in 331–333, thus roughly contemporary with Constantine’s ban, the record is silent on religious matters until Julian, who is hailed in contemporary inscriptions as restitutor sacrorum and restitutor libertatis et Romanorum religionum.Footnote ⁷²

The contemporary evidence appears to vindicate Sozomen’s mid-fifth-century verdict on this period: ‘[The sons of Constantine] confirmed their father’s laws and issued their own against those who would engage in sacrifice, idolatry or other forms of pagan worship. They ordered temples everywhere, in cities and country regions, to be closed; some of these they gave over to the churches when they required either land or building materials.’Footnote ⁷³ This is the ‘maximalist’ view in brief, and it can boast a firm grounding in the legal and anecdotal evidence.

The period after Julian’s brief reign, under the Christian emperors Jovian (363–364) and then Valentinian I (364–375) and Valens (364–378), was characterised by some pagan contemporaries as a period of religious tolerance.Footnote ⁷⁴ This verdict appears to be justified, but it did not take long for some of the old restrictions to return. According to Socrates in the mid-fifth century, Jovian banned public animal sacrifice and closed the temples, but this is almost certainly inaccurate.Footnote ⁷⁵ Our better evidence indicates that nothing occurred on this front until the joint reign of Valentinian and Valens. During their first year in power, they issued a law banning nocturnal rites, including nocturnal sacrifices.Footnote ⁷⁶ According to Zosimus, writing in the early sixth century but presumably relying here on Eunapius’ lost history, Vettius Agorius Praetextatus, who was governor of Achaea at the time, petitioned Valentinian for nocturnal rites to be allowed in this province (evidently for the Eleusinian Mysteries) and the emperor assented.Footnote ⁷⁷ There may be some slight corroborating evidence for this exemption.Footnote ⁷⁸

Committed ‘minimalists’ maintain that this law was the only restriction on pagan cult (re)introduced by Valentinian and Valens.Footnote ⁷⁹ However, this position requires one to ignore or deny the explicit testimony of Libanius, who claims that they eventually reinstated the ban on sacrifice. He says: ‘Animal sacrifice remained legal for a while [scil. after Julian], but after the occurrence of some rebellious acts, it was banned by the two brothers; incense, however, was not banned.’Footnote ⁸⁰ This law, like Constantine’s, does not survive, but it is implausible that Libanius would have invented a recent political fact of this magnitude, with such circumstantial detail, especially when it was also purportedly the current legal reality at the time of writing. If we can identify the ‘rebellious acts’ that he mentions, we may be able to refine the date. There are two primary candidates: either the attempted usurpation of Procopius (therefore c. 366) or the magic and treason trials (therefore c. 372).Footnote ⁸¹ For the purposes of this article, it does not really matter which. The chief point is that sacrifice was prohibited again during the joint reign of Valentinian and Valens. And the scope of the ban is clear from this passage: it applied to sacrifice, not merely secret haruspicy, but it also included a general exemption for offerings of incense. (Note that a ban on private haruspicy with allowance for incense would be nonsensical.) Incidentally, this testimony also vindicates Socrates’ claim of a renewed ban on public blood sacrifice, even if he placed it at least two years too early.

Some would cast doubt on Libanius’ claim that the reinstated ban made allowance for incense.Footnote ⁸² These scholars point out that no distinction between bloody and bloodless offerings is documented in the surviving laws until Theodosius’ comprehensive ban of 392.Footnote ⁸³ But an argumentum ex silentio on its own cannot overturn the positive contemporary testimony of Libanius. Indeed, he doubles down on his assertion by thanking Theodosius for confirming the exemption for incense, bringing his claim for this legal reality from the recent past into the present:Footnote ⁸⁴

This appears to be an excellent guide to the legal situation at the time of composition: the temples were open and offerings of fire and incense — that is, not just frankincense but any pinch of an aromatic — were permitted on public altars, but all other forms of sacrifice were prohibited.Footnote ⁸⁵

The law of Theodosius to which Libanius refers here is probably extant and seems to confirm what he says:Footnote ⁸⁶

This may have been issued — early in Theodosius’ reign but after the military crisis that dominated his attention upon accession had become less pressing — merely as a statement of continuity in religious policy. In any case, ‘minimalists’ seize on the passing reference to divination to claim that this law restricted only private haruspicy, apparently disregarding the reference to the public temples.Footnote ⁸⁷ As a result, they take the ‘illegal sacrifices’ (vetita sacrificia) to be secret sacrifices with a divinatory purpose. That is one way to read the law, but it is governed entirely by the ‘minimalist’ hypothesis and it is incompatible with what Libanius says. If we allow the evidence of Libanius to inform our reading, something else entirely emerges. Haruspicy, of course, had always been a constitutive component of public blood sacrifice Romano ritu — arguably the component that would have been most revolting to Christians and most troubling to emperors — so it is no surprise to find reference to it in some of the laws that prohibit sacrifice.Footnote ⁸⁸ Perhaps most illuminating, Theodosius’ comprehensive ban of 392 explicitly (and uncontroversially) outlaws all manner of sacrifice — notably with a section on the burning of incense (§2) — and yet it lingers over consultation of the ‘living entrails’ (spirantia exta), which is equated with treason regardless of the content of the consultation (§1).Footnote ⁸⁹ Likewise, while there is an allusion to haruspicy in the law of 381, this appears to be in the context of a general ban on sacrifice. What, then, of the vetita sacrificia? These are the sacrifices that had been banned by Valentinian and Valens, that is, everything except for fire and incense. Condemnation of anyone who engaged in ‘illegal sacrifices’ was the legislators’ way of quietly confirming the general exemption for incense, just as Libanius says.Footnote ⁹⁰

