C. first promised a book called The Last Pagans of Rome in 1981, and references to its progress recurred intermittently in his scholarly works over the following three decades – a period in which he was far from idle. In fact, in the acknowledgments of this vast, brilliant, unusual book, C. places its origins still further back, referring to articles of 1977 (Entretiens Hardt 23, 1–30) and 1966 (JRS 56, 25–38, on Macrobius); he could equally have mentioned his 1964 article on Ammianus and the alleged circle of Symmachus (JRS 54, 15–28). This long gestation has given scholars foreknowledge of C.'s overall approach and of many individual ideas and insights, and a few earlier articles are adapted into the book. C.'s penchant for creative destruction is well known, and occasionally the ‘standard views’ polemicised against have already so wilted under his attack and that of others that readers may think them made of straw; very occasionally, works cited as ‘recent’ are anything but. But the great majority of the material is published for the first time, and the work manages to be impressively coherent and up to date despite its formidable length and despite containing discrete studies that could have made separate books.

The pagans of the title are western, mostly Italian, senatorial aristocrats of the late fourth and early fifth centuries: they are viewed through their authorship of and appearances in numerous literary works of that period, as well as through material evidence. The title appears to reference Herbert Bloch, who in 1963 published the most explicit and extreme statement of the argument that the usurpation of Eugenius in 392–394 was supported by aristocrats – whom Bloch dubbed ‘the last pagans’ – as an act of resistance to Christianisation. Taking a title from his opponents is emblematic of the essentially polemical orientation of C.'s work. Disproving the alleged ‘pagan revival’ is the springboard for a wider argument, that in political practices and literary productions wherein previous scholars have seen organised pagan opposition to the encroachment of Christianity in social and political life, there is in fact nothing to be seen. Writers or individuals who are thought to embody pagan resistance do nothing of the sort; those who were pagans (Symmachus and Nicomachus Flavianus, for example) were not fanatics, and many of them were actually Christians (Macrobius, above all); literary revivals of ‘pagan’ literature and scholarship on it had nothing to do with paganism and were as likely to be the work of cultured Christians. The argument is replayed repeatedly for an exceptional range of sources and approaches. In the course of the book, C. provides something not unlike a general literary and cultural history of the Roman west between c. 350 and 430 – with the crucial difference that rather than summarising knowledge, he offers fresh insights on almost every topic.

One way in which the title may remind readers of the book's prolonged genesis is that, since C. started writing, the very terms ‘pagan’ and ‘paganism’ have come under attack, both for representing a hostile characterisation and for implying non-existent homogeneity of belief and aims. C. repeatedly demonstrates the second point, but the first chapter argues on the basis of an exhaustive lexical study that paganus came in the mid-fourth century to be applied to non-Christians (and non-Jews) not in the contemptuous sense of ‘bumpkin’, but as a relatively neutral term: previously used for rural as opposed to urban and civilian as opposed to military, it ‘takes its precise color from an antonym’ (p. 22). His argument will be cited by anybody hereafter who prefers sensitive use of ‘pagan’ to fashionable alternatives such as ‘polytheist’. The second and third chapters cover the condition of pagan public religion under Christian emperors from Constantius II to Theodosius I (with Gratian's withdrawal of funding from the cults in 382 seen as more significant than the surviving anti-pagan laws of Theodosius in 391–2), and the usurpation of Eugenius in 392–4, where the creation of the legend around the battle of the River Frigidus as a pagan/Christian clash is studied in close detail. Then two prosopographical chapters: Chapter 4 on the decline and disappearance of the traditional priesthoods, without which there was no formal public means for aristocrats to be pagan; Chapter 5 on pagan converts, attempting to set the parameters within which the population shifted from one set of religious beliefs to another. Arguing that too much stress has been placed on the rigorists on both sides, he conceives the population as consisting of up to five groups, including moderates on each side and undefinable individuals in the middle (incidentally, in this chapter a reference to Sandwell's work on Antioch [p. 175] prompts C. to deploy the term ‘identity’ in the fashionable sense for the only time in over 800 pages of text – a choice which I leave to readers to deplore or applaud). C. accepts Barnes's argument that in the fourth century Christianisation of the high aristocracy proceeded faster than generally accepted, and himself argues that its comprehensiveness has also been understated in the early fifth.

