This learned book investigates the way in which the three main Roman historians writing in Latin viewed the role of the gods in Rome's past and in the threads running through the development of the empire. It takes religion in the Roman world entirely seriously, and presupposes that both Livy, Tacitus and Ammianus Marcellinus had their own working ‘models of religion' in the back of their mind while attempting to make sense of the vicissitudes of Rome's history.¹ An introduction is followed by two chapters on Livy and one chapter each on Tacitus and Ammianus. Conclusions, a comprehensive bibliography, a full index locorum and a subject index close the book which is as well produced as one may expect from its publisher.

The introduction sets the guidelines. Davies puts his own work in direct contrast to an approach to religion in the classical world which views the ancients with the help of Christianising, hence anachronistic terminology: too often the three historians have been labelled ‘unbelievers' or rather ‘sceptics', for the simple reason that scholars could not accept the premise that they actually ‘believed' in their gods and in the value of adherence to and participation in their cults.² Instead of categorization along such lines, Davies opts for interpretation as his main tool in studying ancient religion. In this, he places himself in the camp that aims to "understand the dynamic creation of identity and systems of meaning by and within a society" (9). One ought not to look for the ‘kernel of Roman religion', but to view it as "a woven pattern of positions in a particular relationship to the gods and all the other religious positions represented within Roman society" (9). Building on the work of Horton,³ Davies argues that Roman religion, like any system of knowledge, is comprised of different levels of explanation that constantly overlapped. As often in the book, a modern comparison is called to the rescue. "It may be less ‘contradictory' to speak of mortal efforts and the gods' actions in the same sentence after all, or, at least, no more so than when we speak of atomic bonds and superglue, or magnetism and gravity: if I drop an iron nail onto a magnet, why does it fall?" (11).

If Roman religion works on different levels simultaneously, another factor is added to the problem of what sort of information reaches us in what format: genre. "A historiographical god in action will not necessarily resemble a poetic divinity, because genres are not simply static types of literature, but strategies in representation" (12-3). Quoting Servius' commentary on Virgil's Aeneid 1.235 (historia est quicquid secundum naturam dicitur, sive factum sive non factum), Davies warns against easily introducing our own value systems upon categories such as ‘history' and ‘nature' in use by the ancient authors:⁴ "these categories do not exist in isolation but are defined to a large extent not just by what they include but by other categories: we would not, for instance, see a distinction between nature and chance, but they did" (13).

Rejecting that the religious references in our three authors are neutral and seeing the relevant material instead as forming "part of a wider strategy" (16), Davies' questions are far removed from the traditional contrast between ‘believing' and ‘being sceptic': "how complex can Roman religion be in a historian? How nuanced was religion as a system of interpretation?" (17). The space for such a discussion is provided by the ancient authors themselves in as far as they produced for audiences much more acquainted with what we could call ‘ancient religious frameworks' than modern scholars can ever wish to be, hence not feeling the need to sketch those frameworks in full. Davies' investigation isolates each historian in his own textual context: "importing contextless material is more likely to mislead us than clarify matters" (17). Only after the individual, fresh examinations is comparison between them allowed to enter the picture.

Livy receives two chapters. In the first, ‘Livy and the invention of Roman religion', Davies commences his turning of the ‘rationalist' Livy of modern scholarship into a main source for our understanding of the construction of ritual practice and its interpretative reading that moderns like to call Roman ‘state religion', into an historian also who seems doubtful whether his lecture on proper religious behaviour is taken to heart by his Roman readership.⁵ This chapter basically proposes a new reading of the way in which Livy presents any aspect relating to the divine world, providing us with "a nuanced series of preferences centred on identity, ‘practicality' and propriety" (27) on the part of an author whose religious history is fully integrated into the Rome of his own time. In a lengthy excursion on Livy's coverage of what we would nowadays perceive as paranormal events, taking into account historiographical issues too, Davies argues that the historian is not sceptic about the category of prodigies as such: "The issues of absolute credulity and scepticism are simply not present; far more pressing are issues of explanation, expertise and interpretation within the traditional framework" (46). Multiple examples show how, in Livy's presentation at least, in Roman ‘state religion' it was the Senate that had "ultimate jurisdiction over religious interpretation" (73), purportedly leading Romans to act "with proper Roman dignity, which drew on the ‘knowledge' that Roman religion was sufficient to deal with the situation" (83).