The same legal situation is presupposed in Libanius’ encomium of the January Kalends, delivered precisely on 1 January 392, which concludes with the following notice: ‘Altars of the gods do not now possess everything that they did in former times, due to the law that prohibits it; indeed, before the prohibition, the first of the month yielded much fire, much blood and much aroma of burnt fat rising from every region to the heavens.’Footnote ⁹¹ The altars ‘do not now possess everything’, so they do possess something, namely, incense. The prohibition of public sacrifice, but with allowance for incense, prevailed from whenever Valentinian and Valens renewed the ban (c. 366/372) until Theodosius removed any remaining allowances on 8 November 392 (less than a year after Libanius delivered this oration).Footnote ⁹² Incidentally, and once again controverting the view that anti-pagan legislation applied merely to secret and dangerous rites, the January Kalends was an empire-wide public festival, and yet, on the evidence of Libanius, it was subject to the legal restrictions on pagan cult.

To summarise: after Constantine’s law (perhaps of 331), public sacrifice was illegal, except for an interval stretching from the end of 361 to c. 366/372; for the period from 356–361, temple cult in its entirety was prohibited and access to the temples denied; for the period from c. 366/372–392, incense offerings were permitted.

II

The ‘minimalist’ hypothesis runs very much against the grain of the legal evidence considered thus far, and it also requires one to dismiss the contemporary testimonies of Eusebius (for the 320s–330s) and Libanius (for the 350s–390s). But this rearguard action is not entirely irrational, for ‘minimalists’ are able to draw on positive evidence of their own — material that puts ‘maximalists’ on the back foot. This contrary evidence is of two types: (1) extant laws indicating that public sacrifice was legal; (2) anecdotal testimonies that seem to support this conclusion. The legal evidence itself is the more critical and will be treated first.

However, this is the point at which to introduce a new hypothesis that may make sense of the evidence in its entirety: the cities of Rome and Alexandria were exempt from the restrictions on temple cult that were in effect everywhere else. In fact, this is not really a new hypothesis at all, because Libanius testifies explicitly to this reality in his Pro Templis.Footnote ⁹³ After registering the point that all blessings flow from the traditional gods, he continues (§33):

After dilating on this last point for a while, Libanius returns to his argument (§35):

It certainly appears that Libanius is speaking here of allowance for full civic cult in its traditional form, including public animal sacrifice, in Rome and Alexandria.Footnote ⁹⁴ And since this is an explicit claim for a current political reality, made to an original audience who could instantly verify or falsify it, one should have confidence in its veracity. Indeed, universal assent to the premise was required if the broader argument had any chance of succeeding: (a) sacrifice is permitted in Rome and Alexandria, (b) because of the material benefits ensured by their cults; therefore, (c) the same allowance should be universal, because all cults contribute to the public good in some way. The whole argument collapses almost before it begins — and Libanius forfeits his credibility entirely — if everyone scoffs at (a).

If it is agreed that this testimony should be taken seriously, a question must be posed. Who are the unnamed people who would have liked to ban sacrifice in Rome and Alexandria but did not? The flow of the polemic beginning in §30 and continuing into §37 suggests that the people intended are the shadowy group that is conjured throughout the Pro Templis, people who imposed on and abused the emperors’ good graces.Footnote ⁹⁵ Of course, while these people could whisper in the emperor’s ear, they did not possess the formal authority to ban sacrifice. This was the emperors’ sole prerogative.Footnote ⁹⁶ The passage is clearly a comment on imperial policy, dressed up in rhetorical garb.Footnote ⁹⁷ If sacrifice was not prohibited in Rome and Alexandria, as Libanius claims, this is because the emperors had declined to do so. This amounts to a legal exemption.Footnote ⁹⁸

Peter Van Nuffelen has objected to Libanius’ claim on the grounds that no such local exception appears in the preserved laws.Footnote ⁹⁹ As in the case of the universal allowance for incense established by Valentinian and Valens and discussed in Section I, Van Nuffelen offers here an argumentum ex silentio against the positive contemporary testimony of Libanius. The latter should prevail. But it is worthwhile asking why the exemptions for Rome and Alexandria would not be reflected (explicitly, at least) in the Theodosian Code. In formal terms, a local exemption of this sort would have been an indulgentia, that is, a special imperial grant that excused an individual or group from what was required under the law but did not establish a legal precedent.Footnote ¹⁰⁰ There were two broad types of indulgentiae: (1) imperial interventions for select people who had already violated the law, that is, pardons ex post facto; (2) legal protections established in advance to immunise select people from liability, that is, exemptions.Footnote ¹⁰¹ These are closely related legal concepts, but the second is the relevant type in this case, and it could be established in more than one way. One possibility was an indulgentia embedded within a general law intended for wide circulation.Footnote ¹⁰² However, it is unlikely that Christian emperors would have wanted to broadcast to everyone, gratuitously, an exemption for the civic cults of Rome and Alexandria. Much more plausible in this case are targeted laws that were not distributed far and wide. If the exemptions were a matter of imperial policy from the outset, the emperors could have issued epistulae expressing their indulgence to competent local magistrates (the urban prefect in Rome, say, or the provincial governor stationed in Alexandria). And these epistulae could have been candidates for inclusion in the Code, although that is certainly no guarantee of inclusion. The other possibility — and the most plausible one — is that the exemptions were established from below, at least formally, through rescripts issued in response to civic petitions. In this case, as with the rescripts to the cities of Orcistus and Hispellum (discussed in Section I), the expectation is that the compilers of the Code would have categorically excluded them as lacking generalitas.Footnote ¹⁰³

