Imagine historians in a future where all physical sport has virtually disappeared from society, the preserve now of a few marginalized extremists who, thankfully, don't live round here and have not really had the chance to infect us with their views. How to address the influence and impact of sport in late twentieth/early twenty-first-century Britain, from (let's say) the vast yet incomplete current YouTube footage? They could be forgiven for drawing up a taxonomy that saw a strong connection between all ball sports (football, cricket, tennis, golf … surely all much of a muchness …), and for postulating a causal link between the proportion of a soccer game that a goalkeeper stood still and his team's overall success in Cup competitions. We might also feel unable to dismiss entirely a repeated assertion that the enormous effort and expense put into ‘proper’ grass as a playing surface was probably ‘really’ about something our putative scholar could relate to — the specifically visual impact of a particular shade of green.

In other words, we can see how we might get conclusions that are vaguely plausible (and that occasionally throw up promising if slightly bizarre angles) drawn from a meticulous and diligent study of the evidence that would nonetheless fail to do it justice.

Michael Lipka has set out to explain Roman religion (including Christian and Jewish phenomena) by subduing it to a taxonomy that extends work by John Scheid (9). My sense of Scheid's work is that he is noting important attributes of religion, but L. explicitly narrows the focus from religion to deity and promotes these attributes to being principles of ‘the single concept of god’: these ‘constituent concepts’ are space, time, personnel, function, iconography and ritual.

This methodology initially demands a cycle of surveys of suitable material from each perspective. While this reminds the reader that these are important (and are indeed nearly all that we have), it did not have the effect of convincing me of the value of L.'s extended (and altered) use of Scheid's discussions where the ‘concept’ of a god consisted of implementing these ‘constituent conceptions’. It also forces L. to give us a slightly dizzying account as we circle repeatedly around the same objects. This is perhaps unfairly encapsulated in the thought ‘oh, Isis again’ (though this sequence of presentation is perhaps unavoidable and is certainly logical within his scheme).

But it is the scheme I have serious difficulties with: apart from the fact that the meaning, relevance and status of ‘concept’ are assumed, nowhere is the idea of treating deities as concepts really problematized (the discussion on pp. 160–1 is not enough). Closing the book by restoring religion's place in society is too little too late, since there is nothing left except to see how these groups (well, the élite) ‘manipulated’ the concepts.

There is no doubt that time, space, personnel and so on were important, and that they are not unreasonable headings for our smorgasbord of evidence: but making these causal is like someone deducing from our historical record that an important motivation for securing paid university employment is to pay tax. It is true that religious activity happened somewhere, was ordered in time, had people attached to it (etc.), but that barely introduces (and certainly does not exhaust) its historical meaning. The scheme displaces its object: thus, after implying that Christianity's great early strength was its evasion of these restrictions and consequent mobility, L. seems almost disappointed that Constantine ‘paganised’ Christianity by creating buildings for it (192). Similarly, his conclusion offers that imperial cult weakened traditional cult and thereby contributed to the ascent of Christianity. Not for the first time, one notes the almost complete absence of references to civic paganism and the respect shown to older methodologies of religion: civic paganism, as a framework, may be contested but ignoring it is not the same thing as demolishing it. L. therefore misses the warning signs that recent scholarship has put up around some abandoned mineshafts.

Religion becomes an abstract game whose rules L. claims to have deciphered, an exercise to which the Romans for some unspecified reason felt they should devote centuries of energy and time, interrupted regularly by an agent's injunction ‘not to forget your sponsor’ (the cue for the familiar interloper ‘political manipulation of religion’).

Of course, treating religion as a system of some internal coherence (but overall irrationality and irrelevance) is a widely-held contemporary position, and this study might therefore be treated as an ‘explanation’ of Roman religion. Its alienness to modern secular culture justifies to my mind a particular scholarly generosity — we should encourage no little experimentation with methods and patient indulgence towards approaches that might yield even partial insights. It is, however, my sense that any account that loses sight of religion as something people did for themselves, on their own terms, to their own satisfaction and as meaningful to them, is one that we should be cautious of. L.'s careful and detailed exposition of his conceptual approach mistakes noting some of the rules of the game for its playing.