After coverage of a miscellany of pagan writers in Chapter 6, including those like Pacatus who turn out not be pagans, Macrobius (Chapter 7), subject of one of C.'s earliest groundbreaking articles, is shown to be a Christian. Writing c. 430, he is nostalgic for but far removed from the pagan past – an interpretation fully accepted by the latest editor of the Saturnalia, R. Kaster. Chapter 8, on the Carmen contra paganos, provides the book's most brilliant display of philological and historical fireworks, demonstrating first that this attack on a pagan prefect cannot refer to Nicomachus Flavianus at the time of the Frigidus, next that it must instead refer to Vettius Agorius Praetextatus ten years before, and finally, on the basis of an attribution in a medieval library catalogue from Lobbes and of metrical, stylistic and intertextual comparison, that the poem was written by Pope Damasus before his own death on 11 December 384 (text and translation of the poem are included in an appendix). Chapter 9, on other anti-pagan verse invectives of the period, includes discussion of the centonist Proba, slightly misplaced alongside the Carmen ad senatorem (the senator is tentatively and unpersuasively identified as Domitius Modestus), the Carmen ad Antonium and Prudentius' Contra Symmachum. In Chapter 10, C. reconstructs the circle of Symmachus that he long ago deconstructed, using the letters to build up a picture of the alleged champion of paganism as a practical politician keen to use his letters to show himself a broad central figure, corresponding collegially with friends of whatever religion. There follows a brief and authoritative history of the fourth-century revival of early imperial Latin literature, which will be an immensely useful starting point for future studies, even if only loosely connected to the theme of the book by the demonstration that this revival has nothing to do with paganism: another chapter of which many scholars would have made a monograph. Chapters 12–14 incorporate an actual monograph, drafted in the 1980s, on the phenomenon of the late-antique subscriptions that survive in many manuscripts. These usually refer to ‘emending’ or ‘reviewing’ the text, and have often been associated with pagan aristocrats courageously saving classical civilisation. C. uses a comprehensive collection of subscriptions to show that they are far from exclusive to aristocrats or pagans or classical texts, and that they refer not to editing in any modern sense but simply the practice, vital before printing, of correcting texts, usually against their exemplars. Romantic nonsense, like the supposed ‘edition’ of Livy laboured over by the Nicomachi and Symmachi, is punctured. Chapters on knowledge of Greek and on the alleged pagan nature of Virgilian scholarship are followed by a pair of chapters on the lost Annales of Nicomachus Flavianus. Their inclusion reinforces the sense of C. as a scholar willing to cross the road to knock down a bad argument: since the Annales are attested only in two inscriptions, and nothing solid is known about them, most scholars who do not believe that they were an important and influential work of pagan historiography which influenced many other histories of the period have not confronted the arguments of those who do. A chapter on pagan art and its patrons reminds us of the remarkable breadth of C.'s skill. The last full chapter before the substantial and thoughtful conclusion treats the Historia Augusta (another case where C. goes out of his way, since few serious scholars believe that the HA has a serious anti-Christian agenda). To remark in a footnote that ‘little or nothing written since [Dessau 1889] has added anything important to the sum of knowledge’ is mischievous and, if understandable, somewhat unfair. Re-examining the unquestionable intertextual relationship between the prefaces of Jerome's Life of Hilarion and HA Vita Probi, he makes Jerome the imitator, and backdates the HA into the 370s or the 380s. On the first point he is very likely right. All the circumstantial detail of the intertextuality is on his side, and the argument is certainly stronger than the alleged allusions to Ammianus or Claudian or other authors that have been used to argue a later date. But his dating of Jerome's VH as early as 385/6 (p. 770) is unconvincing; the life postdates the Vita Malchi, which refers (2.1) to Jerome's friend Evagrius as papa, bishop, putting it after 388, in turn retarding C.'s terminus ante quem of the HA.

In such a large work some arguments will prove less convincing. Paschoud has already pointed out (in his review in Antiquité Tardive 20 [2012], 359–93, at 362 n. 5) that a dramatic date for Macrobius' Saturnalia in 382 immediately before Gratian's disestablishment of pagan cults only works if that disestablishment took place in precisely the last week of the year. Better not to seek an exact dramatic date in a work written half a century later. In the third chapter of the mini-monograph on subscriptions, C. misinterprets a subscription to Livy's first decade (emendavi Nicomachus Flavianus v.c. ter praefectus urbis apud Hennam), though with little harm to the overall argument. Since ter means not ‘for the third time’ but ‘on three occasions’, this implies that the correction took place not during Flavian's third prefecture – as C. argues, while acknowledging the strangeness of a prefect going as far from the city as Sicily while still in office – but afterwards. These are minor points. My largest doubt is whether in the period following 395, not too well attested by narrative sources, C. is overly influenced by the model he has destroyed in Chapter 3, of the civil war against Eugenius as a religious conflict (esp. pp. 187–95). He is right that the evidence adduced for widespread paganism among high office holders in the reign of Honorius is illusory, but his counterargument, essentially that the mere fact of being high office holders after 395 makes them likely to be Christians (which would certainly not have been true in 390), seems nearly as presumptuous, and out of kilter with the undramatic fizzling away of paganism that he persuasively presents elsewhere.

C. has given the thesis of aristocratic pagan resistance the treatment that Hercules gave the Hydra, though plenty of room for debate remains across the work's full range, as already illustrated by a thoughtful collection of essays by distinguished Italian scholars (R. Lizzi Testa [ed.], The Strange Death of Pagan Rome [2014]). Some forceful responses have come in areas which might seem tangential, such as the HA or Flavianus' Annales (see Paschoud's review, op. cit., also reprised in Lizzi Testa). This book will stimulate much more besides in the coming decades. It offers a virtuosic breadth of coverage and approach that must in the end justify its length. It is also wonderfully readable – a fact which in part (whatever one might think of this feature otherwise) is down to the polemical tone.