A variation on this last mechanism is reflected in the closest documented analogue to the exemptions for Rome and Alexandria described by Libanius. When Valentinian and Valens prohibited nocturnal rites in 364, the people of Achaea (or of Athens at any rate) sought indulgence for continuation of the venerable Eleusinian Mysteries, but rather than relying on a direct appeal to the emperors, they apparently routed their plea through the proconsul of Achaea at the time, the influential senator Praetextatus, whose preces to Valentinian was favourably received.Footnote ¹⁰⁴ One may well imagine that this would have been the precise mechanism in the case of Rome and Alexandria, too, especially the former, which surely would have routed its plea through the urban prefect, who had a direct line of communication with the emperors and was also a permanent member of the interested community.Footnote ¹⁰⁵ Of course, we only know about the exemption for Achaea from Eunapius (via Zosimus). It was not included in the Theodosian Code.

A parallel can be observed in the case of tax exemptions. A few relevant laws are preserved in the Code, but (with one interesting exception) these are all broad statements, issued to a praetorian prefect or comes sacrarum largitionum and either confirming or annulling previous grants en masse.Footnote ¹⁰⁶ Here, then, is an obvious question: where are all of the local tax exemptions, granted as imperial benefactions, that are known for the relevant period from literary sources? Why are they not recorded in the Code? It is scarcely credible that they were all invented. The likeliest explanation is that these local exemptions were usually rescripts issued in response to corporate pleas from civic or provincial bodies (whether the pleas were direct or routed through a magistrate). As a result, they were categorically omitted by the Code’s compilers as lacking generalitas.

Because the Theodosian Code preserves only a small fraction of the total legal output of the imperial scrinia, the absence of any given law is never a cogent argument that it did not exist, especially in the face of positive contemporary testimony that it did. And in the case of the exemptions for Rome and Alexandria, for technical reasons related to the likely legal mechanism, there is every reason to think that these laws would have been categorically excluded by the Code’s compilers.

The hypothesis derived from Libanius of local exemptions for Rome and Alexandria needs to be tested against the legal evidence, and in particular the laws that are crucial to the ‘minimalist’ hypothesis of universal allowance for temple cult. The first one to consider was issued by Constans on 1 November 342, not long after the law of 341 that confirmed Constantine’s general ban on sacrifice:Footnote ¹⁰⁷

This law explicitly places the extramural temples of Rome under imperial protection. A fortiori, the intramural temples were also protected (unless one is willing to countenance a scenario in which the emperors would have allowed the destruction of the Temple of Capitoline Jupiter but not some minor shrine ten miles outside the walls). The circumstances that occasioned this law are not difficult to reconstruct, speculatively of course. Some Italian Christians, perhaps emboldened by the reiteration of the ban on sacrifice from the preceding year, had attacked a temple or temples — a regular occurrence in the fourth century — somewhere beyond the Aurelian Wall. Catullinus, in his capacity as chief magistrate for Rome and its periphery, sought from the emperor explicit protection for the extramural temples, too. He received it, and it seems likely on this evidence that the protections extended to public cult as well as to the temple buildings. The sollemnitas at the law’s conclusion refers to ritual observances and may encompass public sacrifice; traditionally, at least, it certainly would have.Footnote ¹⁰⁸ Perhaps even more telling, the legislators did not take the opportunity to reiterate the existing blanket ban on sacrifice. In fact, they appear to allow it. Earlier legislation had defined sacrifice as superstitio (see above, Section I), but the concessive clause at the beginning of this law, while expressing the view that superstitio ought to be abolished, is in fact a concession that it need not be in this case. Compare with this a law issued by Theodosius that places a temple building in Osrhoene under imperial protection, clearly also in response to a request, but with the explicit stipulation that ‘prohibited sacrifices’ — that is, anything other than fire and incense — not occur there.Footnote ¹⁰⁹ By contrast, it appears that full temple cult was allowed in Constans’ law of 342, even if it was also disparaged as superstitio. However, this constitution was applicable only to Rome and therefore can be taken as evidence in support of the hypothesis of an exemption for the City. If so, the exemption apparently (and expectedly) extended to the entire jurisdiction of the urban prefect, that is, a radius of 100 Roman miles.

On 23 November 353, after vanquishing the western usurper Magnentius, Constantius issued the following law:Footnote ¹¹⁰

Nocturnal sacrifices and other secret rites had been banned universally (without exception) much earlier. Magnentius, presumably courting pagan sentiment to shore up his base of support, had apparently allowed for their return in Rome.Footnote ¹¹¹ Since it is highly unlikely that regular public sacrifice was banned under Magnentius while nocturnal sacrifices were allowed, Constantius’ order banning only the latter is powerful evidence that he was allowing the former (regular public sacrifice).Footnote ¹¹² However, this law was addressed to the urban prefect and therefore applied only to the City of Rome. It can be taken, in fact, as evidence for Libanius’ claim of an exemption.