The next chapter, ‘Gods and men in Livy', follows this up with close scrutiny of his religious material: "as a working principle, coherence and intelligibility is maximised, not least as a corrective to previous discussions" (86). It is shown to what degree the divine world is present, and actively so, in the Ab Urbe Condita, though naturally our modern categories of ‘god' vs ‘man' are not to be put onto the ancient world in order to decide upon the absolute nature of the divine sphere, and furthermore "Livy's gods are not those of Homer", but instead "visible in the behaviour of the people concerned" (103). Three interconnected terms - fors ("there is no polarity between fors and the gods" [116]), fortuna (which "only makes sense when understood relative to the jurisdiction of the gods" [118]) and fatum ("even the gods work within its strictures" [108]) - are studied as key parts of the interpretative system of Livy, for whom "mastery of Roman religion was correct diagnosis of the varying factors in a given situation in order to produce an efficient and effective solution" (105).

Davies goes on: "the complexity of deductive principles means that, for the Romans, any statement about religion is interpretative rather than empirical" (123), which opens the door for a discussion of how such interpretations could have been and were received, and of how resulting perceptions played their part in a creative though competitive field of force. Davies' attractive suggestion that Livy's treatment of Scipio Africanus' claims to religious authority - which went well beyond proper adherence to traditional methods - should be understood in the context of the religious claims made by Augustus at the beginning of his reign, is carefully assigned to a footnote (132, n.103). Livy's model of Rome's religion, not to be confused with ‘how it really was', is thus presented to us as "a template for a religious as much as moral and political restoration" (142).

As regards his successor, in the chapter ‘Tacitus and the restoration of Rome', Davies sets out to investigate what had changed and what continued in the way a model of Roman ‘state religion' could be built up and accordingly reflected in historical writing (which itself had undergone serious changes since the days of Livy). The modern scholar is in a position "to see whether [Livy's] labours bore fruit" (142), thus to explore whether Rome had learnt its religious lesson. In this chapter, Rome's most famous pessimist is turned into an author much more positive than Livy. Writing to "an audience who will have to look to their memories to find suffering", Tacitus' works "claim to serve as warnings ‘lest we forget'" (225).⁶ His religious history according to Davies' interpretation is as elegant and highly-developed as his political account, in which it is embedded in the first place.⁷

Davies presents Tacitus as "radically conservative" (222) as far as religion, and with it proper religious behaviour, is concerned. Again, it is not a matter of whether or not Tacitus ‘believed'. Thus the weight given in the Annales and the Historiae to the hype surrounding the appearance of supernatural phenomena is explained here by referring to "his ‘lament' on the decline of prodigies" (163). Tacitus' contempt for magic as a form of superstitio does not mean that - in the eyes of the historian - it was futile, "but that it was unregulated" (165). And the priesthood plays its traditional part in the narrative.⁸ Davies draws special attention to the passage recording the burning of the temple of Jupiter Optimus Maximus on the Capitoline hill (Hist. 3.72.1): "in religious terms it is one of the key moments of the entire account and the logical conclusion of the decline of the previous decades" (207). The consecutive reconstruction of the temple under the Flavians (Hist. 4.53) "is the first major religious act that is not juxtaposed with indications of hypocrisy or corruption in either text" (209):⁹ it is at this moment in the historical narrative (which was, as Davies points out [148], of course not written in order of time) that Rome turns the corner with regard to her dealings with the divine world.¹⁰ According to this reading, it is not the first two Flavians who interrupt a negative spiral, but the reign of the third that acts as only a temporary setback within a process of recovery (cf. 151). The chapter is closed with an analysis of Tacitus' use of fatum in the story of Rome, supporting Davies' interpretation of the text as one in which the religious material plays a vital role.