During Valentinian I’s reign, a spate of prosecutions on the charge of magic roiled elite circles in Rome. This is the context in which he issued the following on 29 May 371:Footnote ¹¹³

Valentinian here stresses the freedom of religion that characterised his reign, but the attentive reader will notice that this did not entail unlimited freedom of practice. He does not condemn haruspicy or any other religious activity that was allowed by his predecessors. Of course, his Christian predecessors had in fact banned some pagan rites, including private haruspicy, which was outlawed already by Constantine early in his reign (with legal precedent going back to the first century) and is obliquely prohibited here in the final clause. Beginning later in Constantine’s reign, and then under his sons, sacrifice in its entirety had been prohibited — a ban that was reinstated by Valentinian and Valens themselves (but with allowance for incense). And yet, Valentinian here allows for public haruspicy, which presupposes public animal sacrifice.Footnote ¹¹⁴ Note, however, that this letter is addressed to the Senate. This is another law that powerfully attests the legality of public blood sacrifice in Rome, but this does not necessitate the conclusion that the same was true universally. (It should be observed in this case that, while this law certainly addresses the situation in Rome, it is uncertain whether Valentinian and Valens had yet reinstated the general ban on sacrifice, which may have occurred as late as c. 372; see Section I.)

If even one of these three laws had been addressed to a praetorian prefect, or to any magistrate or body outside of Rome, it would supply solid evidence for the view that public sacrifice was universally legal at the time of issue. But as it is, all three laws are addressed to Rome, and therefore can be taken to confirm Libanius’ claim of an exemption for the City. The earliest of them is from the year 342, suggesting ex hypothesi that the exemption had been in place for almost half a century, at least, when Libanius composed his Pro Templis. It seems likely, therefore, that it was established by Constantine when he issued the first general ban on sacrifice.

If we have a probable starting point for the exemptions, namely, Constantine — but at any rate prior to 342 — we also possess a reasonably certain date for their termination. In 391, Theodosius issued two laws that prohibited sacrifice and closed the temples — one addressed to the chief magistrate in Rome and the other to the chief magistrates in Alexandria.Footnote ¹¹⁵ Due to constraints of space, these two laws cannot be quoted and discussed in full here, but they are clearly of local application.Footnote ¹¹⁶ As a result, they provide compelling circumstantial evidence in support of Libanius’ claim: these targeted laws terminated the exemptions that had been enjoyed by the people of Rome and Alexandria for their civic religion.Footnote ¹¹⁷ In fact, in the case of the latter, the legislators explicitly say that anyone who defies the new prohibition ‘shall know that he is not to be excused by any indulgentiae’.Footnote ¹¹⁸ It is unusual for an imperial constitution to define a crime and then immediately state that the emperors will not protect those who break the law. This should go without saying, unless, of course, there was reason to expect such protection. In this case, that reason is plausibly that the people of Alexandria had enjoyed a longstanding indulgentia exempting them from the universal bans.

The hypothesis derived from Libanius of standing legal exemptions for Rome and Alexandria is capable of explaining the post-Constantinian laws that make allowance for public sacrifice. Indeed, the hypothesis appears to be positively confirmed by the legal evidence.

III

There remains the matter of the supporting anecdotal evidence for the ‘minimalist’ view, which must now be examined on the hypothesis just introduced. This is a messier business, because the laws did not guarantee compliance, nor do they illuminate the manner of enforcement on the ground. These general matters require brief treatment first.

At the outset, one may state an indisputable fact: the emperors’ subjects did not always respect the prohibitions. This is exemplified in a case that has been mentioned already in passing. Libanius reports that the uncle of his friend Crispinus was truly divine, offering as proof the fact that ‘he went through life scorning a wicked law and its impious legislator’.Footnote ¹¹⁹ The passage is deliberately vague, but since the implied date is 339/340, this is almost certainly a reference to Constantine’s ban on sacrifice, which the ‘divine man’ flouted.Footnote ¹²⁰ Even the emperors’ highest-ranking representatives might flout the law, as did Anatolius, praetorian prefect of Illyricum, in Athens, probably in 357 or 358.Footnote ¹²¹ Delmaire takes this anecdote to indicate that public sacrifice was legal at the time.Footnote ¹²² However, Eunapius says that Anatolius sacrificed and visited the Athenian temples ‘daringly’ (θαρσαλέως) — an oblique notice, to be sure, but only sensible if he was violating the law. Indeed, when Eunapius adds that this was done ‘in the manner commanded by divine law’ (ᾗ θεσμὸς ἱερὸς ἐκέλευεν), one should probably see an implicit contrast with the human law that forbade it. Eunapius records something similar of another imperial magistrate, Justus, an otherwise unknown Roman senator and sometime vicar of Asia who erected altars in Sardis and performed public animal sacrifice there, probably under Theodosius I.Footnote ¹²³ Eunapius is coy again, but this is said to have occurred ‘when the project of the Christians was prevailing and holding everything in its grasp’.Footnote ¹²⁴ To these may be added Proculus, who was certainly a prominent figure under Theodosius. While he was governor of Syria Phoenice in 382–383, his act of sacrifice at Heliopolis was recorded on stone, but in this case there is no comment from a Libanius or a Eunapius to indicate (however obliquely) that this was in violation of the law.Footnote ¹²⁵ These examples of prominent individuals who flouted the prohibitions are only the tip of the iceberg. In fact, long after 392 — the point at which most would agree that all pagan cult was universally outlawedFootnote ¹²⁶ — there remained pagan strongholds here and there, especially in North Africa and the Near East, where entire communities continued at least some of their ancestral rites in plain view, in some cases into the sixth century.Footnote ¹²⁷ To state the obvious, then, because compliance with the law was not universal, anecdotal evidence for sacrifice (unless the source contains an indication of its legal status) is not determinative of anything.

The matter of occasional non-compliance raises another complicating issue: how were the laws enforced? In the first place, the absence of evidence for convictions on the charge of sacrifice or idolatry is decisive. If even one individual in the fourth century had been executed by the state merely for engaging in ancestral rites, this would have been a cause célèbre for Libanius and many others.Footnote ¹²⁸ The bans were evidently not enforced through the courts, despite the threats contained in the laws.Footnote ¹²⁹ This appears to be positively confirmed by the case of the elderly philosopher Demetrius, who in 359 — therefore, at a time when Constantius’ extremely restrictive legislation was in effect — was hauled before a tribunal that had been established to investigate acts of treason, reluctantly confessed to performing sacrifice but denied any treasonous activity, and was acquitted and released.Footnote ¹³⁰ Delmaire takes this as an indication that sacrifice was legal.Footnote ¹³¹ As noted above, this requires an implausible reading of Constantius’ laws. But in any event, the case of Demetrius is illustrative merely of the fact that the government had no interest in executing people for following their ancestral customs (treason was naturally a different matter).Footnote ¹³² There were no pagan martyrs, at least not at the hands of the state.

Of course, prosecution and state violence were only one means of achieving compliance, and a means, moreover, that had failed spectacularly when employed against Christians during the Great Persecution. Christian emperors opted for a two-pronged mode of persuasion that was less direct (and less brutal) but arguably more effective. The first prong was simply the publication of imperial legislation that disparaged pagan cult and declared aspects of it, at least, to be illegal.Footnote ¹³³ This undoubtedly had a chilling effect. Even if everyone was reasonably confident that performing sacrifice would not lead to prosecution in court, there must have been many who opted to stay on the right side of the law and in the emperors’ good graces.Footnote ¹³⁴ The second prong was described many years ago by Garth Fowden, although his account has not achieved universal assent, despite the overwhelming evidence: once sacrifice was criminalised, private citizens could enforce the law by interrupting it or by commandeering temples and private property where it was thought to be occurring, and they could do this with impunity.Footnote ¹³⁵ This is how the law was made effective so broadly and quickly, as is amply attested to have happened, especially by Libanius (see above, Section I). Public sacrifice in major cities became all but impossible once a crowd of Christians was on hand to make sure that it did not happen. These people had the emperors and the law on their side; resistance was futile. This mechanism is precisely what Libanius complains of in his Pro Templis; or rather, he complains of such paralegal acts occurring even when there was no evidence that the law had been broken.Footnote ¹³⁶ Nor was any of this new in the 380s. The same dynamic also existed under the House of Constantine.Footnote ¹³⁷ In an opposite iteration, it was prominent during the Great Persecution alongside judicial torture and execution.Footnote ¹³⁸ And it was Julian’s method of applying pressure to the churches, as he admits in his own words.Footnote ¹³⁹ Indeed, the mobilisation of popular violence in lieu of following the procedural channels of government is a hallmark of authoritarian regimes past and present.

Because the mechanism of enforcement was informal and ad hoc, the effects of the laws were felt differently in different places. Much depended on local factors, such as the relative security of the pagan community, its level of commitment to its cults, the state of its relations with Christian neighbours, and the local bishop’s appetite for conflict and risk (because attempting to commandeer or destroy a temple, for example, was a potentially fatal undertaking). One may be tempted to think that this dynamic could explain the freedom enjoyed by the temples in Rome and Alexandria (to be documented immediately below) — that their pagan communities were simply too powerful for the local Christian populations to enforce the law — but this would be mistaken. Libanius claims that sacrifice in these two cities, uniquely, had not been prohibited, and these exemptions can be observed in the legal evidence. As indicated in Constans’ law protecting the extramural temples of Rome, the sorts of paralegal attacks on the temples that were permissible elsewhere were not allowed in the jurisdiction of the urban prefect. Any attempt by the local Christian community, or by imperial officials operating on their own initiative, to disrupt civic cult in Rome or Alexandria would have been contrary to the imperial will and therefore plainly illegal.Footnote ¹⁴⁰

Despite the complicating factors just addressed, a survey of the anecdotal testimonies to public sacrifice in 331–361 and c. 366/372–391 reveals that most of them — including all of the evidence for full civic religion — pertain to Rome and Alexandria. Let us begin with Rome.

During the joint reign of Constantius and Constans (340–350), no earlier than 343, Firmicus Maternus published a polemical work in which he called on the emperors to issue ‘the harshest laws’ to eradicate ‘daily blood sacrifices’.Footnote ¹⁴¹ These words were penned not only after Constantine’s law banning sacrifice but also after its confirmation in Italy by Constans in 341. And yet, Firmicus seems to imagine a world full of bloody sacrifices that were still in need of legal suppression. Some have taken this to be clear evidence that laws prohibiting sacrifice did not yet exist.Footnote ¹⁴² Alternatively, it could reflect the situation in Rome rather than elsewhere. After all, Firmicus was a Roman senator from Italy, and Mattias Gassman has recently documented the Roman orientation of his De errore, as well as his earlier Mathesis.Footnote ¹⁴³ The more limited scope of his exhortation appears to be confirmed when he also urges the emperors to confiscate temple possessions.Footnote ¹⁴⁴ It is uncontroversial that Constantine had already done this on a large scale, but apparently not in Rome.Footnote ¹⁴⁵ Firmicus was urging the current emperors to do in the City what their father had done elsewhere, namely, to plunder the temples and ban sacrifice.