Finally Ammianus, who wrote in the second half of the fourth century, when Rome was firmly en route to become a Christian empire. Obviously, his Res Gestae were the result of a historiographical process very different from those which had led to the works of his predecessors. He also had a very different audience. In many ways this is the most hypothetical of the chapters, partly due to the multiple, often contradictory, interpretations that the author has undergone in the course of modern scholarship. As before, Davies makes his point in occasionally very abstract and dense language: "Ammianus must … construct a religion for Rome within the vastly increased and polarised options realistically available to the denizens of the empire. By now we are accustomed to the recommendatory embodiment of religious habits and protocols; the historian shapes our knowledge both of what works for Rome and what is appropriate. What we will find is that the hallowed pagan practices of Rome are consistently defended" (236). That is not to say that Ammianus was anti-Christian per se in Davies' reading of his work, but Christianity receives its place in the religious periphery, far from state religion, for which it was "unsuitable as a candidate" (242). Davies views the Res Gestae as consciously civilising and educating, and finishes this chapter with a discussion of Ammianus' use of numen, fortuna and fatum. Where Livy and Tacitus can be said to have cogitated about religious aspects which in the end will have enjoyed widespread agreement, the fourth-century author "is creating religious standpoints, negotiating with traditional claims and contemporary concerns" (280). Rome's world had changed, and when trying to understand Ammianus it is important to realise that "we are dealing not with a change of belief, but a strategic change of emphasis" (283).

Doubtless connected with the fact that he holds his research fellowship at UCL at the Department of Education and Professional Development, the author describes his book in its very opening lines as "an exploration of the form(ul)ation of knowledge in a given context - a process which might well be called education" (1). In the final chapter, ‘Conclusions', Davies brings together the authors whose works are under scrutiny in the book, arguing how each of them sent out to its readership a serious message of warning against the waning of religious standards. Different as Livy, Tacitus and Ammianus may be, all three are presented as exemplary cases of historians for whom "even to avoid mentioning the gods was a historiographical statement" (290).

The density of considerations, and of the various steps of the argument, makes it difficult even for a lengthy review to do Rome's Religious History entirely justice. I would encourage anyone with an interest in Roman religion, society, politics or historiography to learn from this book. An easy read, however, should not be expected (as may indeed be guessed from the various quotations above). Davies' language is highly abstract, although modern and concrete comparative patterns are brought in at various points to explain matters (e.g. 11, 96). The argument is much further subdivided (sections even within 2.2.1.2) than the table of contents (2.2) would make one expect. For the reader who, like Davies, takes religion in the ancient world entirely seriously, his book will nonetheless be a real pleasure to read. Whether those who at the outset are more sceptic will be that easily convinced too, is another matter. The author himself seems to presuppose this when he states, at the end of the chapter that has revealed Tacitus much more of an optimist than Livy, that "the conservative reader will be suffering by now" (224). Davies expects quite detailed knowledge of the three authors with which he deals, though not all readers will be familiar with Livy, Tacitus and Ammianus in the same degree. Many undergraduates will therefore, I fear, be left at their peril.

Davies' respect for textual issues is very welcome. Much of the Latin vocabulary, itself of course limited compared to that of modern languages, carries certain overtones that moderns may like to refer to as ‘religious'. Due to the fragmented state of the historical works under discussion, none of which has been preserved in its entirety, textual studies will always depend to a certain degree on hypotheses, as Davies is the first to admit (e.g. 222-3). As a showcase of learning, the book's accessibility has suffered mildly, but its manifold provocation and well-constructed argument should still ensure that for the beneficiary reader Roman historiography will never be the same again.

Ted Kaizer, University of Durham
ted.kaizer@durham.ac.uk