A decade or so later, in the spring of 357, Constantius famously visited Rome for the first and only time. The only documented act that he undertook to the disadvantage of traditional religion in Rome on this visit was the removal of the Altar of Victory from the Curia Julia.Footnote ¹⁴⁶ This was the secularisation of a political space that, after all, hosted a religiously mixed body of senators. Otherwise, it appears that civic cult in the City received imperial endorsement. At least, according to Symmachus, Constantius toured the temples in the company of senators, expressed admiration for the buildings and their histories, co-opted men into vacant pagan priesthoods and affirmed imperial investment in the temple cults.Footnote ¹⁴⁷ Understandably, scholars have had difficulty reconciling this irenic picture with the violence of Constantius’ recent anti-pagan laws that had prohibited all public cult and closed the temples. Some have suggested that Rome’s grandeur must have induced in the emperor a change of heart and religious policy.Footnote ¹⁴⁸ It may indeed be the case that something had changed between the law of December 356 and the visit of May 357.Footnote ¹⁴⁹ But this is not required to explain a stark discrepancy between the general policy and the one that applied in Rome, because Rome was legally unique: its cults were under imperial protection. Constantius therefore cordially dispensed his duties as pontifex maximus, however repugnant he may have found the City’s public rituals to be.

Two years later, in 359, the urban prefect Tertullus was sacrificing at the Temple of Castor and Pollux in Ostia when the grain ships, which had been delayed, arrived.Footnote ¹⁵⁰ This aligns with what is implied in the law protecting the extramural temples: the exemption for Rome was in fact for the entire jurisdiction of the urban prefect, that is, a radius of 100 Roman miles.Footnote ¹⁵¹ This also accounts for what otherwise may appear to be a surprising inscription from Ostia dating to c. 376: ‘Our Lords Valens, Gratian and Valentinian Augusti ordered that the temple and portico of the goddess Isis be restored under the supervision of Sempronius Faustus, vir clarissimus, prefect of the annona.’Footnote ¹⁵² Just as Constantius had filled vacant priesthoods and affirmed imperial funding of the temples in Rome, so here Gratian was fulfilling his duties as the nominal head of Roman religion, which was as yet unencumbered by the restrictions that existed elsewhere. And this bubble of exemption extended to the entire jurisdiction of the urban prefect, including Ostia and Portus.Footnote ¹⁵³

From Ostia, there is also one fourth-century inscription attesting a taurobolium, although the performance of it presumably occurred in Rome.Footnote ¹⁵⁴ From Rome itself, there are many more such epigraphic attestations.Footnote ¹⁵⁵ The most significant of these are inscribed altars deriving ultimately from the Phrygianum on the Vatican. These commemorate the sacrifice of a bull in the initiation of a named individual into the mysteries of Cybele and include consular dating formulae. In other cases, our knowledge that an individual had undergone this rite derives from Roman inscriptions in which the honorand is identified as tauroboliatus (or similar). These fourth-century initiates, almost exclusively from the senatorial nobility, many of them with known priestly roles in the City’s civic religion, would emerge from the concealed ritual, bespattered with blood and holding the bull’s testicles, to shouts of acclamation from their clients and others. As Neil McLynn has noted, this rite had become another notch in the competition for honours that characterised the senatorial class, although this is not to say that it was devoid of genuine religious meaning.Footnote ¹⁵⁶ It may be relevant that the last dated taurobolium is from 390, just one year before Theodosius revoked Rome’s exemption.Footnote ¹⁵⁷ And it is surely relevant that a ritual that had once been attested far and wide is known in our period almost exclusively from Rome.Footnote ¹⁵⁸

The famous sequel, a generation later, to Constantius’ removal of the Altar of Victory from the Curia Julia need not detain us for long. However, it is worth noting that when Gratian ordered the altar’s removal again in 382, he also took steps towards withdrawing imperial support from the cults of Rome. According to Zosimus, he renounced the role of pontifex maximus, although he probably did not.Footnote ¹⁵⁹ According to well-informed contemporary sources, he withdrew imperial funding from the cult of Vesta (or all of the civic cults); prohibited testamentary gifts of estates to the cult of Vesta and its officers (at least); and cancelled the immunity from munera that had attached to some priesthoods. The uncertainty is due to the fact that this is yet another crucial law that is not preserved.Footnote ¹⁶⁰ The impact of this largely symbolic severance can be registered in Symmachus’ plea to Valentinian II in 384 seeking restoration of the emoluments to the cult of Vesta (at least) as well as restoration of the Altar of Victory to the Curia Julia.Footnote ¹⁶¹ But there is no indication in his plea that Gratian had taken steps to interfere with the City’s cult practices; he had ordered the removal of the Altar of Victory and had withdrawn some material supports.

Ambrose, in his interventions into the proceedings, provides more explicit indications of the exceptional position occupied by Rome’s cults. He notes that the fiscal privileges of the temples had been abolished in almost the entire world very long ago, but at Rome were only recently withdrawn by Gratian.Footnote ¹⁶² In this respect, at least, Rome had enjoyed an exemption from the universal disadvantages. And in the matter of cult, with the exception of the Senate House, which had been secularised again by Gratian, there was complete freedom. At least, Ambrose claims that ‘they celebrate their sacrifices everywhere’ (sacrificia sua ubique concelebrant), including at the Temple of Victory.Footnote ¹⁶³ Alan Cameron is quick to note that Ambrose does not here specify that these sacrifices taking place all over the City included bloody ones.Footnote ¹⁶⁴ But in the same letter, the Bishop impersonates Roma: ‘Why do you stain me daily with the blood of innocent animals?’ And again: ‘Your sacrifice is a ritual of splashing yourselves with the blood of animals. Why do you seek the voice of God in dead cattle?’Footnote ¹⁶⁵ Cameron waves all such language away as ‘trite clichés’.Footnote ¹⁶⁶ However, just a decade earlier, as revealed in a law of Valentinian I, public animal sacrifice with haruspicy was occurring in Rome with the consent and protection of Christian emperors.Footnote ¹⁶⁷ Anecdotally but more spectacularly, in a contemporary letter to Praetextatus, Symmachus casually reports that the magistrates had performed eight blood sacrifices to Jupiter and eleven to Fortuna in an attempt to expiate a prodigy from Spoleto; these were unsuccessful and the next step was to call a meeting of the pontifical college.Footnote ¹⁶⁸ Ambrose was not indulging in mere rhetorical ornamentation in 384. Animals were still being slaughtered on the City’s public altars, and it was perfectly legal.

One could adduce additional anecdotal testimonies to the freedom enjoyed by Rome’s temples under Christian emperors in the fourth century.Footnote ¹⁶⁹ However, the review of the evidence undertaken to this point is sufficient. These impressive testimonies cohere with the legal evidence for Rome, which indicates that public cult in its entirety was permitted there prior to the 390s. And this only further confirms Libanius’ contention that the City possessed a legal exemption.

The anecdotal evidence for Alexandria is less copious but still illuminating.Footnote ¹⁷⁰ To begin with something roughly contemporary with Libanius’ claim, Eunapius reports that the philosopher Antoninus established a sort of pagan monastery at a temple of Serapis in Canopus, a suburb of Alexandria, where he devoted himself to the sacred rites and attracted acolytes who served as priests.Footnote ¹⁷¹ Eunapius claims that Antoninus avoided theurgy because it was illegal. Naturally, but sacrifice was illegal for much of the fourth century, too, at least outside of Rome and Alexandria. Reportedly, Alexandria was full of visiting worshippers, so that its population was double the number of native residents; ‘it was a sacred world’ (ἱερά τις ἦν οἰκουμένη) unto itself (6.10.8–9). It would not be surprising if pilgrim traffic to Rome and Alexandria had increased on account of their exemptions from the legislation that restricted public cult elsewhere. Of the early 390s, shortly after Antoninus’ death, Eunapius says: ‘Worship in Alexandria and at the Serapeum was scattered to the winds — and not only the worship but also the buildings’ (6.11.1). He does not indicate whether temple cult had been completely unconstrained prior to this, but the impression that he gives is of a vibrant public religion that was suddenly snuffed out. This seems to be confirmed by Sozomen, who reports the general fear that the Nile would fail to rise during the very year when Theodosius banned sacrifices in Alexandria (391);Footnote ¹⁷² the obvious inference is that the sacrifices had been in place until that point.

To step back one generation, probably in 360/361, an anonymous individual published the Expositio totius mundi et gentium, a little guide to Roman geography and commerce that was composed in Greek but survives only in Latin translation.Footnote ¹⁷³ In the chapters devoted to Alexandria, one finds this description:Footnote ¹⁷⁴

This strongly suggests that the temples of Alexandria were undiminished at a time when public cult in other major cities had apparently ceased. In fact, the contrast between Alexandria and elsewhere is presumably the very reason for the author’s evident wonder, for what he describes here would have been typical in many cities of the empire just a generation or two earlier. Moreover, it is probably not accidental that his other digression on marvellous religious practices is in the case of Rome: its citizens, too, are deos colentes, and there is praise for the City’s temples, priests and haruspices (§55).

Gabriele Marasco has already very plausibly connected the passage just quoted with Constantius’ repressive religious policy at the time (but without noting the exemption for Alexandria).Footnote ¹⁷⁵ And he finds a further reflection on the contemporary political situation in the passage that immediately follows:Footnote ¹⁷⁶

Marasco notes that the notorious volatility of the Alexandrian populace is here unusually given a positive valence. They justly, even if violently, oppose Roman officials who step out of line. Since this follows immediately on the reported vitality of Alexandria’s cults, Marasco argues that there is a plausible connection to be made with known events around the time of the Expositio’s composition. Space does not allow for a full account, but these events pertain to the tumultuous tenure of George of Cappadocia as the Homoean bishop of Alexandria — a bishop formally, but also operating informally as an imperial proxy.Footnote ¹⁷⁷ George was present in the city in 357–358, during which time he oversaw a persecution of Athanasius’ orthodox flock and also attempted to suppress Alexandria’s civic cults, not least by enlisting the dux Aegypti and imperial troops to seize the Serapeum. Universally hated by both constituencies, George barely escaped an attempt on his life before fleeing the city on 2 October 358.Footnote ¹⁷⁸ Is Marasco correct that these events lie behind the Expositio’s praise of Alexandria’s cults and of the city’s fierce opposition to Roman officials who overstep? The argument is very attractive.

But what do these events mean for the matter of Alexandria’s exemption? Sozomen reports that George not only ‘despoiled their cult images, votive offerings and temple ornaments’ but also ‘prevented them from sacrificing and celebrating their ancestral festivals.’Footnote ¹⁷⁹ This seems to confirm that public cult in Alexandria had been unconstrained prior to 357. And the natural inference is that Constantius had revoked the city’s exemption.Footnote ¹⁸⁰ In fact, this legal action is probably preserved in the harsh law of 1 December 356, which ordered the temples to be closed ‘in all rural locations and cities everywhere’ (omnibus locis adque urbibus universis).Footnote ¹⁸¹ Indeed, one may render ‘cities without exception’; these presumably included Alexandria (and Rome, too). Within weeks of the date of this law, George was despatched to Alexandria (arriving on 24 February 357), almost certainly with imperial permission to suppress public cult there.Footnote ¹⁸² This appears to be confirmed by Sozomen, who notes in a free-floating anecdote that ‘letters from the emperor were delivered to the magistrates [in Alexandria] ordering that the pagans were prohibited from assembling at their temples and observing their customary rites and festivals.’Footnote ¹⁸³ This sounds very much like the law of December 356, which George himself may have delivered to the magistrates when he arrived in Alexandria the following February.Footnote ¹⁸⁴

It appears that Constantius had revoked the exemption and George was arriving to oversee the implementation of a ban on the city’s cults. However, the attempt met with violent resistance. Julian merely hints at this in his famous letter to the Alexandrians, but the disturbance was evidently serious enough that the dux Aegypti had unleashed the army on the public.Footnote ¹⁸⁵ The civic turmoil seems to have convinced Constantius to abandon the attempt at suppression and revert to the status quo ante — a policy reversal that may have occurred within mere weeks of George’s arrival. At any rate, the testimony of the Expositio, which was published late in Constantius’ reign, indicates that Alexandria’s temples had recovered quickly.

These events necessitate a brief return to Constantius’ nearly contemporaneous visit to Rome. If early news of the debacle in Alexandria had arrived already at the imperial court in Milan in March or April of 357, this may have shaped the tenor of Constantius’ visit to Rome in May and the religious policy that he pursued there.Footnote ¹⁸⁶ Indeed, he may have been waiting for just such a report before deciding whether the suppression of Rome’s cults, in accordance with Cod. Theod. 16.10.4, was viable. It was not. Whatever else he hoped to achieve on this trip, part of the agenda may have been damage control for his premature attempt to prohibit temple cult universally, including even in Alexandria and Rome itself. A chastened Constantius may have needed to provide assurances to the Senate and Roman people that the sacra publica would continue as they always had. At least, this reconstruction would explain fully the deferential and conservative character of the imperial visit as reported by Symmachus and Ammianus.

The evidence for Alexandria, while inferior in quantity and quality to that for Rome, suggests that public cult was unconstrained long after the universal bans had been put in place. There was a temporary disruption in 357, probably reflecting Constantius’ attempt to revoke the exemption, but the emperor seems to have abandoned this policy quickly. Moreover, after the period of universal allowance under Julian and his immediate successors, once the ban on sacrifice was reinstated c. 366/372, Alexandria appears once again to have enjoyed an exemption for its cults.Footnote ¹⁸⁷ The reports from Eunapius and Sozomen considered above are suggestive, but there is no need to rely exclusively on inference, for we have the explicit testimony of Libanius that the ancestral cults of Alexandria were protected when he composed his Pro Templis. The other anecdotal evidence coheres with this testimony.Footnote ¹⁸⁸

IV

The hypothesis of an exemption for Rome and Alexandria is capable of explaining all of the seemingly contradictory evidence — without recourse to special pleading or selective blindness — and indeed appears to be positively confirmed by some of the evidence. It would be otiose at this stage to address the few remaining anecdotal testimonies that have been proffered in support of the ‘minimalist’ hypothesis but were not treated above. These fall into one or more of three categories: unconvincing as testimonies to sacrifice; of insufficiently precise date; or, in one case at least, probably to be explained as non-compliance.Footnote ¹⁸⁹

It should be noted that these exemptions are also eminently sensible as a matter of policy. Sacrifice, especially blood sacrifice, was abhorrent to ancient Christians. It is unsurprising that Christian emperors would have prohibited it. But Rome and Alexandria presented unique challenges for the implementation of a ban. This is not, as Libanius disingenuously supposes, because Christian emperors and their advisers believed that the religious rites of these two cities were efficacious. That is a transparently rhetorical imputation of motive, as Libanius acknowledges (μοι δοκοῦσιν). However, he is quite right that the security of the empire and Egypt’s grain supply were at issue. The religious sensibilities of the pagan populations in these two cities could not be simply disregarded, because the Roman Senate was still the most powerful political force outside of the civilian and military branches of the imperial government, fully capable of backing a serious threat to the sitting emperor, and because Alexandria controlled a large proportion of the empire’s grain. Any attempt to restrict civic religion in these two cities, in contempt of their pagan factions, would have been certain to spark mass social and political upheaval in the notional centre of the world and the most important hub in its food supply chain. Any emperor would have been cautious when faced with that prospect. Extension of the ban to these two cities would have to wait for a time when their pagan communities were too small and too weak to upset the political order. 357 was too soon, as Constantius discovered. Theodosius determined that the time was finally ripe in 391. Despite some ensuing turmoil, history vindicated his judgement.

***

To return at last to the radical divergence of scholarly opinion with which this article began, one may now conclude that everyone is correct, but in a qualified way. The broad middle is correct that the truth of the matter is more complicated than scholars at either extreme pole have been willing to allow, but not so complicated that it defies explanation. At one extreme pole, Delmaire is also correct that all temple cult remained legal until 391, but only in Rome and Alexandria with their peripheries. And at the other extreme pole, Barnes is finally correct that severe legal restrictions on temple cult were in effect during most of the period under consideration, but only in the other 99% of the empire.